cover JS3
Questioning Life and Cognition: Some Foundational Issues in the Paradigm of Enaction

John Stewart's book is a life achievement. It looks at three foundational issues for Enaction envisaged as a tenable paradigm for Cognitive Science: at first, the question of a “missing link” between the first living organisms – which, logically, have been dissipative structures simple enough to arise by spontaneous generation – and the simplest extant organisms that exhibit too complex a DNA-based genetic system to have arisen in that way; secondly, a relatively specific area with the cardinal virtue of being open to empirical refutation, i.e. the primitive immune system of vertebrates. Finally, the author tackles the social dimension of human cognition, presenting some of the basic concepts of sociology that typically need to be integrated into a potential paradigm of Enaction.

Continue reading
John Stewart
Tom Froese
Mattéo Mossio
arrow to header


John Stewart's book is a life achievement. It looks at three foundational issues for Enaction envisaged as a tenable paradigm for Cognitive Science: at first, the question of a “missing link” between the first living organisms – which, logically, have been dissipative structures simple enough to arise by spontaneous generation – and the simplest extant organisms that exhibit too complex a DNA-based genetic system to have arisen in that way; secondly, a relatively specific area with the cardinal virtue of being open to empirical refutation, i.e. the primitive immune system of vertebrates. Finally, the author tackles the social dimension of human cognition, presenting some of the basic concepts of sociology that typically need to be integrated into a potential paradigm of Enaction.


arrow to header


Enaction – Autopoïesis - Immune system - Origin of life - Genetic system – Ontogenesis – Phylogenesis - Punctuated equilibria - Social systems – Technology - Anthropological constitution


Enaction - Autopoïèse - Système immunitaire - Origine de la vie - Système génétique - Ontogenèse - Phylogenèse - Equilibre ponctué - Systèmes sociaux - Technologie - Constitution anthropologique
arrow to header
  1. Aspects of biological enaction
  2. The enaction of social life
decrease font size
reset font size
increase font size
arrow to header


The chapters in this book thus cover a wide span of questions, ranging from the origin of life (more precisely, the issue of the origin of a genetic system, see Aspects of biological enaction) to the advent of human society (see The Enaction of Social Life). In spite of possible appearances, however, they are not at all unrelated. Firstly, of course, they all fit in to the grand scheme of Enaction as an overarching paradigm for cognitive science which has the attractive feature of bringing together natural science, in particular biology, and the human and social sciences (in particular psychology and sociology, which themselves stand in need of mutual articulation).

Secondly, they are related in terms of my biographical career as a research scientist, which I will now sketch briefly. After being passionately interested in physics and mathematics at grammar school, which enabled me to gain entrance to Cambridge University, for the part II of my degree I switched to physiological genetics (combining physiology with population genetics), an area in which I continued for my PhD and my first post-doctoral positions in the United States and then in France. ... I then took some “time out”, interesting myself among other things in the “Radical Critique of Science”, being associated with the English Radical Science Journal and the French ImpaScience, and in left-wing politics (notably with the French PSU which developed the theme of self-management). One of the features that was associated with my disaffection for science during this period was the official ideology according to which science is, and (even worse from my point of view, should be) “value-free”; the traditional point of view was that achieving scientific objectivity required eschewing all forms of subjectivity. My feeling was, and still is, that if science really were “value-free”, there would be no good reason for spending time and effort on it. The paradigm of Enaction, particularly Varela's later development of “neurophenomenology”, brings subjectively lived experience back into the realm of questions that can be addressed from a scientific point of view. Moreover, the theme of auto-poïesis has definite resonance with the socio-political current of self-management.

My way back into science began in 1987, when as mentioned above I met Francisco Varela and Nelson Vaz at a summer seminar at Cérisy-la-Salle. What immediately appealed to my rebellious nature at the time was their idea that there must be something fundamentally wrong with classical immunology, which is based on the concept of a “self versus non-self” distinction where what is recognized (and therefore destroyed by the immune system) is “non-self”. From a basic “cognitive” point of view, the reason this has to be wrong is that what is perceived and recognized must be an aspect not of “non-self” but precisely of “self”. This led to the concept of a “VRM system”, on which I subsequently worked with Francisco and Antonio Coutinho at the Pasteur Institute in Paris, and which forms the subject of Aspects of biological enaction in this volume. This led me progressively to become interested in cognitive science more generally, and in particular the Maturana-Varela concept of autopoïesis. The next significant development came when I joined the Department “Technology and Human Sciences” (TSH) at the University of Technology of Compiègne (UTC). This University, created relatively recently in 1972, owed to the vision of its founder, Guy Deniélou, the feature of having a strong Department of Human and Social Sciences (including Philosophy) in an institution where all the students are in training to become engineers. The challenge, thus, is to make philosophy and human science relevant to the practice of an engineering career. One prong of the response to this challenge is to give an important place to Cognitive Science: by its nature, this is an area where human and natural sciences necessarily combine. The other prong of the response is even more original: it consists of thematizing technology as “anthropologically constitutive” (this will be presented in some detail in 4.4.1 of The Enaction of Social Life). Thus, my hope is that in spite of the apparent diversity, the reader will comprehend that the different chapters in this book do belong together.


The French have a saying: “Qui aime bien, châtie bien”. The closest English equivalent is probably “Spare the rod, spoil the child”; but for once, what I want to express is better conveyed by a literal word-for-word translation: “(S)he who loves well, chastises well”. I say this because in this book I will be critical of some aspects of the paradigm of Enaction; I will certainly be posing more questions than answers. But if I do this, it is not because I think this paradigm should be discarded; quite the contrary, it is because I consider that it provides the future for Cognitive Science as a whole, that I think it is worthwhile to look at some of its weaknesses and deficiencies. I propose to do this in the serene conviction that it will indeed possible to correct them; and that in the end these basic concepts can only emerge strengthened from the trial.

One serious weakness of the concept of autopoïesis and more generally the paradigm of Enaction is, in my view, the fact that so far at any rate it has remained rather abstract and schematic. More precisely, it has not been very prolific in generating local hypotheses that can be tested with respect to empirical data. Popper's criterion of “refutability” – a theory is only properly scientific if it is open to refutation by experimental evidence – may not be sufficient to guarantee good science; but it is, I think, a necessary condition. In this book, I will therefore recount an episode right at the start of my interest in this area. Together with Francisco Varela and Nelson Vaz, whom I had just met at a Colloquium at Cérisy-la-Salle, we developed a clear hypothesis concerning the immune system that was definitely refutable. This story did not have a fairy-tale ending, because our hypothesis was not only refutable, but it turned out be refuted! I recount this episode, nevertheless, because I feel that it does illustrate the sort of work that we need more of.

A notable effort to put more empirical content into the notion of autopoïesis is the work of Pier Luigi Luisi (2003[*]), in particular the study of reverse micelles and liposomes. This raises the question of the origin of life on Earth. It might seem that this is a difficult area for empirical work – since we do not have a time-machine that would enable us to go back and observe directly. However, this is an area where theoretical, conceptual issues can join up with empirical investigations. The point is that in all the current approaches to the question, there is a major “missing link”. All currently extant organisms have genes made of DNA, which function to enable protein synthesis. However, this current system is far too complex ever to have arisen by spontaneous generation – whereas, by definition, the very first living organisms must have arisen by spontaneous generation. I address this issue in Aspects of biological enaction that I call “the origin of a genetic system”.

Finally, in my current “shopping-list” of things missing from the paradigm of Enaction in its present state, there is the big question of the social dimension of human cognition. There is a basic requirement for any paradigm in Cognitive Science, and Enaction solves this by recourse to the “Cognition = Life = Autopoiesis” equation which is by now well-known. This is fine as far as it goes; but it undeniably tends to leave an impression that more than being grounded in low-level cognition common to all living organisms, it tends to get stuck there. Of course, one antidote is Francisco Varela's “neurophenomenology” which focuses on consciously lived human experience. This is also fine; but tends to remain individualistic. What is missing is the social dimension; after all, one of the things which distinguish human societies from animal communities is the existence of social structures. I attempt to make a start of remedying this deficiency in The Enaction of Social Life, in which I present some of the basic concepts of sociology which seem to me relevant for the paradigm of Enaction.

Aspects of biological enaction

The problem of the origin of a genetic system

A key moment in the history of biology was Pasteur's demonstration that nowadays, no living organisms arise by spontaneous generation. In other words, life always comes from life. With our current knowledge of the molecular complexity of even the simplest bacteria – a system where DNA directs the synthesis of proteins via messenger-RNA and ribosomes – we furthermore well understand how and why it is that such organisms cannot arise directly by spontaneous generation.

So: life always comes from life. Taking this as a basic premise[1], it follows logically that if we go back and back in time, we must come to a period when the very first living organisms on the planet Earth did nevertheless arise by spontaneous generation. This is only possible if these very first organisms – or “proto-organisms” – were very much simpler than even the simplest contemporary forms. We will be taking a look at what these original organisms may have been like; but the key conclusion for the moment is that life-forms on Earth must have undergone a process of complexification in the course of evolution. There is documentation in the fossil record that prokaryotic organisms such as archeo-bacteria (which still exist today) have indeed complexified – firstly to give rise to eukaryotic organisms with a nuclear membrane, and later with the advent of multicellular organisms. However, what I will be discussing here is an earlier process of complexification which must have occurred, between forms of life simple enough to have arisen by spontaneous generation, and organisms with a DNA-based genetic system.

The aim, then, is to identify an organism simple enough to have arisen by spontaneous generation. A straightforward approach to this is to start from an existing organism, and to simplify by taking away elements one-by-one. Genetic engineering has its limits; but it does make it possible to do this by taking away genes one-by-one. Craig Venter has actually generated considerable publicity around his project to do just this, in order to identify a “minimum genome”. The definitive results are not yet in; but it is already amply clear that this “minimum genome” will still have around 250 genes, compared to 500 or so for the simplest free-living bacteria that were taken as the starting-point. For our purposes, the verdict is unmistakeable: the reduction is only 50%, and 250 DNA-genes are still far too many for any such organisms to have arisen by spontaneous generation. In fact, the problem is actually even far worse than that. Even if the “minimum genome” had turned out to be only 10 genes – in fact, even if it was just 1 or 2 genes – there would still be a problem. Qualitatively, any organism with a DNA+protein system is too complex to have arisen by spontaneous generation.

So we would seem to have come up against a brick wall. The problem is that we are trying to get an idea of an organism simple enough to have arisen by spontaneous generation. Starting from present-day organisms, we simplify by taking away elements one-by-one; and we get to an organism such that no single element can be taken away without the organism collapsing. So, if we want to go back in time from that point, what could the preceding stage in phylogenesis have been like?

The principle of a solution: scaffolding

A principled solution to this conundrum has been proposed by Cairns-Smith  (1982[*][2]) He makes an analogy with the “problem of the arch”. Consider an arch, in which each stone holds others in place (see figure 1).

Here we have a structure such that if we were to take away any single element, the whole structure would indeed totally collapse. So the conundrum is: how could such a structure ever have been brought into existence, supposing that it must have been built piece-by-piece? Cairns-Smith answer is: by using an intermediate stage of “scaffolding”.

This idea of “scaffolding” is illustrated in figure 2.

The first stage 1 is to build a pile of stones. Once the pile is there to serve as a scaffold, it is now possible to place the stones of the future arch on this pile; there is no problem doing this one-by-one (stage 2). This can continue until all the arch-stones are in place (stage 3). Now, once all the arch-stones are in place, it is possible to start removing the stones in the scaffolding-pile (stage 4). Once all the scaffolding-stones have been removed; we do indeed arrive at the free-standing arch (stage 5). Thus, if we run the scenario illustrated in figure 2 backwards, we see that the way out of the “impasse” where we can take nothing away, is to add the stones of the “scaffold”.

Applying this metaphor to our problem of the impasse of the « minimum genome », we see that before the current DNA+protein system, there was not something less but something more: to wit, the elements of something functioning as a “scaffold” which made it possible for the previous organisms to evolve to the point where the DNA+protein system could be put in place. What could/might/should this “scaffolding” have been like? I will call this the “missing link” problem: what could have been the intermediate stage(s) between proto-organisms simple enough to have arisen by spontaneous generation, and the simplest current organisms with a DNA+protein system?

I come now to the main hypothesis that I want to put forward in this chapter: it is that this “missing link” must have involved some sort of “genetic system”. By “genetic system” I mean a system capable of encoding variations in the phenotype of the organism. We can call this a “primordial genetic system” to indicate that it was not based on DNA (nor even on RNA or polynucleotide acids of any sort). I do indeed consider that a genetic system of this sort is the key to the “missing-link problem” that we are trying to solve; the reason is that genetic systems make it possible to evolve solutions to difficult problems, which they do by instituting and organizing a process of random variation and (natural) selection.

A theme that I have introduced in the general introduction to this book is the importance of developing specific hypotheses that are open to refutation, in the sense advocated by Karl Popper. Thus, in this chapter, I will seek to examine the hypothesis that the “missing link” between the spontaneous generation of the first forms of proto-life, and the simplest extant organisms which all have a genetic system based on DNA and proteins, was some sort of “primordial genetic system”. Before that, however, there are a number of conceptual issues that it is important to clarify. Thus, in the next section I will illustrate the important point concerning the effectiveness of a “genetic system” (in this generic sense, not necessarily based on nucleic acids), by looking at several examples from the field of “Artificial Life”. This will show that challenging inverse design problems can indeed be solved by a process of “artificial evolution” employing “genetic algorithms”. In sections The Epistemology of Genetics and Biology and Genetics and information, I will look at the epistemology of the relation between genetics and biology, attempting to clear up some widespread misunderstandings, and specifying how the concept of “information” (and genetic information in particular) can be correctly construed. In section Autopoïesis and Biological Individuation, I will look at the basic invariant features of living organisms, notably autopoïesis, that are a pre-condition for differential genetic information to come into play.

Genetic algorithms and artificial evolution

In this section, I will start on the positive side, by presenting in sub-sections Autonomous robotics to Creatures several examples which illustrate what can be achieved by using “genetic algorithms”. I will defer until sub-section The limitations of genetic algorithms a critical appraisal, indicating the limitations and above all the external conditions that are actually essential for genetic algorithms to be able to function and to “deliver the goods”. These limitations only too often go un-noticed in the enthusiasm of the moment; but they will be crucial when we return to actual biology and our understanding of the origin and evolution of living organisms on Earth.

In general, the technique of “genetic algorithms” consists of encoding the different possible forms of an “organism” in an artificial genome (in Artificial Life, this “genome” is usually simply a bit-string). One starts with a population of organisms where the genomes, initially, are determined at random. At each generation, the organisms with the “best” characteristics – those which are relatively close to exhibiting the desired characteristics – are selected; the genomes of the next generation are generated by recombination between these “good” genomes, together with a certain rate of mutation. After a sufficient number of generations (maybe hundreds or even thousands – so that this sort of study in generally done in the form of computer simulations), experience shows that one does rather generally come up with a population of organisms that do exhibit the desired characteristics.

Autonomous robotics

The Artificial Life robots that are relevant here are “autonomous” in the sense that they have not only a motorized “body” equipped with means of action, but they have their own “sensor organs” whose output can be used to guide these actions. In a little book that has become a great classic, Braitenberg  (1984[*]) shows how “neuronal” connexions, even very simple ones, between the sensory inputs and the motor actions can generate behaviours that are apparently complex. He remarks, in general, that starting from a certain structure of connexions as given, it is relatively simple to deduce the behaviour that will result (he speaks of « downhill synthesis »); but by contrast, if one starts by specifying a certain sort of desired behaviour, it is often very difficult indeed to discover the structure of connexions that will actually generate that behaviour (he speaks of « uphill analysis »).

In order to illustrate these considerations, Braitenberg examines the behaviour of small autonomous robots with two photo-sensors and two wheels. With very simple sensori-motor connexions – each sensor is directly connected to the motor of one wheel, either on the same side or on the opposite side, and with an effect that is either stimulatory or inhibitory – four sorts of behaviour are produced when the robot is placed in the proximity of a light source (see figure 3).

  1. When the sensors are linked in stimulatory mode to wheels on the same side, the robot moves towards the light source, accelerating as it approaches but turning away, and slowing down again as it moves away.

  2. When the connexions are stimulatory but crossed, the robot turns towards the light, accelerating all the while, and charges into it at high speed.

  3. When the connexions are on the same side, but inhibitory, the robot turns towards the light, as in (ii), but slows down as it approaches the light and finally comes to a stop facing the light.

  4. When the connexions are inhibitory and crossed, the robot turns away from the light as in (i), but this time slowing down as it approaches and then speeding up when it has passed by the light.

In a stroke of genius, Braitenberg associated emotions with each of these behaviours: he called them (i) “fear”; (ii) “hate”; (iii) “love” (or “adoration”); and (iv) “curiosity”. Of course this attribution of “emotions” is purely heuristic; no-one seriously thinks that these little robots actually feel anything at all. Nevertheless, the spontaneity and the plausibility of these projections do indicate the intimate relation which exists between emotions and embodiment.

If “meaningful” behaviours can be generated already by such rudimentary “nervous systems”, we may anticipate that more complex neuronal networks linking sensors to actions would give rise to even more sophisticated behaviours. This will already be the case if the organisation of the network is simply linear and “feed-forward”: sensory input neurones being connected to a first layer of “hidden neurons”; then possibly to a second hidden layer, and so on; and finally to the output layer which determines the motor actions. In this case, the overall (sensory) input – (motor) output relation can be modified by adjusting the weights of the connexions between neurones in successive layers; but each time, once the connexion weights have been fixed, each sensory input will be associated with a definite motor output. Robots employing this sort of network will therefore be necessarily “reactive”: for any particular configuration of sensory inputs, they will always produce the same behaviour.

The situation can become even more complex if the organisation of the neuronal network is “recurrent”. In networks of this type, known as « Hopfield networks », each neurone is connected to each of the others. Networks of this sort have their own intrinsic dynamics: even in the absence of any particular sensory input, the internal state of the network will evolve to one or more “attractors” (fixed points, limit cycles or chaotic) according to the structure and connexion weights of the network. If certain neurones determine the motor actions, such robots will display an “autonomous” type of behaviour that can be modulated, but not entirely determined, by the sensory inputs[3].

In any case, already for feed-forward networks, and a fortiori for recurrent networks, the field of autonomous robotics runs up against the difficulty identified by Braitenberg: “downhill synthesis” and “uphill analysis”. It is indeed relatively easy to determine the behaviour generated by any particular network (it is enough to put the robot in the situation and simply observe what it does); but if the aim is to produce a certain sort of behaviour, it is very difficult indeed to specify a network which will produce that desired behaviour.

It is in this sort of context that the technique of genetic algorithms (which can be generalized, as we will see later) is very useful. As we have said, this technique consists of encoding the different possible structures and connexion weights of the network in an artificial “genome” which is initially set at random; of selecting the “best” the robots at each generation; of introducing some novel variation in the genomes of the next generation by recombination and mutation; and continuing for a sufficient number of generations. This technique has given interesting and convincing results. To take just one relatively simple example, it has been possible to produce robots which navigate towards a triangular target but not towards a square or a circle (Husbands et al, 1994[*]). An interesting feature of this experiment was that the architectures of the neuronal networks of the most advanced robots looked, at first glance, “strange” and very different from what human engineers would have invented; although, on retrospective analysis, it was possible to understand how that architecture had indeed generated the observed behavior. Another feature, even more interesting, was that if the experiment was repeated a number of times using different initial populations, it usually ended up with a population of robots that scored well, but the solutions were not the same between the different experiments. For example, in one experiment, the advanced robots were “blind in one eye,” using the inputs of only one of their two receivers, and they moved backwards! This is an illustration of a well-known property of what engineers call “inverse problems”: the desired behavior is well specified; from a given structure, it is straightforward to derive the functional behavior that results from it; but there is no direct way of “working backwards” from the desired behavior to a structure that will produce it. Inverse problems have an awkward property: there is no guarantee that a solution exists; but, if at least one solution exists, there is probably a multiplicity of fairly equivalent ones. In other words, these are problems that do not have a unique optimal solution. With “convenient” problems, those that have only one solution, classical methods (called “hill-climbing”) are efficient; but with inverse problems, classical methods are helpless. The technique of genetic algorithms, on the other hand, works well with these inconvenient inverse problems, which explains why it is so useful for engineers.


A second example illustrates the fact that genetic algorithms can be used to evolve not just control structures, but the actual morphology of artificial “organisms”. This example is that of the “Biomorphs” invented by Dawkins (1986)[*]. Dawkins used a 9-dimensional “genome” which encoded, in combinatorial fashion, over 500 billion possible forms. From a few simple starting points, a wide variety of results are obtained, reminiscent of biological forms (see figure 4). Dawkins's initial goal had been to generate tree shapes, starting from a trunk and developing into branches and sub-branches. But his work with “Biomorph” soon showed that the algorithm was in no way limited to producing different kinds of trees (a fir tree, an apple tree, etc.), but that it could be used to generate all kinds of forms, whether biological or not. Dawkins was sincerely surprised, and quite delighted, to discover a biomorph that looked like an “insect”, and later on planes, bats, chandeliers, etc.


The third example of what can be achieved using genetic algorithms combines morphology and behaviour. It comes from the work of Karl Sims (1994[*]), whose goal was to produce “virtual creatures” with bodies composed of a certain number of “blocks” articulated by flexible joints and animated by “muscles” under the control of “neuronal” circuits. These creatures were then immersed into various environments that were simulated to correspond to a suitably simplified version of physical reality, such as water or solid surfaces. The creatures were then selected on the basis of their ability to perform certain tasks, such as swimming, swimming after a light source, moving across a surface, jumping on a surface, or taking possession of a target in competition against another creature. A few examples are shown in figure 5.

These experiments produced sets of “creatures” that were both fascinating and striking. Some looked familiar, such as a fish with fins or a skate that swam with an undulating movement. Other creatures were efficient in accomplishing their tasks, but their forms and their style of movements were “strange” and very unusual.

This exuberant proliferation of forms is strongly reminiscent of what occurred at the time of the “Cambrian explosion” with the first multicellular animals. As Gould (1991[*]) has pointed out, a great number of “experimental” forms existed at that time, of which only a fraction have survived to become the forms that are familiar to us today. Consistent with this, the style that Sims adopted to present his results was that of a natural historian. The aim of his experiments was not to produce a unique, “optimal” form, but rather to exhibit a variety of solutions – a goal for which, as we have already remarked, the technique of evolution through genetic algorithms is particularly appropriate.

The limitations of genetic algorithms

Having gained some appreciation of what can be achieved by using genetic algorithms, it is now at least as important to be perfectly clear about what they are unable to do. More precisely, it is essential to identify the conditions that are necessary for evolution through genetic algorithms to occur. To state the key point right away: the “genomes” in these experiments certainly contain encoded “information”, in the strict Shannonian sense of the term; but, like all such information, the genomes do not and cannot specify what this information is encoding for. Let us look at this for the examples we have seen in sub-sections Autonomous robotics, Biomorphs and Creatures.

In the case of the autonomous robots that we examined in Autonomous robotics, what the genetic information coded for was the structure and connexion-weights of the neuronal network that controlled the robot's behaviour. But nothing in the genetic information itself specified that this is what it was coding for. Indeed, it is salutary to make a list of all the things that, in these experiments, were not specified by the genetic information:

  • First, as we have already just said, the fact that the sequence of 1's and 0's that make up the “genotype” actually “codes” for a features of a neuronal network. Intimately related to this, there is the very existence of a neuronal network that can be coded for.

  • The very existence of a robot, and the fact that the neuronal network is located inside a robot with the function of connecting sensory inputs to motor command outputs.

  • The fact that each robot is located inside an environment; and the fact that when the robot is placed in this environment, the structure of the robot is going to generate a behaviour.

  • The choice of a criterion for attributing a “score” to each behaviour.

  • The entire setup that enables the evolutionary process to take place effectively: in particular the reproduction, at each generation, of a population of robots, and the rules that specify the conditions for such reproduction.

  • Finally, there is the “code” itself: the manner in which each possible genotype corresponds to a well-specified neuronal network.

This last point is so general for all sorts of encoded information, genetic or otherwise, that it is worth taking another example. As is well known, the Morse system is used in telecommunications, an area of predilection for the deployment of Shannonian information. Here, the « information » is contained in the non-random sequence of dashes and dots: - - . - . . etc., a form that is particularly suitable for tele-transmission by radio ever since the rudimentary devices of the 19th century. What is instructively clear here is that this information: a) does not even specify what it is encoding for; nor b) does it specify what the coding relationship is. These points are determined by the Morse code, which a) specifies that what is coded for is letters of the alphabet; and b) specifies the correspondences, between groups of three dashes-and-dots on one hand and a letter on the other. For example: the group « ... » codes for « S »; « --- » codes for « O ». Now the key point is this: the impossibility of specifying the code itself within a Morse message is nicely illustrated by the impossibility of saying, in Morse: « the code for the letter ‘S' is ‘...' ». This would give: « (the code for the letter) ‘...' (is) ‘...' » – which is doubtless true, but singularly uninformative!

Coming back to our case of genetic algorithms and evolutionary robotics, there is of course a fairly good reason why the things listed in (i) to (vi), those that are not specified by the genetic information, tend to fade into the background. This reason is, that during the course of an evolutionary simulation, all these features remain invariant; whereas what is going on, what naturally enough engages all our attention, concerns the variable features that are precisely encoded by variations in the “genomes”. In all these experiments – and this applies equally to Dawkins' Biomorphs and Sims' Creatures – these invariant features are in fact put in place by the human beings who set up the experiments. In other words the invariant features, that are actually essential for the systems to function, are supplied from outside the systems themselves; in the event, by humans functioning as a sort of “Deus ex machina”. We may anticipate that this will raise some questions when we return to real living organisms, for in this case there is no “divine engineer” working from outside the system to set it up.

For the moment, however, I retain the feature that genetic algorithms do make it possible to evolve interesting solutions to challenging “inverse design problems”. I take this as evidence which supports (although it does not prove) the hypothesis that the “missing link” in the origin of life, between systems simple enough to have arisen by spontaneous generation and those that we know today, might have been some sort of genetic system. However, before taking a closer look at what might actually have happened in the early history of life on Earth, it is important to gain some further clarity about the nature and operational modes of genetic systems in living organisms.

The epistemology of genetics and biology

Precisely because I consider that the concept of a “genetic system” is a key feature in identifying the “missing link” in theories of the origin of life, it is important to clear away what I consider to be some serious misconceptions on the role of genes in living organisms (another case of “Qui aime bien, châtie bien”). In virtually all contemporary organisms, the genes are made of DNA; and there is no denying that these DNA genes are essential components. If they are taken away, the organism simply collapses. Unfortunately however, ever since the heyday of Molecular Biology (Watson & Crick's double-helix model of DNA, the discovery of m-RNA's role in protein synthesis, and Jacob & Monod's discovery of “operons” and the mode of control of “gene expression” via protein synthesis), there has been a lot of « hype » about DNA. At the present time this is only getting worse, with the need to justify the costly but rather mindless project of systematic nucleotide sequencing. Before we can get down to looking at what genes actually do, notably in terms of making it possible for organismic complexification to occur in the course of biological evolution, there are some widespread misconceptions that we first need to get out of the way.

In contemporary biology, “Genetics” has become a dominant force, to the extent of largely eliminating any genuine biology of organisms as such. This is clearly revealed by two telling phrases, both from leading scientists. François Jacob, a major biologist remarkable for his perspicacity and historical awareness, explicitly acknowledges the fact: “Today, life is no longer an object of questioning in the laboratory[4]. Henri Atlan, with characteristic lucidity, confirms the diagnosis: “Today, a molecular biologist has no use for the word ‘life' in his work... This means that biology studies an object, the object of its science, which is not life!” (Atlan & Bousquet, 1994[*]). And the fact is that the principal theoretically constituted object of contemporary biology is indeed not “life”, but “the gene.” To show that this is no exaggeration, and to flesh it out a bit, it is useful to cite a longer passage from Jacob (1970[*]):

“Heredity is described nowadays in terms of information, messages, and codes. An organism's reproduction has become the reproduction of the molecules that compose it. This is not because every chemical species has the ability to produce copies of itself; but because the structure of macromolecules is determined, down to the smallest detail, by sequences of four chemical radicals contained in the genetic inheritance. What is transmitted from generation to generation are “instructions” that specify molecular structures, the architectural plans of the future organism, and also the means of putting these plans into practice and of coordinating the system's activities. Each egg contains, therefore, in the chromosomes that it has received from its parents, all of its own future, the stages of its development, the form and properties of the being that will emerge from it. The organism thus becomes the realization of a programme prescribed by heredity.”

There are so many things wrong with this idea that “genes do everything”, that it is difficult to know where to start in order to criticize and ultimately to dismantle it. To get some bearings, here is a list of the major points:

  • Genes do not produce organisms (they do not even produce or re-produce themselves);

  • Although the regularities in the developmental process leading from a fertilized egg-cell to an adult multi-cellular organism make the idea of a “genetic programme” superficially attractive, there is no good reason to consider that this “programme” is “genetic” – in fact the very idea of a “programme” of any sort is seriously misleading; and finally

  • Genetics is not a science of heredity, for reasons that we will explain in due course.

However, before addressing these major points directly, I will first present three relatively secondary cases that can break us in more gently to the view that genes are maybe not the be-all and end-all of living organisms: the case of Red Blood Cells; an exemplary case of morphogenesis, the process whereby biological forms are created; and the question of the number of genes.

Red blood cells and haemoglobin

I said above that DNA genes are essential components of all contemporary living organisms, and that if they are taken away, the organism “simply collapses”. It is indeed true that in most organisms, if the genes are taken away the organism does not go on functioning for very long. It is nevertheless salutary to consider the case of the red-blood cells (RBC, also known as erythrocytes) in mammals. These cells play a vital role, that of transporting oxygen from the lungs to all the tissues in the body that need it to fuel their metabolism. Numerically, they constitute 25% of all the cells in a mammalian body. They are produced in the bone-marrow, by a process which takes about 7 days; after this, they are released into the blood-stream. Now the point I wish to make is that the RBC which circulate in the blood do not have a nucleus, and therefore do not have any DNA at all. Nevertheless, they go on living for 100-120 days, performing their function of transporting oxygen, and producing ATP from glucose (ironically enough, by an anaerobic process that does not use any of the oxygen they are transporting) in order to meet their energy needs. These facts, while admittedly something of a curiosity, may nevertheless give pause for thought; they demonstrate that DNA genes are not absolutely necessary for all the basic processes of living cells.

While we are on the subject of red-blood cells, it is interesting to look at the case of haemoglobin molecules. These molecules, which are the main component of the red-blood cells, have the very peculiar property of binding, but also releasing, oxygen molecules with great ease. This property, of a purely physico-chemical nature, is of great biological importance since that is what enables red cells to perform their function of transporting oxygen from the lungs towards the tissues. Now we may ask the unthinkable question: do the determining elements of this property actually reside “in” genes?

Haemoglobin is a protein; and as with all proteins, it is perfectly true that in normal conditions the linear sequence of amino-acids that constitutes the protein's “primary structure” is, quite literally, “coded” by the sequence of nucleotides in the corresponding gene. However, the property that interests us here is dependent not upon that primary structure as such, but rather upon the way in which this linear string of amino-acids folds back onto itself in order to form the molecule's tri-dimensional structure. Particularly noteworthy is the fact that this structure is such that iron atoms can fit into it. This is absolutely essential since these atoms, with their characteristic ability to shift between the Fe++ and Fe+++ forms, play a crucial role in the molecule's ability to bind reversibly with oxygen. Now, the folded configuration of the string of amino-acids is strongly dependent on the environment in which the molecule finds itself. For instance, in an aqueous environment of a given ionic strength and pH (which represents the usual case), hydrophobic amino-acids are attracted toward each other and form “clusters” that keep a particular kind of tri-dimensional structure in place. But in an aromatic environment (such as, for example, benzene) – or even in an aqueous environment of different pH – things would be very different, and the molecule's configuration would be changed completely. Given this, one can understand how it is that there is no identifiable nucleotide sequence that “encodes” iron. While it is true that the significant property of haemoglobin depends on a gene, it also depends, to an equal degree, on certain physical and chemical properties of matter, such as water and lipids that together establish the hydrophilic-hydrophobic polarity, the iron atom with its particular size and variable valence, etc. All these physical and chemical properties of matter are, precisely, invariant; and none of them are encoded in the nucleotide sequence of the DNA gene.

If we wanted to try to extend the scope of “genetic information” beyond the sequence of amino-acids and include in it the biological properties of the haemoglobin molecule, we would have to take into consideration all these invariant aspects. In other words, if one insists on saying that haemoglobin as a biological molecule is determined by “information,” one would have to admit that this information is not located exclusively in genes; which would mean, then, that it is distributed over the relationships between all the material elements that interact in order to determine the molecule's properties. But that is hardly an advisable way of proceeding, since it would end up making the notion of information quite meaningless. It is better to recognize, more simply, that genetic information does not determine everything.


Let us take an example, originally given by D'Arcy Thompson (quoted by Saunders, 1993[*]), and imagine that we wish explain the form of the small jellyfish shown in figure 6 (right side):

If one were to take seriously the idea that “genes do everything”, we would have to identify the genes (presumably a considerable number) which specify all the features of this apparently complex shape, and which then act as homuncular “sculptors” to actually produce this shape out of the available material.

Now, however, let us the left side of figure 6. What is shown here is an object that bears a remarkable resemblance to the jellyfish. What is it? – well, it is no living organism, but simply a drop of oil that has fallen into paraffin. How do we explain the form of this oil droplet? Surely not by appealing to the remarkable capacities of genes: an oil droplet has no genes at all. The fact is that at the moment when the droplet entered into the paraffin, purely physical forces acted upon it in such a way that it took on this complex form, without any “information” having been brought into the process from the outside. Properly explaining the morphogenesis – in terms of surface tension, the inertial movement of the oil droplet, and so on – is certainly not a trivial task; but genes clearly have nothing to do with it.

Let us now return to the right side of the figure. Does the form of the jellyfish, apparently so elaborate, still look like the result of genes “acting”? Certainly not. The core of the explanation of this form lies quite simply in the fact that it can be produced by a relatively simple physical process. But what, then, of the elaborate “genetic” scenario? It becomes superfluous and useless, to the point of being silly.

How many genes?

If one considers that “genes do everything”, it is legitimate to ask: how many genes are necessary? It is revealing, in this connection, to notice the dismay that has been provoked in the community of molecular biologists by the recent discovery that the genome of mammals (including primates and human beings) contains “only” about 30,000 genes. Some authors have suddenly come to their senses and begun to question whether 30,000 genes are enough to cover the scope of all that the “genetic programme” is supposed to accomplish. And the reality is truly alarming. For example, the human brain contains 1011 neurons with 1015 synaptic connections; compared to such astronomical figures, 30,000 genes does seem rather small. In addition to this, the process of ontogenesis is so complex that one may have serious doubts as to the ability of a mere 30,000 genes to specify it. This is not a definitive argument, of course, but it may at least serve to pose the question. Ontogenesis is such an important biological phenomenon that we shall be looking at it more closely below (see Ontogenesis). But just to give a foretaste, the problem is aggravated by the fact that the “genetic programme” is supposed not only to ensure the regularity of ontogenesis in normal conditions, but also all the variations that occur regularly under different conditions. Oyama (1985[*]) gives an ironical example of this. If laboratory rats are placed in an overcrowded cage during a heat-wave, one will observe, very regularly, that some of them choose to sleep hanging from the latticed roof of the cage by their teeth. Yet, given this observation, could anyone seriously imagine that the behavior “sleeping hanging by the teeth” had somehow been providentially inscribed in the rats' “genetic programme” right from its origin? This would clearly be absurd – for at the time when the genetic constitution of rats was established, several tens of millions of years ago, there was not a single human being around, let alone scientists putting rats into laboratory cages. The conclusion is obvious: if the “genetic programme” is entrusted with taking care of everything, its mission is impossible. In fact, this somewhat tardy awakening is itself far too limited in its scope: would 100,000 genes, or even a million, be able to do what 30,000 cannot? The real question is not quantitative but qualitative: what is it that genes do really do?

Do genes produce organisms?

We come now to the first of the major questions noted above: do genes actually produce organisms? We will come to the question of ontogenesis, the process by which adult multicellular organisms are produced starting from a single fertilized egg-cell. But even in the case of single-celled organisms such as bacteria, it is simply not true that the organisms are produced by genes. Fundamentally, all living organisms are not so much “things” as processes; they are “dissipative structures” which arise through being coupled to a continual flow of matter and energy. When we come back to the central question of this chapter concerning the origins of life-forms on Earth, we will see that “biological individuation” and “autopoïesis” can arise without any genes at all. I will indeed be arguing that a “genetic system” is indeed actually essential: by coding for differential variations in the form of autopoïesis, it is a genetic system that makes it possible for evolution towards progressively more complex organisms to occur. But the genes are not what produce the organism – neither in the case of the “primordial organisms” that I will be discussing later, nor in the case of contemporary organisms endowed with DNA-genes.

In fact, while we are on this subject of “production”, not only do genes not produce the organism; they do not even produce (or re-produce) themselves. DNA molecules left alone in a test tube do not reproduce themselves[5]. True, they can be copied, which is a highly important property from the point of view of information theory; but that is not at all the same thing. Let us take a rather crude analogy to drive the point home. A text written on a sheet of paper can be copied – whether by hand, if it is in alphabetical writing, or by a photocopy machine. However, it would never occur to anyone to say that the text “reproduces itself[6]” – because in this case, the need for an entire copying apparatus is too obvious. But after all, the same applies to DNA. In order for a DNA molecule to be copied, there is a whole array of very specific conditions which all have to be present: there needs to be a set of precursors, a source of energy, certain particular enzymes, etc., with everything being configured in a particular way. This can now be done in the laboratory, although even there by far the most convenient way of reproducing DNA genes is to “clone” them in a suitable organism. Under natural circumstances, the only place where these conditions are realized is inside a living cell. In other words, for DNA to be reproduced, a living organism already has to exist.

Is the “programme” genetic?

We opened this section with a quotation from Jacob (1970[*]) which concluded in grandiose fashion: “The organism thus becomes the realization of a programme prescribed by heredity”. The notion of a “genetic programme” is indeed one of the cornerstones of contemporary biology. Now just as it is sometimes salutary to ask whether the King is not naked, it might be useful here to ask an apparently naïve question, which runs so contrary to what everyone automatically takes for granted that whoever dares ask it risks being covered in ridicule. The question is this: how do we know that the “programme” is “genetic[7]”? In the literature on this subject, three types of answer to this question can be found:

  • A character is considered to be “genetic” if a difference in a gene, all other things being sufficiently equal, is the cause of a difference in that character. In this sense, the opposite of a “genetic” character is an “environmental” character, in which a difference in a character can be attributed to a difference in the environment, with the other factors and principally the genes being sufficiently equal.

  • A character is considered to be “genetic” if it is “innate,” i.e. if it is invariant within a species.

  • The soma of a multicellular organism is formed by the process of ontogenesis. In the course of this process, the cells resulting from consecutive divisions of the fertilized egg become gradually differentiated. In animals, at the end of gastrulation, three types of cells form: the endoderm, the ectoderm and the mesoderm. Later, the endoderm gives rise to the stomach and intestinal tract cells; the ectoderm becomes the skin and nervous system cells; and the mesoderm develops into muscles, bones and blood cells. Now, we know that following mitotic divisions, the genes of all cells are identical. Therefore, if such differences occurred between different individuals, there would be no hesitation in calling them “environmental differences” (see (a) above). But from the moment these differences occur inside a single organism in the course of its ontogenesis, they are declared to be “genetic” – the only reason being, it seems, that ontogenesis is supposed to be a result of the deployment of the “genetic programme.”

Each of these three reasons, taken separately, can be honestly and openly examined, and it is not a priori impossible that it may be acceptable. But before we even begin to discuss each of them in detail – which we will do – we should note that they are completely incompatible with each other. One cannot rely upon a definition that makes sense only if characters are variable, as in (a), and, at the same time, take phenomena of invariance as a basis, such as in (b). As for definition (c), it is in direct contradiction with (a). The fact that these incoherencies generally escape notice is itself a reason for concern; and it already indicates that the King of the Land of contemporary biology may, perhaps, not be quite as well dressed as some would like to believe[8].

Let us, then, take a closer look at each of the reasons that are usually offered for considering that the “programme” is genetic.

Definition (a) of what is “genetic” (and what is not), a differential definition, is perfectly consistent with basic Mendelian genetics. It is this definition that I shall retain, and indeed we will come back to it when we try to identify the “missing link”.

Definition (c) is in direct contradiction with (a), which is already a sufficient reason for discarding it without further ado. As if this were not enough, definition (c) has the additional flaw of being circular: it amounts to saying that the “genetic programme” is “genetic”... because one has decided in advance that it is “genetic”. This is not acceptable in scientific discourse. I shall take a closer look at this in the next section 4.7, when I will address more fully the question of ontogenesis.

Definition (b) is incompatible with the fundamentally differential epistemology of Mendelian genetics. For there must be variations in an observable phenotypic character in order for us to make an analysis of variance, and to attribute the variations either to genetic differences, or to environmental differences. If the character in question is invariant, it does not mean that it is “genetic,” nor even that it is “environmental,” but simply that the very “genetic versus environmental” distinction no longer makes sense.

This point has created so much confusion that it deserves closer examination. The “heritability” of a character is defined as a proportion of the total variation that is due to genetic differences. Thus, if variations are primarily due to genetic differences, the heritability will be close to 1.0; conversely, if variations are caused by environmental differences, heritability will be close to 0. But heritability is not a “good” property of any given character, since it depends on the particular population under consideration. The same character can have a heritability that is either close to 1, or close to 0, depending on which population is examined. Let us take, for example, the case of the character “skin colour” in human beings. If the population in question comes from interbreeding between black Africans and white Europeans, and if all the members of that population are living in a relatively stable environment, variation in skin colour within that population will be almost entirely “genetic,” and the heritability will be close to 1. On the other hand, if a population is racially homogeneous (for example, if all its individuals are North Europeans), but it so happens that some individuals have exposed themselves to sunlight more than others, so that some are tanned and others are not, then the variations in skin colour within that population will be almost entirely “environmental”; therefore, the heritability of that very same character will be close to 0.

In fact, if one wishes to qualify the various different characters of a certain species of organism by a property that is stable and that does not depend on the particular population that one is observing, “heritability” would be just about the worst choice one could make. A much better candidate would be the distinction between characters that are “plastic” (i.e. that are relatively easy to alter) and others that are “rigid.” We will return to this point when we undertake our discussion of ontogenesis. For the time being, let us simply take a few examples. In human beings, it is not infrequent to find adult individuals who weigh less than 100 lbs, while others may weigh over 200 lbs; the character “weight” is, therefore, a “plastic” one, since it can easily vary by a factor of two. By contrast, the character “number of fingers” is fairly “rigid”: almost everybody has 10 fingers, it is very rare to find an individual with 9 or 11 fingers, and even those cases are usually the result of an accident that is clearly exceptional. Now, if a character is “plastic”, in a natural population where both genetic and environmental differences exist, it is likely that both will cause differences in that character, so that the heritability of this character will be somewhere around 50%. On the other hand, if the character is rigid, neither genetic differences, nor environmental differences will produce any variations in it, so that its heritability will be neither high, nor low, but will be 0/0, i.e. undetermined. Now it is important to understand that everything that is invariant falls within the “blind spot” of differential genetics. The conclusion is clear: from the moment one tries to use the notion of a “genetic programme” to explain regularities and invariances, there is no valid reason left to consider that the “programme” is genetic.


An area of contemporary biology where the unfortunate notion of a “genetic programme” is particularly deeply entrenched is that of ontogenesis. As promised above, we will now take a closer look at this question.

Ontogenesis is the extremely complex process that leads from the fertilization of an egg-cell to young adulthood, passing through the different stages in the formation of an embryo. Ontogenesis, in fact, continues beyond maturity to include ageing and death at the term of a “life span” that is characteristic of the species[9] . Now, the regularity of this process is quite remarkable: it is, one might say, one of the most astonishing facts in biology. Foetuses that abort and monstrous embryos are extremely rare; this is all the more impressive in view of the great intricacy of the ontogenetic process[10]. In addition, this regularity is clearly a dynamic one: an embryo can survive many disturbances, one of the most spectacular being its division in two (which can happen up to a surprisingly late stage), producing not two half embryos, but a pair of monozygotic twins that are very justly called “identical.” This stunning regularity of ontogenesis is precisely what makes the concept of a “genetic programme” superficially so attractive: how, indeed, can such regularity be explained if the process is not “programmed” in advance?

However, we have already taken a preliminary look at difficulties involved in the concept of a “genetic programme.” We have seen in Is the “programme” genetic? that there is no basis (other than simply begging the question) that justifies asserting that this “programme” is “genetic.” In the present section, we will see that the problem is even more profound: it is the very concept of a “programme” (whether genetic or not) that is problematic.

The reason why the processes of ontogenesis have until now resisted scientific explanation (which would render the pseudo-explanation of a “programme” superfluous) is deeply rooted. As Oyama (1985[*]) has most shrewdly observed, this reason lies in a very basic assumption, which runs through the whole of Western thought since the time of Plato and Aristotle, concerning the relation between Form and Matter. This assumption consists in presupposing that matter in itself is essentially inert; that at the very best, it can only be the site of chaotic processes. It follows from this that any “organized” material process must be “in-formed” from a source that is essentially external to the process itself. The image is that of clay which must be fashioned by a sculptor. Applying this schema to the case of a developing multicellular organism, there are two potential reservoirs of external information: first, the environment (which is obviously external to the organism); secondly, genetic information[11] . If significant morphogenesis really did require “in-formation” from an external source, (a view which I am radically criticizing here), it would then be perfectly consistent to postulate a priori that informational resources are distributed between genes and the environment; and to ask the question as to their respective proportions. That is why the “analysis of variance” between genetic differences and environmental differences has seemed so attractive (in spite of the fact that “heritability” is not a stable property of a character, as explained above). It is also the reason why the opposition of “innate” versus “acquired”, or “nature” versus “nurture”, has become such a popular theme. But now, what becomes of this sort of opposition if the material processes that occur in living organisms are neither inert, nor even chaotic, but possess intrinsic capabilities of self-organization?

In order to understand how morphogenetic regularity can exist without a “programme” of any sort at all, it will be useful to take some examples, precisely, from the non-organic realm. We have already seen one example, the oil-droplet in Figure 6 above. But this is such an important question that it is appropriate to take another example. Snowflakes possess a remarkable inner structure. Every snowflake has six branches, each of which has an extremely intricate structure; so elaborate, indeed, that of all the trillions of snowflakes which have existed ever since the Earth was formed, a rapid calculation shows that no two can ever have been exactly identical. Yet, within a given snowflake, each of the six branches is remarkably similar to the other five (even if the six branches are not absolutely identical – see figure 7). How is this possible? How can each branch in the course of its formation “know” what form is being adopted by the others, in order to match it so closely? The temptation is great – almost as great as with biological ontogenesis – to assume that there must be a “programme” somewhere, external to the branches themselves, that “informs” them about what morphology they should adopt. But of course, in the case of the snowflake, we know only too well that no such “programme” exists anywhere: neither at the core of the flake, nor in the environment that surrounds it.

In its main lines, the real explanation seems to be as follows (Begley S. & Carey J., 1983[*]). Snowflake crystals form when molecules of water pass directly from water-vapour to a solid crystal, without passing through the intermediate liquid phase. This means that the process of ice-crystal formation takes place close to the singular “triple point” where the three phases – gaseous, liquid and solid – meet. Consequently, this process is extremely sensitive to the precise and combined conditions of temperature, pressure and humidity. The six branches are almost identical, then, because each snowflake is so small that all six branches share the microclimate; consequently, they share the same history of fluctuations in physical conditions as the growing snowflake swirls through the cloud. The unique character of this history is multiplied by the fact that there is a fourth critical factor for the morphology of each increment, to wit: the pre-existing shape of the branch at that moment; and this fourth factor is also (from one instant to the next) identical for all of the branches, but (progressively) different from one snowflake to another. To sum up, the stunning similarity of the six branches is due to nothing more nor less than the meticulous application of a very basic scientific principle: the same causes produce the same effects.

This example of snowflakes gives a striking illustration of the fact that material processes can be highly self-organized, and thus regularly give rise to complex morphologies. In view of this, it appears that there is no reason why the physical foundations of material morphogenesis should be fundamentally different between living multicellular organisms and non-organic processes. Indeed, upon reflection, it is clear that they must actually be the same. In both cases, the only forces that are capable of moving matter are physical forces – mechanical forces, such as hydrostatic pressure or viscosity; electrostatic attraction or repulsion, or electromagnetic forces; Van der Waals's forces; the hydrophilic-hydrophobic polarity; and so on. I will examine later what it is that distinguishes living organisms from purely physical and chemical processes; but we can already assert that nothing, in the processes of the living, relies upon the intervention of any other forces outside the laws of physics and chemistry. In particular, genes – whatever remarkable properties they may possess in other respects – are in no way little “sculptors” capable of materially shaping the forms of living organisms using special forces that they alone would be able to wield[12].

Our analysis of the remarkable regularities in the morphogenesis of snowflakes leads us to two important conclusions. First of all, if there is a “programme,” it is not located in any specific place; it is “distributed” over the entire set of elements that interact during the process, without any one of them being favoured over the others. Secondly, this “programme” does not pre-exist before the processes in question; the “information,” if one wants to insist on keeping this notion, is created as the process develops, in real time, by the very process that “expresses” it. In fact, a proper explanation of morphogenesis in physical terms makes the whole idea of a “programme” quite superfluous. What becomes, then, of the biological “programme” that was supposed to “inform” the ontogenesis of multicellular organisms? As soon as we take a closer look at the real processes of ontogenesis, it appears clearly that the reason why the efficient causes of ontogenesis are linked together in such a reliable chain is essentially that their organization is based upon regularities that are produced in a reliable way by the developmental process itself[13].

The first stages of embryogenesis

The very first stages of embryogenesis, which are common to a great number of multicellular animals (in particular, vertebrates and echinoderms[14]), are represented schematically in figure 8. Even a relatively simple analysis of these stages reveals that this is a “historical” process that creates by itself, as it proceeds, the conditions for its own further development.

The very first cellular divisions give rise to the morula, a cluster of almost identical cells that globally has a spherical shape (see Figure 8b). Why is the morula spherical?[15] Essentially for the same reason that a droplet of oil in suspension in water is spherical: since the amount of free energy in contacts of the cells (or oil molecules) with each other is less than the amount of energy in the contacts between the cells and the aqueous medium, the shape that minimizes the amount of overall free energy is the one that minimizes the surface/volume ratio in a three-dimensional space; and that shape is a sphere. This mechanism is nowhere inscribed “in the genes”; it is an invariant that results simply from the purely geometrical properties of 3-dimensional space. Therefore, the resulting shape need not, and indeed cannot be, inscribed in the genes either. Furthermore, the interactions among the cells themselves and with their surrounding medium that lead to the effective implementation of the shape, although they are perfectly predictable, are produced by the embryological process itself and cannot, therefore, be pre-existent to it.

This “historical” characteristic of the embryological process becomes even stronger in the later stages. Precisely because of the spherical shape of the morula, some cells will be located at the surface, in contact with the outside medium, while others will be inside, surrounded by other cells. This difference will emerge only under the condition that the morula is indeed spherical (rather than a flat, two-dimensional sheet, or a one-dimensional string). From the point of view of the self-organization of the morphogenetic processes at work in the embryo, this difference in the local environment of the various cells can therefore serve as a perfectly reliable signal for triggering an appropriate kind of differentiation between these two types of cells. In this particular case, the inner cells react by secreting a fluid, whereas the outer cells cohere tightly together to form a “skin”. And that explains how it comes about that the embryo thereafter takes the form of a blastula, a hollow sphere with an epithelial wall (figure 8c).

Resulting in this way from the previous form of the morula, the blastula in turn becomes the pre-condition for the formation of the next stage. The fact that the blastula has the shape of hollow sphere creates the possibility for a particular movement called “gastrulation”: some of the cells that were initially located in a region on the outer surface migrate into the centre of the sphere, which produces the very characteristically shaped gastrula with the topology of a torus. As can be seen in Figure 8d, the cells that migrated into the centre make up the endoderm and the basic form of the intestine; the cells that remained on the surface form the ectoderm, which will later develop into skin and also nervous tissue; and the cells located between the ectoderm and the endoderm form the mesoderm, which will serve as a basis for the skeleton, muscles, and blood. Once again, it is the task of embryology to determine in what way the signals that determine this three-fold cell differentiation derive from the specific aspects of their respective positions inside the embryo. In a sense, the topology of the relations between the endoderm, the ectoderm and the mesoderm is contingent; but it is a sort of contingency that is reliably produced by the embryological process itself. In other words, the fact that the “in-formation” necessary for the process's organization is not pre-existent to that process, but is created on an on-going basis by the process itself, provides a scientific explanation of the robust regularity of ontogenesis.

Going beyond the “internal versus external” opposition

The range of possibilities for self-organization of this type, far from decreasing, multiplies as the embryo further develops and its complexity increases. It must be noted that among these relational regularities, there is no essential distinction between regularities that are “internal” and regularities concerning the organism's “external” relations with its environmental niche. Also, as Jacob (1981[*]) himself has pointed out, biological organization typically results from “tinkering” of an essentially contingent and opportunistic nature. These aspects are important, not only for questions related to the location and pre-existence of a “programme,” but more generally with regard to the whole issue of “innate” versus “acquired” (or “nature versus nurture”); it will therefore be useful to illustrate them with an example from a much later developmental stage.

This example, taken from Oyama (1985[*]), concerns the organization of a critical moment in the development of a certain kind of fruit fly: the hatching of the cocoon and the emerging of a young adult. It so happens that because of the climate of the region where this species lives, this is a delicate matter. In this particular geographical location, the nights are extremely cold, so that if the young fruit fly happens to emerge during the night, it freezes to death. On the other hand, the days are extremely hot and dry, so that if it emerges during the course of the day, it will “burn” to death before its wings and body have time to harden through contact with the air. In order to survive, therefore, the fly must come out of the cocoon at a very precise time in the early morning hours, just when it is getting a little warmer, but before the dry heat kicks in. At first, it may seem that there is a relatively simple and logical solution to this problem: the hatching process only needs to be triggered by a thermo-sensitive receptor. However, as it turns out, such an organization would not be viable. It takes some time for the hatching process to get started before the cocoon actually opens; so that if it were triggered by a marked increase in temperature, by the time the fruit fly was out of the cocoon, the air would already be so hot and dry that the fly would still be burned. So, how is the problem resolved? It so happens that, in this location, day breaks one hour before the air temperature begins to rise. Therefore, if the hatching could be triggered by photo-receptors, the fly would emerge at the ideal moment. And that is, indeed, how this species functions. But one can appreciate the full extent of this organization's contingency: light, as such, has no inherent relationship with what is relevant to the fruit fly's survival. This is proven by the fact that, if the species is moved to a place where the same contingent relationships between luminosity, heat and humidity do not exist, this mode of organization is no longer viable.

This example illustrates well how the very process of ontogenesis is capable of creating situations where available opportunities for self-organization are abundant. Following the logic of the organization of its ontogenesis, it is the fruit fly itself that confers semiotic significance to the light of dawn, by using it as a signal heralding the advent of warmth. Without the fly, the “environment” is nothing; at least, nothing of all the above. Conversely, the organizational setup that ensures the regularity of the fly's development cannot be confined inside the fly's organism: the fly uses certain relationships (the time lag between daylight and heat) that are sufficiently reliable to build upon them the organization of its own development.

Going beyond the “genetic programme”

These examples we have discussed are, of course, extremely sketchy and do not constitute a proper theory of ontogenesis. In fact, a genuine theory of ontogenesis has yet to be developed. The fault lies with the widely-held presuppositions about the relationship between Form and Matter, according to which all regular processes must be “programmed” by “information” external to the processes themselves. These presuppositions, deeply rooted in Western culture as we have noted but equally deeply erroneous, have led scientific research astray from its real purpose. What genetics (as a science of differential heredity) is constitutively blind to, is that self-organized invariances can be constructed. In the case of ontogenesis, the invariant processes are based, on one hand, upon the laws of physics and chemistry, and on the other, upon relationships that are essentially contingent, but, when placed in context, are regular and robustly reliable. And these invariant laws are not “encoded” anywhere.

We have made the point that the “information” that “guides” ontogenetic processes is not located anywhere in particular; rather, it is distributed over all the material elements that interact among each other in the course of the process; and that, in addition, this “information” does not pre-exist the process in question. Now it should be clear that a “programme” that has no spatial location, and does not even pre-exist the events that it is supposed to direct, explains nothing at all. In fact, abandoning the notion of a “programme” (whether genetic or other), as we propose, in no way means abandoning the quest for a true scientific explanation of the regularities of ontogenesis – quite the contrary. As we have already noted, there is a sense in which morphogenesis in living organisms is no different from morphogenesis in the non-organic world; locally, it lends itself to exactly the same types of explanation in terms of dissipative structures and physical and chemical forces. There is, of course, a difference in degree, if not in kind: in the case of ontogenesis, the morphogenetic processes lead on one from another in a long chain, so that the multicellular organism which finally results is vastly more intricate than a snowflake. And there is another difference too, even more significant: the initial boundary conditions that, as in any physical system, are the cause of morphogenetic events are also, in the case of an embryo, a result of actions performed by the embryo's adult parents.

This is most obviously the case for the fertilized egg-cell, which is clearly produced by the parents. It is less obviously the case, but just as important, that the parents place the developing embryo in a context such that the self-organizing developmental process can proceed normally. This is most markedly the case for human babies and children, who require a nurturing environment to continue their ontogeny after birth; it is also the case for all mammals, who provide a womb for their embryos; but it is generally the case for all multicellular organisms: even plants distribute their seedlings in such a way that at least some of them will fall in an ecological niche where they can develop. As Susan Oyama (1985[*]) has pointed out, what is inherited from one generation to another is not (just) the genes, but the developmental system as a whole; in her words, we may best define heredity as “the passing on of all developmental conditions, in whatever manner”.

To sum up: a crucial feature of ontogenesis is that it has a circular organization, on the time scale of a generation. We shall see later, when we look at “biological individuation” and autopoïesis, that “circularity of organization” is not restricted to multicellular organisms; at an even more basic level, on the immediate time scale of moment to moment, a circular organization is the fundamental feature of “life itself”.

Two final remarks are necessary. The first remark concerns the so-called “homeobox” genes which are currently the object of much interest. The basic fact is that differences in these genes cause remarkable differences in the body-plan of the developing embryo. One of the first of these genes to be discovered was the “bi-thorax” gene in the fruit-fly Drosophila: as the name implies, flies carrying this mutation have a “double thorax”. More recently, a number of other genes of this sort have been discovered; sequencing shows that they are remarkably similar in many other species, from jelly-fish to insects, reptiles and mammals. More than that, in each of these genes a particular DNA segment 180 bases long is virtually identical. This DNA sequence, called the homeobox, translates into a protein sequence 60 amino acids in length. Now even in molecular terms, this finding is weird and hard to explain, because it is known that hundreds of different genes participate in the formation of a body segment; yet here we have single mutations that seemingly create new body parts and eliminate others. In molecular terms, an answer to this conundrum probably lies in the fact that the protein sequence involved acts as a “transcription factor”, binding to DNA and switching on and off the process of transcription, the expression of genes into proteins. These homeobox genes may thus act as “master switches”, turning on and off whole arrays of other genes. But if we consider the actual morphogenesis involved, following the whole thrust of the arguments developed in this section, it should be clear that the problem is deeper: the remarkable effects of differences in these genes is not so much an explanation of ontogenesis, but rather something which remains to be explained. And this explanation will require a better understanding of the actual physical processes involved in the morphogenesis.

The second remark is this: I have been particularly critical of accounts which attribute to genes the role of “architects” which actually fashion the developing embryo. It should go without saying, but even better in saying it, that this does not mean that it is the “environment” which plays the role of “sculptors” which fashion the embryo; that would be to fall back into the “innate versus acquired” dichotomy that I have attempted to dissolve completely.


In this section, we have subjected the unfortunate notion of a “genetic programme” to a critical analysis. In Is the “programme” genetic? we have seen that to the extent that there is a biological “programme”, there is no valid reason to suppose that it is “genetic”; and in Ontogenesis, more radically, we have seen that the very notion of a “programme” of any sort, “in-forming” biological processes from the outside, is to be rejected. However, we are under no illusion: in spite of this criticism, this notion of a “genetic programme” will continue to exert a seductive appeal. The only hope of eventually coming to put it aside lies in proposing something to put in its place.

Now it is more than probable that the notion of a “genetic programme” gains much of its apparent plausibility from the analogy with computer programmes. We do indeed live in a society where computers play an important and ever-increasing role; and even a passing acquaintance with computer science is sufficient to know that in this domain, the notion of a computer programme is perfectly valid and operational. In the next section, we will therefore take a closer look at the actual operation of computer programmes. We shall see that the distinction between differential aspects and invariant aspects, on which we have laid such emphasis, does indeed apply in full to the operations of computers.

Genetics and information

A leitmotif that ran through much of the previous section 4 was the necessity of clearly distinguishing between two situations that must not be confused: those where the phenomenon under consideration is invariant; and those where what is of interest is the differential variation that is occurring. This distinction is essential if we are to properly construe the key term of information. The term “information” is used in an improper sense when it is applied to the explanation of phenomena that are invariant. A prime case of this is when it is thought that regularities in natural morphogenetic processes have to be “in-formed” by an external source which in some sense “injects” the form into the process. By contrast, the proper use of the term applies to situations where it is a question of explaining differences; this is very nicely expressed by a formula due to Gregory Bateson: “an 'information' is a difference which makes a difference”. This is the sense in which it is perfectly correct to speak of “genetic information”, and it is this sense that I will take up when we get back to the question of a primordial genetic system prior to DNA. These are, however, areas where so much confusion reigns that it will be worth spending the time to look at the basic theories in question. In section 3.5.2, I will start by recalling the essential features of the classical Shannonian notion of “information”; and I will then recall the key features of classical Mendelian genetics, which do indeed accord with the fact that we are in a situation of differential variations. In each case, I will also emphasize the features are indeed invariant in these situations where the focus of attention is on the variation; these are the features which are “behind the scenes” so to speak (precisely because they are not variant), but which are nevertheless essential for the variational information to function.

Shannonian information

This theory is based upon the concept of a message which is produced by one agent, the sender, and transmitted to a second agent, the receiver. From a formal point of view, the message consists of a string of symbols taken from a finite alphabet (for example, A1, A2, A3, ... An). It is then possible to give a precise and quantitative definition of the amount of information contained in a message. The idea is basically as follows: the smaller the probability of the message, the greater the amount of information. This seems to fit in relatively well with one intuitive sense of the term “information.” If someone tells me: “It was drizzling this November morning in London” (an event whose a priori probability is very high), that person will not be bringing me a lot of information; but if they were to tell me that “There was bright sunshine and a temperature of 25°C”, I will be very surprised, and because of that the message will have brought me a lot of information. If we apply this to the symbols of a message, we can write, mathematically:

I = 1/pi

where I is the amount of information, and pi is the probability of symbol Ai. For a sequence of symbols A1, A2, A3, etc., if we consider that the probability of each symbol is independent of those that precede or follow it, we obtain:

I = (1/p1) × (1/p2) × (1/p3) ...

It is common practice to use logarithms, which transform multiplications into sums (meaning that the amount of information brought by an event that is 100% certain, whose probability is 1.0, will be equal to zero):

I = – Σ log (pi)

Since each Ai symbol appears with a frequency of pi, the efficiency of a telecommunication system is represented by the formula:

I = – Σ pi ⋅ log (pi)

This formula is of invaluable use to telecommunications engineers whose work consists, in great part, in optimizing the flow of information defined in this way. We can note that, in this context, an act of communication is composed of three parts: i) the emission of a signal; ii) the (physical) transportation of the message from the sender to the receiver; and iii) the reception of the signal. Now, in the (general) case where the emitted signal is not directly perceptible as such by the receiver, the signal must be encoded in a form that is adequate for its transmission at a distance, and the message then has to be decoded in order to retrieve the initial signal. For example, Morse code consists in encoding normal alphabet letters into short or long “beeps” (“S” is encoded as “⋅ ⋅ ⋅”, “O” is encoded as “¬– – –”, etc.), a form that is adequate for their transmission by radio ever since the rudimentary devices of the 19th century. For the telephone, the sound waves produced by a voice must be encoded (by a microphone) in the form of variations in the intensity of an electric current, which enables transmission of the message through conducting wires and its decoding through a loudspeaker. The same applies to radio or television broadcasts.

What should be noted is that all these operations are purely formal, i.e. they absolutely do not take into consideration any possible semantic aspect of the message in question. But then we have to ask: how can it be that “communication” thus defined still possesses a meaning, i.e. how can it still make sense?

The way in which “formal semantics” basically operates (as inherited from the research programme on the foundations of mathematics initiated by Hilbert) consists in setting up a system of term-to-term correspondences between the symbols, on one hand, and their “referents” on the other hand. Let us apply this scheme to the situation that interests us here, i.e. to communication by the transmission of information. The sender must have a way of categorizing the various situations of the world that makes sense to them; they must also have a way of encoding the situation in which they find themselves in the form of a string of symbols. Now, for the partners to be able to understand each other through their act of communication, the receiver must share the same way of categorizing the situations of the world, and they must also share the same way of encoding and decoding these categories into and from symbolic strings. The whole question, then, is this: by what kind of process can the partners come to share the same systems of categorization and coding?

This question is all the more daunting since in a normal situation, when communication works more or less well (i.e. when there is no blatant “mis-understanding”), the fact that semantic communication depends so completely upon this sharing of the same codes disappears from the partners' conscience. Their attention – and this is perfectly normal – is captured entirely by the variable aspect of their messages, which is indeed what really matters to them, since it is the variations which contain actual information. Precisely because the categorization and coding systems are invariant, they are automatically taken for granted, and receive no attention.

However, from a scientific point of view, this “blind spot” is of capital importance. When one asks the almost unthinkable question as to how this sharing the same codes came about, one reaction is to suggest that there must have been a communication process exactly on that subject. But upon further reflection, that turns out to be impossible. In order for “communication” (in the precise sense that we define here) to occur about a way of categorizing, there would have to be a shared meta-categorization... about ways of categorizing, and a corresponding coding system. But that, then, would require a shared meta-meta-categorization... The problem only gets worse and worse, escalating into a vicious infinite regression. It would be better to confront the problem at the start; but then the original question has to be asked quite bluntly: how could a system of shared categories and codes systems ever be set up without the use of an information transmission system that was already in place? In the case of human telecommunications, of course, the answer is quite straightforward: the shared systems are set up by telecommunications engineers, operating from outside the systems themselves, as a sort of “Deus ex machina”. But we can already anticipate that there will be a problem when we turn to living organisms, for in this case there is no “divine engineer” working from outside the system to set it up.


What becomes of these considerations when the nomadic notion of “information” is transposed from its field of origin – telecommunications engineering and theoretical cybernetics – and re-deployed in the field of genetics and biology? They remain completely relevant. First, at the level of the role of DNA genes in protein synthesis, we can see that Shannon's scheme works well. There is indeed a “genetic code” that specifies term-to-term correspondences between nucleotide sequences in DNA and amino-acid sequences in proteins. More precisely, a sequence of 3 nucleotides (referred to as a “codon”) specifies a particular amino-acid. As there are 4 nucleotides (A, T, G, and C), there are 4 × 4 × 4 = 64 different codons; this is somewhat more than what is necessary for coding the 22 amino-acids used in the metabolism of all terrestrial organisms, so that the “genetic code” is partially redundant.

Obviously, this code has to be invariant; and the question arises, therefore, as to how this code itself is established. In the case of our discussion, we do move up one level, since the “genetic code” depends on molecules called “transfer RNA” and t-RNA synthesis is itself (partly) coded. But that only transfers the problem another step up, since we then have to ask how this “meta code” itself was established. We arrive, inevitably, at a level where codes and meta-codes depend on structures whose organization is, and must be, invariant but whose invariance is not, and cannot be, “coded” in genes. And in our case, this “invariant organization” is that of a living cell, with its membrane, its metabolism and the precise relationships that hold between its components (including ribosomes and chromosomes). The question then arises, of course, as to how this “invariant organization” can be produced and maintained if it cannot be “coded” in genes. That is a very fundamental question, and we will return to it in the section 3.6 that discusses autopoïesis.

Mendelian genetics

The epistemology of genetics is based upon the classical Mendelian experiment shown diagrammatically in figure 9. Some preliminary remarks are called for here.

The first essential pre-requisite in setting up the experiment is to have available two “parental” lineages displaying a systematic difference between them when the individuals are raised in comparable conditions. This means that in the absence of differences, genetics is quite powerless. In other words, genetics has a “blind spot” for all invariant properties, i.e. properties that do not vary from one individual to another. The “difference” in question may be of any kind (size, weight, colour, shape, smell, etc.); it might also very well be a behavioural difference, or even a mental characteristic. The only necessary condition is that this difference must be reliably observable, thus allowing the experimenter to categorize any individual as belonging unmistakably to either one or the other parental lineage. In the diagram shown in figure 9, we assume that this difference is one of size, so that the individuals can be categorized as either “tall” (or "large") versus “short” (or "small"). The fact that formal genetics is a differential science is an absolutely central consideration: as already indicated, it is a leitmotiv which runs through all our consideration of genetic systems.

The second pre-requisite is that it must be possible to cross the individuals of each lineage with each other. It is therefore assumed that these are sexually reproducing organisms. This means that Mendelian genetics, in the strict sense, applies only to differences between individuals of one and the same species. As a corollary to these two pre-requisites, taken together, all individuals obtained through crossing within one or the other parental lineage for an indefinite number of generations (in practice, about twenty generations are sufficient) must reproduce the difference between the two lineages that allows for categorization of the individuals.

We can now proceed to the Mendelian experiment per se. This experiment consists, firstly, in crossing an individual of the P1 lineage with an individual of the P2 lineage so as to obtain individuals of the first hybrid generation (called the “F1” generation). The result (sometimes referred to as “Mendel's first law”) is that the individuals of generation F1 are uniform, and that they all resemble the individuals of either one, or the other parental lineage. In the figure below, we assume that the F1 individuals resemble those of the P1 lineage, i.e. that they are “tall.”

Secondly, we now cross the F1 individuals with each other. The result (“Mendel's second law”) is that in the F2 generation (the second hybrid generation), the individuals vary; they are (statistically) distributed in the ratio of one quarter of “short” individuals (resembling the P2 lineage), and three quarters of “tall” individuals (resembling the P1 lineage).

Individuals from two parental lineages, P1 and P2, are crossed in order to produce the F1 generation. F1 individuals are then crossed with each other to produce the F2 generation.

Based on these results, we can reason as follows.

1) The individuals of generation F1 must have received “something” from their P2 parent. This is deduced from the fact that when these F1 individuals are crossed with each other, they produce “short” descendants, i.e. individuals showing the P2 lineage type; whereas individuals from the P1 lineage crossed with each other never produce any “short” descendants. It is clear that F1 individuals must also have received “something” from their P1 parent – not only because they have “tall” descendants showing the P1 type, but also because they themselves resemble the P1 type.

2) It follows that a distinction has to be made between an individual's “phenotype” and its “genotype”. The “phenotype” corresponds here[16] to the individual's outside appearance – the fact of being “tall” or “short,” which allows the experimenter to categorize each individual as belonging either to type P1, or to type P2. The “genotype” refers to the individual's genetic constitution, which is revealed by its behavior in crosses. The individual's genotype is made up of these “somethings” that it receives from each of the two parents and transmits to its descendants. The distinction between phenotype and genotype is necessary, since F1 individuals have the same phenotype as their P1 parent (they are “tall”), but their genotype is obviously different (they have “short” descendants, while P1 individuals never have “short” descendants when crossed with each other). The importance of the two initial pre-requisites can now be appreciated: they are necessary to ensure that all P1 individuals, when crossed with each other, will always have “tall” descendants only. Without this, the whole argument would not go through.

We can now summarize the relations between genotype and phenotype in the form of a table:

Table 1




“Something” of the P1 type only



“Something” of the P2 type only



“Something” of the P1 type plus “something” of the P2 type



3) Let us now consider those F2 individuals that have the “short” phenotype. We can see, from the table, that their genotype is necessarily composed of “somethings” that are only of the P2 type. Otherwise, if their genotype contained “somethings of the P1 type” in addition to the “somethings of the P2 type”, they would have manifested the “tall” phenotype. It follows that each of their two F1 parents must have transmitted to them only “somethings of the P2 type”. This leads us to a conclusion of the utmost importance. These “somethings”, which can be either of the P1 type or of the P2 type, have a discrete nature: they are like “particles” that can emerge uncontaminated by cohabitation with the “somethings” of the other type (as was the case within the genotype of F1 individuals). This would not be the case, for example, if these “somethings” could blend, like water and ink; after mixing, pure water cannot be recovered. This phenomenon, which demonstrates the discrete nature of these “somethings” that constitute the genotype, is called segregation: it is the single most important phenomenon in the whole of Mendelian genetics. These discrete “somethings” that can be either of the P1 type, or of the P2 type, are the “Mendelian factors” that later became known as “genes.”

4) In order to proceed further, we will need to introduce some kind of notation. Circumlocutions such as “somethings of the P1 type only” and “somethings of the P2 type only”, although they may sound quite natural at an early stage, will soon become cumbersome if we start using them often – and as we shall see, this will indeed be the case. That is why geneticists, including Mendel himself, decided to replace the expression “somethings of the P1 type only” with a symbol (for example, “T”) and the expression “somethings of the P2 type only” by a related symbol (such as “t”). The reasons for such a choice are easy to guess. The “T” symbol is a reminder that it is associated with the “Tall” phenotype, while the “t” symbol reminds us that in the T + t combination, the phenotype will always remain “Tall” (the t “something” is said to be “recessive” while the T “something” is “dominant”). Nevertheless, the choice of this system of notation had extremely serious consequences later on, since it suggests that the “T-something” is itself Tall, or that a “something of the P1 type” carries in itself, encapsulated like a homunculus, the “Tall” characteristic. But the fundamentally differential nature of Mendelian genetics means that this is absolutely unjustified. The fact that, all other things being equal, a difference between “something of the P1 type” and “something of the P2 type” may cause a difference between a “tall” phenotype and a “short” phenotype in no way entails that these “somethings” themselves are “tall” or “short[17]”. This disastrous confusion lies at the root of the unwarranted hegemony of genetics, and continues to nourish it. It has led to such unfortunate expressions as “the gene for schizophrenia” or “a gene for intelligence” (see pp. XX-XX). The need for a convenient notation is, however, real enough. In what follows, I will suggest a compromise: instead of “T” or “t”, I will use symbols such as “T/” or “/T”, “t/” or “/t”. The “/” sign is there to serve as a reminder that these “somethings” – in other words, the Mendelian factors or genes – do not carry within themselves the actual phenotypic character.

5) After this parenthesis on notation, let us take up again the reasoning based on the Mendelian experiment. So far, the argument has been quasi-deductive; but now, in order to take the next steps, we need to introduce some hypotheses. The first question is: how many “T/”or “t/” factors are needed to make up a genotype; and how many factors does an individual receive from each of its parents? We shall examine the simplest hypothesis possible: an individual receives a single factor from each parent, and the genotype is therefore made up of two factors. When an individual becomes in turn a parent, he (or she) will transmit one of its two factors to each of its descendants. The second question is: how is the choice of the factor to be transmitted determined? Our hypothesis here will be that the choice is random: the paternal factor is transmitted with a probability of 50%, and the other (maternal) factor is also transmitted with a probability of 50%.

It should be noted that since each individual has only two parents, but can have an unlimited number of descendants, these hypotheses imply that a gene may be copied many times so as to generate an unlimited number of copies[18].

We can now consider table 1 again, in a more elegant form:

Table 2







T//t (=t//T)


These two hypotheses are sufficient to generate quantitative predictions, first of all concerning the F2 generation. Let us create the following table (the resulting genotypes are indicated in bold type) :

Table 3

Factor transmitted by the first F1 parent

Factor transmitted by the second F1 parent

50% t/

50% T/


25% T//T

25% t//T


25% T//t

25% t//t

Using Table 2 as a reference, one can see that the t//t genotype corresponds to the “Short” phenotype, that the T//t and t//T genotypes (which are identical) correspond to the “Tall” phenotype, and that the T//T genotype also corresponds to the “Tall” phenotype. The prediction is, therefore, that in generation F2, there will be 25% + 25% + 25% = 75% “Tall” individuals and 25% “Short” individuals. This prediction is in agreement with the results of figure 9.

6) So far our theory and hypotheses have only given us back the initial observations that served to create it; this is of course the least they should do. But, as Popper has pointed out, in addition to accounting for the initial observations, a scientific hypothesis should always provide novel predictions which make it possible to test the hypothesis by exposing it to possible refutation. Mendel carried out a large number of subsequent experiments of this type. For instance, he crossed F2 individuals among themselves so as to produce generations F3, F4 etc. It should be explained here that Mendel was working on peas, which are hermaphrodite organisms so that an individual can be crossed with itself. Under such conditions, the predictions are not too difficult to make. By looking at Table 3, one will see that if those 25% of individuals that have the t//t genotype (and, therefore, the “short” phenotype) are crossed with themselves, they will have 100% of t//t descendants with the “short” phenotype. This will, of course, perpetuate itself in generations F4, F5, etc... We have thus returned, so to speak, to the P2 parental lineage – which illustrates once again the discrete nature of the t/ factors, which have in no way been contaminated by their passage through generation F1, where they were side by side with T/ type factors within a single genotype. What happens, now, if we cross with themselves F2 individuals that have the “tall” phenotype? Table 3 once again tells us that 25% out of 75%, i.e. one third of these individuals, will have the T//T genotype; the prediction is, therefore, that 100% of their descendants will have the T//T genotype and the “tall” phenotype (we have come back, so to speak, to the P1 parental lineage, and we can see that the T/ factors have not been contaminated either after having found themselves side by side with t/ factors). 50% out of 75%, i.e. the remaining two-thirds of “tall” individuals in generation F2, will have the T//t genotype (like the F1 individuals): if crossed with themselves, they will therefore have a variety of descendants, with 75% being “tall” and 25% being “short”. These predictions – which are quite precise and detailed, and far from trivial – were verified extensively by Mendel. We will leave it to the reader to imagine other types of crosses, and to make the corresponding predictions; this exercise is highly recommended to those who wish to fully grasp the Mendelian scheme.

We will now end this section with a particular kind of cross that will be much used in the rest of this chapter. It is the cross of an F1 individual with an individual from the P2 parental lineage (this is called a “backcross”). The table that serves as a basis for making predictions is as follows (resulting genotypes again in bold type, as in table 3:

Table 4

Factor transmitted by the T//t F1 parent

Factor transmitted by the t//t P2 parent

50% T/

50% t/

100% /t

50% T//t

50% t//t

Since T//t individuals have a “tall” phenotype and t//t individuals have a “short” phenotype, the prediction is that the individuals produced from a backcross will be 50% “tall” and 50% “short.” This prediction was also verified by Mendel.

7) Today, the term “phenotype” is widely used to denote any measurable trait in an organism. However, one can never insist enough on the fact that such usage does not conform to the conceptual framework of formal genetics as it was initiated by Mendel. In that framework, a measurable trait is truly a “phenotype” if and only if: a) it exhibits a categorical difference, and b) if this difference behaves in “Mendelian” fashion in breeding experiments (see figure 9 and the other breeding experiments illustrated in point 6 above). Only under such conditions can we establish table 2, where the relations between phenotype and genotype are specified. In other words, the two key concepts of “phenotype” and “genotype” form an inseparable pair; a “phenotype” is defined by the fact that it makes it possible to follow the segregation of Mendelian factors. In order to stress this essential point, I propose the neologism “Mendelian phenotype” to refer to a phenotype so defined. This will allow us to say that most “phenotypes”, in the loose sense of any measurable trait, are not “Mendelian phenotypes.” Mendel discovered a certain number of “Mendelian phenotypes” among his peas[19] ; he did not say much about his failures with “non-Mendelian” phenotypes, although there must have been some, particularly when he was trying to generalize his results and extend them to other species. But science would have gained nothing – on the contrary, it would have suffered a great loss – if Mendel had obscured his great discovery concerning Mendelian phenotypes in an ocean of uninterpretable findings.

With this, we come to the end of the exposition that is relevant for our present purposes. The rest of the story concerns “linkage groups” – the fact that when two Mendelian characters are inherited together, the segregation of the two genes is not always independent; the correlation between these “linkage groups” and the observable structures in the cell nucleus that are called “chromosomes”, which leads to the conclusion that the genes are physically and materially situated in the chromosomes; the biochemical analysis of the chromosomes, and the identification of the genes with the DNA component; then Watson and Crick's discovery of the molecular structure of DNA, and the discovery of the “genetic code”. But for our present purposes, which concern genetic systems in general and not just the system which has come to be universal in all contemporary living organisms (i.e. the DNA system), these details are not directly relevant. The general point we want to take away from this consideration of Mendelian genetics, is the differential nature of genetic information.

Genetics is not a science of heredity

The critical analysis in the preceding sections consolidates a number of conclusions, both positive and negative.

From the positive point of view, it confirms :

  • that genetics is clearly differential in nature;

  • that this framework of thought – not only formal genetics itself, but including its extensions into molecular biology, and its affinities with cybernetics and the theory of information – is remarkably consistent.

On the negative side, this analysis confirms that what is lacking in genetics, and in all its extensions, is everything that is related to those aspects of the system which are invariant – and which must remain invariant in order for the information coded in the genes to exist and make sense. It might be useful to summarize this important point in a simple formula: contrary to what is generally assumed, genetics is not a science of heredity. Quite specifically, we may note that this is in flat contradiction with the concluding sentence in the quotation from Jacob that we have taken to introduce this whole section: “The organism thus becomes the realization of a programme prescribed by heredity”.

It is true that genetics is perfectly appropriate for studying the transmission from generation to generation of phenotypic differences that can be encoded in genetic differences. And it is also true that genetics is perfectly relevant to the Darwinian theory of evolution, where the focus is placed on natural selection and, in the last analysis, on differential reproduction – which is why genetics is so easy to integrate into the neo-Darwinian frame of thought. But, on a more fundamental level, the biological phenomenon of heredity is first and foremost about the reproduction of identical organisms, not of differences. As we mentioned right at the beginning of this chapter, no living organism currently existing on the planet Earth is the product of spontaneous generation from non-organic matter. The great rule is that life originates from life; each living organism has its origin in the reproduction of its ancestors, who were also living organisms. A pair of cats produces offspring in the form of kittens that will develop into adult cats that resemble their parents. As Pasteur so painstakingly demonstrated, the same rule applies even for microbes: no microbe is ever generated spontaneously, but always through the reproduction of other living microbes that lived before it.

It is important to underline that this observation – the fact that life is generated by life – is not at all an endpoint, but rather a starting point. It sends us back to this question: what is this entity called life such that it is able to generate life? As we will see in the next section, the key to this enigma resides in the fact that, even before we can talk about its ability to reproduce anything from one generation to the next, every living organism is a process that generates itself from one instant to the next. Once the question of informational differences – which is really the preferred domain of genetics – is set aside, and the question of identical reproduction is asked, the helplessness of “genetics” becomes evident. What, indeed, can contemporary biology tell us in answer to this question of identical re-production? The conventional answer comes in two parts: (i) genes “reproduce themselves” identically; and (ii) the “genetic programme” contained inside the genes generates the organism. However, in particular in Do genes produce organisms? to Ontogenesis above, we have already shown that “answers” of this sort are at best vacuous, and are in fact deeply mistaken. We will now make a new start on the question of how it is that living organisms really do generate themselves.

Autopoïesis and biological individuation


What is life such that it is able to generate life? As Maturana (Maturana & Varela, 1980[*]) recounts it, from his childhood onwards he had asked himself over and over again the same question: “What is the essential characteristic of living organisms? What kind of systems are living systems that they may die?” The usual approach to this type of question consists in starting from a common-sense definition – i.e. to consider that, after all, we already know enough about what a living organism is, at least enough to be able to tell without hesitation that a dog is a living thing while a stone is not – and to examine “empirically” the properties that are common to all entities categorized as “living” in this way. But that approach is not sufficient. Maturana recalls how, at a certain stage of his quest (and particularly when trying to answer questions from his students) he was forced to accept that one could recognize living systems when one encountered them, but without being able to say what they were:

“I could enumerate features of living systems such as reproduction, heredity, growth, irritability, and so on; but how long a list was necessary? When would the list be completed? In order to know when the list was completed, I had to know what a living system was, which was, in fact, the question that I wanted to answer in the first place by producing such a list. I could speak about adaptation and evolution, development and differentiation, and show how all these phenomena were tied together by the phenomenon of natural selection; but the question: “What was the invariant feature of living systems around which natural selection operated?” remained unanswered. Every approach that I could attempt and that I did attempt left me at the starting point.”

It almost sounds like Alice Through the Looking Glass, when Alice keeps trying to reach the top of the hill, but just as she thinks she is getting there the path takes a sudden twist and turn and she finds herself walking back into the house every time! We might add that not only does the path of the “list” not lead to any solution, but the same problem arises when we try to look closer at any single item on the list. Let us take, for example, the first characteristic on Maturana's list (which is on many other people's list too): reproduction. First objection: mules, for example, do not reproduce; does that mean they are not living animals? But this objection, after all, is not very serious; it could be “the exception that confirms the rule,” and it is indeed true that living organisms that do not reproduce are exceptions. Much more profound is the same objection as the one that invalidates the list-based approach: unless one already knows what a living organism is, the fact that it is an entity that “reproduces itself” does not inform us any better. For example, under certain conditions, crystals – and also, as we know today after the mad cow disease epidemic, prions – do “reproduce themselves”; is that sufficient for them to qualify as “living” beings?

After asking himself these questions over and over again, Maturana realized that he had to change his approach radically. However, unlike Alice, the answer did not come immediately. It was only gradually that he came to think that living systems had to be characterized not by reference to their environment or their context, but with relation to themselves, as autonomous entities. In 1969, he wrote for the first time that living systems were constituted as entities by the circularity of the production processes of their own components. Once the idea is stated explicitly, it does seem intuitively obvious. If one asks what produces a living organism, clearly it is... the organism itself. Whether in an animal, a plant, or a micro-organism, tissues and organs are the result of an ongoing dynamic process of production; the molecules that compose an organism are continually renewed by the metabolism of the organism. And that is true only of living organisms. For example, a machine manufactured by human beings (even a machine tool, or a whole factory) produces something other than itself; and it is also produced by something other than itself. This “self-referential” circularity, therefore, does seem to be an essential characteristic of living organisms. Maturana, in collaboration with Francisco Varela, looked for a more adequate formulation of the concept of “circular organization” and coined the term “autopoïesis,” from the Greek autos (self) and poiein (to produce). The canonical definition is as follows (Maturana & Varela, 1980[*]):

“An autopoïetic machine is a machine organized (defined as a unity) as a network of processes of production (transformation and destruction) of components that produces the components that i) through their interactions and transformations continuously regenerate and realize the network of processes (relations) that produce them; and b) constitute it (the machine) as a concrete unity in the space in which they (the components) exist by specifying the topological domain of its realization as such a network. It follows that an autopoïetic machine continuously generates and specifies its own organization through its operation as a system of production of its own components, and does this in an endless turnover of components under conditions of continuous perturbations and compensation of perturbations. Therefore, an autopoïetic machine is a homeostatic (or rather a relations-static) system which has its own organization (defining network of relations) as the fundamental variable which it maintains constant.”

We have reached the heart of our subject: a living organism is a system whose fundamental invariant is its own organization. It may be worth remarking here that the “circularity” and “feedback” involved here is not simply that of negative-feedback dynamic systems; we are concerned not just with “self-regulation”, but more fundamentally with self-production.

The tessellation automaton

The notion of autopoïesis is so fundamental that it may be helpful to illustrate it with an example in order to make it less abstract. Varela has proposed what he considered as a “minimal model” of autopoïesis, which can be simulated on a computer. We are presenting it here in a slightly modified form (Bourgine & Stewart, 20[*]04) that avoids certain complications attached to the original version (McMullin & Varela, 1997)[*].

This simple “tesselation automaton[20]” is represented in figure 10. It is defined as follows:

The enlarged inset shows the processes taking place in a thin layer just below the membrane: the production of B components through catalysis, and a B component entering the membrane and becoming a C component. The C → D reaction corresponds to the disintegration of a C component, which leaves a hole in the membrane. B components are normally confined inside the membrane, but can escape through holes.

  • The automaton has a membrane, M. This membrane is closed upon itself, so that it defines an intra-cellular space. It is formed by C components, which self-assemble to form a bi-dimensional surface enclosing a volume (the automaton under consideration here is tri-dimensional, shown in a two-dimensional section in the Figure).

  • The C components located in the membrane disintegrate spontaneously, forming a D product: C → D. In mathematical terms, the speed of this reaction is dependent on a parameter kp that corresponds to the rate of spontaneous disintegration per unit of membrane surface. The D product does not integrate itself into the membrane: it escapes into the extra-cellular environment, leaving behind it a hole in the membrane (or, if the C component that just disintegrated was already on the edge of a hole, the hole becomes larger).

  • B components are produced by a reaction between two molecules of the substrate A: A+A → B. This reaction is catalyzed by the inner surface of the membrane. In mathematical terms, the speed of this reaction is dependent on a parameter ks that corresponds to the efficiency of the catalysis.

  • The substrate A is freely available in the outside medium, at a fixed concentration of a0. A diffuses freely across the membrane. As the concentration of A inside the membrane is decreased by the chemical reaction A+A → B, there is a net flow of A towards the inside.

  • The membrane is impermeable to B components, which therefore accumulate in the intracellular space. Two B components cannot combine together; however, if a B component meets the edge of a hole in the membrane, it binds to the surface and becomes a C component that repairs the hole. The repair may be complete if the hole was caused by the disintegration of a single C component; or, it may be partial if the hole was larger. However, if the hole is larger than a single C component, there exists a finite probability that the B component may pass through the hole without binding to the edges. This probability increases with the size of the hole. If a B component passes through a hole, it is lost to the outside medium.

The behavior of this type of automaton can be simulated directly, as has been done in particular by Varela (1989)[*] and Mc Mullin (1997)[*]. It can also be described mathematically[21], and the relationships between the three parameters kp, ks and a0 that allow for equilibrium can be translated into an expression. In qualitative terms, it is fairly easy to understand that in order for a dynamic equilibrium to exist, the parameters ks and a0 (which control the rate of hole repair) need to be great enough relatively to kp (which defines the rate of hole formation). If this is the case, the automaton can continue to “live” indefinitely, and even to survive external disruptions that lead to the formation of additional holes. On the other hand, if ks and a0 are not sufficiently large compared to kp, the dynamics of the automaton can no longer compensate for the formation of holes. What follows is easy to imagine. Some holes are no longer repaired in time; this accelerates the loss of B components, which makes hole-repair even more difficult, so that the holes become even greater. The system accelerates on its way to complete disintegration as an autopoïetic system. In other words, it “dies.”

The fundamental role of the membrane should be underlined: it confines the interactions that take place within the intracellular space. These interactions continually renew the components and the network of processes that produces them, which is in line with the definition of autopoïesis given on 3.6.1. Above all, the confinement leads to an accumulation of B components, ensuring repair of the membrane itself and hence maintaining the vital property of semi-permeability to A, but not to B. In other words, the system's organization is defined by the circular relationship that exists between the membrane and the system's “metabolism.”

This circularity is what gives rise to two well-distinguished domains in the phase-space of the autopoïetic system's operation. Above the condition of equilibrium, the system maintains itself and can even grow. Below this condition, the system enters a loop of positive feedback and accelerates its course towards total destruction[22] (figure 11).

Thus, the circularity of the process's organization accentuates the separation between the two situations and gives way to two modes of operation that are qualitatively different. Like the candle flame, the system is either “alive,” or “dead.” If it is alive, it may “flicker” at times (several holes may form), but it is able to recover and to continue its existence as if nothing had happened. If the system is dead, on the other hand, nothing can bring it back to life; it completely collapses and disintegrates[23].

It should be noted that the formulation of the tessellation automaton given here focuses on the transformations of the material A components, the free B molecules, the C molecules that integrate themselves into the membrane, and the end-product D. We have, therefore, concerned ourselves primarily with flows of matter. However, since we are considering that all of the sequential processes (A+A → B, free B molecules integrating themselves into the membrane B → C, and the disintegration of molecules C → D) take place spontaneously, the overall result of A+A → D also corresponds to a process of energy dissipation. Therefore, an autopoïetic system, of which the tessellation automaton is only a prototypical and minimal example, is not, strictly speaking, just a “thing.” It is rather a pure process, a flow of both energy and matter. And it is indeed as a process that an autopoïetic system has self-organizing properties and does not need to be “in-formed” by external information (see 3.6.1). An autopoïetic system becomes a “thing” again, composed of relatively inert matter, only when it dies.

Conclusions on autopoïesis

What is invariant in any living organism is its autopoïetic organization. That is the absolutely essential and fundamental property that defines what life is. The distinctive property of an autopoïetic system consists, precisely, in neither more nor less than dynamically maintaining its own organization as invariant. It is, in a certain sense, a meta-invariant that encompasses all other invariants.

Since autopoïesis is an invariant – the most fundamental invariant of all – it follows that this is an aspect of living organisms that genetics does not, and cannot, apprehend. So, after all, it is actually quite logical that molecular biology (derived from the neo-Darwinian synthesis, which itself originated from Mendelian genetics) would contend that “life does not exist[24].” But, clearly, we cannot accept such a statement unless we admit that genetics and biology have been completely and irrevocably separated from each other.

We opened this section on autopoïesis by asking again the question: “What is this thing called life that is able to generate life?” The theory of autopoïesis suggests that even before they reproduce themselves, living organisms are basically pure processes that generate themselves from instant to instant. This fundamental property has two important corollaries. The first is that living organisms have the property of being able to repair themselves. As illustrated by the example of the tessellation automaton, the process of self-repair (in this case, the self-repair of holes in the membrane) results from the process of autopoïesis itself.

The second corollary is that autopoïetic systems have the ability to truly reproduce themselves (and not merely to be copied, like genes). An autopoïetic cell only needs to grow (which is a process of exactly the same type as its ongoing self-production) until it doubles in volume, and then divide itself in two. Of course, this division itself must be organized in such a way that each of the new entities preserves the same autopoïetic organization as its parent. But given this condition (which should not be too difficult to satisfy, whether in the model automaton or in reality), the end result is that there are now two autopoïetic cells where there was only one before; in other words, there is reproduction. Thus, identifying life with an autopoïetic organization does indeed explain how “life is able to generate life.”

Biological individuation

The concept of autopoïesis is absolutely fundamental to our enquiry concerning the nature of life. As we have just seen, it makes it possible to understand how it is that “life is able to generate life”; and even how, under suitable conditions, living organisms are able to repair themselves. But in accordance with our motto “qui aime bien, châtie bien”, these virtues do not blind us to the fact that this concept does have certain limitations and weaknesses[25]. These weaknesses do not disqualify the concept of autopoïesis, because it is quite possible to correct them by adding additional developments; but precisely in order to contribute these additional elements, it is necessary to lucidly identify the weaknesses in question. The weakness we will address here concerns the issue of spontaneous generation, which is a logical necessity if we are to account scientifically for the origin of life. Now on the basis of the simulation studies, in particular the “tessellation automaton” shown in Figure 11, it is quite clear that the circular organisation which is the defining feature of autopoïesis could never arise spontaneously de novo. The topologically circular membrane (or in 3-D, the spherical membrane), while it is continually dynamically re-produced, and while it is even possible for small “holes” to be repaired, would never arise spontaneously if it were not “primed” by a complete membrane.

What, then, are the sorts of processes that could plausibly arise by spontaneous generation? In their book entitled The New Alliance, Prigogine and Stenghers (1979[*]) introduce the concept of dissipative structures: these are dynamic structures that are indeed produced spontaneously, through symmetry-breaking, in a wide variety of systems with the common feature that they are far from thermodynamic equilibrium. Such structures already exist in the non-organic world. Whirlpools may serve as a simple prototype. Artificial examples include Bénard convection cells (figure 12), or candle flames. Among natural phenomena, cyclones are a good example. These examples all involve convection currents, i.e. mass displacement of fluid matter, which aids their visualisation and hence an intuitive understanding of their functioning.

However, dissipative structures also exist in the purely chemical realm, notably in the form of so-called “reaction-diffusion systems”. Two examples, taken from the work of McGregor & Virgo (2011[*]), are shown in figure 13 a and b.

What is common to all these examples is that they exhibit a phenomenon that Simondon (1989)[*] has called “individuation”. These dissipative structures typically exhibit an ability to resist perturbations: for example, drafts can make a candle flame flicker; but as long as it does not blow out, the flame then resumes the course of its dynamic existence as if nothing had happened. Moreover, each of these entities – from hurricanes to reaction-diffusion spots – is formed as a spatially distinct individual, separate from other entities of the same sort.

This can be compared to the definition of autopoïesis by Maturana & Varela (1980[*]) as a network of processes that constitute a unity. However, as Simondon also pointed out, entities that demonstrate this sort of physico-chemical individuation are intrinsically ephemeral: they last only as long as certain external conditions, over which they exert no control at all, just happen to be maintained. Thus, Bénard cells disappear as soon as the liquid enclosed between the glass plates is no longer heated from underneath. In the same way, cyclones can last for weeks; this illustrates very well their capacity for individuation, to such an extent that they are sometimes given names. But as they drift uncontrollably, passively driven by the prevailing winds, they eventually move to areas where the fairly precise climatic conditions that are necessary for them to “function” – in particular, conditions of humidity and temperature gradients – are no longer fulfilled; whereupon they disappear. There is nothing in the functioning of Bénard cells, or cyclones, which even tends to maintain the external conditions in the range necessary for their continued existence as dissipative structures.

This characterization of physico-chemical individuation provides a backdrop that leads on to a definition of what Simondon has called “biological individuation” – in other words, living organisms. From this point of view, the essential property of living organisms is that their own functioning does exert systematic control over their own boundary conditions, in such a way that these conditions remain within the limits that are necessary for their individuation to continue. Living organisms can always die from one moment to the next, since they always remain vulnerable to a serious and exceptional disruption of the outside environment that can interrupt their individuation irrevocably (just as the candle flame, too, can be blown out instantaneously by a strong draft). This is, in fact, a necessary condition for defining “the living”; for if there were an entity (or a process) whose intrinsic nature were such that it could never die whatever happened, it would be meaningless to say that it is, at any given time, “alive” (Maturana & Varela (1980[*]). At the same time, however, it follows from our definition that any living organism is potentially immortal: its activity, which consists in maintaining its own boundary conditions within the limits that are necessary for its individuation could, in principle, continue indefinitely[26].

An important question is of course to identify when and how the transition from forms of “physico-chemical individuation” that could plausibly have arisen by spontaneous generation, to the first forms of “biological individuation” as just defined, actually occurred in the course of the natural history of Life on the planet Earth. Even before seeking to garner relevant empirical evidence, there is still some more conceptual work to be done. Did this transition involve the creation of a cell membrane, enclosing and catalysing the (bio)-chemical metabolism as in the schematic tessellation automaton? And did it involve the emergence of a rudimentary form of “cognition”, in the sense of a minimal sensori-motor system enabling these proto-organisms to actively remain in areas where the environmental conditions are adequate for the maintenance of their existence as dissipative structures? As a working hypothesis (whose epistemological function is of course to stimulate attempts to refute it!), we propose to consider that the answer to both of these questions is “yes”. Accepting this (even if only provisionally), the stage is now set for us to return to the central question of this chapter: the origin of a genetic system.

A primordial genetic system

Recapitulation of basic considerations: invariance and variance

We are postulating – as a working hypothesis – that in the early stages of the natural history of Life on Earth, there were first of all some dissipative structures which exhibited “physico-chemical individuation”. These dynamic entities were indeed able to arise by spontaneous generation. Then, in a second stage, there was the emergence of forms which exhibited “biological individuation”; these forms were probably not able to arise directly by spontaneous generation; but we are postulating that the transition(s) from physico-chemical individuation to biological individuation occurred without the assistance of an identifiable “genetic system” of any sort. Terminologically, it will be convenient to call entities of this sort “proto-organisms”. Accepting this, at least for the moment, let us take stock of the situation in order to consider the origin of a primordial genetic system.

The first point to note is that for any class of “proto-organisms” of this sort, there is an absolutely fundamental feature that is radically invariant: that feature is neither more nor less than their very existence. This means that whatever the set of requirements that we shall specify for a primordial genetic system (we will come to that), the very first “genes” did not have to take care of producing the organism that housed them.

The second point is that although certain features of the class of proto-organisms as a whole are invariant, the individual organisms are not strictly identical in all respects. In other words, there is a range of variation which can and does exist between the individuals. Since, in addition to the very existence of the organisms, this variation also existed before the introduction of the genetic system, this means that the very first “genes” did not have to specify or to produce the phenotypic variations that they were going to code for.

At this stage, one may be tempted to ask: well, if these first “genes” did not have to this, and did not have to do that, then what did they have to do? The answer is, that they had to be able to code for certain phenotypic differences between the individuals in the population, including functionally significant differences. We can deduce a first requirement: these primordial “genes” had to exhibit a certain range of structural variation. Some additional requirements follow on from this. The second requirement is that these “genes” had to be such that when they were copied (they did not – necessarily – have to be able to “copy” or to “re-produce” themselves), the “copies” could be reasonably (if not absolutely) faithful to the originals, so that the genetic variations could be re-produced. And thirdly, it had to be possible to set up a reasonably systematic coding relationship between the genetic differences (i.e. the structural differences in the genes themselves), and the phenotypic differences between the individual organisms.

As this exercise of setting out explicitly the requirements for a primordial genetic system reveals, the institution of a population of organisms endowed with a genetic system also involves some additional requirements on the proto-organisms themselves, over and above the basic feature of being endowed with a capacity for biological individuation as envisaged above. These organisms must be such that they exhibit a range of functionally significant phenotypic variation that can then be coded for; and moreover, the organisms must be capable of setting up a coding relationship between these phenotypic differences and the structural differences in the putative genes. In other words, the requirements for the institution of a genetic system that we are working to elucidate do not concern solely the “genes” themselves, not the autopoïetic proto-organisms themselves, but the relationship between the primordial genetic system and the organism that is to house it.

In order to make any further progress in attempting to identify this “primordial genetic system”, we will therefore have to start from both ends – from the side of the proto-organism, and from the side of the primordial genes – bearing in mind that aim is to meet up somewhere in the middle. The time has now come to take a look at the literature, re-interpreting it in accordance with these considerations concerning the putative identification of a primitive genetic system.

Genes first?

In 3.1, we argued that the very first “genes” could not have been bio-chemically akin to the DNA genes of virtually all contemporary organisms. In 3.2, invoking the “scaffolding” principle, we argued that the first “genes” need not have been made of DNA. Putting these two arguments together, it is fair to conclude that the first “genes” were not made of DNA. But this raises an obvious question: if not DNA, what then were they made of?

In order to rehearse our conceptual scheme, we will get started by examining whether DNA, if the synthetic metabolism of primitive organisms had been up to the task of producing it, would fit the bill for the requirements for a genetic system that we have just summarized in Recapitulation of basic considerations: invariance and variance. Firstly, DNA does indeed exhibit a range of structural variation: the well-known sequence of the four nucleotides A, T, G and C. Secondly, DNA does have the property that when it is copied, the copies can be essentially identical to the originals: again, it is well-known that the double-strand structure, with complementarity relations A-T and G-C, allows precisely for this. And thirdly, there is indeed a coding relationship between this structural variation, and functionally significant phenotypic differences: this is the well-known relationship between nucleotide triplets and specific amino-acids, so that a gene codes for a protein. Thus, DNA would indeed fit the bill perfectly – except for the fact, already noted, that the actual synthesis of DNA, and its translation into protein, is metabolically so complex that it could not have arisen in the first stages of the origin of life. So we come back to our question: what could the first “genes” have been made of?

To my knowledge, there is only one author in the literature who has made out a serious case for the existence of primordial genes made of a substance other than DNA. Cairns-Smith (1982 and[*] 1985[*]) has put forward the highly original and commendably audacious hypothesis that the first genes were a certain type of clay crystals. He argues, quite rightly in my view, that there can be no accumulation of appropriate accidents, no kind of progress in biological evolution, without the means to remember. He makes an analogy with throwing dice. What are the chances of throwing 140 sixes in a row? (140 is the number of unit operations that have to be appropriately sequenced in order to produce primed nucleotides). The answer is 6140, which is about 10109: this is the number of trials that would have to be made to have a reasonable chance of hitting on the one outcome that represents success. Throwing one dice a second for the entire period of the Earth's history would only allow about 1015 trials; so one would need 1094 dice, which is far more than the number of electrons in the observed Universe. The impossibility is blatant. However, things change radically if there is memory of success and failure in the past, so that success can be built on. In terms of the dice-throwing analogy, this corresponds to continue throwing at each step until a six comes up, and then going on to the next step. That way 140 sixes could be reached with one dice in about 140 x 6 seconds, or about a quarter of an hour. Hence the need for a primordial genetic system, in order for the evolutionary appearance of DNA to be possible. It is in this context that he develops the principle of “scaffolding” presented in section 2 above. This is the idea expressed in the phrase “genetic takeover”: primary organisms were displaced by secondary organisms of a kind that would have been quite unable to generate spontaneously.

In seeking to identify the material nature of these primordial genes, Cairns-Smith focuses his attention on crystals – more precisely, on the process of continuous crystallisation. This in turn leads him to examine clay minerals, for which the whole Earth is a massive, continuous crystalliser. Considered as dissipative structures (Simondon himself explicitly remarks that the surfaces of growing crystals are indeed dissipative structures), the energy source for this process is the water cycle powered by the Sun. However, a crystal gene cannot be just any kind of crystal: it has to have reproducible irregularities in order to be able to hold information. In order for shape and size to be a replicable property of a crystal, it is hard to see how an arbitrary three-dimensional shape and size could be replicated; but a two-dimensional shape and size – a cross-section – might very well be. All that is required is that the crystals should grow in one direction only and break up (only) across that direction. In this way, the first two requirements on our “shopping list” for can be met: these “crystal genes” do indeed exhibit structural variations, and these variations can be replicated. What now about the third requirement: that this variation should code for meaningful phenotypic differences?

Cairns-Smith proposal is that the “phenotype” of these primordial genes was neither more nor less than the growth and replication of these crystal genes themselves. In other words, the first organisms were no more than naked genes. Thus, these primary crystal genes evolved by direct action to begin with. However, Cairns-Smith goes on to suggest that they may have moved towards a more indirect control of their “phenotypes”. He writes:

“Different kinds of clay mineral crystals may grow in collaboration, one kind affecting the conditions for others. This may happen ... indirectly through one clay altering general conditions, such as flow rate or local acidity, which then favours the formation of other clays that might not otherwise have formed. In so far as replicable features (e.g. shapes, sizes, surface patterns) of crystal gene clays could affect the formation of other non-gene clays, and in so far as the non-gene clays might be helpful to the gene clays, then the non-gene clays would be properly described as phenotypes of the gene-clays. Picture these now somewhat evolved organisms as consisting of masses of crystal genes embedded in a watery matrix of other clay or clay-like material. It is not difficult to imagine uses for such a matrix material. It might provide mechanical protection against damage (growing crystal genes must break, if you remember, but only in the right way); or it might act as a glue, holding the genes in the right place. Or the matrix might provide protection against the effects of fluctuations in concentrations of nutrient solution: if the helper materials were to grow and dissolve more quickly than the gene materials, then the helper materials would have such a stabilising effect on the waters in their surroundings. Or again the matrix material might tend to hold on metal ions that would interfere with the growth of the crystal genes...” (Cairns-Smith A.G. 1985[*], pp. 105-106).

This is quite clearly, and deliberately, a “genes-first” scenario. Cairns-Smith discusses in a rather cursory fashion the subsequent evolutionary incorporation of organic molecules into the “phenotype”. This probably started with small simple organic molecules such as formic acid and the amino acids; then gradually complexifying to nucleotides; then macromolecules such as polysaccharides and polynucleotides (initially with a purely phenotypic role); until finally the stage is reached when nucleic acids, which are already there, come to “take over” the role of genes. After this, the clay-crystal genes, having played their role as scaffolding, would disappear completely. Throughout, this scenario is characterised by its gene-centred approach. In this respect it is reminiscent of Dawkin's “Selfish Gene” concept, according to which “the organism is merely the genes' way of making more genes”. In Cairns-Smith's account, the phenotype (the organism, the metabolic system) is considered principally in terms of its effects on the genes and their reproduction.

As a final note at this point, it may be worth pointing out that “genes” (here, putative primordial genes in the form of a certain sort of crystals) are maybe not as opposed to metabolic processes as it may seem. In Biological individuation, we discussed the processes of dissipative structures and physico-chemical individuation. Now it is interesting to note that Simondon (1989[*]), the author of the concept of “individuation”, considered that crystals – or more precisely, the surfaces of growing crystals – are also instances of “individuation”. In this respect, it follows that if the primordial genes were (as Cairns-Smith suggests) a certain sort of crystals, they were actually a type of “dissipative structures” – not so much “things” as a certain sort of process, coupled to the water-cycle powered by the Sun. In this sense, such “genes” would not really be entirely “naked”, because they would already be coupled to a sort of “metabolism” capable of re-producing them. The key question is whether such a start would provide an adequate basis for subsequent complexification, which would have to lead to the key stages of cellularization and the emergence of a metabolism capable of producing nucleic acids and proteins. The existence of a primordial genetic system, right from the start, is of course a trump card in this respect. However, in spite of all Cairns-Smith's persuasive arguments to the contrary, crystals are not usually propitious for complexification. In order to progress further, it will be necessary to elaborate and to test more explicit hypotheses concerning the putative metabolic processes which could develop in this context.

Metabolism first?

A highly contrasting view on the origin of life has been put forward by Günter Wächtershäuser[27] (1990[*], 1997[*], 2000[*], and 2006[*]). The key idea in his work is that an early form of metabolism predated genetics. Metabolism here means a cycle of chemical reactions that produce energy in a form that can be harnessed by other processes. His model is known as the iron-sulphur world theory, by contrast with the RNA world hypothesis.

In fact, Wächtershäuser's main target of criticism is not even genetics as such, but theories of a heterotrophic way of life. This theory of a heterotrophic origin assumes a primitive ocean of slowly accumulating amino acids, bases, sugars, lipids, and other organic compounds. These are seen as self-organizing to the first reproducing entity. The chemistry of this speculative process is pictured along conventional lines: solution reactions with adsorption-desorption equilibria and heterogeneous catalysis on minerals. This concept, according to which life arose from a prebiotic soup or primeval broth that covered the Earth, is generally attributed to Oparin (1952) [*]and Haldane(1947)[*]. The theory received support from Miller's demonstration that organic molecules could be obtained by the action of simulated lightning on a mixture of the gases CH4, NH3 and H2, which were thought at that time to represent Earth's earliest atmosphere. The organic compounds that were measured included hydrogen cyanide (HCN), aldehydes, amino acids, oil and tar. Additional amino acids were produced by Strecker synthesis through the hydrolysis of the reaction products of HCN, ammonium chloride and aldehydes, and in later experiments polymerization of HCN produced the nucleic acid bases adenine and guanine. Over the past 60 years, these notions have come to be very deep-seated. Nevertheless, they suffer from a serious defect in principle. As stressed in Do genes produce organisms? and The tessellation automaton above, living organisms are not so much “things” as processes; they are “dissipative structures” which arise through being coupled to a continual flow of matter and energy. This raises the question as to the nature of the source of this energy-flow. Flashes of lightning in a primitive atmosphere, giving rise to organic molecules in a “prebiotic soup” do not fit the bill; they are intrinsically sporadic, and do not correspond to the source of a continual flow of energy. This is what Wächtershäuser is getting at when he characterizes these theories as “heterotrophic”.

The alternative proposed by Wächtershäuser, who clearly appreciates these thermodynamic considerations, is that living organisms, right from their origin, are autotrophic. We come back to the question: what then is the energy source? Virtually all contemporary organisms are coupled to energy from the sun, either directly (plants which are able to use photosynthesis to power their metabolism and to produce organic molecules) or indirectly (animals which eat plants). But the photosynthesis of organic molecules is an operation which is difficult to achieve. Contemporary plants use chlorophyll; but this is a complex protein, which requires nucleic acid templates to be synthesized. We are in fact faced with another “scaffolding” problem: for all these processes to have got under way, there must initially have been a much more primitive system. Wächtershäuser postulates that the “pioneer organism” was powered by the reducing potential of undersea volcanic exhalations. It has a minimal substructure–superstructure organization. The superstructure consists of low-molecular weight bioorganic compounds. They are bonded to the surfaces of an inorganic substructure, i.e. as an interphase between the inorganic substructure and the water phase. I use here the name pioneer organism for the totality of the organic molecules that are bonded at any given time to the inorganic surfaces plus the surface regions of the inorganic substructure. Extension and curvature of the surfaces of the substructure are indefinite. In the direction normal to the inorganic surfaces, the superstructure has an essentially monomolecular extension and a vectorial orientation. The water phase constitutes the environment and the source of nutrients. Wächtershäuser remarks that such a pioneer organism may be seen as having a remarkable combination of three capabilities: for growth, reproduction and evolution. These central aspects of life coincide in the pioneer organism within one single type of process. This process may be briefly stated as follows. The thermodynamic driving force is provided by the chemical potential of the volcanic exhalations in the water phase. The kinetic reactivity is provided by the catalytic activity of the transition metal centres in or on the surfaces of the inorganic substructure.

The combination of these thermodynamic and kinetic aspects has the following effects:

  • Synthetic carbon fixation reactions generate organic products that become bonded to the surfaces of the inorganic substructure in statu nascendi, which means growth.

  • Some of the synthetic organic products exhibit an autocatalytic positive feedback into the synthetic reactions whence they arise, which means reproduction.

  • The autocatalytic feedback effect exhibits variations, which is the basis for evolution.

In the years following the initial publication of this iron-sulphur-world theory (Wächtershäuser, 1990[*]), a number of experimental results have been obtained which have made it possible to improve and to consolidate the theory. We shall not present the details of these results here, which are available in Wächtershäuser's publications (2000[*] and 2006[*]). In outline, Wächtershäuser presents some remarkably precise suggestions concerning these “pioneer organisms”, covering the following aspects:

  • their geological “habitats”;

  • their basic biochemical conditions, based on biochemical retrodiction from extant biology;

  • their synthetic reactions, in particular the formation of C2-structures, the formation of C3-structures, reductive amination, the activation of amino acids and a peptide cycle, and phosphorylation.

Wächtershäuser then goes on to envisage the process of “cellularization”, through the stages of surface lipophilization, semi-cellular structures, the origin of chemi-osmosis, culminating with Kandler's pre-cells. Of particular relevance for the theme of this chapter is Wächtershäuser's view of the origin of what he calls the “genetic machinery”. In line with his principally “metabolism first” approach, he addresses this question in terms of the (bio)chemical synthesis of nucleotides and nucleic acids on one hand, and of peptides and oligo-peptides on the other. He then examines how these processes could be tied together. He writes (Wächtershäuser, 2006[*]): “We assume that in the earliest phases of evolution, these two processes were essentially insensitive to or non-selective for particular nucleic acid sequences. Only later, with the appearance of more diversified amino acids, and longer nucleic acids and peptides, the question of a control of base sequences in nucleic acids and of amino acid sequences in peptides came into the picture. Sequence-controlled (template-directed) nucleic acid synthesis, whereby a sequence is copied from an old strand to a new strand of nucleic acid, is called ‘replication' and sequence-controlled peptide synthesis, whereby a nucleic acid sequence is copied into a peptide sequence, is called ‘translation'.” (Wächtershäuser, 2006[*]). This raises a basic question, as to the relation and the temporal order of ‘replication' and ‘translation'. On the “prebiotic broth” theory, replication is clearly the first event giving rise to an ‘RNA world'; translation would have appeared much later. By contrast, Wächtershäuser's chemo-autotrophic theory leads to a drastically different conclusion: here, “the two processes of sequence control, those of replication and translation, must have become established jointly, by coevolution whereby the emergence and evolution of replication slightly trailed the emergence and the evolution of translation.” (Wächtershäuser, 2006[*]).

Reconciling genes and organisms

In the history of biological thought, there has been an undeniable tendency to consider that the “genetic” and “environmental” aspects of any biological phenomenon are mutually exclusive and/or diametrically opposed. The debates on this question often take a polemical turn, for example in the case of the “IQ debate” (Dumaret & Stewart, 1989[*]); they go by variety of labels (“innate versus acquired”, “heredity versus milieu”, “genetic versus environmental”, and so on), but the basic issues are common to all of them. Now if we put the proposals of Cairns-Smith and Wächtershäuser side by side, it appears that this unfortunate relation of mutual exclusion is at work again. As we have seen, Cairns-Smith pays primary attention to the origin of a genetic system, and relegates metabolism to a subordinate role. Conversely, Wächtershäuser starts out with metabolism, and only brings in a genetic element with the advent of nucleic acids and protein synthesis, which are themselves introduced in the first place as elements of metabolism.

In a previous publication (Stewart, 2004[*]) I have argued that this whole area is riddled with a number of epistemological confusions; moreover, and more importantly, that when these confusions are properly sorted out there is no longer any need to oppose “genes” and “organisms”. On the contrary, they belong together. On the one hand, genes do not exist as such in isolation, so that DNA all by itself in a test-tube is not properly a gene; it only becomes one when it is integrated into an organism, the organism itself being a member of a population existing in an environment, and where the population is evolving over phylogenetic time (of the order of millions of years). Conversely, living organisms could never have become what they are today without the contribution of genes which have played a vital role in structuring their evolution. We have already mentioned the perceptive observation of Oyama (1985[*]) that much of the trouble lies in conception of the relation between Form and Matter, according to which any “organized” material process must be “in-formed” from a source that is essentially external to the process itself. Oyama made this observation primarily in the context of the remarkable regularities in the ontogeny of multi-cellular organisms. However, it is a very fundamental observation which applies equally to the basic processes of the autopoïesis of single cells. Once it is fully realized that the material processes that occur in living organisms are neither inert, nor even chaotic, but possess intrinsic capabilities of self-organization and indeed of self-production – and this, more than ever at the most basic level close to the very origin of life through spontaneous generation that we are discussing here – then there should no longer be any question of an opposition between a differential genetic aspect, and the basic processes of self-production.

In this respect it is interesting to note that when speaking of basic “chemoautotrophy”, Wächtershäuser is explicitly critical of the notion of “information”. He writes:

“In my papers I have deliberately avoided the use of the "information" metaphor because of its concept-narrowing effects and because it is not needed in chemistry. In chemical terms the process of reproductive multiplication is a synthetic chain reaction with branching, which may be represented as an autocatalytic production cycle. From this vantage point the mechanism of evolution may be given a very simple formulation: the appearance of branch products with a dual catalytic feedback – a feedback into the production cycle and a feedback into their own branch pathways. This is evolution by autocatalytic expansion loops. Each expansion loop is induced by the low-propensity de novo formation of a catalytic branch product. It latches onto the production cycle by the auto-catalyzed high-propensity formation of the catalytic branch product. In conventional chemical terminology this is called a memory effect. It is the physical basis for heredity. The information metaphor is neither needed nor helpful for an understanding of this mechanism.” (Wächtershäuser, 1994[*]).

This can be construed as making exactly the same point as our criticism of “in-formation” in the context of invariant, auto-productive processes. However, this does not necessarily mean that Wächtershäuser is correct in altogether excluding any sort of differential “genetic system” before the evolutionary stage of nucleic acids and protein synthesis. Wächtershäuser does clearly recognize that the emergence of a “genetic system” marks a crucial stage in the evolution of living organisms. Thus, in a later publication, he writes:

“To sum up, the early, direct, deterministic, chemical mechanism of evolution by ligand feedback is the precondition for the later emergence of an indirect, stochastic, genetic mechanism of evolution by sequence variations. This conversion is truly an evolution of the mechanism of evolution.” (Wächtershäuser, 2006[*]).

The question we are examining here is whether there may have been a “primordial” genetic system, as a sort of scaffolding to reach the stage of the nucleic acid genes of contemporary organisms. More specifically, the question is whether the complexification involved in passing from physico-chemical individuation, the sort of process that can indeed arise (as a dissipative structure) by spontaneous generation, to fully-fledged biological individuation as identified above in The tessellation automaton, could really be achieved without the contribution of a correctly construed, differential genetic system. In The tessellation automaton, we proposed that this transition to “biological individuation” involved two critical steps: a) the creation of a cell membrane, enclosing and catalysing the (bio)-chemical metabolism; and b) the emergence of a rudimentary form of “cognition”, in the sense of a minimal sensori-motor system enabling these proto-organisms to actively remain in areas where the environmental conditions are adequate for the maintenance of their existence as dissipative structures. We will now conclude this chapter by examining whether either or both of these critical steps could have involved the contribution of a “primordial genetic system”.

The appearance of a cell membrane is certainly a key stage in the origin of life. Compared to the chemical processes which can occur in the proximity of an extended surface (cf. figure 14), the existence of a membrane enclosing an intra-cellular compartment has the immense advantage of making it possible to accumulate significantly higher concentrations of intermediate metabolites, and thus indeed to promote the complexification of metabolism. The question is whether the membrane itself, or certain membrane components, can be considered as “primordial genes” in the sense defined in this chapter. Let us recall the list of requirements for these hypothetical primordial genes, as we identified them in Recapitulation of basic considerations: invariance and variance. :

  • firstly, these primordial “genes” had to exhibit a certain range of structural variation;

  • secondly, they had to be such that when they were copied, the “copies” could be reasonably faithful to the originals, so that the genetic variations could be re-produced; and

  • thirdly, it had to be possible to set up a reasonably systematic coding relationship between the genetic differences (i.e. the structural differences in the genes themselves), and the phenotypic differences between the individual organisms.

Do the first membranes (or maybe certain components of the first membranes) fit the bill? If we consider the tessellation automaton illustrated in figure 10, we have already remarked that this gives rise to a circular relation between the membrane and metabolism, as shown in figure 11. The membrane is itself re-produced, but on a slower time-scale than the metabolism. This is as it should be if membrane components are to function as “genetic elements”; but it is not sufficient. The key question is: did the membranes of early organisms exhibit structural variations that could be reproduced? If we are not to engage in uncontrolled speculation, we cannot immediately give a definite “yes or no” answer to this question; we will leave it as an open question for future work.

The second line of thought for developing a hypothesis about a putative “primordial genetic system” involves the realm of “cognition”, and takes its cue from the work using genetic algorithms to evolve robot behaviour that we discussed in Autonomous robotics. In general terms, the connection between sensory inputs and motor actions is open to arbitrary variation, since there is nothing intrinsic to the activation of a sensor organ which means that it necessarily has to be linked to any particular effector organ; and this means that the connection structure is particularly suitable for coding into genetic information (in the precise sense that we are using the term here). In order to explore this line of possibilities, we have to specify what might be the sensor organs, and what the effector organs, involved at a very early stage in biological evolution[28] . In order to get some clues on this question, we may consider the phenomenon of chemotaxis, i.e. the system used by bacteria which enables them to approach a source of nutrition[29] . Some bacteria, such as E. Coli, have a number of flagella on their cell surface which can move in two ways: A) they can align into a single rotating bundle, which from the point of view of an outside observer causes the bacterium to swim in more or less a straight line; B) the flagella bundle breaks apart, each flagellum pointing in a different direction, which (again from the point of view of an outside observer) causes the bacterium to “tumble”, i.e. to reorient itself while remaining in place. These bacteria also have sensory receptors in their cell surface, which for their part can also be in one of two states: i) if the concentration of sugar (a nutrient) in the medium is increasing, they are in state 1; ii) if the concentration is constant or decreasing, they switch to state 2. The key point for our present discussion is now the relation between the sensory states 1 vs 2, and the “action states” A vs B. If sensory state 1 triggers action A, and state 2 triggers action B, then (as shown in figure 15) the bacterium will eventually end up approaching the source S. If the sensory-motor connections had been made differently – if 1 triggered B and 2 triggered A – then (from the point of view of an outside observer, but also for the forces of natural selection) the bacterium would move away from S. This would of course be inappropriate (and selected against) if S were a source of nutrient; but if S were a poison, then it is this pattern of connexions that would be appropriate, and would be selected for.

Now in the case of real contemporary bacteria, the sensory receptors and the flagellar motors are proteins, and the key feature of the precise form of “signal transduction” (the “connections”) are also mediated by proteins. Thus, this actual example requires that there

should already be nucleic acid genes. The question is whether some analogous sort of situation could arise earlier in evolution, in the context of our hypothetical “primordial genes”. Now it is interesting to note that “dissipative structures” of the sort discussed in The tessellation automaton and illustrated in figure 13 a and b can already, under certain circumstances, exhibit “chemotaxic” behaviour (McGregor and Virgo, 2011[*]), moving towards areas where the concentrations of “nutrient” chemicals are higher. In this case, however, this happens for intrinsic reasons related to the dissipative reactions themselves; so that there are no variations, no degrees of freedom that could be coded for by genetic information. As in the case of membrane components that we discussed above, it must therefore remain an open question as to whether “primordial genes” could make a contribution to the achievement of “biological individuation” through this sort of mechanism.


In this chapter we have put forward the hypothesis that there was a “missing link” between the first proto-organisms simple enough to have arisen by spontaneous generation, and the simplest known contemporary organisms which have a complex genetic system involving DNA and proteins. The hypothesis is that this missing link was a primordial genetic system, with a material substrate other than nucleic acids; this genetic system may have coded for membrane components crucial for the process of cellularization, and/or elements of a sensori-motor system corresponding to a primordial form of cognition and enabling the primitive organism to remain in areas favourable to its ongoing autopoïesis.

A leitmotiv in this book is the importance of elaborating scientific hypotheses that are genuinely open to refutation, in accordance with Popper's epistemology. We may note here that Wächtershäuser's explicit and deliberate adherence to this methodological principle has been a major factor in promoting the considerable development of his “iron-sulphur world theory” over the period of 20 years since its initial formulation in 1988. So how could our hypothesis of a primordial genetic system best be laid open to potential refutation? All questions relating to the evolution of the early forms of life on Earth would be much easier if we had a time-machine which would enable to go back and look directly; but for better or worse we do not have any such machine, even for going back one year let alone the thousands of millions of years that would be necessary. All hypotheses concerning the evolution of life thus have to rely on a combination of indirect methods, which include notably: fossil records, and more generally reconstructions of the probable conditions on the early planet Earth; laboratory simulations of these conditions, in particular to see what chemical reactions could occur under those conditions; and comparative studies of contemporary organisms as a basis for working back to presumptive common ancestors. An important resource, fully in accordance with Popperian epistemology, is the serious formulation of theoretical hypotheses including an examination of their internal consistency. Finally, in relation with this last point, it is important to set out as clearly as possible the various alternative hypotheses, which can lead to an identification of the crucial empirical evidence for differentiating between them.

So what then are the main alternatives to the hypothesis we propose here, i.e. that of a “missing link” in the form of a primordial genetic system? If we refer to the literature presented inA primordial genetic system , there are two main alternatives: i) that there was indeed a “primordial genetic system”, but that it was there right from the start and indeed provided the backbone for the subsequent progressive elaboration of metabolic processes (cf. Genes first?); and ii) that an original autotrophic chemistry had capabilities for endogenous complexification such that it could go through to cellularization and the emergence of a genetic system based on nucleic acids and proteins without any need for a “primordial” genetic system (cf. Metabolism first?). The agenda then, as I see it, is to elaborate these contrasting hypotheses in such a way that it will become possible to refute some (if not all!) of them.

On the enaction of an immunological self

In the general introduction to this book, I aired the view that in order to further the development of the paradigm of Enaction, it is important to elaborate hypotheses in sufficient specific detail that they are open to possible refutation, in Karl Popper's sense of the term. In this chapter, I propose to illustrate this view by recounting the story of some collaborative work in which I participated right at the beginning of my involvement with the paradigm of Enaction. The hypothesis we imagined was indeed refutable – the proof being that it ended up being refuted! The story is nevertheless worth telling, both for the elegance of our hypothesis (which could have worked – if God had been a bit smarter he would have taken advantage of the possibility we imagined), and because it lead on to currently viable hypotheses concerning the function of the immune system which do illustrate an auto-constitutive dynamic network view of biological systems.

I will start with some general considerations concerning life and cognition. Living organisms are characterized by their circular organization: they are metabolic processes, pure fluxes of matter and energy, which have the very special property of producing themselves. This is expressed in the concept of “autopoïesis” invented by Maturana and Varela. Thus, living organisms are dynamic dissipative structures; but unlike purely physical and/or chemical structures which are intrinsically ephemeral (cf. Simondon's concept of “physico-chemical individuation”), living organisms function in such a way that they can go on producing themselves indefinitely (cf. Simondon's “biological individuation”). In other words, although living organisms can of course die at any moment if an accident befalls them, they are potentially immortal.

Cognition is also characterized by a circular organization : the objects of cognition (for example, colours) are not pre-given, and do not even exist as such independently of the perceiving subject; these objects are “brought forth” together with the subject, they are “enacted” as we say now, and they are thus inseparable from the organization of that subject. Along with Humberto Maturana, Francisco Varela had the deep insight that these two forms of circular organization are, at root, one and the same: this insight can be summed up by the formula «life = cognition = autopoïesis».

These were, and still are, heady ideas - of the stuff that dreams are made of, with a capacity to evoke an immediate intuitive conviction that they just have to be right. However, it is also true that in order to make a substantial contribution to science these ideas need to be deployed quite concretely; they have to do some real work in relation to precise, detailed experimental observations, and as I have said they have to be made refutable. In 1987, at a week-long seminar at Cérisy-la-Salle, I had the chance to meet Francisco Varela, together with a Brazilian immunologist, Nelson Vaz. Nelson thought that the immune system offered a particularly promising field for putting these ideas concerning autopoïesis and cognition into practice. It is a historical fact that immunology is an area where working biologists themselves spontaneously have long had recourse to “cognitive” metaphors: “recognition”; a “repertoire” of recognized molecular shapes; “learning” and “memory”; and of course, most striking of all, a “self versus non-self” distinction that I will come back to. And it did indeed turn out that the application of the concept of autopoïesis to the immune system lead with extraordinary rapidity to a radical renewal of classical perspectives. This is what I shall now set out to explain.

Classical immunology and a potential paradigm-shift

In classical immunology, the immune system is conceived as a linear input-output system. The inputs are antigens: substances, generally foreign to the body of the animal, which trigger the production as an output of antibodies, each of which specifically recognizes the antigen that evoked it. The specificity of this recognition is impressive: the number of different antigens that the immune system of a mammal is able to recognize has been estimated to be of the order of 10¹⁷. Moreover, immunologists consider that the repertoire of antibodies is complete: the mammalian immune system is capable of recognizing the totality of all possible molecular shapes (of an appropriate size), including molecules that have never before existed in the course of biological evolution because they were synthesized for the first time by human chemists. Classically again, immunologists consider that the function of the immune system is to protect the body against foreign antigens (typically, those that belong to pathological micro-organisms); more precisely, it is considered that the function of an antibody is to trigger the destruction of the antigens that it recognizes (figure 16).

This classical scheme can be summed up by saying that the immune system can potentially perceive everything; and that it triggers the destruction of everything that it actually perceives. Now this bald schematic formulation raises a troublesome question: what about the relation between the immune system and the body in which it is housed? This body is composed of molecules, many of which are the appropriate size for being “perceived” by the immune system; the proof of this is that if such molecules are injected into another animal, they do indeed provoke a destructive immune response. It is for this reason that grafts of tissues or organs are practically impossible in mammals, whereas they are easy in plants or invertebrate animals such as insects. But now, if we apply the classical schema literally, we arrive at an astonishing conclusion: the prediction is that the immune system should systematically destroy the body in which it is housed.

It is clear that this is not what happens; or more precisely, when something of this sort happens one speaks of “auto-immune disease”, but auto-immune diseases are both relatively rare and, when they do occur, they are far less catastrophic than the scheme predicts. In fact, quite apart from empirical observations, it is obvious that this prediction cannot be fulfilled, because systematic auto-immunity would be horribly dysteleological and quite incompatible with the constraints of viability and natural selection. The classical immunologists invented a term, “horror autotoxicus”, to indicate that this prediction of their theory was not and indeed could not be systematically verified. In order to patch up the situation, classical immunology circumvented the problem by a straightforward adjustment to the theory: the immune system perceives everything, except its own body. One suspects that this bald-faced adjustment, which is quite shamelessly ad hoc in order to avoid what would otherwise be a straight refutation, caused a certain unease; an unease which may be subliminally expressed by the very term “horror autotoxicus”, quite horrible indeed in its mixture of Greek and Latin roots. However that may be, the conclusion - that the immune system perceives everything except its own body - is quite inescapable, given the premises of the argument; and it is for this reason that the so-called “self versus non-self” distinction plays such a central role in classical immunology.

What light does the perspective of autopoïesis shed on this rather murky situation? According to the notion of autopoïesis, the objects of cognition are specified, constituted, by the organism itself. This can be summed up very neatly by a verse of the Portuguese poet Pessoa:

What we see

Is not what we see

But what we are

The point is, in a way, a very simple one; but it runs so radically counter to ordinary common-sense that a few words of explanation may not be amiss. The usual point of view is to consider that the objects of perception are ontologically primary: they exist, and are what they are, quite independently of any perception that there may or may not be concerning them. A perception by a cognitive subject then corresponds to an internal “representation” of the referential object; ideally, the representation will tend to be more or less isomorphic with the pre-existing referent. Technically, this point of view is termed “objectivist”; we may note that it corresponds to the point of view of an external observer who is presumed to be omniscient, and is thus able to examine both the object and the representation and to check out on the degree of correspondence or isomorphism between them. The perspective of autopoïesis is radically critical of the objectivist position. The point is that, from the point of view of a cognitive subject, there is no way of getting “out of its skin” and perceiving the “object in itself” directly as such. The only thing that an organism can know, is the effect that its interaction with an object has on its own functioning; having access to only one of the two terms, there is just no way that an organism can judge whether the content of its percept is, or is not, an adequate “representation” of an external object. Thus, the “percept” is not separable from the cognitive subject. This is reminiscent of Berkeley's position, “to be is to be perceived”; with the difference, that an organism is subject to a viability constraint, so that its percept cannot be an arbitrary hallucination but must bear significantly on the organism's interactions with its environment.

In other words, in this new perspective, whatever one perceives is, by the very fact of being perceived, the “self”; and whatever is not perceived is, ipso facto, “non-self”. This amounts to an exact reversal of classical immunology, according to which the immune system does not perceive, ignores, the “self” (otherwise it would destroy it), and perceives only the «non-self». Francisco Varela was the first to develop these considerations, in close collaboration with Nelson Vaz; together, they proposed that in order to denote this radical reversal in perspective, it might be better to speak of a distinction between “self” and “non-sense” [Vaz & Varela, 1978)[*].

Before going further, it may be well to say a few words to try and dispel some confusion that has quite understandably arisen. It is abundantly clear that a major (if not necessarily exclusive) function of the immune system is to protect the organism against infectious diseases caused by pathological micro-organisms. This is demonstrated quite straightforwardly by the simple observation that severely immuno-deficient mice – and humans – do indeed die of uncontrolled infectious disease. It is equally clear that in order to do this, the immune system must make a distinction between pathological micro-organisms, and the body of the organism itself. The perspective of autopoïesis does not gainsay any of this. What is at stake is the choice of an appropriate nomenclature in order to designate the two terms of this distinction. The key here is a remark that Humberto Maturana never tired of repeating: “Everything said is said by an observer”. In particular, whenever a distinction is being made, we should always ask: “who is making the distinction?” In the case of classical immunology, the distinction is being made by an external human observer: it is the immunologist who can see the difference between the organism on the one hand and pathological micro-organisms on the other; it is the immunologist who designates molecules from the body of the organism as “self”, and molecules which he knows came from a micro-organism (or another source external to the organism) as “non-self”. By contrast, when we employ the conceptual framework of autopoïesis, we are in a certain sense looking at things from the point of view of the immune system itself[30] . It is from this point of view that a “self” versus “non-self” distinction is impossible, because unlike a human immunologist, the immune system itself has no means of knowing where the molecules came from. The immune system is composed of cells, the lymphocytes; at the level of a local interaction between an individual lymphocyte and a molecule, there is nothing that distinguishes what the immunologist calls a “self” molecule from a “non-self” molecule. The conceptual framework of autopoïesis is thus the more appropriate one if we are trying to understand the mechanisms and mode of operation of the immune system itself. As we shall see, the immune system as a whole, considered over the history of its development, is capable of making a distinction which, to all practical intents and purposes, does roughly coincide with the immunologist's “self” versus “non-self” distinction. However, when we identify the mechanisms whereby the immune system is capable of making this distinction, we shall appreciate that calling this a “self” versus “non-self” distinction is a misleading misnomer. We shall return to this point; but the time has come to present the understanding of the actual functioning of the immune system that comes from adopting the conceptual framework of autopoïesis.

A mathematical model of the immune network

A key element that made it possible to deploy the framework of autopoïesis to the workings of the immune system was the work of the great Danish immunologist, Niels Jerne. The starting-point of Jerne's theory is this: if the repertoire of antibodies really is “complete” or “open-ended” as he called it, then it is logically inescapable that the antibodies themselves should be included in this repertoire. After all, antibodies are protein molecules of the same size as many other antigens. In other words, there are strong a priori grounds for supposing that the set of “antibodies” form a connected network, where each “antibody” is recognized by other antibodies in the system. We have put the term “antibody” in scare-quotes, because it is clear that in this perspective that recognition does not necessarily lead to complete and immediate destruction. For this reason, it is preferable to use the term “immunoglobulin” to designate the molecules produced by the lymphocytes. In order to designate this sort of interaction between immunoglobulins we employ the term “idiotypic”; consequently, the sort of network predicted by Jerne is an idiotypic network.

When Francisco came to Paris in 1985, he worked with Antonio Coutinho (himself a student of Jerne) to set down the basis of a mathematical model of idiotypic networks. This is not the place to enter into technical details and the mathematical equations (see Varela et al., 1988)[*], (Varela & Stewart, 1990[*]). Qualitatively, in natural language, the basic idea was the following. The survival of a lymphocyte, its proliferation and its capacity to secrete immunoglobulins, depended on the “field” that it received as a result of its idiotypic interactions with other immunoglobulins. The dynamics of the network were highly non-linear: if the received field was below a lower threshold, the lymphocyte perished (Nelson Vaz expressed this by saying that the lymphocyte died of “loneliness”); if the field was above an upper threshold, the lymphocyte died of “suffocation”; however, if the field was in the favourable “window” between the two thresholds, the lymphocyte could survive and/or or proliferate and produce immunoglobulins (figure 17).

Now, the immunoglobulins which provide for the field itself, in the form of idiotypic interactions with the immunoglobulin receptors of the newly emerging lymphocytes, are produced as a consequence of these very interactions. Hence the field determines itself and its maintenance; in other words, it is autopoïetic, and it selects newly formed lymphocytes into its own operation, such that these dynamics are supplemented by a process of “meta-dynamics”. We have already indicated that a lymphocyte could disappear from the system (if its received field lay outside the window); but at each time-step, it was also possible to recruit new lymphocytes into the network from a sample of “random” lymphocytes freshly produced by the bone-marrow. Specifically, a candidate lymphocyte would be recruited if but only if its received field was situated within the window between the lower and upper thresholds. There was thus a circular relationship between the dynamics and the meta-dynamics. On the one hand, the dynamics of the concentrations of the different lymphocytes resulted in their disappearances and recruitments, and so the dynamics determined the meta-dynamics; conversely, the meta-dynamics gave rise to the structure and connectivity of the network at each instant, and so the meta-dynamics determined the dynamics.

This “circularity” between the dynamics and the meta-dynamics was quite deliberately in the spirit of the “circular organization” so fundamental to the general theory of autopoïesis. Whether the circularity as expressed in this particular mathematical model is really adequate to fully capture the “circularity” of autopoïesis is an open question; but this is the price to pay for developing a hypothesis sufficiently explicit and precise for it to be refutable. The model described here aimed at a reasonable compromise between biological realism on one hand, and maximal simplification on the other. All the components, properties and relations in the model were based on entities and processes that were known to exist on grounds of actual biological observations; on the other hand, these elements were represented in the model in the simplest possible form that would still give rise to interesting emergent properties of the system as a whole.

Morphogenesis in shape-space

In 1988, I joined the group that had formed around Francisco and Antonio Coutinho at the Pasteur Institute in Paris. My contribution was to run actual computer simulations, based on the model that just described, in order to examine the emergent properties of an idiotypic network. The immediate result of this was to reveal the necessity of a method for specifying the structure of connectivity of an idiotypic network; more specifically, of defining the matrix of all possible pair-wise affinities between a given set of immunoglobulins. The experimental data on this point were (and still are) scanty and quite insufficient; in addition, we needed a mode of representation that would render the evolution of the connectivity structure as the system matured over time graphically visible and comprehensible. We solved this problem by adopting a modified version of the “shape-space” concept originally suggested by Perelson & Oster (1979)[*] and developed by Segel & Perelson (1989)[*]. According to this concept, the universe of stereo-chemical shapes which determine intermolecular affinities can be represented as points in a multi-dimensional shape-space. In our version, we used a 2-dimensional space (for obvious reasons of graphical visualisation); and each point in shape-space was taken as representing a pair of two perfectly complementary shapes with maximum affinity (conventionally, the members of a pair are labelled “black” and “white”). This mode of representation has the following advantages: firstly, relations of similarity in molecular shape (and hence affinity profiles) are immediately perceptible as the proximity of corresponding points of the same colour in the shape-space; and secondly, relations of complementarity (and hence high affinity) are also immediately perceptible in the form of proximity between “black” and “white” points. The generation of random immunoglobulins as candidates for meta-dynamical recruitment was then quite straightforward: it was effected by generating black or white shapes at random positions in the total shape-space.

Using this procedure in conjunction with the “window” model, we very rapidly obtained some promising results. Firstly, we showed that under these conditions, a self-sustaining idiotypic network could arise – without either collapsing or exploding. We quite deliberately started by studying the behaviour of the system in the absence of external antigens, in order to characterize its “eigen-behaviour”. The idiotypic network did indeed exhibit interesting properties of self-organization: as can be seen in figure 18 (left), the combined dynamic and meta-dynamic process gives rise to clear patterns in shape-space: there are “chains” of lymphocytes of the same colour, and the chains of complementary shapes mutually sustain each other. We can consider that these patterns correspond to the identity of a “molecular self” as defined and indeed constituted by the autonomous dynamics of the immune system itself.

Secondly, we studied the modulation of this eigen-behaviour when the system was perturbed by the introduction of external antigens, modelled here as points in shape-space which produced a field for the lymphocytes, but whose own concentration was constant irrespective of the field they themselves received. Typical results are shown in figure 18 (right). What we see is that the idiotypic network adjusts smoothly so as to integrate the antigens harmoniously into its own pattern of behaviour; in other words, the antigens are effectively assimilated as a part of the “molecular self” constituted by the network. More precisely: we see in Figure 18 that all the antigens of a certain colour (black or white) are surrounded by lymphocytes of the same colour. We know that all the lymphocytes in a "chain" of a certain colour receive a "field" (from lymphocytes in the facing chain of the opposite colour) that is within the limits of the window (if this were not the case, the lymphocytes would already have been eliminated). Since, under these network conditions, the antigens receive the same field as the lymphocytes of the same colour that surround them, there is an important corollary: the field received by the antigens remains within the limits of the “window”, i.e. this field is at most equal to the upper threshold. Thus, if we suppose (as is reasonable) that the destruction of antigens is only triggered by fields well above the upper threshold, in the presence of an idiotypic network antigens will not be destroyed but will be “tolerated”.

Thirdly, we can compare these results with what happens if the idiotypic network is abolished. This would be difficult to realize experimentally in a real biological situation, but in the model it can be achieved by a stroke of the pen – for example, by recruiting lymphocytes of only one colour, which have zero affinities with each other. In this case, the lymphocytes are activated only by the antigens; consequently, lymphocytes complementary to the antigen are recruited without limit, and the fields received by the antigens increase indefinitely until they reach levels that we may suppose do trigger destruction of the antigens.

The results of these computer simulations contributed to a renewal of interest in network ideas. When Jerne first presented his concept of an idiotypic network in 1974, the idea received quite a favourable reception from the community of immunologists. However, over the years, the idea gradually fell into disrepute. What seems to have happened is this. As we have seen, classical immunology is centred on the phenomenon of strong, destructive immune responses to external antigens. Thus it was quite natural that in the “first generation” models of the immune network, the aim was to make the network produce immune responses. However, the result of these attempts was general failure: there seemed to be just no way that an idiotypic network could be got to produce a good immune response. Retrospectively, it seems clear that this “failure” stemmed from the fact that the first-generation models were trying to make the network do exactly the wrong thing. In order to produce classical immune responses, Burnet's mechanism consisting of the selection of unconnected clones is both straightforward and perfectly adequate; at this level, a network organization is not only unnecessary but actually counter-productive, because the network prevents the development of a strong immune response. A much more appropriate role for the immune network, for which its natural emergent properties are an advantage rather than a handicap, is to promote tolerance by protecting the antigens of the body from attack by the immune system. These considerations lead Varela and Coutinho (Varela & Coutinho, 1991)[*] to make a proposal for “second generation immune networks”, whose distinctive feature is that the immune system is composed of two complementary compartments: the “Central Immune System” and the “Peripheral Immune System”.

The central immune system and the peripheral immune system

It is important to note that these theoretical considerations were carried out in close conjunction with the ongoing experimental work in Coutinho's laboratory at the Pasteur Institute. In particular, there was great interest in the so-called “natural antibodies”, the circulating immunoglobulins that are found in the sera of all normal vertebrates even when they are secluded from all antigenic contacts with the environment. These natural antibodies are produced seemingly in a spontaneous manner, and thus appear to be the result of the autonomous internal activity of the immune system. Coutinho's group quickly found that these antibodies bind to “self” (i.e. antigens of the body) and are often multi-reactive. Further work demonstrated that increases or decreases in the concentration of certain specific natural antibodies had an influence on the natural antibody repertoire as a whole. The natural antibodies are produced by “naturally activated” lymphocytes, which represent about 10% of total lymphocyte numbers; the remaining 90% are resting cells which are mitotically inactive, do not secrete immunoglobulins and are thus devoid of effector functions.

The second-generation immune network model arose by putting together these empirical observations with the theoretical considerations outlined above. Every day, the bone-marrow produces a large number of new lymphocytes, each of which carries a unique immunoglobulin receptor. The numbers produced are so high that the total population of lymphocytes can be replenished in a few days; in addition, the repertoire of these new lymphocytes is “complete”. If these new lymphocytes are not stimulated, they remain in a “resting” state, and die after 2 or 3 days. However, the rate of production is such that at any one time, these resting cells make up 90% of total lymphocyte numbers. These resting cells constitute the “Peripheral Immune System” (PIS); they have no functional idiotypic connections. The “Central Immune System” (CIS) is composed of the 10% of lymphocytes that are “naturally activated”; according to the model, this activation is primarily the result of idiotypic interactions between these lymphocytes, so that they form a connected network. As predicted by the computer simulations, the repertoire of the CIS incorporates all the antigens of the body of the organism that are permanently present. Also in line with the computer simulations, it is the fact that body antigens are included in the repertoire of the CIS with a network organization that protects them from immune attack and thus accounts for the phenomenon of “tolerance”.

It is to be noted that according to this model, the two compartments CIS and PIS are complementary. The CIS is composed of lymphocyte clones that have been “rescued” from death within 2 or 3 days (their fate if they had remained in the PIS) by their meta-dynamical recruitment into the CIS. We may recall that the repertoire of lymphocytes freshly emerged from the bone-marrow is “complete”. Hence, by construction, the repertoire of the PIS is “complete minus the repertoire of the CIS”. Since the repertoire of the CIS includes all the body antigens, it follows that to a first approximation (but we shall have occasion to return to this point) the repertoire of the PIS is “complete minus body antigens”. Since the lymphocytes in the PIS are isolated, unconnected by network interactions either with each other or with the CIS, if they are stimulated by a novel antigen (for example belonging to an invading micro-organism) they will mount an unfettered immune response. Thus, the PIS is ideally constituted both by its repertoire and by its mode of functioning to the role of protecting the organism from pathological micro-organisms.

We may now compare this second-generation network model with the scheme of classical immunology. In a certain sense, the distinction between the CIS and the PIS corresponds to the classical distinction between “self” and “non-self”. This is comforting, and means that the new network view renders unto Caesar that which is due to him. However, there remains a major difference with classical immunology, which is interesting. The difference is that in classical immunology, tolerance to body antigens results from eliminating all the lymphocytes that interact with them – the so-called “clonal deletion” theory. According to the second-generation network model inspired by the concept of autopoïesis, however, tolerance to body antigens is the result of a positive process: the lymphocytes that interact with the body are not eliminated, but on the contrary they are activated by being incorporated into the dynamics of an idiotypic network. This has a number of consequences – conceptual, experimental and practical – which are worth spelling out.

The conceptual difference is that the category of “self” is not defined by an external human observer on the basis of knowledge that is intrinsically inaccessible to the immune system. In this new conception, “self” is first and foremost defined by the immune system itself on the basis of its autonomous functioning as a self-sustaining idiotypic network. It is only subsequently that the body antigens are incorporated into the repertoire of this network. The body antigens are not intrinsically “self” as such (and even less because they are decreed to be “self” by an immunologist); they become self by virtue of being assimilated into an “immunological self” that has already been constituted by the autonomous operation of the immune system. We may note here that even if we accept that the body antigens are normally incorporated into the immunological self, it does not follow that the “immunological self” reduces to just the body antigens. As already mentioned, we shall have occasion to return to this point.

The experimental difference is this. On the classical view, tolerance is due to the elimination of lymphocyte clones that interact with the antigen in question. Thus, if a “hybrid” immune system is produced experimentally, “tolerance” should be recessive (i.e. a hybrid between a tolerant and a non-tolerant system should be non-tolerant). On the new view, tolerance is due to the positive effects of an functional network; thus, on condition that the hybridisation is carried out in such a way that the network is not disrupted, “tolerance” should be dominant (the hybrid should be tolerant). Without going into details, many experiments have been performed which amply demonstrate the phenomenon of dominant tolerance. It suffices to say that, after a few decades of great predominance of “recessive tolerance” theories, it is now widely accepted that “natural tolerance” is dominant indeed.

The practical, clinical difference is this. On the classical view, auto-immune disease arises because the immune system is functioning over-zealously; it is therefore quite logical to treat auto-immune disease by immuno-suppression. The results are generally not very satisfactory: immuno-suppressive treatments are, at best, symptomatic, and may have serious side-effects. To date, there are no observations indicating the cure of autoimmune patients by such treatments. On the new view, auto-immunity arises from a deficiency in the normal ongoing activity of the immune system (and quite apart from treatment, it is a widespread clinical, and experimental observation that immuno-deficiency is indeed often associated with auto-immunity); the logical treatment thus consists in an (appropriate) activation of the immune system. In line with this prediction, the treatment of auto-immune disease by the injection of a balanced mixture of normal serum immunoglobulins has had some very positive results (Kazatchkine & Morell, 1997)[*].

This model is therefore promising; but before it can be properly opened up to the possibility of experimental refutation, there is another step to be taken. How is it that the distinction between a CIS and a PIS actually comes about? What the model showed was that if the immune system functioned in a “network” mode, then the antigens that fall within its repertoire will be integrated into the network dynamics and will hence be tolerated (this is the basis of the “CIS”); whereas if the immune system functioned in a “non-network” mode without idiotypic connections, then the antigens that fall within its repertoire will provoke an immune response leading to their destruction (this is the basis of the “PIS”). However, this left quite unresolved the question as to how the distinction between the CIS and the PIS actually came about. In our first simple models, illustrated in figure 18, the “network” spread over the whole available shape-space (the self-organized patterns only arose when the whole shape-space was saturated), thus leaving no room for a residual “PIS”. It is true that if we abolished the network interactions by simple fiat, then the system would function in a PIS mode. However, this sort of intervention, coming from outside the system, would be quite contrary to the spirit of autopoïesis where the whole point is to explain phenomena as resulting from the autonomous operation of the system itself. To be more precise, what was missing was an account of how the CIS versus PIS distinction (the successor to the classical “self” versus “non-self” distinction) could arise through the autonomous ontogeny of the system itself.

This problem was tackled with great energy and imagination by Jorge Carneiro, at that time a PhD student at the Pasteur Institute. Jorge came to the conclusion that this problem could only be solved by extending the model to include not only the B-lymphocytes which produce immunoglobulins, but also the T-lymphocytes which (among other functions) provide “help” to B-lymphocytes. Without going into the details (Carneiro, Coutinho, Faro & Stewart, 1996)[*], Jorge came up with an aesthetically beautiful model which exhibited the emergent properties we were looking for. The first condition was that the ontogeny of the system should start with a sufficiently numerous and diverse set of antigens (these plausibly correspond to the body antigens already present when the immune system first starts developing in embryonic life, but we will refer to them more prudently here as “initial antigens”). On this condition, the immune system went through several phases in its ontogeny. Firstly, T-cells were activated and these in turn recruited B-cells. On condition that the initial set of antigens was sufficiently numerous, this process continued until the repertoire of the B-cells was “complete”, and so the B-cells started to exert a regulatory feedback influence on the T-cells. At this stage, the second phase commenced: now that the system was “saturated” and complete, competition set in amongst the B-cells for the T-cell help that they themselves were limiting. Under these conditions, the repertoire of the B-cells “shrank” until only those that with a B-cell receptor (BCR, alias an immunoglobulin) that directly recognized a T-cell receptor (TCR) remained. It is aesthetically pleasing to note here that since the TCRs are a “mirror” of the initial antigens, and these BCRs are a “mirror” of the TCRs, the BCRs are a “mirror of a mirror” – i.e. a sort of “internal image” of the initial antigens. Thus, the B-cell repertoire in the mature CIS was far from complete, and in fact was restricted to an internal image of the initial antigens. This set the stage for the third phase in the ontogeny. If a single novel antigen, that the immune system had not seen so far, was now presented to the system after the maturation of the first two phases, then the global regulatory processes that occurred with the initial set of antigens was not reproduced. The reason is that the B-cell repertoire was now tightly restricted to an internal image of the initial antigens, and had virtually no chance of percolating to include a BCR that would regulate the T-cell activated by the novel antigen. The result was that the immune system would mount an unfettered immune response to the novel antigen.


The model had now matured to the point where it was open to refutation by experimental evidence. As it turned out, the proof that it was refutable is that... it was indeed refuted (at least in its initial form as presented here). One straightforward prediction of this model is that μ-knockout mice (that have no B-cell receptors at all) should exhibit uncontrolled T-cell proliferation; but they do not. Moreover, this model did not give rise to experimental observations that would validate it positively. The most serious problem was that it is not at all sure that idiotypic interactions between the highly variable BCRs and TCRs have any physiological significance at all. Jerne's initial argument in favour of idiotypic networks was purely logical and qualitative: if the immunoglobulin repertoire is complete, then logically interactions between immunoglobulins must exist. Incontrovertible as far as it goes, this is not enough to show that the concentrations and affinities involved are quantitatively sufficient for these interactions to have real physiological effects; and when the calculations were done on the basis of this initial model, it seems that they are not. Thus, the model was indeed refuted.

The methodological principle that it is important to develop predictions that are open to refutation may seem, to outsiders, somewhat masochistic. But in practice, a refutation of this sort is not catastrophic; it is not the end of the road, it just means that there is some more work to be done. Jorge Carneiro continued working on this question with his co-workers, and quite recently they have come up with a revised model that is indeed compatible with all the experimental evidence (Carneiro et al., 2007)[*]. As is so often the case, this new model is somewhat more complex than the original one: it involves distinguishing two separate classes of T-cells, the conventional effector T-cells (TE), and regulatory T-cells (TR). According to this new model, the persistence and expansion of TR cell populations depend strictly on specific interactions they make with antigen-presenting cells (APCs) and the effector TE cells. Mathematical analysis of the dynamics of this three-partner cross-regulation system reveals that it is bi-stable: either i) the TE cells dominate and the TR cells are virtually absent; or ii) the TE and TR cell-types both co-exist in a stable balance. This latter state, in which the TR cells feed on the specific immune activities that they suppress and regulate, provides a basis for tolerance to self-antigens. Thus, the peripheral dynamics sort the T-cell repertoire into two subsets: a less diverse set of small clones of auto-reactive effector and regulatory cells that regulate each other's growth (this corresponds to the CIS in the previous model); and a more diverse set of barely auto-reactive TE cell clones, whose expansion is limited only by APC availability (this corresponds to the PIS in the previous model). To sum up, this model provides an operational mechanism which explains how the repertoire of the immune system can be partitioned into two components, one corresponding to the “self” and the other to the “non-self”; more precisely, how this demarcation could actually come about in the course of an autonomous ontogeny, without the providential help of an omniscient external agent. Perhaps the most significant point is that tolerance to body antigens is dominant, and so it cannot possibly be explained on a “one-by-one” clonal basis and certainly not by the elimination of lymphocytes that interact with the body. On the contrary, tolerance is primarily the result of a positive, active process involving lymphocytes which interact productively with the body. Francisco Varela liked to say that the central, primordial function of the immune system is that of “self-assertion”; and this insight, a basic renewal of perspective that came from the deployment of the concepts of autopoïesis in the field of immunology, has now been substantiated. To harp back to the leitmotif that is one of the main themes of this book, this has been possible because the work involved in developing and refining empirically refutable hypotheses has been done.

The enaction of social life


In the general introduction to this book, we noted that in certain circles there is a feeling that the paradigm of enaction, based on the “cognition = life = autopoïesis” theme, may be all very well for “low-level” sensori-motor cognition; but that when it comes to “high-level” human cognition (innuendo: real cognition), then the classical Computational Theory of Mind, dealing with the manipulation of symbolic representations in the brain, is still the only game in town. If Enaction is to gain the status of a fully-fledged paradigm for Cognitive Science as a whole, it is clearly necessary to remedy this situation. There are several possible routes towards this goal. One is the scheme of “neurophenomenology” inaugurated by Francisco Varela (1996)[*], which squarely addresses the question of human consciousness, and relates it (in a non-reductive fashion) to neurological states using recent techniques of brain-imaging. I shall not say anything more about this route here; instead, I wish to explore a different route, complementary to that one, which concerns the domain of human society.

The fact is that, to date, Cognitive Science in all its guises has not yet really done justice to the social dimension of human cognition. The now traditional Computational Theory of Mind (CTM) does treat characteristically human high-level cognition; but it does so by considering that cognition takes place principally in the heads of individuals, involving the manipulation of symbolic representations. Thus, when it attempts to address the social dimension, it does so by adopting a form of “methodological individualism”, the doctrine which “insists that the ‘behaviour' and the ‘actions' of collectives, such as states or social groups, must be reduced to the behaviour and to the actions of human individuals.” It is indeed a category mistake to attribute to entities (such as collective institutions) properties and capacities (such as desires, intentions and actions) which properly belong only to individuals. The problem is that it is also a mistake to slide from there to the position that only individuals are ultimately real, social collectives being “mere fictions”. Thus von Mises (1949[*]) writes that “The hangman, not the state, executes a criminal. ... For a social collective has no existence and reality outside of the individual members' actions. The life of a collective is lived in the actions of individuals constituting its body. ... There is no substratum of society other than the actions of individuals.” All of these quoted statements are ontological rather than methodological in character. Other authors, notably Dupuy (1992[*] and 1994[*]), argue for a “Complex Methodological Individualism” which takes into account complex emergent phenomena. However, the basic weakness remains; and as a result, the CTM does not properly take into account the reality of social structures which are, as we shall see below, actually constitutive pre-conditions for the genesis of human cognition.

The fundamental importance of social structures as a condition for human cognition is vividly illustrated by the cases of human infants abandoned by their peers and raised by wild animals. Even though the reports of such cases (be they avowedly fictional or purportedly authentic[31]) are not totally reliable, the over-riding impression is unmistakeable: not only are such children unable to speak, but their general behaviour and cognition is much closer to that of their animal foster-parents than to that of individuals in any known human society. Another indication of the profound relation between forms of cognition and forms of social life, less dramatic but more certain and still very striking, is provided by the comparison between “primordial” prehistorical societies (represented today by Australian aborigines and certain Amazonian tribes) and our own “civilized” societies, which are notably characterized by the growth of scientific knowledge which is inseparable from a certain form of social relations. As I have said, the CTM is demonstrably weak in this area; but the truth is that up until now, the paradigm of Enaction has not done much better. The grounding of cognition in autopoïesis and embodiment does of course occur at the level of the individual. The phenomena which have been studied at a “collective” level form a rather motley list of phenomena such as: collective intelligence in ant colonies; wolves hunting in packs; communities of apes; relations between a domesticated dog and its master; empathy between a mother and her child; conversations between friends; traffic jams; interactions between artificial creatures in experimental or artificial environments; and so on. These phenomena involve varying numbers of individuals, ranging from two to several to many; but what they have in common is the feature of focussing on the interactions between these agents. What is conspicuously lacking is any substantial examination of social structures – institutions, norms and so on – which as we will see below are the salient feature of human societies. This deficiency is all the more disappointing because the very concept of Enaction – the process of bringing forth a realm of reality – is intrinsically particularly well suited to the study of human society. “Social reality” is par excellence a phenomenon which is enacted. In this chapter, my aim is to make a start towards remedying this deficiency.

Fundamental theories of society

I consider that it is important to start out by formulating a fundamental theoretical definition of “society”. It is one of the strong points of the paradigm of Enaction that it is grounded in a strong theoretical definition of Cognition as an essential feature of living organisms; and in turn, on a strong theoretical definition of “Life” as basically an autopoïetic from of organisation. Already in the field of biology, one can appreciate the importance of this if we look at the current state of research in this domain. François Jacob (1970[*]), with characteristic lucidity, has famously noted that “life is no longer an object of study in the laboratory”. The consequence of this is that the central object in contemporary mainstream biology is DNA, and notably the rather mindless project of sequencing genomes. Living organisms, as such, are indeed no longer an object of study; and this is because living organisms as such are not properly constituted as scientific objects. Coming back to the question of society: it is in analogous fashion that if, as we have seen, “society” as such is not properly an object of study in cognitive science, this is related to the lack of a strong theoretical definition. Hence the importance of this question.

Now in the endeavour to provide a strong theoretical definition of “society”, it would be misguided to try and reinvent the wheel from scratch. The social sciences in general, and sociology in particular, have addressed this question for over a century; in order for Enactive Cognitive Science to do justice to the social dimension, which is constitutive of many characteristically human activities, it will be necessary to incorporate insights from the sociological tradition. To give an immediate idea of what is involved, we may note the following points:

  • A truly social domain is always defined by a set of structural norms ;

  • Moreover, these social structures are not only a set of constraints, they are also and above all enabling: they actually constitute the possibility of “enacting worlds” that would just not exist without them.

  • Interactions between two or more agents are never properly “social” unless they take place in the context of an environment of social structures or norms which give meaning to the interactions, not just for an external scientific observer but for the actors themselves.

  • This feature of being actively embedded in a shared normative order, independent of any particular individual minds, is a major element in any fundamental theory of the social.

Two currents of thought in sociology

Is there a fundamental, qualitative difference between the « natural sciences » and the « human sciences », not just in their objects but in their methods? Ever since the inception of sociology as a scientific discipline towards the end of the 19th century, there have been two major strands in sociological thought, which adopted different stances on this question and gave rise to what was known as the “Methodenstreit” or the “quarrel of methods”. These two currents are:

« Explanatory » sociology

Historically, the leading figure in this current was Emile Durkheim (1858-1917), who had the avowed ambition of founding sociology as a scientific discipline fully on a par with the natural sciences (although Marx may also be considered as more than a mere precursor). In line with this, Durkheim advocated the method of “explaining” (Erklären in German), which is indeed akin to the method of the natural sciences. In his major book “The Rules of Sociological Method”, Durkheim(1894[*]) famously proposed that “The first and fundamental rule [of sociology] is to consider social facts as things... A social fact is every way of acting, fixed or not, capable of exercising on the individual an external constraint; or again, every way of acting which is general throughout a given society, while at the same time existing in its own right independent of its individual manifestations”. As an example, Durkheim takes the case of suicide (1897[*]) – deliberately chosen because on the face of it, suicide would seem to be an intimately personal, individual act. But Durkheim uses statistics to show that the suicide-rate is very significantly different between Protestants and Catholics, and argues that this is the result of a difference in cultural norms and the level of social control.

« Comprehensive » sociology

The major figure in the other current was Max Weber (1864-1920). According to Weber, human beings interpret their situation, and they themselves attribute subjective meanings to their actions. This feature is ontologically constitutive of human action (in a social context); and because of this, sociologists cannot afford to deprive themselves epistemologically of the resource of mobilizing their own subjective understanding of the situations and the actions. The counterpart to the Erklären of the first current is thus the method of Verstehen, understanding. Another important figure, generally considered as belonging to this “interpretative” current, is Georg Simmel (1858-1918).

A synthesis

Largely because of this historical background context of the « Methodenstreit », for over half a century there has been a tendency to consider that these two currents in sociological thought were opposed to each other. This tendency was shared and indeed promulgated by the protagonists themselves, and spread over to the human and social sciences in general. However, taking a step back from these polemics, it appears that an interest in individuals and small-scale interactions is in no way incompatible with taking into account large-scale social structures and their historical evolution; on the contrary, these two aspects can be seen as complementary. Each of them taken alone is incomplete; but actually there are no insuperable problems in putting them together in an articulated way. As shown in figure 19, the basic scheme for doing so is to thematize a circular relationship between large-scale social structures on the one hand, and small-scale interactions or individual actions on the other. Thus, the social structures do not exist in themselves, in isolation; they are the result of permanent processes of renewal and construction which can indeed be ultimately traced back to individual actions (arrow (a) in figure 19). In other words, without the actions there would be no social structures. But at the same time, these social structures (norms, institutions and so on – we will take a closer look at some specific examples below) provide the very condition of possibility for the small-scale actions (arrow (b) in figure 19). These social structures often go unnoticed, being taken for granted by the individual actors; but in fact they not only constrain the actions (an aspect which is relatively noticeable) but actually and the actions of individuals or small-scale interactions.enable them.

The vast majority of human actions would not even be conceivable without this large-scale background. A synthesis of this sort has been explicitly proposed by various authors, in particular Berger and Luckmann (Berger & Luckmann, 1966)[*] and Giddens (1984[*]), which has given rise to what is known as the “new sociologies”.

In his book The Constitution of Society, Giddens (1984[*]) notes that there is a difference in kind between society and nature: nature is not made by man, is not produced by man; whereas society is. This production of society is a performance which requires a great deal of skill; it is only possible because each member of a human society is a practical social theorist who draws on resources that cannot be “corrected” by sociologists; on the contrary, sociologists themselves draw on these resources in each of their studies. This is clearly in resonance with Max Weber's “comprehensive sociology” mentioned above. But Giddens' whole effort consists of articulating this aspect – understanding and reflexivity on the part of human individuals – with the question of large-scale social structures. He writes:

“The constitution of agents and structures are not two independently given sets of phenomena... According to the notion of the duality of structure, the structural properties of systems are both medium and outcome of the practices they recursively organise. Structure is not ‘external' to individuals: as memory traces, and as instantiated in social practices, it is, in a Durkheimean sense, more ‘internal' than external to their activities. Structure is not to be equated with constraint but is always both constraining and enabling. ... The routinized character of the paths along which individuals move in the reversible time of daily life does not just ‘happen'. It is ‘made to happen' by the modes of reflexive monitoring of action which individuals sustain in circumstances of co-presence.”

In their book The Social Construction of Reality, Berger and Luckmann (1966[*]) also note that social reality has a dual aspect. On the one hand, social world – the one that we know and perceive as « obvious » in our everyday life – is in fact the objective product of human activity over the course of history. It is thus clearly « constructed ». On the other hand, however, each of us routinely perceives this social world as exterior, as a « social reality » which has an existence independent of our will; in other words, it appears to us as « non-constructed[32] »! Berger & Luckmann ask how it is that human activity produces a world of « things »; and what is the role of knowledge in our perception of this world as « natural » and « obvious »? They answer by making the following points:

  • Daily life is the « supreme reality »; in practice, as a socialized human being, I cannot doubt it. I engage in regular routines, which become un-noticed habits.

  • Moreover, I share this daily life with others (which is not the case with dreams, for example); the meanings and the practical certainty in the reality of this world are shared.

Berger and Luckmann consider that these two features – routines, and the fact that social life is shared – explain how it is that this world of social reality comes to be perceived as natural, as taken for granted. This makes it possible to understand the process of the « social construction of reality ». The process has three aspects, which are dialectically inter-related. Berger and Luckmann sum this up by the formulae:

  • Society is a human production (externalization);

  • Society is an objective reality (objectivation);

  • Man is a social production (internalization).

We can sum up by saying that this sort of synthesis between the two traditional currents in sociological thought means that there is no call to oppose « macro-social structures » as against « micro-social actions and interactions ». On the contrary, these two aspects belong together; so we can serenely take a closer look at social structures, not only because they are important and interesting in themselves, but because they constrain and above all enable individual actions and small-scale interactions in human society.

This completes what I have to say here as a brief introduction to the theme of “Fundamental Theories of the Social”. We shall now take a closer look at various sorts of social structures.

Institutions, roles and norms

The most basic type of social structure is an institution. In our present societies, the primary institution is the nation state, with the accompanying governmental machinery of the various ministries, which typically cover the areas of foreign affairs (diplomacy and armies); interior affairs (police, judicial and penal systems); financial systems (taxes, the regulation of currencies, banks, etc.); educational systems (schools and universities) and scientific and technological research; health-care systems and social security; industry and agriculture; transport and housing systems; culture, communications systems, media, literature and art, sport; and so on. This is accompanied by a system of local government, typically centred around a town-hall and a local council. At a more general level, institutions include social structures such as cultures and traditions; languages; mechanisms for organising the division of labour, including market economies; religions; and so on, this list makes no attempt at being exhaustive. An important aspect is that these all these institutions create a large number of social roles: being a mayor, a citizen, a spouse, a criminal, a believer, a general, a student, a borrower and so on. Being a seller, a cheater or a guest are social roles delimited by constraints that are resources at the same time. By occupying a social role, individual behaviours become acts that are generally meaningful and intelligible. Social roles are often complementary: many roles are defined by their relations to other roles.

In the sociological literature, another important aspect of social structures lies in their functions as norms, which prohibit and/or prescribe a wide variety of social phenomena: ways of speaking, norms of understanding, principles of action, habits of thought, modes of behaving and patterns of public behaviour, models of interacting, categorical types, frames of interactions, common sense and common knowledge. It is the fact of following such norms which transforms what would otherwise be merely “behavioural interactions” into socially meaningful acts endowed with meaning. This is the most basic form of “sense-making” that we engage in; right from the start it is thoroughly social, because the very frames, categories or norms that we use to identify actions are themselves entirely social, being made of constraints and “oughts”. Nothing that we do, as humans, is exempt from this: like a shadow, it follows us wherever we go. Right from “getting up in the morning”, to “getting dressed”, to “going out” and so on throughout the day - everything we do makes sense precisely because it is interpretable as relying, exploiting and respecting shared structural constraints. The “behaviours” of socialised humans are not responses to external constraints; they rather involve normative responsibilities, where we rely on structural constraints that give us freedom to achieve meaningful and mutually intelligible actions. Other examples are: being invited for dinner; waiting in a queue; buying stamps in the post office; using public transportation; engaging a conversation; attending a conference; going to the doctor; writing a scientific paper... Often (although not necessarily) these acts involve common values: what “ought” to be done, according to the norm, may be justified by what it is good to do. These norms are not natural or law-governed constraints; it is perfectly possible for agents not to respect them – but at the price of being social outcasts... and no-one wants that!! These norms are not mere regularities, since they have a crucial prescriptive dimension; but they generally give rise to regularities that seem natural to those following them. When they are made explicit, norms may take the form of prescriptive/proscriptive/permissive statements; but most of the time they are implicit (and all the more powerful for that). We are actually most aware of norms when some behaviour goes against them and when we feel that something else should have been done. How do the “sanctions” that enforce this normativity actually work? It cannot be a question of actual physical constraints – no police force could begin to be sufficient if the norms were not already massively “interiorized” as a question of “moral values”. Ethnomethodology (Garfinkel, 1967)[*] has shown that social normativity begins at the grass roots, in the most mundane and ordinary events of everyday life. The basic sanction is (simply!) that if you do not conform to a social norm, people will look at you as “weird” – and no one wants that!

We will now look in more detail at three social structures in particular, that in our view are particularly important for deepening our understanding of the social dimension of cognition. These are: language (Language, writing and inscriptions), technical systems (Science, Technology and Society) and monetary systems (The history of social synthesis by market mechanisms). Each of these dimensions is clearly fundamental in making human life in society qualitatively different from anything that happens in animal worlds; and it is thus essential to examine them if we wish to develop an adequate approach to “high-level” human cognition.

Language, writing and inscriptions

A theoretical definition of language

Language is not by any means the only constitutive institution in human society; but it is such an important medium of human practical activities, and above all it is so pervasive in its effects, that we must look at it more closely. We will start out by providing a definition. Language is a form of communication; but it is quite different from “animal communication”. Maturana and Varela (1980[*]) define “communication” as: the emission of a “signal”, and the modulation of their behaviour by other organisms on the reception of such signals, and the fact that the conditions which trigger the signal emission of signals together with the effects of the modulated behaviour, lead to a coordination of the behaviours that contributes to the viability of the interacting organisms. The evaluation as to whether the coordination “contributes to viability” is made by an external observer. Thus, the organisms do not necessarily themselves have any understanding of what they are doing; arguably, in the vast majority of biological cases at least, the organisms “know how to communicate”, but indeed do not know that they are communicating. In “animal communication” thus characterized, the emission of signals and the behaviour triggered by perception of them are both stereotyped reactions that are typical for all normal members of the species; they are functionally adapted and effective because they have been subject to natural selection over an evolutionary scale of time.

This contrasts strongly with the situation of human social life. Human language is dramatically not stereotyped. Firstly, because of the combinatorial mechanisms at work (phonemes or letters into words, words into sentences), the number of different “signals” is stupendous. The number of different semantic meanings is even greater. Considering the word as a unit, the meaning of a word can vary according to its linguistic context (the neighbouring words with which it is combined) and even more according to its pragmatic context. Taking this into due account, one could seriously put forward the hypothesis that no word has ever been used twice to mean exactly the same thing.

This, however, immediately raises a problem. Animal communication, as characterized here, functions (without necessarily involving any conscious understanding) because it is stereotyped. If human language is not stereotyped, how do human beings ever communicate correctly by talking? A part of the answer is that in general we probably understand each other far less than we fondly imagine. Garfinkel (1967)[*], in his foundational work in ethnomethodology, impishly pointed out that in the course of normal conversation, the socially acceptable thing to do is to accept to have only a very vague and imperfect understanding of what is actually being said, and riding the wave of good faith that things will become “sufficiently clear” as we go along. Arguably, some of the most significant moments of communication occur when speakers identify a misunderstanding; paradoxical though it may seems, what happens is that they then realize that up until that point, they had been misinterpreting each other (with the best of intentions, of course). My point here is not nihilistic; I am not saying that we do not understand each other at all, only that our understanding is not, and cannot be, “100% perfect” as the “information-transfer” model in classical CTM-based cognitive science would suggest.

If we accept that a verbal utterance radically underdetermines the meaning to be communicated, how can some degree of communication nevertheless occur? This is where the intention to communicate comes in. Firstly, the hearer puts great creativity into inventing, imagining, guessing what the speaker might be trying to say. Of course, this is (at best) a hypothesis; the communication can only be consolidated if there is some feedback. This is why such phrases as: “Do you mean that....” (followed by a paraphrase); or “I don't understand what you mean at all, please say it again”; or (sometimes) “Yes, yes, I see, go on...” are so common in ordinary conversation. It is to be noted that these meta-linguistic messages – absolutely vital for linguistic inter-comprehension, on this account – are often operationally performed by facial gestures and mimics: a frown, a deliberate silence, a nod of the head, winking the eyes, and so on. Such gestures are not usually counted as “linguistic” (they are not words); but if this theory is right, such meta-linguistic signals are actually at the core of what is characteristically linguistic. To sum up, linguistic communication is governed by a (mutual) intention to communicate. As Maturana and Varela (1980[*]) have clearly explained, language is a second-order communication; “communication about communication”.

There is thus a sharp contrast between language as thus defined, and “animal communication” where by and large the participants do not even notice whether they are effectively coordinating their actions. Common parlance uses the term “language of the bees” to denote the “dance patterns” whereby a bee indicates the direction and distance of a source of nectar to other members of the colony; but with respect to the term “language” as we define it here, this is a misnomer. We may note that it is also common to employ the “social insects” to designate colonies of bees and ants; but in relation to the way we are attempting to define the term “social”, this is also a misnomer. There is such a gulf of difference between “animal communication” in insect colonies, which is characterised by a lack of reflexive understanding on the part of the interacting individuals, and the “social life” of humans which is constitutively characterised by interpretative understanding and meaning, that we consider the use of the same term introduces, quite unnecessarily, damaging confusion. Animals just do not engage in the sort of normatively constituted activities outlined above.

Language is not simply a typical social structure; it is central, and pervades everything else. It is nevertheless clearly a “social structure” as we define the term here; and it does provide a nice illustration of the relationship between the individual and the social. « Living languages » such as English and French only exist because there is a community of individuals who speak them. (cf. figure 19, Actions continually re-producing the Social Structure). And under the influence of speakers, a “living” language does evolve over time. At the same time, for each individual speaker, the language in question is « always already there », providing a set of constraints that the speaker cannot afford to ignore on pain of not being understood (i.e. failing to speak). Moreover, it is very clear in this case how this set of structural constraints is also enabling, opening up a whole realm of possible actions (and giving meaning to those actions) that would simply not exist if the social structure in question were not in place.

This all-pervasive socialisation of human cognition is more apparent than ever when we consider social reproduction, the passage of generations. Human infants are, in an important sense, “socialised” before birth, and indeed before they are even conceived, because they are thought about and talked about and already integrated into a meaningful set of social relations. The process of socialisation through language becomes even more salient from the age of about four or five, when children themselves begin to understand what is going on - and, of course, to incessantly ask questions: “Mummy, what is that man doing?”, which is a request for exactly the sort of interpretation, construing behaviours as actions, that we see as one of the central features of social life.

Language as it develops in an oral culture is, however, only the beginning of the story. We shall now look at the changes, both cognitive and social, which are brought in with the invention of writing.

The domestication of the savage mind: the invention of writing

One of the most remarkable events in the intellectual history of mankind is the one that has been dubbed the “Greek miracle”. In the space of only a few centuries, there was the invention of philosophy (Thales, Parmenides, Heraclitus) with its passage to maturity (Socrates, Plato, Aristotle); and there was the invention of mathematics (Pythagoras, Euclid). These intellectual exploits were not isolated; they occurred in the context of other major inventions: a market economy, with the invention of coined money (a point I shall return to later); the invention of democracy as a political system; and finally the invention of alphabetic writing (a theme which will form the core of this section). There is a very real sense in which what occurred then was the “entry into history”, not just because of the great historians (Thucydides, Herodotus), but because it seems clear that these various exploits form a systematic whole; and this system is the very basis of our own contemporary culture, so that everything that preceded it appears, literally, as belonging to “pre-history”.

This feeling was reactivated and reinforced two thousand years later in the historical movement explicitly known as the “Renaissance”. The context was that of the great explorers, Columbus, Magellan and other Europeans, who “discovered” the rest of the world and opened the way to colonisation. The reference to the “Greek miracle” then served an ideological function, both explaining and justifying colonial domination. This gave rise to the theory of a “Great Divide”, an insuperable chasm between the culture of “civilised men” and that of “wild savages”. Goody (1977[*]) quotes Lévi-Strauss and notes the following list of dichotomies which are held to distinguish “domesticated” from “savage” forms of social life and thought:

Table 5





Science of the abstract

Science of the concrete

Scientific knowledge

Magical thought



Abstract thought


Using concept

Using signs


Atemporality, myths and rites

Today, political sensitivities have shifted and the ideology of a “Great Divide” between “superior” and “inferior” human beings (which, it must be remembered, justified slavery and the direct or indirect genocide of native populations) leaves an uneasy feeling. This has lead to a “cultural relativism” which, if it does not deny any differences at all, seeks to minimize their import. The aim in this section is to take a step back from these politically value-laden questions, and to situate “domesticated” and “savage” forms of social life and thought as stages in historical development, in particular by seeking to analyse the mechanisms which made it possible to pass from one to the other. We will largely follow the proposal of Goody (1977[*]), in his attempt to disarm the Manichean simplifications by looking more closely at the cognitive effects of the invention of writing.

Writing was invented not by the Greeks, but around 3000 BC in Mesopotamia. In its earliest form, in Uruk, it consisted of clay tablets that were attached to objects in order to identify their owners; later, the objects were represented by signs, which made it possible to detach the tablets from the objects. This system owes its origin to administrative and economic needs. The gains in productivity resulting from State-controlled systems of irrigation gave rise to an agricultural surplus; this surplus had to be stocked in warehouses in town, and redistributed; hence the need for a system of accounting. Two things are worthy of note here. Firstly, writing was right from the start inseparable from a system of social relations themselves structured by technology: this social system was both the reason for the invention of writing, and in return writing aided its development. Secondly (and quite contrary to what many contemporary linguists assume), writing is not a simple derivative transcription of spoken language; right from the start, writing is an integral part of cognitive operations which would be simply impossible on the sole basis of spoken language.

The earliest form of written language is thus the list – catalogues and inventories of persons, objects and events. One can distinguish three types of list: retrospective, prospective and lexical. Retrospective lists of events can be organised either by reference to episodes (chronicles) or by reference to the calendar (annals); together, they constitute the documentary archives which are indispensable for the emergence of “history” in the modern sense of Thucydides and Herodotus. Prospective lists are a key tool for an important form of cognition: programming and planning action. One of the important functions of reflexive consciousness is that it creates the possibility of imagining several possible future scenarios, and to choose one after envisaging their probable consequences; the practical development of programmed action is multiplied by use of the written list, as is illustrated by the contemporary examples of shopping lists and cookery recipes. In Mesopotamia and Egypt, there were annals of astronomical observations combined with the height of the rivers – which aided the development of irrigation, yet another example of the synergy between writing and technological development.

Lexical lists were initially less frequent than administrative lists; but later they became very important, notably in Sumer and in Egypt. This third type of list is particularly interesting, because it illustrates the endogenous dynamics set in motion by the development of writing. These lists typically appear in educational institutions, for teaching purposes; in other words, there is an effect of decontextualization with respect to immediate practical needs. It is to be noted that the simple fact of writing a list of words induces cognitive effects, in particular categorisation.

How can the words in a list be grouped? The grouping can be thematic, related to the properties of the objects. Thus, in the temple school of Nippur, there are many such lists of trees (84), stones (12), gods (9), officials (8), farm animals (8), reeds (8)... Landsberger considers that the large number of these lists results is a consequence of the nature of the Sumerian language, which has a transparent and non-ambiguous structure particularly suitable for classifying the world. Goody suggests that the relation is at least as much the reverse: it is the practice of constituting lists which influenced the language by forcing it to become less ambiguous.

Alternatively, the grouping can be organized on the basis of the form (and not the semantic content) of the words themselves. This principle is clearly at work in the Mesopotamian and Egyptian lists. The first systems of writing were pictographic or hieroglyphic; but in this case, it is difficult to order the signs on the basis of their graphical form. In the archaeological record, it is found that the signs were ordered according to the similarity of the sound (in particular, the initial sound). The key point is that this manner of ordering induces an evolution in the system of writing itself: from pictographic to syllabic, then consonantic, and finally alphabetical. In other words, alphabetic writing is not the simple result of phonetic transcription; it is the result of a systemic evolution internal to the process of writing itself. Thus the invention of alphabetic writing – generally attributed to the Greeks – is not a sudden “miracle” mysteriously fallen from the heavens; rather, it can be understood as the logical result of a process spanning thousands of years.

At the same time, without being miraculous, the invention of alphabetic writing had in turn some very important effects. Since it is, indeed, a phonetical coding, it brings spoken and written language much closer together; and because of this, writing extends beyond the rather narrow and specific limits of its origins to cover virtually all the domains of language. This extension was helped by the relative simplicity of alphabetic writing. Pictographic, hieroglyphic or ideographic systems contain thousands of different signs that must be memorized; because of this, the access to these systems is inevitably restricted to a small caste of scribes, whereas alphabetic writing can (in principle at least) be made accessible to all members of a society. Thus, it is with the advent of the alphabet that writing fully invests the genres of narration and story-telling (fictional or otherwise), of poetry, of dialogue and monologue (including the interior monologue/dialogue that we call “thought”). Now however paradoxical it may seem, it is at the very moment when written language comes the closest to spoken language that the originality and the specificity of its contribution to cognition become most apparent. Without attempting to be exhaustive, I will examine this contribution in three domains of enormous cognitive import: history, philosophy and mathematics.

The contrast between “myth” and “history” is one of the major headings in the series of oppositions between “savage” and “domesticated” thought. Goody argues that this opposition is in large part due to the contribution of writing. We have already remarked that the accumulation of documentary archives – that are only possible because of writing, obviously – is the basic condition which makes the work of a historian (in the modern sense of the word) possible. We have already noted that the narrative dimension is made possible in written form by the invention of the alphabet; it is then possible to juxtapose different versions of the “same” story, and to compare them in detail to identify on one hand the convergences and confirmations, on the other the divergences and contradictions. In the heat of the moment, in the real time of the chanting of a ballad by a gifted orator, all sorts of collective emotional effects are possible. By contrast, writing is structurally individual and private, both during reading and the writing itself; it is thus an instrument which induces a critical distance, which dissipates collective emotion and promotes what is called “objectivity”. It is not a question of reducing the difference between “myth” and “history” to one simple cause; it is not writing as such, considered in isolation, which mechanically and inevitably produces all these effects. The causal relations are not linear, but circular and complex: the introduction of writing, limited in the first instance, produces certain results which are also limited; but these results have feedback effects on the practice of writing, and lead to its extension. At the culmination of this process, writing appears as an integral and essential part of a major cultural mutation.

These multiple cognitive effects of writing are not limited to the passage from “myth” to “history”; they also illuminate what has been called the “invention of philosophy”. Goody notes that writing has the instrumental effect of spatializing language and, thus, of rendering linguistic statements visible. This change in modality produces a qualitative effect. When one observes a written inscription, one can look at it in all directions, for as long as one desires; this is not at all the case with a phonetic statement whose trace is intrinsically transitory. It is thus writing which gives the force of apodictic conviction to a syllogism: it is, indeed, the fact that one can continue a sceptical examination of each step in the argument for just as long one likes, and come back to the argument to re-examine it at will, which in the end overcomes all resistance and conveys a free adhesion to the argument. There is a strong affinity between Greek philosophy and mathematics: the motto “None enter here who are not geometers” is well known. It was by reflecting on the constitution of mathematical entities as pure ideals that Husserl recognized the essential role of writing (Derrida, 1978[*]). We may note that this role functions both at the individual level, but also at the collective, social level: the fact that alphabetic writing can be reliably and controllably copied independently of its interpretation, contributes decisively to the formation of a common conviction that is freely shared. Concerning mathematics, Goody notes the central role played by equations. An equation crucially conserves its truth-value if the same operation (addition, subtraction, multiplication or division) is applied to both sides of the equation; now the point is that this sort of procedure is so greatly facilitated by writing that, in practice, it depends on it. In the same vein, the heart of Greek philosophy resides in the nexus where questions of Truth and Idealities meet; and it is from this base that ethical, moral and existential questions are addressed. In all these areas, the hallmark of alphabetic writing – its capacity to follow all the meanders of thought-processes – is decisive. Western philosophy is par excellence an exercise in the clarification of thought, where the identification of ambiguities and contradictions is structurally essential. This style of thought is, quite literally, inconceivable without the contribution of writing.

This is a key point in the paradigm of enaction as a whole. The question which is posed here is whether material technologies, which are classically considered as empirical and therefore as only constituted, can reach back “upstream” of the transcendental conditions of possibility and thus attain the status of contributing to the very constitution of reality. Now it must be noted that writing is a material technology; and the thesis presented here is that writing is indeed the condition of possibility for the constitution of mathematical and logical idealities. Derrida (1978[*]) has built on the germ of this idea in the writings of Husserl, extending it to the much larger domain of what he calls “archi-écriture”; and Stiegler (1998[*]), in turn, has extended this theme to technical artefacts which form a “tertiary retention”, in other words a form of memory which constitutes human society by providing that which is “always already there” for human beings.

Having introduced the idea that a goodly number of the characteristics on the “domesticated” side of the “Great Divide” can be attributed to the constellation of effects that derive from writing, it is interesting now to turn back and take another look at “oral” societies. A good way of testing the validity of Goody's theses concerning the cognitive effects of writing is indeed to examine how thought is structured in a society without writing. To pose the question with all the brutality induced by the myth of a “Great Divide”: can there be genuine intellectuals in an illiterate society?

We may start with the question of mathematics. Oral societies are far from being bereft of arithmetic. A good example is given by the Ghanaian tribe of the LoDagaa, who until 1949 had no contact with literacy; Goody was able to make an in-depth study of their practice. Goody noticed that when it was a question of counting large number of sea-shells (which were used as currency – a dowry could for example be 20,000 shells), the LoDagaa went much faster than he did. Whereas Goody moved the shells one by one, the LoDagaa had a system which consisted of moving a group of three, and then of two, which together formed a pile of five. The number “five” is both an exact fraction of twenty (the basis of subsequent calculations), and a number that can be verified by a glance even while moving the hand to form the next pile – which increases the reliability of the counting. Four groups of five then form a group of twenty; five groups of twenty make a hundred; and so on, until the dowry was counted.

The counterpart of the rapidity and precision of this procedure is that it is specific and concrete. The first time Goody asked his LoDagaa informers how they went about counting, they asked a question in return: “counting what?” Of course, the procedure for counting cows could not be the same as that for counting sea-shells. Abstract concepts are not completely absent for the LoDagaa; they do posses the vocabulary of a numerical system which applies to cows and to sea-shells; and the method of regrouping four groups of five to form a group of twenty, and so on, does indeed correspond to the “abstract” operation of multiplication. However, these concepts are, each time, deployed at the level of concrete operations in daily life. Abstraction as such only develops in the very particular “context” of literacy learning in a classroom, where the students are required to perform the operation “4 x 5” in the absence of cows, sea-shells or anything else. The example of the LoDagaa is eloquent, because one sees both what is valid in the notion of a “great divide”, but also the sense in which it is mistaken. The difference resides not so much in the “mind” or a “mentality”, but much more prosaically in the material techniques of the intellect.

Another aspect of intellectual activity in oral societies concerns the opposition between “myth” and “history”. From an objectivist point of view, “history” simply (?!) consists of relating the objective facts as to what actually occurred; so that no subjective creativity is required for its elaboration. By contrast, a myth is precisely not “objective”; its elaboration cannot therefore be considered as the transcription of non-existent facts, but must call into play a process of invention, of creative imagination. The question then arises as to the authors of this invention. According to the most widely held view (usually shared by the actors themselves), myths are “traditional”, in other words they are transmitted unchanged, from generation to generation, since the beginning of time; the role of the story-tellers is thus reduced to a simple (?!) question of memory. Now, even putting aside the fact that this “explanation” leaves quite unsolved the question of ultimate origin, it does not resist critical examination. First of all, the fact of memorizing exactly, word-for-word, an epic ballad tens of thousands of lines long, would in itself be an astonishing cognitive feat bordering on the impossible. But the question that arises is then this: in the absence of a written text to serve as a reference, how is it possible to know whether the myth really has been transmitted word-for-word perfect? Certainly, the myth is perceived by the listeners as being “the same”; but since neither the listeners, nor the story-teller, have any sort of actual recording, this does not mean much. Goody (1977[*]) settled this question by making several different recordings of the Bagré, the main myth of the LoDagaa. It turned out that these recordings, made at an interval of only twenty years, revealed that there were in fact very substantial variations: several elements which seemed quite essential in the version recorded in 1951 had completely disappeared from the 1970 version, being replaced by others. Even more surprising, Goody collected a dozen recordings of the first twelve lines, which form an incantation well-known to all members of the tribe; and it turned out that no two of these recordings were word-for-word identical.

After the event, it seems obvious that an “exact” reproduction, even if it were possible, is neither necessary nor even desirable. A study by Lord (1960[*]) of the apprenticeship of young story-tellers in an oral culture shows that it is not a question of memorizing verbatim all the words, but rather of developing a capacity to improvise a story on the basis of phrases, themes and recitals that the learner has already heard. The bard cannot remember enough to sing an entire song, and he does not; he has to learn how to created phrases, and this is indeed what he does. Closer to us, this is indeed what occurs in jazz music: contrary to what one might imagine if one has been trained in classical music where one plays on the basis of a written score (and here one should play each note as it is written), a jazz musician will never play exactly the same piece note-for-note. On the basis of an understanding of the harmonic structure, of cadences, of scales, the music is essentially an improvisation. If one has not practised this art, which is essentially an oral culture, this capacity for improvisation may seem mysterious, almost magic; but in fact we are all virtuosos – if not of music, at least of stories. For example, each of us would be quite capable of improvising a version of Little Red Riding Hood or Goldilocks and the Three Bears. How would we do it? Well, on the basis of a vague memory of the gist of the story, the climax and the way it ends, and fragments of key phrases (“Oh, Grandmother, what big eyes you have – All the better to see you with, my dear.... What big teeth you have....”), one launches forth: “Once upon a time....”. Especially if there is an attentive and expectant audience, the story creates its own dynamic. This sort of improvisation does of course improve with practice – and this is exactly what occurs in oral cultures.

This brings us back to the question of the existence and the nature of intellectual creation in non-literate cultures. Far from being absent, or relegated to mists of time, creativity is in reality omnipresent. In a way, this creativity is less visible precisely because it is distributed everywhere. In fact, the attribution of a work to one well-identified individual author is actually an effect – an artefact – of writing. In an oral culture, there is no distinction between the author and the performer – they are one and the same. Moreover, an innovation introduced by a story-teller will be adopted (or not) more or less in real time, on the occasion of subsequent performances; if the innovation is retained, it will “melt” into the whole without leaving any clear trace; in an oral culture, it is difficult to attribute creativity to any particular individual. The “roles” of author, the performer and the audience are hardly differentiated; in one sense nobody is an author, but in another sense everyone is. Thus, with the invention of writing and the passage to a literate culture, some things are gained but others are irremediably lost. In general, with each technical invention (writing is but an example, albeit a major one), what “the world” is for human beings becomes more or less drastically modified; but it is not correct to consider that there is regular, unambiguous “progress”.

We are far from finished with the question of the cognitive effects of writing and, more generally, of inscriptions. In particular, this will form a major theme when we seek to understand how “science” and “technology” are allied under the heading of “technoscience” (see Cycles of accumulation).

Science, technology and society

In this section we will look at a set of related phenomena that I have grouped together under the heading “Science and Technology”. These phenomena are most often omitted from studies of social structures by traditional sociology; but they play such an important role in fashioning the society we live in today that I consider it essential to take them into account. The reason for this neglect is probably related to the problem that C. P. Snow identified as the breakdown of communication between the “two cultures” of modern society, the sciences and the humanities; it is a fact that engineers and social scientists do not usually work together[33] . However this may be, the question of the relation between technical systems and the forms of social life is of such central importance in the emergence and historical development of specifically human forms of social life that the attempt must be made to address it. When we begin to get to grips with this question, we will find that there are two complementary aspects: on one hand, there is the structural influence of technical systems on the constitution of society; on the other hand, there is the question of a fine sociological analysis of the actual activities of engineers and scientists. In what follows, we shall be looking at both of these aspects – trying not to forget that they only really exist and make sense when they are put together, as in the spirit of figure 19. First, we will look at Technology as Anthropologically Constitutive; in The social construction of scientific facts, we will switch to a sociological analysis of scientific activity; in Cycles of accumulation, we will attempt to draw various threads together, by analysing how scientific activity (generally characterized as functioning on the basis of inscriptions, typically within scientific laboratories) articulates with what is actually going on in the realities of social life in the “outside world”.

Technology as anthropologically constitutive

In order to introduce this topic, we may illustrate the far-reaching effects of technical systems by noting that archaeologists, basing themselves solely on material remains and in particular on the remains of technical artefacts, are able to reconstitute to an astonishing degree the daily life of vanished civilisations[34]. Another way of illustrating the importance of man-made artefacts is contained in Husserl's concept of Begeistete Objekte, i.e. “spirit-laden objects”. Public objects and their spatial disposition – for example roads, signposts, buildings, doors open or closed, uniforms, counters, tables, coins, seats, bells, and so on – already inform agents about the roles they occupy and what they should do, so that they can enter into interactions on the basis of shared common knowledge about the situation.

To begin at the beginning, the first anthropoid tools for which there is solid evidence are the well-known flint tools made by our hominid ancestors. There is a clear sequence in the fossil record, running from the first “choppers”, crudely made just by banging two pebbles against each other, to increasingly sophisticated and finely-chiselled bi-face tools. The pace of this pre-historical change over the period from 3 million years to about 100 thousand years was very slow, in tune with the pace of anatomical change and in particular the increase in brain-size. Leroi-Gourhan (1964[*]) has remarked that during all this time, “man secreted his tools almost as though they were nails or teeth”, by a process that was therefore largely continuous with biological evolution. There were then two crucial events in the process of hominisation which lead to mankind as we know it today.

The first event, which occurred around 100 thousand years ago, was not (as we fondly like to think) a sudden and dramatic increase in the size of the brain. After the shock of Darwin's suggestion that human beings had common ancestors with apes, the question arose as to the nature of the intermediate stages. Since we are so proud of our brains, the expectation was that this “missing link” would be a creature with a human head on an ape-like body. As Leroi-Gourhan said, “we were ready for anything but this: mankind began by the feet”. Thus Lucy, like other early Australopithecus fossil specimens, had a small head; the new distinctive feature occurred at the level of the legs and feet, which were adapted for standing upright with a biped means of locomotion. This new means of locomotion was remarkably efficient from a mechanical and energetic point of view; but what is even more important from the present point of view is that it freed the hands, thus creating an “anterior field” between head and hand which is a key prerequisite for making and using tools. In other words, we are into the theme of “technology as anthropologically constitutive”; what characterizes human beings compared to their animal relatives, right from the start, does indeed seem to have been an aptitude for dealing with technical artefacts.

The second event occurred around 50 thousand years ago; and it can be seen as marking a “break” in the rhythm between biological evolution, which occurs on a time-scale of millions of years, and human pre-history where the time-scale is measured in thousands of years (KY). Our own species, Homo sapiens sapiens, arose in Africa around 150 KY. The “event” at 50 KY marks an acceleration in the rate of cultural evolution which relegates any further biological evolution to relative insignificance. Symbolically, this event is marked by the disappearance of Neanderthal man, the last surviving species of our numerous biological cousins (the hominids Australopithecus, homo habilis, ergaster, erectus, neanderthalensis, etc.) so that are closest biological relatives are now the great apes (chimpanzees, gorillas, orang-utan). Symbolically also, this is the period of the first cave art.

At the level of technical artefacts, which is our theme in this section, this “event” is characterized by a significant innovation in the making of stone tools. Before, the useful part of the tool was what was left behind by the chiselling process; the innovation consisted of using the chips that were broken off from the main body of the silex material. Thus we see the appearance of the exquisite polished arrow-heads of the Neolithic period. In terms of efficiency, the length of useful cutting-edge per kilogramme raw material (the brute flints were a relatively rare and precious resource) was increased by orders of magnitude. Now in order to produce useable chips, a long and elaborate process of preparing the flint is necessary, before the final blow knocks off the desired chip. From a cognitive point of view, this requires a strongly developed capacity for anticipation. Guille-Escuret (1994[*]) has put forward an intriguing hypothesis to account for this technical innovation. By 50 KY the making and using of tools was already established with its own history (and ethologists have claimed that all sorts of animals, not only apes but also crows and others, make and use tools); on the other hand, language was also developed, with its own history. But there is no evidence of any connexion between language and tool-use. As modern stone-nappers have found, talking is neither necessary nor sufficient when one is trying to make a primitive bi-face. It is plausible to suppose that language was used first of all in the context of personal social relations; and as is still true today, this leads to a run-away effect. As our social life becomes more complex, we need linguistic communication to cope with it; but ironically, one of the main effects of talking about our social life is... to make it even more complex! This independent appearance of tools on one hand and language on the other sets the stage for Guille-Escuret's hypothesis: the “event” at 50 KY could have been the meeting of these two strains, of using language to talk about technical artefacts (making, using and especially inventing them). Etymologically, this corresponds exactly to the birth of techno-logos, i.e. technology.

With this, the stage is set for a whole series of other technological innovations – the use of fire and pottery, weaving, the working of metal (leaving behind the “Stone Age”), then agriculture and the creation of towns. The whole quality of human life is radically transformed. In order to analyse how the introduction of tools and technical systems produced a qualitative change in the rhythm of change, pre-historical and then historical change, it will be useful here to enter into some theoretical considerations. With reference to the basic scheme of sensori-motor coupling between a biological organism and its environment, the crucial feature is the extension of this coupling to include its mediation by technical artefacts (see figure 20).

This schematic figure illustrates the fact that what tools do is to increase the range of possible actions, and to increase the range of possible sensory returns. Thus, human beings live in a world that they themselves have constructed: not just in the sense that they modify their environment (buildings, streets, towns etc), important though this is; but because to the extent that their sensory-motor coupling with the environment is mediated by tools, what the enacted “world” becomes for human beings is largely constituted by these tools. This is, of course, more than ever true today.

Now in order to grasp the full import of the invention of tools, the key point is that a tool (unlike a biological sensory organ or motor organ) exists in two distinct modes, that we may call the “in-hand” mode and the “put-down” mode. When it is being used, the tool itself typically disappears from consciousness; attention is quite naturally focussed on the particular sort of “world” that is being brought about by the successful mediation of the tool. This is the “in-hand” mode. But tools, typically, can also be detached from the body and “put down”. It is in this second mode that they become themselves the focus of attention, and can be repaired, or made, and (most important of all in this context) invented. It is also because tools can exist in this “put-down” mode that they can be used by different individuals; and this also prepares the way for a social division of labour with respect to tool-making and tool-use. Tools, and technical systems more generally, thus have a “dual nature”; and the reason why they introduce an acceleration in the rhythm of change comes from the possibility of a constant back-and-forth between these two modes.

To put human tools in the context of evolution in general, animals do make some primitive tools; but there are at least three criteria which distinguish tool-making by humans and by animals (Lestel, 1998[*]). Firstly, chimpanzees do not make polylithes, i.e. objects which result from assembling several other objects. Secondly, no primate in a natural setting has been observed making a tool destined to be used for making other tools. Thirdly, as we just mentioned, human technology involves the co-operative making of artefacts, which is never observed in chimpanzees. These specific characteristics work together to free human technical invention from pre-established stereotypes. It follows on from this that man-made technical artefacts generate ever more artefacts, so that technical invention becomes a process with its own dynamics where “functional over-determination” brings into play characteristic “laws” of technical evolution (Stiegler, 1994)[*]. “As in a phylogenetic series, each stage of evolution contains within itself the structures and the schemas which form the principle of an evolution of forms. A technical being evolves by convergence and internal adaptation; it unifies itself internally according to a principle of internal resonance” (Simondon, 1989[*]). This “technical tendency” is there right from the start, but it becomes particularly pregnant at the stage of industrial production, and confers on technical artefacts a historicity of their own. It follows that it is not correct to speak of a technical object as a simple heap of inert matter that has been shaped purely from the outside, by an independent organising will (Stiegler, 1994)[*]. This concept of technical objects is thus quite different from the usual position, according to which technical objects are simple utilitarian instruments entirely subject to pre-defined goals. In this context, the concept of a clear and distinct “goal” formed by an isolated self-sufficient individual is a myth. On one hand, “goals” emerge from a social process which is distributed over all the actors and which no single individual can entirely master. On the other hand, technical objects systematically escape from preconceived aims. This is already the case because they have their own internal dynamics, as just explained; but added to this, there is the fact that social agents regularly create and seize opportunities for subverting and re-appropriating previously “standard” uses. The upshot of all this is that the purpose of a technical artefact barely exists in any definite form prior to its being made and used.

The social construction of scientific facts

We turn now to the question of science. It is clear, particularly in our contemporary society, that there is a strong mutual relationship between technology and science. At the same time it is not clear exactly what the nature of this relationship actually is. We will be looking a little later at the historical genesis of science as a social institution: the “birth” of modern science at the Renaissance, prefigured in an important sense by the “Greek miracle” that we have already alluded to in The domestication of the savage mind: the invention of writing. We have seen, in the previous section, that human beings were already invested in the invention of technical artefacts well before the advent of modern science, so science is not absolutely necessary for technological innovation which cannot be reduced simply to “applied science”. But in order to look more closely at the role that science plays in technological innovation – and it does play an important and indeed ever-increasing role – it will first be necessary to dispel some common illusions about the nature of scientific research. The most serious of these illusions is that since science is “simply” (!) about discovering true facts, “sociological factors” are merely anecdotal and do not concern the actual content of scientific knowledge. To dispel this illusion, we will mobilize sociology in the way indicated in Introduction, i.e. as a means for proceeding to a fine analysis of what actually goes on in scientific laboratories.

The “founding father” of the sociology of science is generally considered to be Merton (1942[*]), who examined the social norms which regulate scientific research. Thus, he studied questions such as scientific institutions, the structure of the scientific community, career structures, and the attribution of credit for scientific discoveries. However, he quite deliberately avoided discussion of the actual content of scientific theories, considering that that was the reserved domain of epistemology. A further step was taken when the sociology of science did start to investigate the way in which the content of scientific theories was built – but initially, such studies took as preferred objects of study theories which are nowadays considered to be false, such as the phlogiston theory in alchemy. One can understand the rationale behind this: since such theories are “false”, they cannot be explained as representations of an independent reality – for the simple reason that there is no corresponding reality; and so the field is indeed open for a “social constructivist” account of how they nevertheless came about. A decisive step forward came when Bloor (1976) called for a “strong programme” in the sociology of science, which would quite explicitly eschew this sort of asymmetrical approach, in favour of a “symmetrical” approach to theories that subsequently come to be considered as “true” or “false”. There is a fairly simple argument in favour of this position: at the time when the scientists are actively working on their theory, they do not know (nor does anyone else) whether the theory will turn out to be “true” or “false”; and so there is no sense in having a double standard which could only apply retrospectively. But this approach is anything but trivial, because it does of course open up the possibility of a radically constructivist approach – or, to put it in the terms of this book, to study the enaction of a certain sort of reality by means of scientific research.

We thus come to the question of the social construction of scientific facts – which is indeed exactly the subtitle of a classic study by Latour and Woolgar (1979[*]). This book is based on a series of observations made during a stay of two years in a biology laboratory; in the event, the laboratory of Roger Guillemin at the Salk Institute. The work carried out, for which Guillemin would later receive a Nobel Prize, was concerned with the identification of certain brain hormones. Latour and Woolgar describe their observations as “anthropological” in the sense that is fairly standard in ethnological observations, i.e. they take a deliberately “agnostic” attitude to the belief systems of the human subjects they observe. Here, they describe with great care the process by which a scientific hypothesis, at a certain critical moment in its development, undergoes a transformation and becomes a fact. In conventional epistemology, since Popper, it is clearly established that in principle all scientific theories have the status of hypotheses; what we will examine now is the psycho-social process by which some of these theories come to acquire the status of an “objective fact” which appears to be the faithful reflection of an independent reality.

The starting point for any scientific object, then, is its birth in the form of a hypothesis in the minds of the scientists in question. In a scientific laboratory, the majority of statements are indeed hypotheses: more or less vague, more or less speculative, more or less serious. The effect of scientific activity is to modify the status of these hypotheses, tending either to confirm them or on the contrary to invalidate them. But these modifications are neither definitive nor irreversible, so that each hypothesis follows a fluctuating trajectory. There is thus an essential continuity of the hypothesis over the course of these fluctuations.

Now if one invites a scientist to explain the status of a hypothesis at a given moment in time, he will retrace its history since its birth as a speculative idea up to the most recent modifications. In an account of this sort, the scientists themselves engage in a veritable “spontaneous sociology of science”, wherein various sorts of factors – cognitive, social, psychological... – are all indiscriminately mixed up together. Thus s/he will mention the social context and the subjective motivations which help to understand the birth of a hypothesis; going on to an explanation of its potential interest both for the field of “pure” research and for possible applications; and including a critical evaluation of the “reliability” of the experiments carried out by scientific peers and colleagues, and a comparison with possible alternative hypotheses and interpretations. During this phase, phrases stating the hypothesis are usually qualified by modal expressions such as “believes that...”, “fears that...”, “hopes that...”, “thinks that...” etc. Latour and Woolgar give a telling example from another domain: “X, who had not slept for three nights and who was exhausted, thought that he had seen an optic pulsar”. The allusion to psycho-social factors – here the circumstance that “X had not slept...” – is combined with the insistence on the fact that X thought that he had seen..., and together they strongly convey the feeling that the existence of optic pulsars is only a hypothesis.

The great majority of these hypotheses end up by dying. This death can be violent, if the hypothesis is eliminated by the results of an experiment which decisively refute it, so that it is never taken up again. More often it is a lingering death, by lack of sustenance; if no scientists are sufficiently interested and motivated by the hypothesis to carry out experiments designed to test it and hence to modify its status, it simply fades away into oblivion. As Lakatos (1970[*]) has seen with perspicacity, research scientists are generally pretty pragmatic, and they are less sensitive to the possible “truth” of a hypothesis than to its fruitfulness.

However, in a small minority of cases, a different fate awaits the hypothesis. Following an important experiment, or more often a whole series of experiments, that the scientists in question consider as “decisive”, the hypothesis crosses a threshold... and becomes a fact. Latour and Woolgar, with commendable analytical precision, note that the mechanism of this remarkable transformation comprises three distinct stages. The first stage consists of a splitting. The statement – the set of words (or mathematical formulae) which constitute a formulation of the hypothesis – continues to exist. But what happens, is that this statement of the hypothesis projects a “twin copy” of itself into the outside world; and this “twin copy” takes on an independent existence in the form of a “real object”. From this point on, there are two distinct entities: “the object” and “the statement about the object”. The importance of analytically identifying this initial stage is that at this point, the genealogical relation between the “hypothesis” and the “object” is absolutely clear: it is the hypothesis which is the ancestor of the object, not only because it was there right from the start, but also and above all because it is rigorously impossible to say anything about the object which is not a pure repetition (maybe masked by a paraphrase, but which changes essentially nothing) of the terms of the hypothesis.

It is crucially important to note this first stage, the “splitting”, because it is very rapidly followed by a second event, that Latour and Woolgar call “inversion”. What happens here is that the genealogical relation between statement and object is inverted. As we just noted, initially it was the statement which, by splitting, gave rise to the object. However, very soon, more and more “reality” becomes attached to the object, and less and less to the statement about the object. When this process of inversion has run its term, it is no longer the object which is a perfect reflection of the statement, but the statement which has become a reflection of the “real object”. We end up with that marvellous adequatio rei et intellecticus which has fascinated generations of philosophers... but which is so perfect that one should have been suspicious...

These two processes, the inversion preceded by the splitting, generally happen in such quick succession that it is difficult to detect the sleight-of-hand. The effects of the inversion are then consolidated by the third stage, which is neither more nor less than a re-writing of history. Now, if one asks a scientist to explain why one says that “X observed an optic pulsar”, it is because the pulsar really exists, and X has simply (!) seen it as it is. Concomitantly, all the modalities which qualified the expression of the hypothesis disappear to leave the pure statement of fact: “optic pulsars exist”. The reality of this fact is situated outside human space and time, so that the whole social and historical dimension of the construction of this fact – the local conditions of its initial formulation as a hypothesis, the years of work, the personalities of the scientists and laboratory technicians who contributed, the false leads of alternative interpretations, the controversies, even the precise time and place of the final transformation of the hypothesis into a fact – none of that is important any more. The truth has always been what it is, the real object has always existed, waiting patiently for scientists to come along and discover it.

The power of such a systematic re-writing of history is enormous, and it is practically impossible to resist it. One feels somewhat like Winston Smith in Orwell's novel 1984, correcting the unique copy of the newspaper The Times; Smith is reduced to murmuring under his breath “the only proof is in my own mind”. And yet... if we put everything out on the table, this sort of re-writing, discreetly described as “rational reconstruction”, does not stand up to critical examination, for two reasons. Firstly, the “object” only takes on substantial existence after the processes of splitting and inversion, and so (unless one abandons the idea that causes precede their effects) cannot be the cause of them. Secondly, and even more decisively, the history of science disqualifies this sort of rewritten history, because the transformation between “hypothesis” and “fact” is often reversible – which would be rigorously impossible if the objective facts were actually the cause of the splitting and inversion. This point is so important that it is worth taking a little time to illustrate it.

Those who are familiar with “laboratory life” know well that “scientific objects” are subject to marked fluctuations over relatively short periods. At the frontiers of research, the construction and deconstruction of a “real object” is almost a daily affair:

“Tuesday a peak was considered as the sign of a real substance. But on Wednesday the judgement was that this peak was the result of an unreliable physiograph. Thursday, the use of another extract gave another peak that was taken to be “the same”. At this point, the existence of a new object was solidifying, only to be redissolved the next day.” (Latour & Woolgar, 1979[*]).

Even major scientific facts are not immune from possible deconstruction. Latour and Woolgar cite the example of TRF, a brain hormone. In 1969, when Guillemin and Schally attributed it a definite molecular structure, TRF became a “real” object. But Latour, with great perspicacity, reminds us:

“It is quite possible that in the end TRF will turn out to be an artefact. For example, there is no convincing evidence that TRF actually exists in the body in the form Pyro-Glu-His-Pro-NH2 in ‘physiologically significant' quantities... Up until now, the negative results in this respect have been put down to the lack of sensitivity of the assays... However, a slight alteration in context could still lead to the conclusion ... that TRF is after all an artefact.”

This reversibility in the status of scientific theories is just as manifest on a longer time-scale. The history of science abounds with examples of theories which were long held to be true but which have subsequently turned out to be false; and conversely, theories which were once considered false but which are nowadays accepted as true. For example, at the time of the alchemists, it was thought that metals contained a substance, phlogiston, which they lost when they burned and disintegrated into powder. This theory was very widely accepted up until the advent of “modern chemistry”, when the metal and powder were carefully weighed and it was shown that the powder actually weighed more than the metal; hence, the hypothetical “phlogiston” would have a negative weight. For reasons that were as much if not more social and political than purely cognitive (Berman, 1981[*]), this argument was considered to be decisive; and nowadays, no-one thinks that phlogiston exists. Conversely, there is a currently accepted theory that the continents once formed a single block; this block broke up into several pieces which then drifted apart, to give rise to the current shapes of the continents. This theory was first put forward at the beginning of the 20th century; the idea could indeed occur to any schoolboy who looks at a map and notices that the “bump” on the North-West coast of Africa fits very nicely into the concave “hole” of the Caribbean Sea on the East of the American continent. What is interesting for us here is that for over half a century, professional geologists mocked this idea as naïve and ridiculous.... until quite recently, when in the space of a few short years this theory has become generally accepted, and is now dignified by the name “Continental Drift”.

Even more relevant for us, the metamorphosis of “true” into “false”, or vice versa, is not always definitive. A major example, from the history of physics, concerns the nature of light. Ever since the birth of modern science, at the time of Newton, there was a confrontation between two great rival theories: according to one theory, light propagates in the form of waves ; according to the other, light rays are composed of a flux of minute particles. At different moments in the history of physics, one or other of these theories has gained favour. At the beginning of the 19th century, an orchestrated campaign that included a political as well as a purely cognitive dimension (Frankel, 1985[*]) led to an apparently complete and final victory of the wave-theory. However, at the beginning of the 20th century, Einstein – iconoclast here as elsewhere – showed by his studies on the photo-electric effect that there were certain phenomena which could only be explained by the hypothesis that light-rays are composed of indivisible elementary units, the “photons”. The corpuscular theory was resuscitated. Today, physicists have had to get used to the idea that each of these theories, the wave-theory and the corpuscular-theory, is in some sense “true” in spite of the fact that they are apparently contradictory.

Now this reversibility of the status of scientific theories poses an insoluble problem for an objectivist interpretation according to which scientific theories and objects are neither more nor less than “reflections” of a pre-given, independent “reality”. The point is that if such were the case, there could only be a change in the status of theories if there was first of all a change in “reality itself” to be the cause of that change in status. This is such a crucial point in the thematic opposition between objectivism and constructivism that it is worth taking a little time to dramatize what is at stake. To do this, I will base myself on a scenario imagined by Latour (1980[*]) that I have condensed in some respects and elaborated in others. The two main characters in this scenario are a scientist – a palaeontologist – and a young PhD student in the sociology of knowledge. The student is carrying out an enquiry on the question which interests us here, i.e. the conditions under which the status of a scientific theory can change. To do this, he engages in a “field study” where he questions an “informer”, just like ethnologists who “inform” themselves about the belief-systems of “primitive” tribes. An important methodological point is that the student rigorously neutralizes his own judgements, in order to focus on just describing the belief-system and trying to make it intelligible. In this case, the community in question is that of palaeontologists.

At the beginning of the story, set in the 1920's, there was a scientific theory according to which dinosaurs, who lived 100 million years ago, were rather slow and stupid animals (which largely explains their complete disappearance some 70 million years ago); they had small brains, they crawled with their bodies resting on the ground, and they were cold-blooded. The sociologist asks his informer why his community holds this theory as established fact. The palaeontologist answers him: “What a strange question! Look, it's really quite simple, it's just because there really were dinosaurs 100 million years ago, and they really were slow and stupid, etc., and our theory is simply the reflection of that reality.” The sociologist notes down the answer, and writes up an initial report to the X Foundation who has given him his grant: “Scientific theories reflect reality”. Upon which he takes off for a well-deserved week's holiday. End of Act I.

When he gets back from his holiday, the young sociologist finds his palaeontologist informer in a state of considerable excitement. “Ah! What a pity you missed that symposium I just attended, I'm sure that would have interested you. They showed some new fossils footprints, and it seems that the great dinosaurs didn't crawl by dragging their bodies along the ground, they were up on their feet. Besides, the anatomists have reanalysed the data on their skeletons, and it seems that it was quite possible. And they have also done some new calculations on the body-temperature resulting from metabolism, and it turns out that without really having an active thermo-regulation, the body-mass of the dinosaurs was so great that simply by thermal inertia, their temperature must have been pretty constant around 37°C. And they have also redone the calculations about the ratio between the size of the brain and body-size, and it seems that the brains of the dinosaurs weren't so small after all, they were probably quite bright.” The sociologist is just flabbergasted. “WHAT!!! However could I have missed that? I will regret it the rest of my life, I would have given my right arm to be there!” The palaeontologist is enthusiastic, but even so it does seem to him that the disappointment of his young friend is a bit excessive. “Hey, calm down. You know, it was only a symposium, there will be others.” “No, no, no, here's what must have happened. There you all were, comfortably installed in your symposium seats, when all of a sudden: “Boom-boom-boom”, the wall crumbles and a real dinosaur bursts into the amphitheatre. And that's just the beginning; when he entered, the dinosaur was slow, stupid, dragging along the ground and so on; and then, under your very eyes, his blood started to warm up, he got up on his feet, and his brain swelled so that he became quite clever. And indeed he must have been quite clever to have invented that time-machine which he used to come into your amphitheatre, and that he rode again to go back to his place in time 100 million years ago.” Now it's the palaeontologists turn to be taken aback. “What are you talking about? No, no, there was no dinosaur in the amphitheatre”. “But yes, there must have been! You yourself explained to me that your old theory was a reflection of reality, and that you believed in it because it was a reflection of reality. So, the only way that your theory can have changed, is that the reality itself must have changed. The rest is just pure logic – elementary my dear Watson!” “No, no, you don't understand properly. I don't remember too clearly what I explained to you before, but you know, that old theory was never anything other than a hypothesis. It is true that up until last week, that old theory had not been refuted by any of the observations we had available at the time, observations which by the way are still perfectly valid; but you know, there is this philosopher Popper who explains that scientific theories are never positively proved, at most they escape provisionally from being refuted.” The sociologist has a long, hard think; at last he comes up with his conclusion. “Ah, so that's the way it is... In a way, I am quite relieved to hear you say that scientific theories are never anything more than hypotheses. I see now that the theories can change without reality itself changing, and I am glad that I didn't miss the dinosaur in the amphitheatre. But you know, now I will have to rewrite my report to the X Foundation, because that is not what I understood before. To get everything straight this time, just tell me one thing: is your new theory a reflection of reality, or is it just a hypothesis?” Now it's the palaeontologist's turn to have a long hard think, and he is visibly embarrassed. But he ends up by answering: “Well, if you insist, I would say that for the time being it is only a hypothesis. After all, these new elements are all quite recent, and it is quite possible that some inquisitive colleagues will find something to correct in these new calculations”.

The years pass, our young sociologist gains a reputation and, as is the order of things, becomes a Professor. He has the opportunity to set another young research student on doing a follow-up study on the field of the palaeontology of dinosaurs, and gives him the second report that he had written fifteen years ago. The student does his field-work, and reports back to his thesis director. He is clearly embarrassed. “Professor, with all due respect, it seems that the report you wrote fifteen years ago no longer corresponds to what the palaeontologists say nowadays. This theory according to which the dinosaurs had warm blood, that they walked up on their feet, that they were quite clever, and so on, has stabilized, and the palaeontologists told me that now they know that it is true and that it does reflect the reality of 100 million years ago. Besides, they have another sort of confirmation that the dinosaurs were reasonably intelligent: if they disappeared, it was not at all because they were stupid, but because a great comet crashed into the planet Earth, and the resulting cloud provoked a terrible “winter” of several years through which the dinosaurs could not survive.” The old Professor was not vexed that his clever young student dared to contradict him, quite the contrary; but at the same time he could not forget his own youth and the emotion about the idea of the “dinosaur in the amphitheatre”. What is to be concluded from all that? He decides that the ontological question concerning ultimate reality must be definitively put into parentheses. One can then make the following description of the sort of cognition which is constituted in scientific communities. The scientists sometimes consider that their theories are the reflection of an independent reality; at other times, that these theories are only hypotheses or interpretations of which they themselves are the authors. The first configuration corresponds to a situation where the theory in question has been stable for some time (of the order of ten or twenty years, a substantial portion of their active life as research scientists). The second configuration corresponds to a situation where the theory in question is labile and a subject of open controversies. The alternation between these two configurations is a matter of historical contingency. The temptation to “explain” this sort of alternation by saying that the periods of stability occur “because” the theory “really” reflects reality must be firmly resisted; that would amount to mistaking effects for causes, and for giving oneself what is to be explained. To sum up: scientific objects are the result, and not the ontological cause, of a process of construction.

Cycles of accumulation

We come now to the articulation between technology (Technology as Anthropologically Constitutive) and science (The social construction of scientific facts), which we can formulate in terms of a key question concerning the social dimension of scientific knowledge. How does it come about that knowledge produced in the “ivory tower” of scientific laboratories actually has real effects in the “real world” of human society at large? We will address this question in terms of the notion of “cycles of accumulation”, processes which are indeed at the heart of a capitalist social system. Latour(1987[*]) introduces this question by taking the example of the great European explorers of the 15th century – Christopher Columbus, Vasco de Gama, Magellan. The crucial point here is that it was not enough just to go there: if the explorer perished in the foreign country, nothing was gained. It is not even enough to go and to come back: if the explorer returns empty-handed, nothing is gained either. In order to set up a “cycle of accumulation”, it is necessary to go, to come back, and to bring something so that those who will follow do not start from nothing. In a way, the explorer must “bring back” the distant things that he has seen – countries, peoples, events – so that his successor can study them at leisure at home, before setting off in his turn. It is important to keep in mind the totality of all the conditions that are necessary for such a “cycle of accumulation” to succeed. Latour gives the example of an expedition financed by King John of Portugal. The cycle of accumulation that was the aim of this voyage could be ruined at many points: Spanish ships could waylay the Portuguese caravels; an unscrupulous captain could sell elsewhere the precious spices that are to finance the expedition; the wood of a boat might not resist the assault of a typhoon; a miscalculation could cause a shipwreck... A difficulty arises here, because these conditions are transversal to the distinctions that are usually made between economic history, scientific history, technological history, politics, business management, law..... The point is that such distinctions are infinitely less important than what links them together in accomplishing the goal of an expedition: a cycle of accumulation which makes it possible for one place to become a centre with the capacity for acting at a distance on many other points. How is this action at a distance possible? One way or another, it must be possible to bring back distant entities – the events, places, and peoples concerned – to some sort of centre; and for this, it is necessary to invent the means to render these distant entities mobile, so that they can be brought back; stable, so that they will not be altered or damaged; and combinable, so that they can enter into relation with each other.

To sum up, if we want to look analytically at how cycles of accumulation function, there are three aspects: (1) the relevant events and processes in the outside world must be mobilized; (2) they must then be gathered together in centres of calculation so that the key aspects can be abstracted and combined with others; (3) the results of these operations must be re-exported into the outside world, so that actual “action at a distance” occurs. We will now look at each of these aspects in turn.

The mobilisation of worlds: inscriptions

We will take up again now the theme of inscriptions that we have already introduced in The domestication of the savage mind: the invention of writing. As a prototype of the way that inscriptions can render distant places “mobile, stable and combinable”, Latour takes the case of the making of maps. For that, the ships had to become veritable instruments, equipped with everything necessary to identify the longitude and latitude of every place they visited – quadrants, sextants for the latitude, and accurate marine clocks for the longitude, together with reference books so as to integrate the new elements with those that were already known. This is a way of producing what Kant called a “Copernican revolution” in the field of knowledge: before, it was the mind which turned around things which were largely unknown; now, it is the things (here, the geographical world) which turn around the “mind” – i.e., very prosaically, the things turn around the maps (the mind is not “in the head”). The balance between men and the Earth is shifted; a new centre, Europe, is constituted and starts to make the rest of the world turn around it.

To generalize, there is no reason to limit the mobilisation of stable, recombinable traces to places where humans can go in flesh and blood. Probes can be sent in their place. For example, when looking for petrol, it would be precious to know where to start drilling. Thus, in the 1920's, a French engineer, Schlumberger, had the idea of sending an electric current into the subsoil so as to measure the electrical resistance of the layers of rocks. After some technical adjustments, the signals obtained became sufficiently stable and reproducible for the geologists to go back and forth between the new electrical maps and sediments that had already been explored. This marked the beginning of a cycle of accumulation where petrol, money, physics and geology entered into a relation of synergy. The great advantage of this system of inscriptions was that it allowed not only for the mobility of the subterranean structure, and for the stability of the relations between the map and this structure, but above all for new combinations. At first sight there does not seem to be any simple connection between a Wall Street banker, an exploration manager at the Exxon headquarters, an electronics engineer specialized in weak signals at Clamart in the outskirts of Paris, and a geophysicist at Ridgefield. These elements seem to belong to different domains of reality: economics, physics, technology, computer science. But once we consider the cycle of accumulation which employs inscriptions which are mobile, stable and combinable, we see how these elements work together. It becomes possible to understand how the managers of an oil-drilling platform can plan their production, how economists can add in some of their own calculations, how bankers can use these figures to evaluate the value of a company, and how all these inscriptions can be stored in archives so that the government can calculate the oil reserves. One can do a whole host of things with the inscriptions in this paper world that it would be impossible to do directly with the world itself.

Let us take another example. It would seem that nothing dominates us more than the stars; they are quite out of reach, and it would seem that there is no way to invert the roles so that we human beings can dominate the sky. But this changes when Tycho Brahe (Danish astronomer, 1546-1601, Kepler's master) set up a well-equipped observatory, and began not only to record himself the positions of the planets, but to do this on homogeneous maps that he sent out to other astronomers in all parts of Europe so as to collect their observations too. The cycle of accumulation functioned all the more effectively when Brahe was also able to gather together the observations contained in old books of astronomy, that had become available thanks to the invention of printing. It is not Brahe's “mind” that is transformed; his eyes were not suddenly freed from ancient prejudices; he did not observe the sky more carefully than those who had gone before. But he is indeed the first human being who is able to consider together the sky that he sees, plus his own recorded observations, plus those of his contemporary colleagues, plus the books of Copernicus, plus the multiple versions of the great Almagest treatise of Ptolemy; he is the first to be able to place himself at the beginning and end of a whole network which generates mobile, stable and combinable inscriptions.

To sum up: what we have seen here are the conditions which make it possible to produce a “Copernican revolution” in the shape of knowledge, which reverses what counts as the periphery and what counts as the centre. Before this “revolution”, it is the human beings who are peripheral, who turn around a world which they do not master; after the “revolution” it is the world which turns around the centre which the humans have constituted. But this revolution does not occur in the brains of the humans in question; the essential conditions lie with the innumerable inventions, apparently modest, which make it possible to constitute mobile, stable and recombinable inscriptions. This is how entities as apparently diverse as a new sea route to India, oil reserves, the planetary system – and many others, of course, these are just a few metonymical examples that stand for a whole host of other cases – all end up taking the form of an inscription on a sheet of paper (or nowadays a computer screen), that can be stored in archives, brought out again at will, and... combined ad lib with other inscriptions. We will be looking more closely at this process of transversal connections in the next part (

Before that, however, we are already in a position to examine a question of very general interest: and that is the construction of space and time. Much of our difficulty in understanding science and technology comes from the common belief that “space” and “time” pre-exist, and provide a “neutral” framework within which events occur. It is this belief which prevents us from understanding how a whole range of multiple “spaces” and “times” are produced within the networks that are constructed so as to mobilize and recombine the world in the course of cycles of accumulation. What scientific and technological activity does, is to import immense extents of space and time into the scope of these networks where they can be handled and dealt with. You are ashamed that you cannot grasp what it means to speak of “millions of light years”? Don't be ashamed; the astronomer, for his part, understands it because he can measure them with a ruler, in the form of centimetres on a map. You are uneasy about the “nanometres” of living cells? But that only begins to mean anything, for anyone at all, when the “nanometres” become centimetres on an electronic micrograph. It may seem strange to say that new varieties of “space” and “time” can be constructed locally, in the centre of a particular network; but we join up with fundamental philosophical insights when we realize that “space” is neither more nor less than what is constituted by reversible displacements; and “time”, by irreversible displacements. Thus it is actually quite reasonable that each invention of a new mobile, stable and combinable inscription, which will give rise to a new set of possible displacements, will thereby give rise to a new space-time.

This is a key to understanding what we see when we start to follow scientists and engineers “in action”. They circulate within the context of narrow, fragile networks that they have constructed themselves; to use a metaphor to illustrate this important point, they are somewhat like the termites who circulate almost exclusively in the galleries they have built to link up their nests with sources of food. Within these networks, the scientists and engineers work hard at improving the circulation of inscriptions by increasing their mobility, stability and combinability. It is also important to specify that these networks are not made out of homogeneous materials, on the contrary they require a quite extraordinary amalgam between a very diverse set of elements, so that the question as to whether these networks are “scientific”, “technical”, “economic”, “political”, or “administrative” loses its meaning – they are all of these at the same time, which means that they are none of them in particular. We might be tempted to use the label “social” as a general term to cover all these aspects; but this is either too vague to help us much, or else leads us to appreciate more than ever why the question of the relation between “forms of knowledge” and “forms of social life” is indeed a “hard problem”. Picking up again the thread of our analysis, the result of the construction, the maintenance and the extension of these networks is to open up the possibilities for “action at a distance”; to do things at the hub of a network which make it possible to dominate the periphery in space and time. We will now look in finer detail at two aspects in particular: how is it that what is done in the centres, by carrying out operations on the inscriptions that are gathered together there, gives a decisive advantage to those who are placed in the centre (; and what has to be done to maintain the networks so that what is done in the centres can indeed have real effects on what happens at a distance (

The centres of calculation

When we enter these “centres”, these places where mobile, stable inscriptions are gathered together, we come across a problem that we have not mentioned so far. These inscriptions are certainly infinitely more convenient to manipulate than the entities and events that they represent; that is precisely what makes it possible to collect them in the centres. But after a certain time, this proliferation of inscriptions becomes a problem in itself. Faced with these mountains of first-order inscriptions – those that have a relatively direct relation to the peripheral world outside – one ends up being almost as lost as at the beginning. What is the solution? – it is analogous to the very principle of inscriptions: one has to produce second-order inscriptions, that have a relation to the first-order inscriptions that is analogous to the relation between the first-order inscriptions and the world. This process can clearly be recursive: in general, it is a question of constituting inscriptions of order (n+1) from inscriptions of order n. Latour gives two examples: the Mendeleev table; and the Reynolds coefficient.

In 1860, during their first international meeting in Karlsruhe, the community of chemists found themselves in the state of confusion that is typical of those who are overwhelmed by a profusion of first-order inscriptions. Each new instrument and laboratory procedure produced new chemical elements, and hundreds and thousands of new chemical reactions. Most courses on chemistry consisted of presenting more or less chaotic lists of known reactions. Mendeleev was one of those who were attempting to find some order in the chaos. He wrote the name of each element on a card, and tried to fit them in a table with rows and columns. In 1869 he found a compromise that satisfied him: by arranging the elements in order of increasing atomic weight, and by putting in the same column the elements having the same valency, he found an arrangement that allowed him to fit in all the known elements (figure 21).

An interesting feature of this table was that it contained a certain number of empty spaces; Mendeleev had the courage of his convictions, and conjectured that these spaces corresponded to elements which had yet to be discovered. Initially, his colleagues were not convinced by the system; but between 1875 and 1886, three of these elements predicted by Mendeleev's table – gallium, scandium and germanium – were indeed discovered. With the adoption of this system, each chemical element is situated on a new, (n+1)-order inscription at the intersection of a “longitude” and a “latitude”: elements on the same line are close by their atomic weights but different as to their chemical properties; elements underneath each other in the same column are quite different in their atomic weights but very similar indeed concerning their chemical properties. Thus, a new space is created locally; new relations of proximity and distance, new neighborhoods, new families are constituted; there is the appearance of a periodicity which up until then was invisible in the chaos of diverse chemical reactions.

Another example of the passage to a level (n+1) is given by the work of Reynolds, a British engineer specialized in fluid mechanics who made a particular study of the phenomenon of turbulence, culminating in 1883. How can one related the apparently diverse cases where turbulence manifests itself, in gases and liquids, in natural settings and in the laboratory? At the time when Reynolds was working, a certain number of relations had already been noticed, which took the form of statements: “the more X the less Y”, “the more... the less”. Thus, a considerable mass of first-order observations and graphs based on experiments could be summarized by the second-order statements:

  • The Turbulence T is proportional to the Velocity V: T α V

  • T is proportional to the length L of the obstacle: T α L

  • T is proportional to the density D of the fluid: T α D

  • T is inversely proportional to the viscosity Q: T α 1/Q

And as we may imagine, it is now possible to pass to a third-order statement: T α V.L.D/Q (the turbulence is proportional to the velocity V times the length L times the density D, divided by the viscosity Q). After certain adjustments, so that the units compensate each other to give a non-dimensional number, Reynolds obtained a new formula:

CR = V.L.D/Q (this is the “Reynolds coefficient”).

What is gained, is that it is now possible to compare situations as apparently different as a small stream flowing against a stone and a great river stopped by a dam; or even more different, a feather falling in air (air is indeed a fluid, with a very small viscosity obviously) and a body floating in syrup. If two processes, however different they may seem, have the “same Reynolds”, they will behave in the same way as concerns their turbulence: if the “Reynolds” is 2300 or less, there will be smooth laminar flow; if it is 4000 or more, the flow will be turbulent; between 2300 and 4000 there is a “critical zone” where the triggering of turbulence will depend on contingent local factors. CR is now a coefficient which can be used to characterize all possible turbulences, whether it be in galaxies in the sky or the knots in a tree; by this means, all turbulences become one in the physicist's laboratory. Even more important for the questions we are discussing here, knowledge of this coefficient paves the way for re-exporting knowledge of how to act outside the laboratory. We will address this question in section iv); here, we will give just one example in the case of the Reynolds number. In the 1960's, Professor Bijker and his colleagues studied the construction of a new dam in Rotterdam harbour. The problem was to balance the flows of fresh water in the rivers and sea-water: many dams that had been built before limited the flow in the rivers, with the harmful consequence that the salt sea-water, damaging for agricultural crops, penetrated further and further inland. Would the new dam aggravate this problem? How can one know in advance? Thanks to his knowledge of the Reynolds coefficient, Bijker was able to construct a reduced model in a hanger. This allowed a “Copernican revolution” of the type we mentioned above: Professor Bijker and his colleagues were able to master the problem much more easily than the managers of the harbour who were out there in the wind and the rain and who were much smaller than the landscape. With respect to the numerous problems that can arise in the real world, the engineers have already seen them. They can envisage all the various possible scenarios, trying them out at leisure and obtaining inscriptions to document their findings. They have constructed a new space-time – which will enable them to know how to act in the real world.

Before going on, there is an issue which arises here, and which is of the greatest importance for our attempts to understand the relation between forms of social life – what actually happens at the level of the actors out in the real world – and forms of thought. This is the question of “theoretical abstractions”. In the recursive cascades of inscriptions of order n, (n+1)... that we have been examining, we always remained at the level of concrete, local activities. To be sure, each stage had the aim of simplifying and summarizing the stage before; but this activity of re-presentation remained perfectly concrete: it employed pieces of paper, tables, laboratories, instruments. And above all, this activity was always subordinated to its goal of mobilizing inscriptions in order to act at a distance; and therefore, this activity never left the narrow networks that made it possible and which gave it meaning. Now if by the term “abstraction” we mean this process by which each stage in the cascade of inscriptions “extracts” elements from the underlying level in order to improve their mobility and combinability, then we can indeed talk about a process of “abstraction”. But it is important to be clear that this process in itself remains just as concrete as the process by which an oil refinery extracts increasingly “refined” products from the brute oil. To say that the process of abstraction we are interested in happens “abstractly” is perhaps tautological, but it is actually misleading; it is rather like saying that the oil refinery functions “refiningly”. So concerning the process of “abstraction”, a problem arises if we start saying that the scientists and engineers in the centres of calculation function “abstractly”. The actual concrete processes which are involved in producing “abstractions” (i.e. inscriptions of level (n+1) based on inscriptions of level n) lend themselves perfectly well to observation and study. However, if we start to imagine that what is going on is a mysterious process happening in the “minds” of the individuals in question (and/or “in” their brains), we are lost because we will never have access to it. We are indeed at the cross-roads between the “hard problem of consciousness” and the “hard problem of social cognition” that is one of the main themes in this chapter.

This difficulty is multiplied up when we consider the question of theories. If by the term “theory” we designate the cross-roads which makes it possible for the centres of calculation to mobilize, to combine and to relate all the inscriptions that have been gathered by the network, then it is perfectly possible to study the way that these “theories” actually function in reality. They are the focal point of what happens in the centres of calculation, accelerating the mobility and the combinability of the inscriptions. Their power, which can come to seem mysterious if it is considered “in abstract” (!), actually comes from their situation at the heart of the networks – from their relation with the multitude of inscriptions which are gathered together in the centres, and as we shall see next in their possibility of returning into the outside world. This may be the place to emphasize that these networks are social constructions, which will give us a lead in identifying their profoundly social nature. This is also the place to consider a question which can also seem “mysterious” if considered outside its actual social context: how is it the “abstract” forms of pure mathematics apply so well to the “empirical world”? Latour (1987[*]) notes ironically that entire volumes have been written in attempt to answer this “well-known fact”; but that no-one has actually taken the trouble to verify that things do actually happen that way. And when one does actually follow more closely the actual practice of science, it turns out that this “mystery” never actually occurs. What does happen is much cleverer, much more interesting and... much less “mystical”. Typically, at a certain moment in the cascade of inscriptions of order n, (n+1) etc., one obtains a cloud of points on a graph – each time with respect to a specific, local problem, but the form is common. “Pure” mathematics are maybe far from rivers and dams (to take the example we have just cited, with respect to the Reynolds coefficient, which is of course just one example among thousands and thousands); but when the rivers and dams are inscribed on pieces of paper, and they take the form of graphs, they become very close; literally, as close as one piece of paper is to another piece of paper in a centre of calculation. The connection between mathematics and the empirical world is an unfathomable mystery; the superposition of a mathematical form written on paper, and another mathematical form obtained from instrumental observations, becomes understandable.

This way in which mathematics becomes operational in a relation to the empirical world remains a remarkable achievement. What happens in the centres of calculation is that forms written on paper can come from very different domains, but they resemble each other mathematically. This opens up the way to combining forms of order (n+1) from one domain, with forms of order (m+1) from a domain that may seem quite different. In section (i), we mentioned that setting up cycles of accumulation requires relating elements from domains that are apparently very different – in the example we studied there between economics, physics, technology and computer science. Here, we see an analogous process at work within the centres of calculation. In addition to the vertical connections which are the result of passing from inscriptions of order n to order (n+1), mathematics opens up the possibility of establishing transversal connections from one domain to another. A mathematician can intervene everywhere, on condition that the inscriptions have taken the form of a graph with Cartesian co-ordinates. As with the basic cycles of accumulation, this is a phenomenon which amplifies as it goes along: the more the centres of calculation integrate heterogeneous sources of inscriptions, the more they need mathematics to maintain their coherence.

Acting at a distance

In, we have examined how events in the outside world can be mobilized, in the form of inscriptions, so that they can be gathered together in “centres of calculation”. In, we have looked at what is done in these centres of calculation, by abstracting and combining these inscriptions so as to produce an “added value”. But we now have a third question to examine. The final inscriptions produced in the centres of calculation are not themselves the real outside world; they only represent it in its absence. What remains to be done is to re-translate the “added value” into things that can actually be done in the outside world; without this, all that has gone before would count for nothing. This question of the actual application of the knowledge gained in the centres of calculation has been insufficiently studied, in large part because of a belief that scientific knowledge is “universal”, so that it should be possible to apply it without any further ado. Latour emphasizes that this is not at all the case; it is quite possible to re-export the worked inscriptions into the outside world, but... at a cost. The point is that “scientific facts” and “scientific theories” can only survive in the particular sorts of networks that gave birth to them in the first place, those that we introduced in, and which must now be extended to the sites where the “application” of the scientific knowledge is planned. A “scientific fact” would not last 30 seconds outside these networks. There is a simple rule of method: every time you hear about a successful application of scientific knowledge, look for a network that it has been possible to extend; every time there is a failure, look for the element(s) in a network that have failed. Latour gives two examples, one of a failure and the other of a success.

The failure concerns the project for a solar village at Frangocastello in Crete. When the architects, town-planners and experts in solar energy had finished their calculations, they had in their office in Athens a complete reduced model of the village. They knew “everything” (!) about Crete: the local weather conditions and in particular the power and duration of periods of sunshine, the demography of the region, the water resources, the economic tendencies, the methods of agriculture and the use of glass-houses. They had envisaged all the possible scenarios and all the configurations, consulting with the best engineers in the world; they had the enthusiastic support of several European banks for their promising, original project. All that remained to be done was to go “out there” to apply their calculations and to demonstrate, once again, the almost supernatural power of Science. But... when the engineers from Athens turned up at Frangocastello to acquire the necessary land and to adjust the last details, they came face-to-face with a “reality” that they had never even thought of. Not only were the inhabitants of the village unwilling to abandon their lands in exchange for houses in the new village, but they were actually ready to fight rifles in hand against what they thought was a new atomic base of the American military, camouflaged as a solar installation. The “application” of the scientific theory became more difficult with every day that passed, as the opposition gained the support of both the Pope and the Socialist Party. Since it was out of the question to send in the army to oblige the inhabitants of the village to accept the new prototype, the only thing to do was to open negotiations between the “inside” (the centre of calculation) and the “outside” (the field site). But what sort of compromise was possible between the project of a revolutionary new solar village, and a few shepherds who only wanted 3 kilometres of paved road and a service station? The only possible “compromise” consisted of totally abandoning the project. All the plans of the engineers went back into their boxes in the centre of calculation. The “reality out there” had dealt a mortal blow to this example of Science... which had failed to pass the threshold into “Action”.

We will now take a contrasting example which succeeded. On June 2 1881, Louis Pasteur made a dramatic prediction: all the non-vaccinated sheep in the little village of Pouilly-le-Fort were going to die of the terrible anthrax disease; but all the vaccinated sheep would be in perfect health. Was it a “miracle” of science? – no, of course not; it was the successful extension of a network. Let us look more closely at what Pasteur actually had to do. What we find is a fascinating negotiation between Pasteur and the farmers; what was at stake, was to transform the farm into a laboratory. Pasteur and his colleagues had already made several successful trials inside their laboratory; that had enabled them to shift the balance of power between humans and the bacteria responsible for the epidemics, by creating artificial epidemics inside their laboratory. However, they had not yet succeeding in really “acting at a distance” on a real farm. Now Pasteur and his colleagues were not fools; they knew full well that on a real, dirty farm with hundreds of spectators, they would not be able to create the laboratory conditions that had been so favourable to them (they would risk the same sort of misadventure as befell the solar-energy experts in the Cretian village). On the other hand, if they asked everyone to come into their laboratory, no-one would be convinced. What they had to do was to find a compromise: transform enough of the characteristics of the farm into conditions similar to those in the laboratory, but taking enough risks so that the trial would count as being performed “outside” the laboratory. As things turned out, the experiment at Pouilly-le-Fort succeeded, so that it is still talked about today. But what is important to understand is that it may well have failed; and that if it had done so, it would have been because certain crucial conditions in the laboratory had not been correctly exported to the farm; there would have been a failure in the extension of the network. To use a metaphor, techno-scientific facts and accomplishments are like electricity and trains: they can go everywhere... on condition that their networks (like the electrical wiring or the railroads) have been properly put in place. If “scientific knowledge” really was “abstract” and “universal”, it could be extended everywhere without any additional cost. But it is not; putting place the requisite networks, and maintaining them in order, is quite feasible... but it does have a cost. We will illustrate this, by coming back to two of our basic examples from section, involving the construction of space and time.

We will take first the example of time; more precisely, the time of day as it is defined by “science”. If I ask you “what time is it?”, you have to look at your watch. There is no way of reliably answering this question without reading that instrument (for certain purposes, looking at the sun will do, if it is visible – but not if you have to catch a train). Of all scientific instruments, the clock is the one with the longest history, and it has always been highly influential (in fact, we will come back to it very shortly – space and time are intimately interconnected). But sticking with time for the moment: if your watch and mine don't quite agree, we will be lead to seek a third timekeeper (the radio, the clock tower). If we persist to the end, we will be lead into a network: a very sophisticated network, made up of atomic clocks, lasers, communication satellites; it has a name, the International Time Bureau, which co-ordinates the time everywhere on Earth. The time of day is not spontaneously “universal”, off its own bat; it becomes universal, a bit more every day, by the extension of a network which links together, by visible and increasingly invisible means, all the reference clocks in the entire world; and which organizes secondary and tertiary references until we come back to the watch that you are wearing on your wrist. There is a continuous network of readings, lists, inscriptions, telephone lines.... which links together all the time-pieces. If you leave this network, you are immediately uncertain as to the exact time of day; the only way to be sure is to connect up to this network. And need we labour the point? – the creation and upkeep of this network is costly¬ – not absurdly so, not out of all proportion, but it is definitely not “free”.

To complete this discussion of the networks that are necessary for the potential “added value” obtained by operations in the “centres of calculation” to be effectively realized by leading on to actual “action at a distance”, we will come back to the question of “space”, as it is “represented” (cognitively, but externally and not “in the head”) by maps. One might have thought that once a good map has been put together in a centre of calculation, everything has been done; that the re-exportation of the map towards a particular geographical situation follows more or less automatically. But actually this is not the case at all; once again, it is not possible to stray from the networks without getting lost. The essential point is that if one does not know where one is on a map, the map itself becomes useless. When we use a map, it is very rare to compare the map with the landscape itself. No; in practice what happens most of the time is that we compare a position on the map with road-signs that are also marked as such on the map. The “outside world” only lends itself to an application of the map when all its major features have been marked by appropriate signs. An illustration of this was given when the Russians invaded Czechoslovakia in 1968: the Czechs took away all the road-signs... with the result that the Russians, with all their maps, were often seriously “lost”, they just did not know where they were. When “raw reality” is really encountered, when the things “out there” are seen for the first time outside the context of the networks, “Science” is no longer effective; the essential cause of its superiority over other forms of cognition vanishes into thin air. And need we labour the point? – these networks are perfect examples of “social structures, both constraining and enabling, in the fundamental sense that we introduced in Fundamental Theories of Society.

The history of social synthesis by market mechanisms

The social synthesis

We now come to the third instance of social structures that we mentioned at the end of Fundamental Theories of Society, i.e. the institution of a monetary system. In order to appreciate the far-reaching influence of the invention of money – as important in its way as language, pervading as it does all aspects of human social life in contemporary society – it is necessary to introduce the concept of “social synthesis”. Every human society in which there is the least degree of division of labour (one of the main themes developed by Durkheim) must necessarily have a mechanism which provides functional answers the following three questions: (i) What are the productive activities which will be performed? (ii) How is the sum of all the work to be performed to be distributed between the members of society: who will do what? (iii) How are the fruits of this labour to be divided up amongst the members of society? – It is a question of the viability of any form of social life that there should be a mechanism which provides an effective answer to this question (not necessarily explicitly, but in terms of practical results); in the absence of adequate answers, there will be anarchy and the dissolution of the society. Now in very broad terms, one can make a distinction between two major types of mechanism for ensuring the social synthesis, which thereby condition two sorts of human society; I will call them “traditional” societies and “market” societies.

In the great majority of human societies in the past, the mechanism of social synthesis can be designated by the term “traditional”: there is a definite sort of social order, with a specification of the roles of the various members of society, which is reproduced from generation to generation in an essentially unchanged form. In many cases, this social order comprises institutions of discussion and negotiation: the African palaver can serve as a metonymical example. One can also speak of a “communal” mode of production, where the nature of the productive activities themselves integrate in large measure the distribution of their fruits and, upstream, the corresponding division of labour. We may remark that there is some proximity here with animal communities, where in some cases the differentiation of activities necessary for collective survival can be quite sophisticated. However, since no animals have the capacity for language, there is no animal equivalent to the institution of the palaver type.

By contrast with these “traditional” forms of social organization, there are societies (including our own contemporary society) that we can designate by the term “market” societies. Here, a large part of the social synthesis is neither traditional, nor the object of relatively direct discussions, nor integrated with the productive activities themselves, because it is delegated to the mechanisms of a market economy. In this case, the social synthesis is achieved by the famous “invisible hand” of Adam Smith, according to the laws of supply and demand which are balanced by the mechanism of prices. In other words, the social synthesis is not directly achieved as such; it is, rather, the “emergent” result of a whole series of purely local economic decisions, without there being anywhere a coherent vision or conscious will at the level of the whole. It is important to emphasize that in market societies, characterized by a division of labour, economic exchanges play a fundamental role because they determine the form of the social synthesis. The life of each individual depends on the activities of production and consumption; but without the intervention of market exchanges none of these activities would occur. Each economic crisis is an object lesson in the fact that the activities of production and consumption are perturbed precisely to the degree that the functioning of economic exchanges is compromised.

Before going on, it is relevant here to recall the point examined in Science, Technology and Society, i.e. the crucial role of technological systems in determining the form of social life. Tools, and more generally the “technical systems” which articulate them, constrain and enable specific forms of the division of labour and the differentiated roles which correspond to them. In this respect, it is to be noted that if traditional societies often seem resistant to “technology transfers”, it is not just because of simple superstition or “backward” mentalities; it is above all because technological innovations inevitably produce an upset of the traditional social order, and it is this upset which the members of these societies, quite deliberately, decide that they do not want[35][35]. On the other hand, money is an instrument of social synthesis which is so powerful and flexible that technological development, and the social division of labour which goes with it, are (for worse or for better) “freed” from this sort of consideration. It is no accident that it is in market societies that technological and industrial development takes place on a scale beyond all common measure with that which occurs in traditional societies.

The historical genesis of money

A major author on the history of money is Georg Simmel, to whom we shall make frequent reference. In it is interesting to note that Simmel, together with Max Weber, is generally considered as one of the great protagonists of the “comprehensive” current in sociology (cf Two currents of thought in sociology); but his great work, The philosophy of money (Simmel, 1900[*]) is concerned with money which is a macro-social structure par excellence. Be that as it may, we will start by a salutary reminder: it is important to note that market exchanges (which are entirely absent in animals) are not even an anthropological invariant. Even today, there are some human societies where objects only change owners by the rudimentary mechanisms of theft or gifts. For members of these societies, the very idea of “exchange” is quite unknown; and even a long time after the introduction of money (for example, in certain aristocratic circles in Europe) this attitude subsists in the form of haughty disdain for all sorts of commercial practices or bargaining. We may take one example from a thousand: when Ulysses and his companions arrive on the island of the Cyclops Polythemus, the latter asks them whether they are pirates or merchants: these were the only two known categories of foreign travellers. The succulent point here is that in this context, the term “pirate” is not at all pejorative, whereas that of “merchants” is considered pretty despicable (Vidal-Nacquet, 2000, p. 117)[*].

From that starting-point, the evolution was very gradual: gift and counter-gift (Mauss, 1925[*]); direct barter; the appearance of certain commodities which were not quite like the others because they were systematically used as intermediaries in market exchanges (grain is a good example); the use of precious metals; the first instances of coined money, which implies the institution of banks; promissory notes and paper money, which reinforces the importance of the banking system; and this evolution is still progressing today with the invention of credit cards and electronic money, which goes together with the increasing (and disturbing) importance of the financial domain compared to the productive domain. Now this long evolution has indeed been accompanied by a corresponding transformation in cognitive mentalities, social institutions and technical systems.

As Durkheim indicated so clearly, there must be a social mechanism which results in putting in place “transcendental” reference-points, i.e. mentalities and thought-forms which are shared by all the members of a given society, which are considered as “self-evident”, so “obvious” that indeed they are taken for granted and are not even noticed, becoming “invisible” to the members of the society themselves. Now it is crucially important not to commit a sophistry here; Simmel (Simmel, 1900[*]) is explicitly critical of attempts to falsely “explain” this phenomenon by appealing to some sort of “collective unconscious”. In order to avoid this error, we will examine very concretely the co-evolution between forms of money (as a type of social mediation) and forms of thought, by looking at three aspects in particular: i) the question of the materiality of money, whether it is a “substance” and if so what sort of substance; ii) the relation established between means and ends, which involves the question of “technical systems” and the social division of labour; and iii) what Simmel calls “styles of life”.

From a theoretical point of view, there is an important distinction between two sorts of reality : “first nature” and “second nature”. The term “first nature” (erste Natur in German) refers to the ordinary realm of material reality; “second nature” (zweite Natur) is the realm in which money as such exists, a realm which is entirely social and, by its constitution, perfectly abstract. Now coins of money possess both sorts of reality. The “first nature” of a coin of money is its characteristic as a piece of metal, with all the natural attributes of size, weight, colour and so on; its “second nature” is its value as money, i.e. so many Euros or Francs or Pounds sterling or Dollars. Once it is accomplished, the distinction between “first” and “second” nature is perfectly clear and precise; but in terms of psycho-social reality, it was only very gradually that the aspect of “second nature” – which is of course quite fundamental to the essence of money in its ideal form – became differentiated with respect to “first nature”. Simmel poses the question very explicitly as to whether money can ever really be entirely detached from the materiality of its substrate. Certainly, nothing in the ideal functioning of money as a means of exchange requires that it should posses a value in itself. (“In the end”, so to say, bank-notes – and today even more electronic money – illustrate this very clearly). However, an examination of the actual course of the development of money throughout historical ages shows that things were not so simple.

Everywhere, in primitive economies, it is entities which have a real use-value which play the role of money: cattle, salt, slaves, tobacco, skins, grain, etc. Money could not have existed if it were not perceived as something immediately precious in itself, because of the material substance of which it was made. There are numerous examples which bear witness to the strong tendency to treat “money” and “intrinsic value” as quite simply natural and inevitable counterparts. A historical example in the long process leading from barter to money in the proper sense of the term, which illustrates this difficulty of detaching the value of money from its material substrate, is that of the Middle Ages. For medieval theory, there was no conceptual distinction between “use-value” and “exchange-value”; there was just “value” itself which was taken to be fully objective. This lead to the concept of a “fair price”, which could even be fixed institutionally. Thus, the value adhered to the thing in itself, and it was this value which the object brought with it when it entered into an act of exchange. This conception of value is perfectly coherent with the substantialist and absolutist vision of the period; and it can well be understood in the framework of exchanges in kind.

“A plot of ground for services rendered, a goat for a pair of shoes, a jewel for twenty masses for the dead – in all these cases we are dealing with objects which are immediately attached to certain intensities in the feeling of value, to such an extent that their prices seem to correspond to each other quite objectively” (Simmel, 1900[*]).

When barter develops, one sees the gradual emergence of a “general equivalent” – but which nevertheless remains attached to a unit which still has a very concrete use-value:

“Thus the ancient Irish, when they entered into contact with the Romans, laid down their own unit of value, the cow, as the equivalent of an ounce of silver metal. The wild tribes from the mountains of Annam, who amongst themselves practiced only exchanges in kind, have as a fundamental value the buffalo; and when they enter into contact with the more civilised inhabitants of the plain, the unit of value of the latter (a silver bar of a certain size) is considered as the equivalent of the buffalo. One finds the same fundamental characteristic in a wild tribe in the outermost reaches of Laos: they only practice barter, and the basic unit of value is the iron hoe. However, they do look for gold in the river in order to sell it to neighbouring tribes, and this is the sole substance that they weigh. For measuring a weight they have no other unit than a grain of maize; and so they sell a maize-grain's weight of gold for an iron hoe!” (Simmel, 1900[*]).

A further step, which is very important (in particular for its conceptual implications), comes with the principle of bargaining and, inseparably, a quantification of value:

“If the route which leads to money starts from barter, one enters in this direction, without yet leaving barter behind, from the moment when a homogeneous object comes to be exchanged no longer against another homogeneous object, but against a plurality of objects. Exchanging one cow against one slave, one article of clothing against one talisman, one boat against one arm, these are transactions which do not yet divide up the operation of evaluation; such transactions do not proceed by reducing objects to a common denominator in order to multiply up afterwards. However, once one starts to take a flock of sheep against a house, ten trimmed beams against a fine costume, three pints of drink against an afternoon's work.... and once one starts bargaining (for example, is the costume worth twelve beams or only ten, so that the value of this costume, externally indivisible, also becomes measured by the value of a beam, so that it is possible to recompose it by multiplying by eight, by twelve, or finally by twelve), an important change in mentality occurs. What happens is that the values of two objects that are exchanged become commensurable in a quite other sense than when they could not be divided; it becomes possible to express the value of both with one and the same abstract unit. An exchange against money is it not indeed this object of exchange which is eminently divisible, endowed with a unit which appears commensurable with all and any counterparts (however indivisible they may be); and which thereby facilitates – or even supposes – the extraction of the abstract value included in the latter and thereby dissociated from its concrete content”. (Simmel, 1900[*]).

The fact of considering the equality in exchange as a quantitative equality – as Simmel says, “manifestly a solid basis, perceptible by the senses, in the formation of values in primitive peoples” – is attested by many ethnological examples:

“In the islands of New Brittany.... the native people use as money strings of cowry shells, which they call dewarra. When buying, this “money” is used as a function of its length: the length of an arm, and so on. For fish, it is usual to give their own length in dewarra. ... The same measure of two different commodities passes for having the same value; for example, a certain measure of cereals is worth the equivalent measure of cowry shells. Here, manifestly, the immediacy of the equivalence between a commodity and its price reaches its most complete and most simple form. ... In the case of certain tribes in West Africa, the money which circulated took the form of iron bars, which served to designate the quantities of commodities, so that a certain quantum of tobacco or rum was called respectively “a bar of tobacco” or “a bar of rum”. Here, the need to consider the equality of value as a quantitative equality ... has taken refuge in a linguistic expression. The town of Obia on the Dnieper, a colony of Milet, has left us bronze moneys in the form of fish, carrying inscriptions which probably mean “tuna fish” and “fish basket”. We may suppose that this fishing community originally used tuna as a unit of exchange, and... judged it necessary, in adopting money, to represent the value of a tuna by a piece of money which, having the same shape, directly materialized the equality of the money-object with the object that it replaced; in other places, however... the money was simply stamped with an image of the object (cow, fish, axe...) which served as the basic unit at the stage of barter.”

This passage from a qualitative determination to the quantitative symbol facilitates an increasing differentiation between use-value and exchange-value. In this respect, Simmel quotes a story from ancient Russia:

“In the beginning it was marten skins which served as money for exchanges. However, with time, the size and the beauty of each individual skin lost all influence on their exchange value, so that each skin simply had the value of one skin, the same as all the others. Consequently, all that mattered was the number of skins; and what followed was that as circulation increased, it became more practical to use just the tips of the skins as money, until finally just small pieces of skin, probably stamped by the government, circulated as the means of exchange. Here it is clearly apparent that the reduction to a purely quantitative viewpoint is the basis for that symbolisation of value on which the realisation of money, in all its purity, will come to rest.

On the other hand, it seems that a sort of money which would have immediately taken a purely ideal form would not be able to satisfy economic requirements... In the lower stages of the economy one can find the most widely varying monetary values all mixed together; on the one hand one finds a form of money which presents the character of concrete value in stark form, such as the cattle, the silver or the cotton fabrics which circulated in the Philippines by way of large sums of money; on the other hand one finds money as close to the ideal form as cowry-shells, as the money in the form of silk-tree bark discovered by Marco Polo in China, as the pieces of porcelain marked with Chinese ideograms which had currency in Siam. A certain functional evolution is initiated, beyond these categories of money with concrete value, where it is natural articles which serve as means of exchange, but where at the same time these are mainly articles for export: tobacco in Virginia, rice in Carolina, salted cod in Newfoundland, tea in China, skins in Massachusetts. In the article for export, the value is somewhat detached, psychologically, from immediate internal consumption of this commodity-money. Nevertheless the best compromise between the abstract forms of money that we have mentioned and this commodity-money is still represented by jewel-money, principally gold and silver, because this is neither as arbitrary and meaningless as the abstract forms, nor as gross and particular as the commodity-money. This is manifestly the reason which leads silver, most easily and at the same time most surely, towards its transformation into a symbol[36]. There seems to be an obligatory passage by this stage of a link between the two aspects for money to come to its maximal effectiveness, and it seems that in a foreseeable future it will not be possible to completely disengage from this link.”

On a purely theoretical level, use-values (“first nature”) and exchange-values (“second nature”) must mutually exclude each other. However, in the course of actual historical development, it is only very progressively that this mutual exclusion comes to be set up. Simmel quotes the case of pieces of petrified salt which are specially carved to serve as money in Abyssinia: they only become money from the moment when they are no longer used as salt. Another example: on the Somali Coast, pieces of blue cotton fabric two yards long; from the moment when these pieces of cloth served as money, they could no longer be cut up and re-sown at will, which is a clear indication that they could no longer be used as cloth. This is further confirmed in the case – particularly important in the evolution of money – of the precious metals. Their main use consists in making jewels and other adornments. From a theoretical point of view, money as a means of exchange must ideally be characterized by a constant value. We have already noted that the precious metals lend themselves to this, because their “first nature” is such that they do not undergo change. To this, one can add the fact that as fine adornments, they constitute an elastic reservoir: if the metals should become relatively abundant they can be converted into articles of adornment almost without limit; conversely, if the metals become rare, the jewels can be reconverted into money; this flexibility contributes even further to the constancy of their value. But this inter-convertibility only serves to make it even more striking that use and exchange mutually exclude each other: jewels are not money, and reciprocally pieces of money are not jewels.

An important stage in historical development is the appearance of merchants and shopkeepers: in place of relatively direct relations between the various producers, there is now the relation of each producer, on his own account, with the merchant. In conjunction with this, these relations are mediated by money, which thereby takes another step on its road to its ideal form. As the merchant is interposed between the subjects of the exchange, so in exactly the same way money is interposed between the objects of the exchange. We have already seen, with Sohn-Rethel, that it is at this stage of the development of commercial exchanges in the full sense of the term that there is the historical appearance of coined money; and this marks a further step in the “de-materialization” of money. In the Middle Ages in Europe, the function of money was still linked to its materiality. Thus, it was the corporations of silversmiths and goldsmiths that ensured the guarantees for exchange operations. Progressively, the authorisation to strike coined money was transferred to the towns, i.e. to a public institution, with a consequent depersonalisation of money. The extension of a monetary economy goes together with the instauration of the modern Nation-State – i.e. with the dissolution of the relative auto-sufficiency under a feudal regime. Similarly, imperialism is co-extensive with the validity of coined money; it is for this reason that in Antiquity, the coins were struck, symbolically, with the effigy of the emperors; later, at the time of the British Empire, with the head of the reigning sovereign.

This is the stage at which monetary value as such makes its appearance. “Normally”, the value of a commodity consists of the value of the raw materials plus the value of the work necessary to produce it. But in the case of coined money, there is of course a difference between the cost of fabrication and the exchange-value of the coins. This raises the grand question of the guarantee which is necessary so that buyers and sellers accept to place their confidence in the money. This question of confidence, in relation with the de-materialization of money, is of course posed in a particularly acute form with the advent of paper-money. The point is that as long as money remains linked to a material substance – even if it is a precious metal which is relatively close to the ideal form – its development is blocked. Conversely, the more the importance of money as a means of exchange becomes preponderant compared to its substantial material value, the greater is the amount of money which can circulate around the world in a form other than that of a precious metal. Let us take a closer look at the stages in this evolution, which of course did not happen from one day to the next.

A first step, that can be found already in the Middle Ages, is a promissory note by which the signatory recognizes an explicit debt. One can even find, at that period, a clause accompanying the recognition of a debt which allows not only the actual lender, but simply the bearer of the document, to receive payment for the bill; however, this arrangement did not have the function of transferring the value, but merely to facilitate its payment by a representative of the creditor. This purely formal mobilisation of paper becomes more effective with the blank note, in circulation at the Stock Exchange in Lyon. This blank note still referred, formally, to an individual creditor, but whose name was not written on the note; however, as soon as a name was written in the blank space, the creditor was determined individually. Veritable commercial relations with titles that refer only to the bearer, began in the 16th century at Antwerp; we know that at the beginning it often happened that when they were submitted for collection without bearing a nominal specification, they were often refused payment on the stated date; to such an extent that an Imperial Decree was necessary to establish the principle of their validity. This marks a very definite progress. The value in question starts out in the case of an individual recognition of debt by being “squeezed” between the creditor and the debtor; it acquires an initial mobility when it can at least be collected by a third person, even if it is only on account of the original creditor; this movement is enlarged when the blank note, even if it does not suppress it altogether, delays indefinitely the specification of an individual creditor; and finally with the simple title referring to the bearer, which passes from hand to hand just like a coin of money, the value is entirely mobilized.

As Simmel says:

“Money is inscribed in a cultural movement. The principle of an economy of forces and substances becomes more and more active, and this leads to wider and wider use of substitutes and symbols which have absolutely no similarity of content with what they represent. This is what happens when the operations with value are achieved by means of a symbol which loses more and more any material relation with the realities behind its specific domain, and which thus becomes nothing other than a pure symbol. This mode of life requires not only an extraordinary increase, in number, of psychical processes (consider the complicated psychological presuppositions that are required by the simple covering of bank-notes by a reserve of cash money); it requires also that these processes be raised to a qualitatively new level, and that the civilisation in question opts for a fundamentally intellectual mode of operation. That social life should rest essentially on the intellect; that of all our psychical energies it is the intellect which passes for the most precious in practice; all this goes together... with the penetration of a monetary economy... The increase in intellectual capacities for abstraction characterizes a period where money becomes increasingly a pure symbol, indifferent to its own material value.”

With this, we are in the domain where a certain form of social life – here, monetary economies – goes hand in hand with a certain form of cognition, in this case the development of the abstract intellect. We will take up this question of the relation between forms of thought and forms of social life in Forms of Thought, Forms of Social Life. For the moment, we will continue our account of the historical development of money.

The origin of capitalism

In Ancient Greece, the extent of the social synthesis realised by the exchange abstraction was after all fairly limited, because it was restricted essentially to external commercial exchanges. The domestic economy – the daily production of goods of common use in family households – was very largely removed from the market regime. To be brutally precise, this domestic production was mainly performed by slaves; and the institution of slavery is typically “traditional”, the antithesis of a market economy. As we said above, the fundamental rule of market exchanges is that “the fact of appropriating an object and ceding another does not result from a direct or “natural” action (for example, looting), but only by an act of exchange by free and mutual consent”. The least that can be said is the fact of appropriating the labour force of a slave does not concern an act of exchange by “free and mutual consent” – certainly not on the part of the slave. To sum up, then, for the Greeks the regime of the exchange abstraction was restricted to external commerce, and did not order the production of current domestic goods. Market exchanges were important, sufficient in any case to give rise to the unprecedented innovation of coined money; but it was not a capitalist system in the full sense of the term, with cycles of accumulation and reinvestment giving rise to exponential economic growth.

This state of affairs only really changed when, after the interludes of the Dark Ages and the feudal low Middle-Ages, the birth of modern capitalism gave rise to investments in the domain of production itself. As Marx says: “It is only from the moment where there is the free sale, by the worker himself, of his labour-force as a market commodity .... that the production of commodities generalizes and becomes the predominant form of production; it is only from that moment that every product is made in order to be sold immediately, and that all riches pass through the routes of the circulation of money. It is only there where paid work is the basis that the production of commodities asserts itself at the level of society as a whole; it is only there, also, that the production of commodities reveals all its hidden potentialities.” (Cited by Sohn-Rethel, 1978[*]).

In order to prepare for the next section 5, where we will explicitly explore the relations between forms of social life and forms of thought, it will be useful now to characterize the form of thought necessary for capitalism. For this, an essential reference is the great book of Max Weber, The Protestant Ethic and the Spirit of Capitalism (Weber, 1905[*]). I cannot refrain from remarking here that in this book Weber, renowned as a leader of the “comprehensive” school of sociology, addresses here the question of a key macro-social structure. Be that as it may, to illustrate immediately what Weber calls the “spirit of capitalism”, we cannot do better than cite a remarkable passage where Benjamin Franklin (1748[*]) develops the theme “Time is Money”.

“Remember that TIME is Money. He that can earn Ten Shillings a Day by his Labour, and goes abroad, or sits idle one half of that Day, tho' he spends but Sixpence during his Diversion or Idleness, ought not to reckon That the only Expence; he has really spent or rather thrown away Five Shillings besides.

Remember that CREDIT is Money....

Remember that Money is of a prolific generating Nature. Money can beget Money, and its Offspring can beget more, and so on. Five Shillings turn'd, is Six: Turn'd again, 'tis Seven and Three Pence; and so on 'til it becomes an Hundred Pound. The more there is of it, the more it produces every Turning, so that the Profits rise quicker and quicker. He that kills a breeding Sow, destroys all her Offspring to the thousandth Generation. He that murders a Crown, destroys all it might have produc'd, even Scores of Pounds. ....

The most trifling Actions that affect a Man's Credit, are to be regarded. The Sound of your Hammer at Five in the Morning or Nine at Night, heard by a Creditor, makes him easy Six Months longer. But if he sees you at a Billiard Table, or hears your Voice in a Tavern, when you should be at Work, he sends for his Money the next Day. ...

Remember that Six Pounds a Year is but a Groat a Day. For this little Sum (which may be daily wasted either in Time or Expence unperceiv'd) a Man of Credit may on his own Security have the constant Possession and Use of an Hundred Pounds.”

Weber remarks that we are here in the presence of the authentic “spirit of capitalism”. Simple avarice is of all countries and all epochs. Many in the course of history are the Kings, Emperors and other despots who sought by every means to amass great wealth. But with the spirit of capitalism, it is something else again. This obsession of every instant to remain constantly stretched to the limit in order to gain the maximal rate of profit is quite unique of its kind. It is to be noted that the exemplary “capitalist” does not seek to gain riches in order to spend them and thus enjoy himself, which would be humanly normal; no, he seeks to avoid every non-profitable expense, and to accumulate capital... with the sole aim of re-investing it. In other words, the accumulation of capital is here a goal in itself; and we may remark further that the least lapse with respect to this goal is not a simple stupidity, but truly an inexcusable misdeed with respect to an ethical duty.

This mentality, so strange and remarkable, raises two questions which Weber takes great care not to confuse: on one hand, the question of its origin; on the other hand, the question of its maintenance and the fact of its contagious extension. I will examine these two questions in the reverse order.

First of all, then, let us look at the question concerning the maintenance, and the proliferation, of the capitalistic mentality once it has arisen. Weber is very clear on the point that capitalism in no way requires the continuation of the rather special circumstances (which we shall examine below) which were necessary for its origin. Whatever its origin (in the event, it was a form of religious belief), capitalists do not have to adopt themselves those particular beliefs. It is sufficient for the capitalist mentality to exist in a form giving rise to a certain type of economic practice – for example, the form shorn of its religious connotations that we find in Franklin – for its extension to become inexorable. As Weber wrote (in 1905)[*], but it remains more true than ever today), “the capitalist economy of the contemporary epoch is an immense cosmos in which the individual is born, and which presents itself to him as an immutable order of things in which he has to live. This order obliges the individual, to the extent that he enters into the system of market exchanges, to conform himself to the rules and norms of capitalist society. The entrepreneur who would act contrary to these norms will inevitably be evicted from the economic landscape.... Thus, contemporary capitalism, which has developed to the point of dominating economic life, educates and selects the economic subjects it needs by a process of economic selection of the fittest”. To say the same thing in other words, the capitalistic mentality proliferates in tight relation with the concomitant proliferation of capital itself, following the same inexorable logic of exponential growth. I have presented this argument quite briefly, in spite of its importance which is great and indeed practically incalculable, because its very simplicity makes it massively inescapable; in truth, further discussion would only serve to confuse the issue.

Weber continues his argument: “Nevertheless, we may easily perceive here the limits of the concept of selection as a means of historical explanation. In order that a mode of life so well adapted to the particularities of capitalism can have been selected and come to dominate all others, it was certainly necessary that it could arise somewhere; and not only for a few isolated individuals, but as the way of life common to whole groups of human beings. It is this origin which, in the last resort, is what truly has to be explained.” We thus arrive at the question that I will address in second place, that of the origin of the “spirit of capitalism”.

As the title of his great work indicates, Weber finds this origin in the “Protestant Ethic”, and more precisely in the beliefs of the Calvinist sect. Briefly, Weber's explanation is the following. One of the distinctive features of Calvinism is its belief in predestination. In other words, by an inevitable decree of God, all humans are predetermined at birth, as being either in a state of grace in which case they will go to Heaven, or in a state of perdition in which case they are condemned to Hell. It is therefore of the utmost importance for each person to know in which state they are; all the more so as it is the duty of each believer to belong to the faithful elect[37]. But the question is, how can one know? The ways of God are unfathomable, and there is no way of communicating directly with Him in order to have the question settled once and for all. Nevertheless, God (in His mercy?) has provided a means of knowing after all: he rewards his happy elect by their fortune during their terrestrial life. Thus, if one accumulates great wealth, one can be sure of being “on the right side”. Weber explains, with great insight, the extraordinary strength of the psychological motivation engendered in this way. There are other religious doctrines, including Catholicism, where one can obtain eternal salvation by means of good actions. But besides the fact that the identity between “good actions” and “accumulating the maximum amount of wealth” is not guaranteed (to put it mildly!), the quality of the motivation is not at all the same. In the Catholic version, one can always hope that a few lapses can be made up for by some extra effort later; by contrast, in the Calvinist perspective, the least weakness in accumulating wealth is a “sign” that one is... on the wrong side, and that is irretrievable. Hence this obsession, truly remarkable, which consists of doing at every instant everything possible to... accumulate and ceaselessly reinvest capital.

This is how Weber explains the origin of the capitalist mentality. As we have already said, once this state of mind and the economic practice that goes with it are in place, they become irresistible and carry all before them; capitalism becomes a system in itself and no longer requires the particular religious belief which gave rise to it in the first place. Before leaving this question, we may note that Weber himself was in no wise a eulogist of capitalism. Even though he kept a certain critical distance with respect to Marxism – which he criticized in particular for an excess of naivety in interpreting the domain of ideas and mentalities as merely a sort of “superstructure”, a simple “reflection” of the economic infrastructure – he observed the proliferation of capitalism, which he saw as inexorable, with all the pessimism of the intelligence. Commenting upon the major obstacle that capitalism had had to overcome, which he identified with “traditionalism”, he wrote: “The capitalist obtains none of his wealth for himself, other than the sense of having done what he had to do. But it is precisely this which appears to pre-capitalist man so incomprehensible and mysterious, so unworthy and so despicable”. It is not difficult to imagine where Weber's own sympathies lie...

The 19th century

The industrial revolution

Following through on the chronological development of capitalism, we come to the nineteenth century. We have already discussed, inScience, Technology and Society, the question of the coming together of the two strands of technology and science. The nineteenth century is the historical period when this junction occurred; and there was indeed a structural reason for this, related to the intrinsic dynamics of capitalism. The key point here is what Marx identified as the structural tendency for a “fall in the rate of profit”. In a “free market” situation, the rule of no-holds barred competition means that in any particular sector of activity, the rate of profit is constrained by the fact that competitors are “free”, precisely, to enter the market by cutting the profit-margins. In this situation, there develops an imperative for technological innovation: if one has the chance to dispose of an effective technological innovation, the rate of profit can be substantial during the time that it takes for the competitors to “catch up”. The question is, notwithstanding, anything but trivial. “Modern Science” did not immediately give rise to effective techniques of production. A striking example of this concerns the invention of steam engines, one of the motors (literally as well as figuratively) of the industrial revolution. Nowadays, when we have become used to considering that “science” is the source of virtually all knowledge, there is a certain temptation to imagine that “science” was indeed at the origin of the steam engine. However, the historical reality is quite different: the great inventors of the steam engine (Newcombe, Watt and others) were engineers, or rather inspired tinkerers, who did not rely at all on scientific knowledge. In fact, it was rather the other way around: thermodynamics as a branch of Science (the major figure here was Carnot) came after the steam-engines, and to a very significant extent actually used them as a significant resource. The point here is not that the scheme “pure science -> applied science” never applies; however, it is anything but trivial, and historically came later than is often thought.

The ground-breaking example is this respect is the emergence of the chemical industry, notably in Germany, towards the end of the nineteenth century (Aftalion, 1991[*]). In this case, it is perfectly true that the “science” – modern chemistry, from Lavoisier to Mendeleev (cf. – was indeed original and did not rely on any pre-existing industrial practice. Other major examples of this configuration, where it is indeed “pure science” that gives rise to technological innovations, are the cases of electricity (Faraday) and electro-magnetism (Maxwell); and closer to our own day (and for better or worse), nuclear energy which presents its own social and ecological problems in both military and civil uses.


This new mode of creating technological innovations – on the basis of scientific knowledge initially produced outside the sphere of direct economic activity – has a social significance that is sufficiently important to merit specific comment. The issue at stake is clearly illustrated by the so-called “Scientific Management”, developed by Frederick Winslow Taylor in the 1880s and 1890s. Taylor's work is fascinating, because he himself explains with disarming candour the aim of his work. Briefly, here is the heart of the matter (Braverman, 1974[*]). In a situation of salaried work under a capitalist regime, the workers make a bitter finding. If they work with a sincere good will, and invest their own creativity in inventing procedures that result in an increased productivity, they do not benefit from it. More precisely, they receive neither the same salary for less time at work, nor an increased salary for the same time at work; what happens is that the piece-rate they receive is reduced, so that at the end of the day they continue to receive the same salary for the same time at work. The reason for this is that their salaries are determined by the balance of power in the class struggle between capitalists and their employees, who have objectively opposed interests. The capitalist, who has appropriated a certain amount of work-force by paying for it at market-rates, considers that he has every right to exploit this force to maximum . Faced with this impasse, the workers fight back by the practice of “soldiering”, which consists of deliberately working below their maximum[38] capacity, so as not to tire themselves out unprofitably (from their point of view), and to avoid a faster pace becoming the new standard – all the while avoiding also any downright offences for which they could be penalised. The capitalist, for his part, is quite defenceless against this sort of practice; not having himself a sufficient degree of expertise concerning the work-process, he does not have any clear standard as to what he can demand of his workers, and he can only submit to the united front of the workers. Taylor, who had himself a real experience of the condition of workers since he had himself worked for several years on the shop-floor, knew what he was talking about; and he proposed (quite lucidly and explicitly) to betray his previous class-comrades by providing the capitalists with a means for controlling the work-yields that could be required. As is well known, his method for achieving this was to establish a finely detailed description of all the elementary gestures involved in the work-task, and of timing them to a fraction of a second, so that the capitalist manager could specify very exactly what was required from the workers. The computerized “expert systems” that we see today are neither more nor less than the accomplished version of Taylor's project; the aim being to “extract” the knowledge of the workers, so as to convert it into a commodity that the capitalists can be sure of appropriating. At the time, towards the end of the 19th century, Taylor was not able to fully implement his programme; what happened was that the workers were fully aware of what was going on, and attempts to actually put Taylor's scheme into practice led to social conflicts of a rare violence which threatened the stability of the whole system.

It is in this context that we can appreciate the significance of the radically new industrial procedures, based on the application of scientific knowledge. In terms of social relations, scientific knowledge is characterized by the fact that it is elaborated outside any specific context of use, by a process that is essentially intellectual[39] . Because of this there is no longer any necessity for a violent struggle to expropriate, to dispossess the workers of their own knowledge of the work-process, because here what is in question is a sort of knowledge that they never had in the first place. This does indeed shed a fresh light on the social significance of scientific knowledge. The fact that scientific knowledge is elaborated in an “ivory tower”, abstracted away from all immediate concerns with use as such, does not mean that this knowledge is socially “neutral”; quite the contrary.

The exchange abstraction

Finally, in order to prepare for the final section Forms of Thought, Forms of Social Life in which we will examine the relation between forms of social life and forms of thought, it will be useful here to develop the notion of the “exchange abstraction” which is implicit in all market mechanisms. In a word, market exchange is an abstraction because it involves a rigorous relation of mutual exchange between use and exchange. Let us look at this more closely.

The essential point is this: the activities of use on one hand, and activities of exchange on the other, are not simply different; they must take place separately, during different and mutually exclusive temporal periods. The reason is that the exchange activity serves the sole purpose of a change in owner, in other words a change in the purely social status of the commodities as elements of private property. In order for such a change to take place on the basis of a negotiated agreement, the material status of the commodities, their physical condition, must remain unchanged during the whole period of the negotiation – or rather, which is even more relevant here, their material status must be presumed to be unchanged. In order to emphasize this crucial point, I will quote directly from Sohn-Rethel (1978[*]), a German Marxist who spent the last years of his life in London, and to whom I owe the essential part of what follows, and to which I will return in Forms of Thought, Forms of Social Life.

“There, in the market-place and in shop windows, things stand still. They are under the spell of one activity only; to change owners. They stand there waiting to be sold. While they are there for exchange they are not there for use. A commodity marked out at a definite price, for instance, is looked upon as being frozen to absolute immutability throughout the time during which its price remains unaltered. And the spell does not only bind the doings of man. Even nature herself is supposed to abstain from any ravages in the body of this commodity and to hold her breath, as it were, for the sake of this social business of man. Evidently, even the aspect of non-human nature is affected by the banishment of use from the sphere of exchange.” (Sohn-Rethel, 1978[*], p.25).

The practical activity of exchange does not in itself have any meaning in terms of nature; it is purely social by its constitution and scope. Nevertheless, the transfer of ownership that is negotiated under property laws in no way lacks physical reality itself. Exchange involves the movement of the commodities in time and space from owner to owner, and constitutes events of no less physical reality than the use activities which it rules out. It is indeed precisely because their physical reality is on a par that these two kinds of activity, exchange and use, are so mutually exclusive. Thus, exchange is an abstraction because, while remaining inseparable from use (otherwise no-one would bother to exchange the commodities in question), it quite rigorously excludes it. At the same time, it is a real abstraction, because it is a perfectly real event in time and space.

To sum up the argument so far: in market societies, there are two registers of spatio-temporal reality which exist side-by-side, but which mutually exclude each other. This point will be so important for what follows that it will be useful to employ specific terms. We come back to the distinction between “first nature” and “second nature” that we have already alluded to in The historical genesis of money. In German, the register of “use” is designated by the term “first nature” (erste Natur); this register is entirely and substantially material. The register of “exchange” is designated by the term “second nature” (zweite Natur); this register is entirely social and, by its constitution, perfectly abstract. The same term “nature” is employed to indicate that these two worlds are endowed with an equal degree of spatio-temporal reality, and that they are inextricably combined in the fabrication of our daily life in a market society.

As we have seen, the strange relation between “first nature” and “second nature” is brought to its peak by the social institution of coined money. Money is an abstract, paradoxical entity: it performs a decisive function in the social synthesis, but unbeknown to the actors concerned (I will come back to this point in Forms of Thought, Forms of Social Life). But even if the “exchange abstraction” is practically never thought of as such by economic agents, no animal can begin to understand what money is: it is a register that is solely accessible to human beings. Sohn-Rethel makes this point in striking fashion, and I will cite him again:

“Take your dog with you to the butcher and watch how much he understands of the goings on when you purchase your meat. It is a great deal and even includes a keen sense of property which will make him snap at a stranger's hand daring to come near the meat his master has obtained and which he will be allowed to carry home in his mouth. But when you have to tell him ‘Wait, doggy, I haven't paid yet!' his understanding is at an end. The pieces of metal or paper which he watches you hand over, and which carry your scent, he knows, of course; he has seen them before. But their function as money lies outside the animal range. It is not related to our natural or physical being, but comprehensible only in our social interrelations as human beings. It has reality in time and space, has the quality of a real occurrence taking place between me and the butcher and requiring a means of payment of material reality. The meaning of this action registers exclusively in our human minds and yet has definite reality outside it – a social reality, though, sharply contrasting with the material realities accessible to my dog. Here we have the spheres of the ‘first' and ‘second nature' which we distinguished earlier side by side, and unmistakably divided.”

Marx says quite explicitly that the exchange abstraction never receives a mental representation as such, since its sole expression resides in the act of considering that the value of one commodity is equal to the value of another. Gold, or silver, or any other material entity which lends to money its instantiation as a visible, palpable body is only a metaphor for the exchange value, it is not the abstraction as such. In fact, the material instantiations of money do more to mask than to reveal its veritable “second nature”.

Historically, when commodity exchanges spread, becoming multilateral and involving a wide range of commodities, there is an overwhelming practical need to employ one of these commodities as a general means for the exchanges of the others. This new role does not in itself, immediately, confer the commodity in question with an appearance that is different from before; but as a means of exchange, it is invested with the postulate that it should undergo no material change as long as it continues to exert that function. It is therefore easy to understand that the choice of a “standard-commodity” will fall on an entity that by virtue of its physical durability, its divisibility and its mobility is relatively conform to the required properties. In this way, the postulate of immutability, which has its true source in the abstraction of exchange, quite rapidly acquires the appearance of a consequence of its particular properties. The fact that a special “aura” attends this commodity “not like the others” does more to confirm than to refute this misleading appearance.

This confusion reaches a summit when the choice for a “standard-commodity” falls on one of the precious metals. On the occasion of each market transaction, it was necessary not only to weigh the metal, but also to melt it and test for purity; in short, it was necessary to relapse into treating them according to their first nature. And precisely for this reason, they failed in the end to perform their function as a universal means of exchange. This deficiency only found its solution with the invention of coined money: this step, which was to have such weighty consequences, was first taken in Ionia around 680 B.C. With coined money the preceding relation, where its status as exchange-value was subordinated and masked by its material, first-nature status, was overturned. A piece of coined money is stamped in order to signify that it is to serve as a means of exchange and not as a use-object. Its weight and metallic purity are guaranteed by the emitting authority; thus, if it happens that a coin of money has lost weight through wear and tear, the authority in question will replace it free of charge. Its physical matter has become merely the bearer of a social function[40] . A piece of coined money is an entity which conforms to the postulates of the exchange-abstraction; it is presumed to be composed of a substance which is absolutely unchanging, on which time has no effect, and which is thus unlike any material substance which actually exists in nature.

We will take up this question again in Forms of Thought, Forms of Social Life , as it provides a key case for consideration of the relation between forms of social life and forms of thought.

Forms of thought, forms of social life

We come now to the culminating point of this chapter: the question of the relation between forms of social life, which we have examined in the previous sections, and forms of that high-level cognitive activity that we call “thought”. It is because I am intimately persuaded that there is an essential relation of this sort that I consider that cognitive science in general, and the paradigm of Enaction in particular, will only be able to address the question of “high-level” human cognition if it engages seriously in an examination of social structures. Right from the start, I wish to make it clear that in the course of my attempts to address this question, I have been driven to the conclusion that it is actually a “hard problem”, fully on a par with the much better-known “hard problem” of the mind-body relationship. It is a hard problem because on the one hand, there is both a strong a priori presumption and a wealth of more detailed empirical indications that there is indeed a relation between forms of thought and forms of social life; and yet on the other hand, when we try to examine the question more closely, analytically, and to work out in detail how and why such a relationship should occur, the affair turns out to be much more difficult and uncertain than it may have appeared at first sight. In other words, it is a “hard problem” because it is manifestly important, and won't just go away even if it is ignored; but on the other hand it turns out to be difficult to provide a convincing resolution. In the case of the mind-body problem, an essential feature of the difficulty lies in the difference in nature of the two entities one is trying to relate: on the one hand, “the mind” exists in the domain of conscious first-person lived experience whereas “the body” (and in particular the brain) is typically characterized by impersonal third-person descriptions and mechanisms. Because of this, “category mistakes” – the error consisting of attributing to an entity properties or characteristics that by its very nature it could not possibly have – are always lurking. There may well be difficulties of a similar order in our question of the relation between forms of social life and forms of thought; as we have seen when noting the existence of two schools in sociology (see Two currents of thought in sociology), there is the question as to whether social forms should be described in impersonal third-person terms as in the natural sciences (“explanatory” sociology), or in terms involving first-person understanding (“comprehensive” sociology).

The plan of this section, seeking nevertheless to get to grips correctly with this “hard problem”, is the following. In A priori concepts and social life, I present the formulation of the problem proposed by Durkheim (1912[*]). For quite explicit methodological reasons, the empirical evidence that Durkheim cites to support his thesis that conceptual categories have their origin in the form of social life is drawn from observations on aboriginal societies. I consider that Durkheim makes a quite convincing case, and this sets us firmly on the road to a more general theory. The challenge then is to continue this sort of analysis to account for the major forms of thought in our own societies – in particular, the categories of philosophical and scientific thought. This will form the object of the following sections.

A priori concepts and social life

In his book The Elementary Forms of the Religious Life, published towards the end of his life, Durkheim (1912[*]) addresses fairly and squarely what is arguably the central question in the entire study of human thought: the nature and origins of conceptual categories. “At the roots of all our judgements there are a certain number of essential ideas which dominate all our intellectual life; they are what philosophers since Aristotle have called the categories of the understanding: ideas of time, space, class, number, cause, substance, personality, etc. They correspond to the most universal properties of things. They are like the solid frame which encloses all thought; this does not seem to be able to liberate itself from them without destroying itself, for it seems that we cannot think of objects that are not in time and space, that have no number, etc. Other ideas are contingent and unsteady; we can conceive of their being unknown to a man, to a society or an epoch; but these others appear to be nearly inseparable from the normal working of the intellect. They are like the framework of the intelligence.” (op. cit., pp 21-22). So far so good; but now the question is: where do these categories come from?

Durkheim notes that up until his time there had been only two doctrines in the field: empiricism, and apriorism. Empiricists hold that the categories are built up by bits and pieces, on the basis of regularities in perceptual experience, and that the individual is the artisan of this construction. This is the viewpoint that was developed historically by the British Empiricists – Locke, Berkeley and Hume – and it culminated with Hume's famous conclusion that the notion of “causality” could only be an illusion. It was this scandalous conclusion that provoked Kant to “awake from his dogmatic slumbers”, and lead him to propose his “Copernican revolution”: far from experience leading to the categories it was the other around, the categories exist a priori and are themselves the condition of possibility for experience to exist. On this point, Durkheim considered that the Kantian “apriorists” were entirely correct: empiricism completely fails to account for the nature of the categories, the way they impose themselves on the human mind as necessities. But he immediately goes on to remark that apriorism has its own fatal weakness: it does not even attempt to account for the origin of the categories as a natural phenomenon in principle open to scientific investigation. Durkheim writes: “The apriorists are the rationalists; they believe that the world has a rational aspect which the reason expresses excellently. But for all that, it is necessary for them to give the mind a certain power of transcending experience and of adding to that which is given to it directly; and of this singular power they give neither explanation nor justification. For it is no explanation to say that it is inherent in the human intellect. It is necessary to show whence we hold this surprising prerogative and how it comes that we can see certain relations in things which the examination of these things cannot reveal to us. Saying that only on this condition is experience itself possible changes the problem perhaps, but does not answer it. For the real question is to know how it comes that experience is not sufficient unto itself, but presupposes certain conditions which are exterior and prior to it, and how it happens that these conditions are realized at the moment and in the manner that is desirable. To answer these questions it has sometimes been assumed that above the reason of individuals there is a superior and perfect reason from which the others emanate and from which they get this marvellous power of theirs, by a sort of mystic participation: this is the divine reason. But this hypothesis has at least the one grave disadvantage of being deprived of all experimental control; thus it does not satisfy the conditions demanded of a scientific hypothesis.” (op. cit., p 27).

Durkheim concludes this part of his argument by noting that the twin doctrines of empiricism and apriorism have been pitted against each other for centuries; and that if the debate seems to be interminable, this is because these doctrines are more or less equally defective. “If reason is only a form of individual experience, it no longer exists. On the other hand, if the powers which it has are recognized but not accounted for, it seems to be set outside the confines of nature and science.” We may pause here to remark that if cognitive science does not come up with an adequate answer to this conundrum, then its whole project of explaining human cognition as a natural phenomenon goes completely bankrupt.

In this critical situation, Durkheim proposes a radically original hypothesis: this is that the categories have a social origin. Making this hypothesis more precise, Durkheim postulates that there are thereby two sorts of knowledge: on the one hand empirical knowledge, which relates directly to the interactions between the individual and their environment[41] ; and on the other hand knowledge which is framed in terms of the categories, that are essentially social in nature. “Between these two sorts of (knowledge) there is all the difference which exists between the individual and the social, and one can no more derive the second from the first than (one) can deduce society from the individual” (op. cit., p 28). Durkheim concludes his Introduction as follows: “Thus renovated, the theory of knowledge seems destined to unite the opposing advantages of the two rival theories. It keeps all the essential principles of the apriorists; but at the same time it is inspired by that same positive spirit which the empiricists have striven to satisfy. It leaves the (faculty of) reason its specific power, but it accounts for it and does so without leaving the world of observable phenomena. It affirms the duality of our intellectual life, but it explains it, and with natural causes. The categories... appear as priceless instruments of thought which the human groups have laboriously forged through the centuries and where they have accumulated the best of their intellectual capital. A complete section of the history of humanity is resumed therein. ... This is how it is legitimate to compare the categories with tools[42] ; for on its side, a tool is material accumulated capital. There is a close relationship between the three ideas of tool, category and institution.” (op. cit., p 32).

The bulk of this long book is devoted to adducing empirical evidence to support this hypothesis concerning the social origins of the conceptual categories. Obviously here we can give no more than the barest outline sketch. In the Preface, Durkheim presents a most interesting methodological precept: in the study of any natural phenomenon which undergoes evolution, there is an immense advantage in starting with the most primitive form known. He illustrates this precept quite explicitly with the case of living organisms: “Biological evolution has been conceived quite differently ever since it has been known that mono-cellular beings do exist.... The discovery of unicellular beings has... transformed the current idea of life. Since in these very simple beings, life is reduced to its essential traits, these are less easily misunderstood.” (op. cit., pp 18-19). Similarly: “Primitive civilisations offer privileged cases... because they are simple cases... That which is accessory or secondary... has not yet come to hide the principal elements. All is reduced to that which is indispensable, to that without which there could be no (society). But that which is indispensable is also that which is essential, that is to say, that which we must know before all else... But primitive (societies) do not merely aid us in disengaging the constituent elements of (society); they also have the great advantage that they facilitate the explanation of it. Since the facts there are simpler, the relations between them are more apparent. The reasons with which men account for their acts have not yet been elaborated and denatured by studied reflection; they are nearer and more closely related to the motives which have really determined these acts” (op. cit., pp 18-19). Thus, the reason why Durkheim draws mainly on ethnographic studies of Australian aborigines, with supplementary material from studies of Native Americans, is not “simply for the pleasure of telling (the) particularities and ... singularities ... of a very archaic (society)”; but because he hopes thereby to approach the essential constituent elements of human society, and to explain them.

Now it becomes very rapidly apparent that in all these “primitive” societies, there seems to be an anthropological invariant: the very nexus of their social life is provided by religion: but a religion which is in large part foreign to all idea of divinity or god(s). What is at the root of these religious practices is a distinction between the profane and the sacred. Durkheim therefore goes on to ask what could be at the root of this distinction. An Empiricist might suggest that the notion of “sacred” could derive from extraordinary and possibly “supernatural” events, such as cosmological rarities, showers of falling stars and the like (the theory called “naturism”); or maybe it derives from the phenomena of dreams (the theory of “animism”). But Durkheim very properly dismisses both of these suggestions: since all these phenomena, naturist or animist, do actually occur in the realm of “natural events”, they cannot for the life of them suggest the notion of the “sacred” as different in kind from the profane. But, Durkheim continues, there are indeed two different sorts of reality with which human beings are confronted. On the one hand, there is the ordinary everyday reality of perceived objects and processes (which corresponds non-problematically to the class of the profane); but on the other, there is indeed a quite different sort of reality, which is equally non-negotiable by an individual, and that is... social reality!! So Durkheim arrives at the conclusion that the “sacred” is neither more nor less than the form in which “the social” presents itself to the consciousness of individuals in these “primitive” societies. His task then becomes to show that the conceptual categories of time, space and so on have their natural origins in the religious categories by which social life is ordered. He was able to muster an immense amount of empirical data to support this hypothesis.

By means of a very thorough and critical appraisal of the ethnographic literature, Durkheim came to the conclusion that, quite generally, the “elementary form of the religious life” was that known as “totemism”. Each tribe is divided into a certain number of clans (usually a dozen or so). Each clan is identified by its emblematic totem, which is often but not necessarily a particular species of animal or plant (an additional indication of the sacred nature of animals is given by the cave-paintings at Lascaux at elsewhere - Lewis-Williams & Clottes, 1998). The totem is sacred for members of the clan; it is forbidden for consumption (except possibly under special ritual circumstances). This system is inseparably religious and social, confirming Durkheim's theory concerning the intimate connection between the two. We now come a crucial point: for the Australian, everything which is in the universe is considered to be a part of the tribe; consequently, just like men, all things known are distributed between the clans (op. cit. pp 166-168, where Durkheim cites some examples). Naturally enough, things which are attributed to the same clan tend to have some similarities; this is particularly clear in the case of the phratries[43] , where there are just two classes. Thus, if the white cockatoo is in one phratry, the black cockatoo will be in the other; and the moon is regrouped with the black cockatoo whereas the sun is with the white cockatoo. However, as Durkheim notes with insistence, “the feeling of resemblances is one thing and the idea of class is another.... The contents cannot furnish the frame into which they fit.... This is why... the idea of class (must not be confused) with that of a generic image.... The best proof of the distance separating these two notions is that an animal is able to form generic images though ignorant of the art of thinking in classes and species.” (op. cit. p 171-172). Thus, the very notion of “class”, and of systematically and logically classifying entities into a system of classes, is a clear example of a non-empirical, a priori conceptual category. What we see here is that the very first systematic classifications that we meet with in history are “modelled upon the social organization, or rather that they have taken the forms of society as their framework. It is the phratries which have served as classes, and the clans as species.” (op. cit. p 169). One could scarcely ask for a clearer or more direct vindication of Durkheim's hypothesis that the a priori categories have a social origin. Durkheim provides analogous demonstrations for other major categories. We cannot go into the details here, but will have to content ourselves with the barest summary: “it is the rhythm of social life which is at the basis of the category of time; the territory occupied by the society furnished the material for the category of space; it is the collective force which was the prototype of the concept of efficient force, an essential element in the category of causality.” (op. cit. p 488).

Finally, we must remark that this successful demonstration of the intimate and fundamental relation between the form of social life and the form of human thought, at the most elementary level, is not the end of the story; rather, it is just the beginning. Durkheim remarks: “Attributing social origins to logical thought is not debasing it or diminishing its value or reducing it to nothing more than a system of artificial combinations; on the contrary, it is relating it to a cause which implies it naturally. But this is not saying that the ideas elaborated in this way are at once adequate for their object.” (op. cit. p 493). Thus, Durkheim considered that he had laid the foundations for a whole new research programme, consisting of following through the whole evolution of human thought, and of relating this to concomitant changes in the forms of social organization.

The exchange abstraction and conceptual categories

In A priori concepts and social life, we have already noted that it was Hume's scandalous conclusion (according to which the concept of strict causality could not be derived from empirical observation and that it was therefore essentially an illusion), which famously awoke Kant from his “dogmatic slumbers” and lead him to propose his “Copernican revolution” in epistemology: far from being derived from empirical experience, the major conceptual categories are themselves the condition for structured, cognitive experience. We have also noted that Durkheim considered that Kant was entirely correct on this point, but that this only leads on to a new unanswered question: if the conceptual categories do not derive from empirical experience, where do they come from?! As we saw, Durkheim further proposed a radical and audacious solution, i.e. that the categories have their origin in the forms of social life; he illustrated his thesis in eloquent fashion by relating a number of major non-empirical conceptual categories – space, time, causality and the principle of logical classification – to social structures in Australian aborigines. But finally, we also noted that this is not the end of the story but rather the beginning: in particular, what remains to be done is to provide an analogous account for the whole set of Kantian categories, and thence the conceptual foundations of modern Western science.

This is precisely the ambitious project of Alfred Sohn-Rethel (1978[*]); even though he does not refer explicitly to Durkheim, the basic intuition is clearly the same. It is to this end that Sohn-Rethel lays out the Marxist analysis of the “exchange abstraction”; in particular, the distinction between “first nature” which corresponds to the pragmatic, empirical domain; and “second nature” which belongs to the social realm. We have already presented in a characterisation of the exchange abstraction as the core of the social synthesis realized by market economies. We now move on to Sohn-Rethel's original proposal, which consists precisely of seeking the roots of the “a priori” conceptual categories in the exchange abstraction.

Sohn-Rethel introduces this part of his thesis with a pleasant thought-experiment. The leading role is played by a philosophically minded Athenian from Classical Greece, who asks himself searching questions about the coins of money in his pocket: “What sort of substance should these coins be made of?” As none other than the great Plato emphasized clearly, all material objects existing in the world are perishable, corruptible, and unable to resist the ravages of time; and precisely because of this, they are not properly suitable to the function of money. Now Plato also speaks of entities of another sort, which are spotless, eternal, perfectly pure, and always strictly identical to themselves: he denotes them by the honorary title of “Ideas”. So, our Athenian asks himself, “are coins of money actually pure Ideas?” Worried, he takes hold of the coins in his pocket, and thinks hard: “These coins are real things; and they are real not just for me, but for all my fellow citizens who accept them in payment for wares. Might money be immaterial? – what an absurd idea, no coin could properly be money if it did not have material reality”. So he comes to the reassuring conclusion that the substance that his coins are made of is a real substance, as real as any other substance existing in time and space. And yet, this substance is quite different from all these other ordinary substances, because this one is just as immutable as the entities that Plato speaks of. But how can a substance which is immune to the ravages of time exist in time? Nowhere in the whole of nature, and nowhere within the limits of sensory perception, can any such substance be found. But then, how can our Athenian know about this extraordinary sort of substance if he cannot see it or hear it or touch it? This is clearly a question about knowledge, of prime concern for Cognitive Science. He knows about it by thought and only by thought. Never in all his life has he ever come across this sort of entity, something which is obstinately and uncompromisingly real and yet which is detached from any of the sensory qualities by which things are usually real for us.

This reflection can introduce us to more detailed examination of the formal analogies which exist between the conceptual categories of philosophical thought on one hand, and the distinctive features of the exchange abstraction on the other. It is important to emphasize here that what characterizes each and every one of these conceptual categories is their “canonically apodictic” nature: quite generically, each of them has the remarkable property that once they are identified, in their ideality, it appears intrinsically manifest that they could not be other than they are. At the same time, they are radically non-empirical: there is nothing in our daily experience of nature which is sufficiently similar for it to be at the origin of the concept. What Sohn-Rethel is suggesting is that actually, there is something in our daily experience that does fit the bill: however this is not any sort of material reality, but social reality. On this basis, Sohn-Rethel enters into finer detail and draws up a list of “homologies” between conceptual categories and various aspects of the exchange abstraction. We shall examine seven of these “homologies”; they are grouped together in the next section, on the grounds that they all apply to the first historical realization of the exchange abstraction in Ancient Greece.

Characteristics of the exchange abstraction in Ancient Greece: the homologies


The doctrine according to which “I alone exist” (solus ipse) is a leading leitmotiv of Western philosophy. This doctrine reached the summits with Descartes and Berkeley. In Descartes' famous “Cogito ergo sum”, the “self” in question guarantees its own existence – the very idea would collapse if the “existence” in question extended to anything other than the subject of the cogito. Berkeley deliberately pushes this solipsism to a provocative limit with his “Esse est percipi”: to be is neither more nor less than being perceived. In other words, it is not only other subjects but the whole world which only exists to the extent that I perceive it. With his usual clarity, Kant summarizes the apodictic character of solipsism: “there is no foundation in theoretical reason which makes it possible to infer the existence of another subject”. It is true that later philosophical currents – notably the phenomenology of the 20th century – seek to pursue the question of the status of “the Other”. But in a way, this only serves to accentuate the astonishingly non-empirical nature of the solipsistic doctrine; because in ordinary everyday life, no-one seriously doubts for a moment the existence of other subjects, nor the real existence of the external world. So where could this apparently preposterous idea have come from? There is undeniably a certain irony in looking for an origin in the social domain, because solipsism would seem to be the very antithesis of sociability – unless of course it is a psycho-analytically symptomatic denial... Be that as it may, let us now look at the “social” side of the homology.

Since solipsism is a private thought par excellence, the first idea that comes to mind concerning the social sphere is that of private property. This is all the more plausible in that at first sight it would seem that the institutional principle of private property is logically prior to market exchanges. But Sohn-Rethel argues that actually the relation is the other way around: the principle of “private property” is actually only a retrospective conceptualisation of necessities that are already inherent in the social act of exchange. Let us look at this more closely.

During the whole duration of an exchange transaction, the commodity in question must imperatively be withdrawn from the sphere of use. This is what we have already analysed above, where we noted that market exchanges induce a rigorous mutual exclusion between use and exchange. We now have to pursue this analysis, by examining the consequences of this separation for the consciousness of the agents. To do this, we will successively examine the two aspects: first that of use, then that of exchange.

  • Concerning use, we may note that the minds of the participants in the market transaction are each necessarily engaged with what they are planning to do with the merchandise once they have acquired it. If there were no such plans, the motivation for engaging in the transaction would disappear and the exchange would have no reason to take place. It is for this reason that an exchange is an abstraction which, in the last resort, is inseparable from use. But we may also note that these thoughts are essentially private: the specific content that each partner has in mind (whether one wishes to acquire some sodium chlorate for gardening, or to make a home-made bomb, for example) does not enter into the exchange as such.

  • Concerning now the exchange, we may note that it is an action, and that this action is social; but that it is not thought of as such by the agents. In a commodity exchange, whatever the agents think about it (and even if they are not thinking about anything at all other than their private motivations), two principles are tacitly implied: i) that of a mutual exclusion of property (what belongs to A does not belong to B, and vice versa); ii) the fact of obtaining one object and giving up another does not result from a direct, “natural” action (just to take an example, as in looting), but from an exchange involving mutual consent. In other words, the direct relation to nature is suspended, and replaced by a social relation.

To sum up: what the owners of commodities do in the context of a commodity exchange is effectively equivalent to pragmatic solipsism; and this is the case, quite independently of what the agents concerned may actually think or say about it. This establishes that there is indeed a telling correspondence between “solipsism” as a philosophical category, and certain aspects of the exchange abstraction.

The unicity of that which is

The first thinker in human history who attained the sphere of “pure thought”, a style of thought quite different from anything that exists in traditional communal societies, was Parmenides. His central concept is designated, in Greek, by the words το εον, which is generally translated as “the One; that which is”. This entity is intrinsically and perpetually unchanging; it occupies the whole of space; it lacks all the attributes of sensory perception; it is strictly homogeneous and uniform; it is indivisible; it is incapable of any sort of becoming or decaying; and it is forever immobile. Parmenides emphasizes that the reality and the being of this entity are such that it is intrinsically and literally inconceivable to think that it does not exist. This reasoning is central to his whole doctrine; and it marks the first time in the whole of human history that a conclusion is based on purely logical arguments. Thus, the το εον is the starting-point for a thought-process which proceeds by pure reasoning. In other words, what characterizes this style of thought, quite unprecedented at that time, is the fact that this purely conceptual thought grasps the dialectics of truth and non-truth according to the canons of logical necessity which is absolutely binding. Parmenides writes: “The fact of thinking, and the thought ‘it is', are one and the same thing. For you will never find any thought divorced from that which is, from what the thought is about. For there is not, and there never will be, any thing other than that which is.” Hegel was later to recognize himself perfectly in this stance, and comments: “This is indeed the fundamental idea. Parmenides marks the beginning of philosophy”.

We may note that the concept of το εον is a premise for the logical arguments of Parmenides; but the origin of the concept itself is enigmatic. One thing is clear at any rate: it is a radically non-empirical concept. It is indeed totally evident that no-one has ever seen (or heard, or touched, or tasted, or smelt) anything at all which bears the least resemblance to this το εον. In this respect, it is worth noting that neither Parmenides, nor any of the other founders of Greek philosophy, claim to have personally invented their key concepts themselves. Parmenides never suggests, for example, that he arrived at this concept by a process of generalisation on the basis of multiple cases in order to arrive at the level of a universal concept. The abstractions which underlie these concepts are of a quite different sort: one finds them already there, complete in themselves, totally without any process by which they could be derived. They come from elsewhere, outside and independently of any human thought. It is an anecdotal indication, but nevertheless very significant, that in the preface to his work Parmenides describes how he flew off (in a dream? – in a hallucination?) to the abode of Dike[44] , the goddess of knowledge of good and evil, and that it was her who initiated him into the wisdom that he is now proclaiming. And he adds, explicitly, that the goddess gave him a severe warning: “It is only by reason that you should consider and weigh the teaching that I have delivered to you”. Obviously however, in our present attempt to understand the genesis of abstract concepts, we cannot accept at face value this account of a divine gift. In other words, the question of the origin of this concept remains entire.

It is in this difficult situation that Sohn-Rethel proposes his audacious solution to the problem. According to him, the concept of Parmenides corresponds in quite exemplary fashion to a description of the abstract substance from which, ideally, money should be made. A market commodity can be exchanged between two private owners precisely to the extent that it has the capacity to be constituted as the object of a mutual exclusion of ownership. It is this capacity which makes it impossible for such a commodity to belong simultaneously to two different owners: a commodity is essentially one in the context of a rivalry between two owners.

What, precisely, does this “unicity” consist of? It has nothing to do with the indivisibility of the commodity considered as a material entity; it has nothing to do with its actual natural properties. In fact, what is brought into play is not the unicity of the commodities themselves, but the unicity of their existence. The ways in which a commodity can be perceived – as an object and in terms of its possible use-value – are as diverse as the persons who perceive it; but it exists in a single world which is common to all the private individuals, and this is the world of market exchanges.

The unicity of the exchange abstraction is thus absolutely fundamental, because it is this unicity which constitutes it as an instrument capable of realizing the social synthesis; in other words, of conferring on the society in question its coherence and its unity. I consider that Sohn-Rethel is perfectly correct when he points out the astounding formal concordance between this unicity of the exchange abstraction, and the ontological unicity of the το εον of Parmenides which is the founding abstraction of philosophical thought.

Abstract quantity

The work of the formalist school of mathematics, notably following Hilbert, have made quite explicit something which was up until then merely implicit in the whole of “pure mathematics”: this is the perfectly abstract quality of the cardinal numbers. These numbers are indeed defined by nothing other than the relation “larger than” (>), “less than” (<), or “equal to” (=). The fact that the very considerable work of the formalist school was necessary to make these concepts explicit is an eloquent indication of their abstract, non-empirical nature. “Numbers” as we experience them empirically are not at all built in this way (which explains the abstruse, non-intuitive nature of “formal mathematics” which has given such headaches to pupils and teachers alike in schools where a well-intentioned but possibly quite misguided attempt has been made to introduce this new programme of “modern maths”). Numbers as we come across them in daily life are never separated from the objects that are to be counted; what we can actually experience empirically are twenty sea-shells, or twenty cows (cf.The domestication of the savage mind: the invention of writing). But then, if the concept of “pure quantity” cannot be derived from empirical experience, where on earth could it have come from?

Sohn-Rethel, continuing his analysis of the exchange abstraction, sees an answer to this enigma in the following way. The act of exchange contains within itself the postulate that the two sets of commodities to be exchanged are equal. But how are we to define and to characterize this “equality”? It does not reside in the identity of the commodities, because if they were completely identical there would be no point in exchanging them; only different commodities are exchanged. Neither are the commodities considered to be equal in the minds of the agents, because their action would become absurd if they did not see any advantage in realizing the exchange. What is more, this sort of evaluation only exists in the solipsistic register of each individual conscience; from one person to another, such evaluations are not comparable. Nevertheless, it is of the very essence of the postulate of equality that it transcends the gulf of experience between the agents. The postulate of equality does not derive from their experience; the only thing they agree to is that the two sets of commodities can be exchanged. The two sets of commodities are rendered equal by the very act of exchange; they are not exchanged in virtue of any sort of “equality” that they possess in themselves.

An act of exchange of this sort, which ends up by postulating the equality of the sets of commodities, may well be preceded by a negotiation, by a sort of petty bargaining where what is at stake for each agent is “take more” and “give less”. Now it is true that many commodities can be measured in dimensional units (tons, gallons, square metres and so on). But the comparative terms “more” and “less” employed during the bargaining do not involve a quantitative comparison between, for example, tons of coal, gallons of petrol, or square yards of fabric. The relational equation postulated by an act of exchange leaves behind it all such dimensional measure, and establishes a level of pure non-dimensional quantity. At the end of all this we find, very precisely, the level of pure numbers defined by nothing other than “>”, “<” and “=”.

Before leaving the concept of pure quantity, we may note that it has a deep relationship with the concept of unicity that we examined just above in section b). At the level of pure thought, this relationship resides in the unicity of the cardinal numbers: the cardinal “12”, for example, is rigorously identical to itself, whether it results from the operation “9+3”, “4+8”, “2x6” etc., which all reduce to “1+1+1+1+1+1+1+1+1+1+1+1”. In the realm of the exchange abstraction, this corresponds to the fact, well-known to economists, that in the last resort there is only one currency. At the early stage in its development when money took the form of a precious metal, one might have thought that there would be multiple accounting units – gold, silver, copper, nickel and so on. But upon reflection (in the style of our fictional Athenian in a) above), it appears that any one of these metals could be used; and that if at a certain moment several of them are employed simultaneously, there must exist a single, well-defined exchange-rate between them. Besides, we have already seen that the precious metals are only a gross approximation to the ideal abstraction; because of this, they represent an obstacle to the free development of market exchanges, and this is what led to the invention of coined money. But even coined money, when there are multiple currencies (pounds, francs, dollars, yen etc.) does not attain the perfect from. At this stage, which is still an intermediate one, the ideality expresses itself in the fact that there must in principle exist a unique exchange-rate[45], with the result that theoretically all the different currencies communicate to form a single monetary system. We see here also that relative imperfections with respect to the “ideal” constitute a “friction” which hinders the ideal movement (cf. below). This explains the importance of establishing a common currency, as has quite recently been done in Europe with the euro. To sum up: the exchange abstraction is, ideally, both unique and an instance of pure quantity.

Abstract time and space

In the list of categories of synthetic a priori judgement, as Kant set them out, an important place is occupied by the concepts of time and space. This space is that of Euclidean geometry: it is notably characterized by the fact of being rigorously homogeneous and isotropic. As Jaynes (1976) [*]has pointed out with great perspicacity, time is only accessible to reflexive consciousness and indeed to scientific thought if it is metaphorically transposed to this conceptual framework of an ideal space: in this context, “time” is nothing other than a Euclidean point which advances uniformly along a straight line which is also Euclidean. It may not be necessary to dwell at length on the totally non-empirical nature of these concepts, since this thematic leitmotiv is becoming familiar. The space in which we move in the course of our daily life is anything but homogeneous and isotropic. As embodied beings, we are constantly subject to the anisotropic influence of gravity (in fact even this characterization is already idealised with respect to our phenomenologically immediate lived experience). And even the space of our movements in the two horizontal dimensions is not homogeneous. We have no perception of spatiality outside our actions (this is particularly clear in the “enactive” approach to cognition and perception). Now these actions are constitutively dependent on the particularities of our embodiment and of our natural Umwelt; and both of these are anything but homogeneous and isotropic. And as for time, considered as we have immediate lived experience of it, its “framing” by the metaphor of spatiality is in no way empirically given; and on the other hand, it is characterized by biological and psychological rhythms which, once again, are anything but homogeneous and linear. So where could the rigorous ideality of the Euclidean conceptions come from?

As we may expect, Sohn-Rethel sees the source of this ideal abstraction in the switch which comes when the categories of space are applied not at the level of use, but at the level of market exchanges. At the level of use, which we interpret here as covering the totality of all human activities in relation with nature, space and time are inextricably linked to natural events and human activities: as for example in the ripening of harvests, the seasons of the year, hunting animals, the birth and death of human beings, and generally everything that happens in the course of life. Now every act of exchange requires abstracting away from all this, because the commodities are supposed to be quite immutable during the whole duration of the exchange. The transaction does take a certain lapse of time, because one must include the delivery of the commodities and the payment which concludes the exchange. But the totality of this time is emptied of all the material realities which make up its content at the level of use.

Very similar considerations apply to space, for example the distance that the commodities must cover when they change owners. While the commodities are in transit from the old to the new owner, the equality between the two sets of commodities holds at each position and at each instant in exactly the same manner as at any other position and time. It is for this reason that time and space, when they are applied to the exchange, must be perfectly homogeneous. They are also continuous, in the sense that they allow for an interruption at any moment during the transit. In other words, the exchange abstraction excludes everything which makes up history, whether it be human history or natural history. The empirical reality of facts and events, and their descriptions which make it possible to differentiate one local time and position with respect to another, is entirely obliterated. This is how time and space acquire that character of universality and atemporality which must mark the exchange abstraction in each of its traits.

Substance and accidents

It is well known that Aristotelian logic operates a fundamental distinction between the “essential” properties of an object – in brief, the necessary and sufficient properties for an objet to belong to a certain class of objects (for example, being “a tree”, “a cat”, and so on) – and the “accidental” or contingent properties, those that an object can have (or not) without affecting its membership of a class (for example, the fact that a cat is grey or ginger). In its more highly developed form, this distinction becomes that between “primary” properties – in physics, these reduce essentially to the mass, the position and the state of movement of a particle – and “secondary” properties such as its colour, its sound, its small and so on. It is pretty evident that this conceptual scheme – which gives pride of place, need it be said, to the “essential” or “primary” properties – is the exact opposite of the empirical situation, for everything that can actually be perceived (and which are in fact “transductive” objects inseparable from the particularities of the perceiving subject) is relegated to the status of “accidental” or “secondary” properties. But if the “essential”, “primary” properties are non-empirical, where do they come from?

As we may imagine, Sohn-Rethel once again finds an answer in the exchange abstraction. In fact, we have already largely presented what is at stake: the “ideal” substance of which money should, ideally, be made is very precisely devoid of all sensory qualities; all that remains are the properties necessary for it to transit in abstract space and time. Let us recall, once again that we are dealing with an “abstraction” precisely because the use¬-value of a commodity (and without which it would actually not have any exchange-value either) is constituted precisely by its empirical qualities.

The continuous and the discontinuous

One of the grand themes which characterize the whole tradition of Western mathematics is the tense opposition between the continuous and the discrete (Salanskis, 1992[*]). Already in ancient Greece, this gave rise to the paradoxes of Zeno – Achilles who would arguably never quite catch up with the tortoise. Another key moment was the invention of differential calculus by Leibniz and Newton, notably for the analysis of movement; and more recently, there is the non-standard analysis developed by Nelson. Once again, it is a concept that does not arise in the empirical sphere of daily practice; and once again, Sohn-Rethel finds roots for it in the exchange abstraction. On the basis of what we have already said, and summing up, it is clear that an act of exchange must, intrinsically, be described as the abstract movement, in abstract space and time (i.e. homogeneous, continuous and empty) of abstract substances (materially real but devoid of any sensory qualities) which does not undergo any material change and which can only be differentiated in a quantitative and non-dimensional manner. However, on one hand the constancy of the exchange value confers a continuity to the whole process of exchange; but on the other, it must be possible to interrupt the movement of the commodities at any place and time in order to verify the constancy of their value, and this cuts their movement up into a number of discrete packets. This contradictory nature, both continuous and discrete, comes from the social origin of their abstract nature. In an analogous way, in order that the indescribable substance without qualities that money should ideally be made of may serve as an equivalent for every exchangeable commodity, in all possible proportions, this substance must be dividable ad libitum; money must be dividable so that the commodities can be left undivided. However, the continuity implicit in this indefinite divisibility is contradictory with the unicity that we examined in b) above. This contradiction corresponds to the tension, which comes up again and again in science, between entities that can only be discrete and indivisible (atomicity) and, on the other hand, the fact the ultimate entities can only be continuous (wave theories, which are expressed by the mathematics of the continuous).

The transcendental

A final element that can be brought to this list resides in the fact that above and beyond the relatively fine and specific details of the homologies we have examined in a) to f), there is an over-riding, generic characteristic of the conceptual categories. Although philosophers are general silent (not to say evasive) concerning the genetic origin of the Kantian categories, they are all agreed to emphasize the fact that they are both “given” a priori in a non-empirical fashion, and at the same time absolutely compelling in their apodictic normativity. “Logic” in this sense has the property that it could not be other than what it is. This is the meaning of the philosophical term “transcendental”. But where could this remarkable property come from? Once again, we find a corresponding characteristic on the side of the exchange abstraction. This abstraction is indeed founded not on empirical facts, but on social postulates; and there is a sense in which they postulates could not be other than they are, on pain of the entire edifice collapsing (and in this case, in the framework of a market society, all activities of production and consumption would cease and the whole society would materially collapse). We can make an impressive list of these postulates which all have in common this feature that on the one hand they are pure postulates, but at the same time endowed with a sort of intrinsic necessity. Thus: it is a postulate that the use of commodities should be suspended until the action of exchange is completed; that no modification should occur in the physical state of the commodities, and that this postulate must be maintained even if empirical facts would seem to run counter to it; that the commodities which are exchanged should count as equivalent in spite of all their manifest empirical differences; that the fact of acquiring and giving up commodities is bound to a priori conditions concerning their exchangeability; that commodities change owners by transiting from one place to another without being materially affected, and that this movement occurs in an “empty” space. None of these formal concepts invokes any sort of empirical, factual observation; they are all norms that the exchange of market commodities must satisfy in order to implement the social synthesis.

Characteristics of the exchange abstraction at the birth of capitalism

The “homologies” between aspects of the exchange abstraction and major conceptual categories that we have noted in Characteristics of the exchange abstraction in Ancient Greece: the homologies, all refer to the first flowering of the exchange abstraction, with notably the invention of coined money, which occurred in Ancient Greece. There ensued the period of the “Dark Ages”, and then from the 5th to the 15th century the period of the feudal “Middle Ages”. From the point of view of developments in the exchange abstraction, all this corresponds to an interlude. It was with the European Renaissance that the exchange abstraction entered into a new phase of growth and development; more specifically, this was the period of the birth of Capitalism that we have described, from the point of view of the social synthesis, in The origin of capitalism. The question is now whether it is possible to identify a set of corresponding intellectual mutations in the forms of thought; and it is indeed, because the Renaissance is generally recognized as cradle of “Modern Science”. We will now continue our identification of “homologies” between these new social forms and corresponding thought-forms.

Modern science

The first new “homology” is quite general, and concerns as such the birth of “Modern Science” – that is to say, a system of knowledge which combines an explicit theoretical basis with extensive empirical observations and including a deliberately experimental approach. We can better appreciate the originality of “Science” in this sense if we compare it with Greek thought. A major point of contrast is that Greek thought appears as essentially “philosophical”, in the sense of thought which is purely abstract and speculative. Greek philosophy is literally inspired; its fruitfulness in the identification of what we now call the categorical concepts of a priori synthetic judgement is unequalled (it has been said, with justice, that there is hardly any major concept – from the idea that the Sun is the centre of the planetary system, to the atomic theory, to the glories of Euclidean geometry – that the Greeks did not have a dazzling intuition of); however, the fact remains that the Greeks themselves did not really put these concepts “to work” in order to obtain substantial positive knowledge. For that, it was indeed necessary to wait for the Renaissance; as its name indicates, this was the re-birth of the Greek concepts – the nascent modern science did not have to invent de novo many of its concepts, they were essentially all there already in Ancient Greek thought. But what modern science did do, notably with Galileo, was to articulate these concepts with the domain of concrete material production (Galileo was well known for seeking out the company of the artisans at the arsenal of Venice). The parallel with the change in status of the exchange abstraction – limited to the domain of external commerce for the Greeks, invested in the very processes of material production only from the Renaissance onwards – is striking. We may note in this regard that Galileo's science was immediately “applied”, notably for calculating the ballistic trajectories of cannon-balls which he showed to be parabolas. Newton extended the same principles to the movements of the planets, which he showed to be ellipses with the Sun at one of the foci; this brought the astronomical sphere – up until then considered to be “stellar” and radically out of reach, with a nature radically different from that of the “sub-lunar” sphere – within the fold of natural processes as these can be observed and experimented with on Earth.

Dynamic inertia

Another homology, more precise and detailed, concerns a case where in spite of all that we have said concerning the precedence of the Greeks at the conceptual level, there is a purely theoretical innovation that occurred at the Renaissance: this is an important change in the concept of inertia. For the Greeks (and in particular for Aristotle, whose “physics” continued to be dominant until the advent of modern science), inertia was conceived in profoundly static terms. In other words, “inertia” was understood as the state of rest, and only this state of rest, so that movement of any sort required an effort, an impetus, in order to produce or even simply to maintain it. This effort did not reside in the things themselves, but had to be provided by a human being; and even when the movement occurred out of reach of humans, the effort that one imagined necessary to produce it came from material forces endowed with a capacity for action analogous to that of human beings. As Sohn-Rethel remarks, this conception is indeed in conformity with the empirical reality as we experience it in daily life; we may note that this static connotation remains even today for “inertia” as a word in ordinary everyday language.

The concept of a dynamic inertia, which is the signal invention of Galileo himself that Newton would later take up and develop systematically, marks therefore a striking break with tradition. It may be useful to emphasize the radically non-empirical character of this concept, which will play a decisive role in the future development of modern science. The concept of inertial movement can be formulated thus: “A body in movement, left to itself, will persist in its movement, and will continue to move in a straight line at constant velocity, as long as nothing intervenes to prevent it from behaving in this way.” Koyré has very rightly emphasized the incredibly audacious character of this conception at the time. “The principle of inertial motion seems to us perfectly clear, plausible and even obvious... This concept seems to us today so “natural” that we even believe that we have derived it from experience and observation, even though no-one has ever encountered this sort of movement – indefinitely continuing in a straight line at constant velocity – for the good and simple reason that it never occurs and in fact is impossible... We are no longer conscious of the daring and paradoxical presumption of Galileo, which consisted of treating mechanics as though it were mathematics, and of substituting for the real world such as we have experience of it a world of geometry made real, and of explaining the real by the impossible”.

But now, if the concept of a dynamic inertia is non-empirical to this extent, where on earth (!) can it have come from? We come back to the form of questioning that has become familiar in the course of the previous section. The idea of Sohn-Rethel consists of identifying a formally homologous change in the form of the exchange abstraction. His point is that the purely commercial exchanges of the Greeks, even though they were quite highly developed, remained essentially optional and episodic. There was no intrinsic necessity for money to continue to circulate: each new voyage for commercial purposes was in a way a new start, which required a new effort to equip the boat, to gather the necessary funds and commodities. There were no real cycles of accumulation; in this sense, the system was not fully capitalist. Above all, as we have mentioned in The historical genesis of money, the basic system of production was carried out in a “domestic” context, by slaves. By contrast, when money is invested in the processes of primary production of the goods of daily consumption, and above all when the social synthesis includes the labour force itself – and this is what occurred with the birth of Capitalism – the circulation of money cannot cease, on pain of a collapse of social life itself. And this does indeed correspond to the concept of “dynamic inertia”.

The argument of Sohn-Rethel is attractive; but there is one point on which it is not quite clear. This “imperfection” may actually be useful to us later in our own “detective puzzle”, when we shall attempt to “interpret the homologies” (see Interpreting the homologies below) and to better understand how cognitive forms and social forms are actually articulated. This “imperfection” is the following: we may well admit, following Sohn-Rethel's argument, that the movement of money can be characterized by a ceaseless “constant velocity”; but what about the fact that it continues “in a straight line”?! At the level of the exchange abstraction, one says that money “circulates”, which implies that in it is the natural order of things to come back to its own starting-point (for an individual economic agent, money must come back to the starting-point, so that his book-keeping can be properly balanced). It is therefore interesting to note that Galileo himself did not seem to be perfectly clear on this point. In the Discorsi of 1638, Galileo describes inertial movement as the movement of a body which persists in a constant trajectory, at constant velocity, parallel to the surface of the Earth. This gives the impression, at the very least, that Galileo conceived of inertial movement as being, precisely, circular. This brings us to the third level of analysis, in particular the contribution of Newton.

State-determined dynamic systems

Newton clarified the obscure point left by Galileo that we have just noted. By means of differential calculus – an extremely important mathematical tool that he contributed to invent – he showed that circular movement is precisely not at constant velocity, because there is a permanent acceleration in a direction orthogonal to that of the movement at any point in time. More generally, the absolutely fundamental contribution of Newton consisted of elaborating the concept of a State-Determined Dynamic System (SDDS). A system of this sort is perfectly “autonomous”: once it is set up, and the dynamic law governing the temporal evolution of its “state” is specified, everything thenceforth occurs without the least “external” intervention. Laplace, who emphasized the radical determinism of such a system, is reputed to have replied to Napoleon when the latter questioned him about the place of God in his system: “Sire, I have no need of that hypothesis”. Once again, this grand scheme is radically non-empirical; it comes from the integration of practically all the non-empirical categories that we have already mentioned. And so once again, we come to the question of the origin of such a conceptual scheme.

As we may anticipate, Sohn-Rethel proposes to look towards the historical state of the exchange abstraction. We may recall that this is the historical period when the capitalist entrepreneur, who up until then had been content to buy commodities already made in order to enter into purely commercial exchanges, begins to invest in the means of production by investing in the workshop (or a little later the factory), the tools, the raw materials, and above all the labour force of salaried workers. It is interesting to note that it is common to speak of a capitalist of this sort as a “manufacturer” – as though Mr Ford, for example, had really made thousands and thousands of cars with his own hands; but this is misleading. How does the capitalist fulfil his role as a “producer”? He does not accomplish this by his own work: he achieves it neither with his hands, nor with tools and machines that he would operate himself. He achieves it by means of the money he has invested as capital, and with nothing else. “The process of work is a process between entities that the capitalist has bought”, says Marx, “entities that belong to him”. In fact, if ever a capitalist did come to lend a hand himself, that would only show that he had partially failed in this role as a capitalist entrepreneur, and strictly speaking he should pay himself for that manual work. In other words, the role of “producer” falls on an individual who does not perform a single productive function in the work-process. To sum up the essential point: the key characteristic of the production process, from the point of view of the capitalist entrepreneur who invests in it, is that this process should function all by itself.

Thus, the power of the capitalist resides in this postulate of the self-acting or “automatic” nature of the production process. It is important to note that a postulate of this sort does not necessarily correspond to a historical reality; in fact, as we saw in The origin of capitalism, it required centuries before, very progressively and even then imperfectly, the social reality of production relations began to assume the ideal form that we have just described. This only makes it clearer than ever that the postulate of the automaticity of production processes does not come from any empirical source in the actual technology of production; it is rather the other way round, the fact that in the course of the historical evolution of technology the latter progressively come to be conform to the “ideal” in question is a consequence rather than a cause of this postulate. The postulate itself is in no way empirical; it is clearly in the realm of the non-empirical a priori; and what we have seen is that it is formally intrinsic to the social relations of production in a capitalist society. The formal homology between this postulate, which is social through and through, and the Newtonian concept of a SDDS, is I think quite impressive.


Finally, to complete this list of “homologies” between forms of social life and forms of thought, I will consider a major scientific theory where the social implications – both “upstream”, in terms of its sources, and “downstream” in terms of its consequences – are perhaps somewhat easier to read than elsewhere; although even here, as we shall see, they are by no means trivial or obvious. The theory in question is Darwin's great conception of biological evolution through natural selection; to which we shall add a theory of heredity, Mendelian genetics, which was not available to Darwin himself but which was necessary to arrive at the full neo-Darwinian synthesis which is the currently accepted theory of evolution. I shall also allow myself some additional comments on some weaknesses in this neo-Darwinian theory which are not widely recognized, but which I consider to be both scientifically interesting, and also revealing in terms of the relation of this scientific theory to the society in which it has appeared. I will treat this example in considerably more detail than those I have listed above, for two reasons. One reason is that it is a domain in which I have worked myself; and it links up with The problem of the origin of a genetic system where I considered the question of the “origin of a genetic system” from a purely scientific point of view. The second reason is that, as I have just said, the interpretation of this homology may be somewhat more accessible than in most of the other cases; and because of this, in what follows here I will make a start on not only identifying some additional “homologies”, but on moving forward to interpret them.

We may start with an aspect that provides a relatively clear and simple entry-point. Even though this aspect is not really at the heart of the questions involved, this whole area in which we are seeking to identify and above all to interpret “homologies” between social forms and knowledge forms does as I have said constitute a truly “hard problem”, and we have no reason to put ourselves above taking advantage of places where the going is just a little easier. This entry-point here is the fact that Darwin quite openly admitted that in working towards his great idea of “natural selection”, he had been inspired and informed by the practices of artificial selection developed by plant and animal breeders in order to produce varieties appropriate for agricultural practice. A point that particularly interested Darwin was that the breeders were not able to produce at will the forms that they were interested in; they had to wait for forms that they found interesting to arise, and it was only afterwards that they could select those forms. This gave rise to the salient feature of Darwin's theory: biological evolution is directed by natural selection, the variation offered up for selection being itself “random”. We will have occasion to come back to this concept of “random” variation[46] ; but for the moment, the important point is that this allowed Darwin to transpose the scheme of artificial evolution to the natural evolution of biological species, because there was no need for an external “intelligence” (divine or otherwise) for the process to occur. Even the “guiding principle” of “natural selection” was “blind”, because the criterion of selection among the available variants was simply viability, or more precisely reproductive success (for which viability is a necessary but non-sufficient condition). Darwin's theory was of course highly controversial at the time, notably because it went against Biblical accounts of the Creation by God (“In the beginning...”[47] ). If we accept Durkheim's proposition that Religion forms the backbone of the traditional social structures, what we are witnessing here is indeed a radical innovation.

With this, we come to a first level of relationship between this scientific theory, and the historical society within which it saw the light of day. There is indeed a fairly clear “homology” between the process whereby composition of a biological population is fashioned by the predominance of the forms with the highest rates of reproduction, and the process whereby the composition of the firms in a capitalist economy is fashioned by the predominance of those which have the highest rate of profit. And now we can try and tackle the “hard problem”: how are we to interpret this homology? The first possibility is that the scientific thought-forms had an effect on the social forms. This corresponds to a current which did indeed exist historically at that time, which went by the name of “Social Darwinism”; it involved concepts such as “the survival of the fittest”, and even “Nature Red in Tooth and Claw”. This doctrine, developed notably by Herbert Spencer, provided an ideological justification for the social order of a capitalist economy, by arguing that the domination exerted by the rich and the successful, in particular the exploitation of the workers by capitalists, was nothing other than the natural order of things. However, this doctrine was very far from universally accepted, even at the time, and seems quite insufficient to account for the “homology”. The second possibility is that it was because of the social form – a flourishing capitalist economy – that Darwin was able to come to this scientific theory. In very general terms, there is most probably something in this; but the difficulty is to disentangle the routes whereby this “conversion” from social to cognitive forms could have operated.

Coming back to the science itself, there was a serious weakness in Darwin's own original formulation of his theory, which was situated at the level of the mechanisms of heredity. Darwin worked on the basis of an apparently plausible and commonsensical “blending theory” of inheritance, which was widely held at the time, according to which offspring have values for any quantitative trait which are intermediate between those of their parents. The problem with this, from the point of view of population genetics, is that the population would quite rapidly become uniform, and there would no longer be any variation for natural selection to operate on[48] . The first antidote to this problem that came to mind was to invoke a substantial injection of new variation at each generation, to compensate for that lost by blending; and Darwin himself spent quite some time worrying about where this new variation could come from, even going so far as to envisage some form of inheritance of acquired characteristics. What is now considered to be the true solution lay however in adopting a different theory of heredity: Mendel's scheme of particulate inheritance, where the “genetic factors” do not blend at each generation, and “segregate” out uncontaminated by their cohabitation with other factors. The synthesis between the theories of Darwin, Weismann and Mendel gave rise to the “neo-Darwinian synthesis” which is the core of current orthodoxy in evolutionary biology. It is worth taking a little time to expand on the concepts involved.

It is more than possible to relate Mendel's conception in close relation to experimental findings. In this vein, it is important to start out with two parental true-breeding lines which differ qualitatively in an all-or-nothing phenotypic difference: for example, “Red” versus “White”. The first step in the experiment then consists of crossing these two lines to produce a hybrid F1 generation. These individuals are all phenotypically “Red”. The next step consists of crossing these individuals amongst themselves to produce a second hybrid generation, the F2. Remarkably, what is observed is that qualitatively, this F2 generation comprises both “Red” and “White” individuals; quantitatively, they are in a statistical ration of 3 Red : 1 White. It is then possible to argue as follows. The F1 plants must have received “something” from their “white” parental line, because they have white offspring whereas the parental “red” line never did. These F1 plants must also have received “something” from their “red” parental line, because by the same argument they have red offspring, and even more simply because they are themselves “red”. At this point, there are grounds for making a principled distinction between the “phenotype” – the external appearance of the plants – and the “genotype”, their genetic constitution made up of these “somethings” that they receive from their parents and pass on to their offspring; the reason being that the F1 plants have the “red” phenotype, but their genotype is manifestly different than that of the red parental line since when crossed amongst themselves they have some phenotypically “white” offspring. The time is now ripe to put forward an explicit hypothesis, and to introduce some technical terms. The hypothesis is that the genotype of an individual is composed of two “somethings” (that we now call “genes”), one being received from each parent; these genes can be of two types (technical term “alleles”), one characteristic of the red parental line denoted by R/, and the other type characteristic of the white parental line denoted by W/. This makes it possible to set up a table of correspondences between genotype and phenotype:

Table 6







R//W (or W//R)


The hypothesis is completed by specifying that when an individual becomes a parent in its turn, it passes on just one of these two genes, at random (i.e. (50% chance for each), to each of its offspring. Thus, the genotypes of the F2 generation will be 25% R//R (phenotype Red); 50% R//W (phenotype Red, again); and 25% W//W (phenotype White). The hypothesis thus gives the result that phenotypically, the F2 generation should be 75% red, 25% white – this is of course the least it should do, give back the result that it was based on. But the hypothesis is also able to generate novel predictions – for example, that the “backcross” of the F1 with a “white” parental line should give offspring that are genotypically 50% R//W and 50% W//W, hence phenotypically 50% red and 50% white. These predictions, and many others of the same sort[49], were fulfilled; so that the Mendelian theory is now generally accepted.

Thus, as I said, it is indeed very possible to present the Mendelian concept of a “gene” in close relation to empirical observations. But does this justify an “empiricist” view, according to which the concept of a Mendelian gene derives simply from theory-free observations? I will now argue that it does not. An initial point is that although the “Mendelian experiment” as I have presented it here, in deliberately schematic form, can indeed be performed in good faith, it is actually tailor-made to illustrate the core thesis concerning the particulate nature of “Mendelian factors” or “genes” as we now call them. One aspect of this is that the “character” in question, the one that is different between the two parental lines, actually has to be chosen quite carefully; many, indeed most “characters” do not exhibit “Mendelian” behaviour in crosses of this type. This does not decisively refute the Mendelian theory, because retrospectively results that do not conform can be “explained away” by supposing that the difference in question is controlled by two or more genes. But it does reveal that the Mendelian experiment is not as straightforwardly demonstrative as it might seem. In this respect, it is worth noting that those with a pragmatic interest in animal and plant breeding had never, in the course of centuries of practice, found occasion to design and to carry out a “Mendelian” experiment. From this practical, fully empirical point of view, the Mendelian experiment appears extremely “abstract”; and indeed it is known in such circles as “formal genetics”, a phrase which conveys a lot... To argue more positively on the “non-empirical” side of the question, it is legitimate to point out that “Mendelian factors” or “genes” are entities with a set of remarkably ideal properties. Perhaps their most striking feature is that they can be reproduced indefinitely while maintaining a strictly identical form. They are not influenced by the environment (emphasizing that point was Weismann's signal contribution); they are not influenced either by the organism they inhabit (although they are dependent on the organism for their reproduction); they are not even influenced by their cohabitation with other genes in the genotype (they segregate out unaffected). They are not absolutely immutable – in fact there does have to be some mutation to give rise to the allelic variants. But since this variation is not lost (as it would be if there were blending inheritance), the novel variation arising by mutation can be kept to a strict minimum; and indeed virtually all contemporary organisms, even the simplest now extant such as bacteria, have fairly sophisticated mechanisms for “repairing” the bulk of accidental variants that arise. Gene lineages are not absolutely immortal; any particular gene lineage can die out, and sooner or later most of them do. But they are potentially immortal; to illustrate this, all the genes in contemporary organisms have uninterrupted lineages that go back.... not to the very first living organisms (if we admit that autopoïetic dissipative structures are minimally “alive”, nor even to the very first “genetic system” (see part 3), but to the moment when the present system based on nucleic acids staged a successful “genetic takeover”, an event which occurred thousands of million years ago. This is not a bad approximation to infinity, especially when we consider that this will be able to continue indefinitely, as long as there is life on Earth.... To sum up, then, the “gene” as a scientific entity does have some strikingly “ideal” features, with a general flavour of intrinsic, logical necessity to them. This is, once again, reminiscent of the exchange abstraction, and money in particular; so we have another “homology” to add to our list. As we have already amply seen, interpreting such homologies is a “hard problem”, and we are not going to fully solve it here. There is, however, an opportunity that we shall explore. What we have seen in that the Darwinian theory of evolution by natural selection does have some relationship with the functioning of a capitalist society; and now we add to this the necessity for a genetic system that also appears to have its own relation with the exchange abstraction. The opportunity in question consists of pursuing the question of the relation between these two aspects in the case of biological theories of evolution; and then seeing if we can gain some new insights by “running the homology backwards”, i.e. examining what the strengths and weaknesses of the biological theories can tell us about the functioning of the capitalist society we live in.

What then is there to say about this view of biological evolution, which uses basic concepts which do seem to have some relation with the exchange abstraction? Well, one striking feature is that the organisms themselves, as they “enact” their lives in flesh and blood, bringing forth fascinating ecological niches for themselves as they do so, are quite conspicuously absent from the story in its currently orthodox neo-darwinian version. This is perhaps not quite fair to Darwin, who was never a model student at Cambridge University but did become a passionate amateur naturalist, avidly collecting beetles along with fellow undergraduates. And a little later, of course, Darwin embarked on his famous trip on HMS Beagle, where he served as naturalist on a round-the-world voyage which lasted five years. But what posterity has retained of Darwin's work is not his activity as a field naturalist, but the grand theoretical scheme of Evolution, with the mechanism of natural selection; and so what is more fully relevant here are the biological, organismic aspects of this process. And here, as we have already indicated above, a point of entry into the weakness of this conceptual scheme is Darwin's characterisation of the genetic variation in natural populations as “random”. It is quite correct to say that they are “random” with respect to the natural selection process itself; they are not “pre-adapted”, in which case there would be no call for selection. But for a biologist with a “feeling for the organism” (Fox-Keller, 1983)[*], they are not random at all. If one takes properly into account the minute-to-minute physiological processes which constitute the autopoïesis of the organism, the variations which can arise are both constrained and enabled by the intrinsic organisation of these processes, and so they are anything but random.

An analogous point arises with respect to the ontogeny of the organism – the set of processes which start with the fertilisation of an egg-cell, which lead to the events of embryogenesis, and which continue with maturation and ageing until the death of the organism. A “gene-centred” view of the organism qualifies the regularities in ontogeny as resulting from a “genetic programme”; but the very notion of a “programme” as something external to the process itself is a hypostasis, talking about the resulting regularities of this complex, ongoing dynamic process as though it were the cause of the process. Even worse, in a way, is that even if one were to admit the heuristic usefulness of the concept of a “programme”, there are no good reasons at all (other than genocentric prejudice) for considering that the “programme” is genetic. This point is sufficiently instructive that it is worth spelling it out. In order for a “partitioning” of causes into “genetic” and “environmental” to make any sense, one must be in a situation of partitioning causes of variation. Thus, it is properly meaningful to ask whether the specific differences between individuals in a population are due to genetic differences or to environmental differences; technically, the proportion of variation due to genetic differences is termed the “heritability”. It is important to be clear that the heritability is always specific not only for a particular character, but for a particular population: the heritability for the same character can perfectly well be low in one population and high in another. As it turns out, the heritabilities for quantitative characters in natural populations are rather generally of the order of 50% – which does not get us very far! But what is even more important is that if the character in question is invariant in the population, then the heritability of that character is 0/0 – i.e. neither high nor low, but simply indeterminate and meaningless. Thus, attempting to explain the ontogenetic regularities – the lack of variation – by claiming that the “programme” is “genetic” is simply nonsense. The costly and rather mindless “genome project”, consisting of obtaining complete DNA nucleotide sequences, has at least had this significant result: these sequences tell us practically nothing about the organisation of ontogeny.

What do these signal weaknesses reveal if we now turn back and look again at the other side of the “homology”, the “social” side where the exchange abstraction lies at the root of the social synthesis? We find that there is indeed a corresponding vacuity in the role played by the capitalist entrepreneur. We have already noted above that the capitalist does not fulfil his role as a “producer” by his own work; he achieves it by means of the money he has invested as capital, and with nothing else. From the point of view of the capitalist entrepreneur, the key characteristic of the production process is that it should function “all by itself”, “automatically”. In fact, there is a certain irony in the situation: while the attention of the capitalist is focussed almost exclusively on his rate of profit, it is actually quite difficult to evaluate it. Adam Smith (1776) [*]wrote eloquently on this right at the start of the capitalist era:

“Profit is so very fluctuating that the person who carries on a particular trade cannot always tell you himself what is the average of his annual profit. It is affected not only by every variation of price in the commodities which he deals in, but by the good or bad fortune both of his rivals and of his customers, and by a thousand other accidents to which goods when carried either by sea or by land, or even when stored in a warehouse, are liable. It varies, therefore, not only from year to year, but from day to day, and almost from hour to hour. To ascertain what is the average profit of all the different trades carried on in a great kingdom must be much more difficult; and to judge of what it may have been formerly, or in remote periods of time, with any degree of precision, must be altogether impossible.”

In a way, the situation is almost worse than that: one of the more robust generalizations in economic theory is that there is a “tendency of the rate of profit to fall”, commonly abbreviated to TRPF. The reasons for this are much discussed; one plausible explanation, that has the merit of simplicity, is that in periods of relative stability, straightforward economic competition will lead to a decrease in prices to the bare minimum – which corresponds to a profit-rate close to zero. It is interesting to note that a quite closely analogous situation occurs in biological evolution, which is characterized by long periods of stability where the various species simply maintain themselves – which corresponds indeed to a growth-rate of zero.

Now according to the theory of Eldredge and Gould (1972)[*], in the biological realm these long periods of “equilibrium” are “punctuated” by relatively short episodes where there is a proliferation of innovative forms. A major example of this is the “Cambrian explosion” which followed the first appearance of multi-cellular animals (1989)[*]. Coming back to the economic system, Schumpeter (1934[*]) has interpreted the “Kondratiev cycles”, where periods of economic recession followed by new growth occur at intervals of approximately 40-60 years, as the result of technological innovations: the steam engine (opening the way to the industrial revolution, 1790-1850); the railways, which gave a dynamic impulsion to the economy over the period 1890-1930; and the automobile industry after the Second World War. We might well add to this, in the current period, the “NTIC” (New Technologies of Information and Communication). Thus we see that just as it is not possible to understand what happens in evolution by looking just at the genes, but that it is necessary to look at the material processes whereby living organisms are produced, so it is not possible to understand economic history by looking just at monetary capital, but it is necessary to look seriously at the actual material processes involved in the production of social life.

Interpreting the homologies

Is it a pure accident that Plato and Aristotle laid the foundations for Western philosophy at the same historical place and time that coined money was invented? – that Galileo and Newton laid the foundations for Modern Science at the same historical place and time that saw the birth of Capitalism? The question is admittedly rhetorical: our elucidation of the various “homologies” in Characteristics of the exchange abstraction in Ancient Greece: the homologies and Characteristics of the exchange abstraction at the birth of capitalism was carried out under the auspices of a presumption that we are not dealing with mere coincidences; and this presumption was indeed strengthened rather than weakened by the more detailed examination of each case. But this brings us face to face with the question: if the “homologies” are not mere coincidences, how are we to explain them? I have already alluded several times to this question of the relation between forms of social life and forms of thought as being a “hard problem”. One of the features of “hard problems” of this sort, is that there is a real difficulty in knowing how to formulate them correctly. An awful lot of mischief can come from an improper formulation – for example, if it involves a category mistake. In the present instance, where we are trying to get to grips with the relation between social forms and thought forms, what I propose to do is to start out with a relatively straightforward approach – bearing in mind that if it does not work, this may give us some clues as to a better way to formulate the question itself.

In a general way, if there is a significant correlation between two entities A and B, this can be for three reasons: a) because variations in A are the cause of variations in B; b) because inversely, variations in B are the cause of variations in A; or c) because there is a subjacent variable X which is the cause of variations in both A and B, without there being any direct causal relations between A and B. We shall now examine each of these possibilities, referring in particular here to the “homologies” identified by Sohn-Rethel.

  • The first possibility, then, is that the social form (the exchange abstraction) is the cause of the forms of thought (the a priori categories). It would seem that Sohn-Rethel adopts a position of this sort; he clearly identifies himself with the Marxist tradition, according to which ideological “superstructures” are to be understood as deriving from “infrastructures” which are socio-economic in nature. There is, however, what I consider to be a telling objection to this first interpretation, at least in its simple, basic form. The whole cognitive interest of the a priori categories resides in the Kantian critique of empiricism, the fact that these categories cannot derive from their own pre-given objects of which they would be some sort of “reflection”. We have seen how, time and again, Sohn-Rethel emphasizes this non-empirical feature of the “categories”. After this, it would be ironically disappointing if they turned out to be straightforward “reflections”, not indeed of their material objects, but of another sort of pre-existing entity, to wit the forms of social structures. This would seriously undermine all the good work of the Kantian critique of empiricism.

  • What then of the second possibility: might the homologies be explained by the fact that the social forms derive from forms of thought? This is just the reverse of a), but it is not any more satisfactory; in particular, it leaves completely unsolved the question we started with, i.e. the origin of these conceptual categories themselves. There may be something in it; in particular, it is quite conceivable that certain sorts of cognitive and conceptual capacities are necessary for certain social forms to function; but in its simple form it is not a satisfactory answer to our “hard problem”.

  • What then of the third possibility: that variations in social forms and in thought-forms might both derive from an underlying variable X? In general terms, this is quite an attractive possibility; in particular, it may contribute to a reformulation of the original question, which as I have indicated above may well be necessary. However, if this is to be more than a mere abstract possibility, we will have to move towards a more positive identification of what this hypothetical “X” might be. Of course, for this to be anything more than an empty formalism, it will be necessary to make at least a start in identifying what this factor X (or, quite plausibly, what set of factors) this might actually be.

This question is not at all the same as the traditional “hard problem” in cognitive science, that of the relation between body and mind. As another “hard problem” it does, however, bear something of a “family resemblance”; and this being so, we may be able to garner a useful clue by recalling how the paradigm of Enaction addresses the mind-body problem. The key insight here is that “the mind” is not some sort of ghostly “homunculus”. This can be seen most clearly in the case of perception. Perception is not the reception, by some such homunculus, of incoming sensory stimulations which somehow give rise to percepts. Rather, perception is a feature of the sensory-motor dynamics; it is essential to act in order to perceive. Indeed, as O'Regan and Noe (2001) [*]have pointed out, the famous “qualia” of the traditional senses are related not to the sensory organ alone, but to the “sensory-motor contingencies”, the way in which sensory feedback varies as a function of particular actions which are taken. This is also the basis for Bach-y-Rita's “sensory substitution” demonstrations, which give a practical illustration of the fact that it is the sensory-motor dynamics, and not the sensory organ per se, which characterizes a perceptual modality. The upshot of all this is that there is no perception without action; and the key point is that in order to act, a cognitive organism must indeed have a body. Thus, “the body” is not something that has to be somehow added in afterwards: it is there right from the start, with an essential role in constituting “mind” in the first place. This is summarized by the very title of Varela's book “The Embodied Mind” (1991[*]).

Carrying this insight over to our problem of the relation between forms of thought and forms of social life, this means that the “forms of thought” in question are not to be considered purely in the abstract, but must be related to a quite concrete set of activities. As a starting-point, we may consider the exemplary case of an “underlying variable” which does have effects both on social forms and on thought-forms: to wit, the technique of writing. In The domestication of the savage mind: the invention of writing, we examined in some detail how it is that writing is both inseparable from a certain sort of social relations, and at the same time an integral part of a certain sort of cognition. Now the issue we are trying to get to grips with here is the relation between the “exchange abstraction” – to put it simply, money – and the emergence of certain forms of thought. This raises the question, quite naturally, of the relation between money and writing. And as it turns out, there does seem to have been a very close connection, historically, between the invention of money and the invention of writing. Both were invented around 3000 BC in Mesopotamia; and in fact the relation is quite direct, since writing was actually developed as a record-keeping vehicle for commercial transactions. Moreover, there seems to have been a parallel in their historical development. Writing started in the form of pictograms, which became stylised as cuneiform script on clay tablets. The first forms of money, used indeed as the substrate for commercial exchanges of many sorts, were nevertheless “in kind”: grains of barley, and then precious metals from lead and tin to silver and gold. After their initial invention, both money and writing developed progressively over the centuries; and this parallel development reached a landmark stage in Greece around 650 BC, which was the time and place of both the invention of alphabetic writing (so that the individual signs no longer have immediate meaning, but only in specific combinations), and the invention of coined money (so that the value of money became abstracted from the actual value of the material substrate).

In The domestication of the savage mind: the invention of writing, we have already examined in some detail the way in which the technique of alphabetic writing enables the development of philosophical thought, in particular in its mathematical and formal aspects. The question now is, what is added with the inclusion of the institutional technique of coined money into this nexus of elements? Recalling our “clue” from the comparison with the mind-body problem, the key insight was that the “forms of thought” characteristic of society at a particular historical epoch are not to be considered purely in the abstract, but must be related to a quite concrete set of activities. What the institution of money does, is to establish a connection between the concepts proper to the realm of philosophical thought, which may seem to be purely abstract, and the most mundane but ubiquitous dealings of everyday life. We may recall here Sohn-Rethel's thought experiment, concerning a philosophically minded Athenian who asks himself searching questions about the coins of money in his pocket: “What sort of substance should these coins be made of?” He comes to the conclusion that he is dealing with a remarkable entity, something which is obstinately and uncompromisingly real and yet which is detached from any of the sensory qualities by which things are usually real for us. And things begin to fit together here: because this is, indeed, the realm of social reality. We may also recall here another aspect: the use of money, in order to effect the social synthesis, requires the development of certain quite particular cognitive capacities. As Sohn-Rethel remarked with respect to his dog at the butcher's, the meaning of money is quite beyond the ken of animals, and indeed of young children.

Continuing in the same vein, the nexus formed by the association of techniques of writing and of money also illuminate the great civilisational mutation which occurred at the Renaissance: the birth of modern science, and the advent of capitalism. We have mentioned this “homology” as point (h) in the list above. The difficulty in interpreting this homology is to understand what relation there could be between the social activities of the first entrepreneurial capitalists, and the cognitive activities of scientist in their laboratories, seemingly “ivory towers” pretty much divorced from the ongoings of everyday life. It is here that Latour's notion of “cycles of accumulation, which we examined in some detail in Cycles of accumulation, is particularly relevant. If we generalize the technique of writing to include inscriptions, in particular those produced by scientific instruments, we see that a key characteristic of “modern science” involves the re-investment of previously produced knowledge as a basis for producing even more knowledge – in a dynamic which is so closely analogous to that of the reinvestment of capital in order to produce even more capital that it becomes less enigmatic to understand how these two innovations should have occurred at the same historical epoch.


It is as well to be brutally clear that these considerations do little more than scratch the surface of our “hard problem”; working out more fully how it is that relations exist between forms of social life and forms of thought will remain a major unsolved point on the agenda of Enactive cognitive science for years to come. This may be a fitting note on which to end: for if there is one realm, even more than all the others, where “reality” is brought about, is enacted, it is indeed this domain of social reality. As a final note, we may come back to the theme I set off with, “qui aime bien châtie bien”. My hope is that I have said enough about what are currently undeniably weak-points in the paradigm of Enaction, to open up the perspective (and maybe trigger some ambitions?) of tackling them vigorously enough to turn them into growing-points.


arrow to header

Tom Froese

Epilogue to John's book

I would like to conclude my job as a glossator for John's book with some personal remarks. This will help other readers to better assess the underlying social fabric of which my comments are an expression. At the same time I hope that I can also bring some of John's charming personality to life for the reader by relating how I have been experiencing his presence in my own life. At the very least this little epilogue is an opportunity for me to express my gratitude to John for being my friend, colleague and mentor ever since we first met in France in 2006 while I was still a doctoral student.

It is no exaggeration to say that John has been the most important role model that I had the pleasure to meet during my academic career. On a personal level, he is attentive, kind, and he takes what you have to say seriously even when you are just a young student. Moreover, what is brilliant for a student is that you can always count on John to speak his mind loudly during talks and seminars, and that he never shies away from asking difficult and awkward questions, which other members of the audience might have on their mind but are too polite or afraid to ask. Even the most respected scientists get the same probing treatment, often to their own astonishment. But he does not misuse his sharp intellect as an instrument of dominance; he is always willing to listen, and always ready to be shown wrong if his arguments do not stand up to scrutiny. John is a pleasure to listen to, as he continually underlines his commentaries with lots of emphatic gesticulations, and you never know when he will shock the audience with one of his theatrical interludes, garnering laughs and more than a few surprised faces. John has the admirable ability to play the self-defacing clown in order to dispel the gloomy seriousness that all too often descends on academic meetings.

I first came across John's work in 2004-2005, while I was in the first years of my PhD studies at the University of Sussex. At that time I was completely immersing myself in the intellectual traditions of von Foerster's second-order cybernetics, Maturana and Varela's biology of autopoiesis, Husserl's phenomenology, as well as von Glasersfeld's radical constructivist approach to science. I was voraciously reading everything from these authors that I could get my hands on, and I thus quickly noticed that there was this person called J. Stewart, who had continued their research agenda in more recent years (e.g. Stewart 1992[*],1996[*], 2001[*]). During these early years I also once saw John give a lecture during one of the weekly seminars at Sussex, namely on the outstanding big questions regarding the origins of life and of the genetic system. This was the first time that I came across the biological notion of ‘scaffolding' and the differential approach to the genetic system, which John also discusses at great length in this book. At the time I did not understand everything in that seminar, but I left with an intriguing impression of the clever way in which John was able to use scientific reasoning to subvert the received wisdom of the biological mainstream.

I did not properly get to meet John in person until January 2006, when I joined a group of doctoral students supervised by Ezequiel A. Di Paolo in attending the NUCOG/PHITECO Seminar on “Cognition, Motivation, Action” at the Université de Technologie de Compiègne (UTC). John had come to the train station to pick up all the arriving students and take them to dinner on their first night, and I could not resist the temptation of launching into a technical discussion with John about the ins and outs of autopoietic theory. During that weeklong seminar we happened to stay at the same pension with a small group of other philosophy students, and we continued our deep discussions daily.

I fondly remember one heavy night of philosophizing, drinking and smoking with some other students in a bar, and that John – despite his age – managed to keep up with us young students for most of the night. John is a great storyteller, and I was fascinated by his personal stories of how he had once been fed up with mainstream science, of how he therefore temporarily had ‘dropped out' by going to live in India (an idea that I had also briefly toyed with before deciding instead to start my doctorate studies), and of how meeting Francisco Varela had turned him on to science once again. Varela's writings had done much the same for me, but unfortunately he had already died a few years before I first became aware of his work. So I was very excited to hear that John had actually worked with him in France for many years, and I hung on every word as he shared some of his personal memories. I was thrilled to have found in John a living connection to one of my greatest intellectual heroes, whom I had previously only known through his many published writings. That night was a memorable time. As it turned out, John was scheduled to give the first lecture in the morning of the next day, which he proceeded to do with his usual intellectual clarity, but not without showing bodily signs of fatigue stemming from the long night before. I remember feeling slightly guilty for having contributed to his diminished state of health.

During that week I noticed that I shared more of my intellectual fascinations with John than just the traditions following on from second-order cybernetics. During one night I was fortunate enough to talk my way into attending the dinner of the invited speakers of the conference, and I was able to instigate a heated debate between John and François Sebbah regarding the empiricist-transcendental split in philosophy and Husserl's eventual turn from the transcendental subject to the lifeworld. This philosophical debate was of great personal importance to me, because at the time I already vaguely knew myself to be intellectually trapped by the operational closure that defined the center, as well as the very limits, of the second-order cybernetics traditions. For instance, I remember that although only the local students were obliged to write an essay during the seminar, I had volunteered to write a short essay about how the world would appear from the first-person perspective of an operationally closed system. What does it mean to say that everything turns back on oneself, thereby constituting my observer's point of view? How would I be able to even notice if I lived isolated in such an experiential bubble or not? I appreciated that John had tried to find a middle way in the debate with François, by arguing that the very conditions of possibility of such philosophizing are to be found in our communally shared lifeworld, which always already precedes its theoretical articulation.

Nevertheless, at the end of the week I still tried to convince John that absolute observer-relativism, whereby the notion of reality goes completely up in smoke, thus leaving one's self and others in a mysterious limbo, was the inevitable outcome of taking the radical constructivist approach seriously. But he was not to be persuaded by me. On John's understanding, the ultimate fate of radical constructivism is to relativize its own relativity, thereby actually providing an opening for a suitably attenuated notion of objectivity (Stewart, 2001[*]). More importantly, this attenuation of self-centered relativity provides an opening in which the concrete existence of the Other can take root, and is therefore ethically imperative. In this way John helped me to break out of my operationally closed prison, and instead turned me on to the ideas of the social constructivists such as Bruno Latour, eventually leading to my current interest in anthropology and archaeology. He showed me, both in theory and in practice, how embodying a concern for the existence of others puts constraints on the viability of our forms of life and theories, while still staying within the remit of scientific practice. At the same time John did not fall for the extremes of postmodern social relativism, with which I also flirted for some time, but insisted that the observer is also biologically embodied and physically constrained, and that the scientific method can therefore provide us with adequately ‘objective' insights into her underlying mechanisms (Rohde & Stewart, 2008[*]).

After that first meeting in Compiègne we were able to deepen our friendship at the memorable CNRS Summer School on “Constructivism and Enaction: A New Paradigm for Cognitive Science”, which took place for a week in May-June 2006 on the beautiful Ile d'Oléron in France (see photo below). This summer school helped me to really feel part of a small yet vibrant and free-spirited community of researchers united by a common enactive agenda. John used the contributions made at this meeting as a starting point for co-editing a programmatic statement of the paradigm of enaction (Stewart, Gapenne & Di Paolo, 2010[*]). And despite frustrating repeated delays to the publication of this collection, this concerted push towards new directions lost none of its originality for me (Froese, 2012[*]).

Since then John and I have often crossed paths at various academic events that reflect our broad interests in topics ranging from biological autonomy and artificial life to human technology and the social, leading to further stimulating exchanges. In the following I would like to emphasize three areas of research in which John has inspired me to go beyond the pioneering work of Varela and co., namely in the fields of biology, technology and sociology.

Something that we both feel is missing from Maturana and Varela's biology of cognition, and which needs to be rectified if we want to include the concept of autopoiesis in the enactive approach to cognitive science, was a more serious consideration of far-from-equilibrium thermodynamics. As John explains in this book, the process of living belongs to the class of dissipative structures, which are also self-producing to some extent. But only living dissipative structures are able to adaptively regulate their own boundary conditions so as to satisfy their viability constraints (Bourgine & Stewart, 2004[*]). Taking these considerations as our starting point, we regularly met in cafes and bars while I was in Paris for an exchange visit at the end of 2009 in order to more specifically diagnose where Maturana and Varela's original formulation of autopoiesis had gone wrong. We pinpointed their initial reliance on general systems theory as one of the major stumbling blocks that had to be overcome by an enactive approach to life (Froese & Stewart, 2010[*]). As I later discovered to my surprise, John had previously done a much better job of highlighting how a purely dynamical systems perspective as such is unable to do justice to autopoiesis, and how autopoiesis by itself is unable to account for the semantic dimension of life (Stewart, 2000[*]). Nevertheless, our paper at least earned the distinction of having received an extensive response by Maturana himself (Maturana, 2011[*]) although, as could perhaps be expected, he mostly disagreed with our critical analysis, as did other commentators (Bich & Arnellos, 2013[*]). Still, this correspondence with Maturana enabled us to further clarify the unique contributions of the enactive approach, and to highlight how it has been advancing some of the central tenets of Maturana's biology of cognition (Froese & Stewart, 2013[*]).

Also remarkably absent from the initial formulations of the enactive approach was the question of technology. In contrast, John has been a vocal part of what we might call the Compiègne approach to cognitive science, which has long been arguing that technology is constitutive of human life, for instance by modulating our basic sensorimotor interactions (e.g. Stewart, Khatchatourov & Lenay, 2004[*]; Khatchatourov, Stewart & Lenay, 2007[*]; Stewart, 2010[*]). This technology-centric approach was influential on my conception of the Enactive Torch as a minimal and pragmatic tool for the phenomenology and science of perceptual experience (Froese & Spiers, 2007[*]; Froese, McGann, Bigge, Spiers & Seth, 2012[*]). In addition, it was John's practical demonstration of UTC's perceptual crossing equipment (Auvray, Lenay, & Stewart, 2009) during the Enactive Interfaces conference held in Grenoble in 2007, and his interest in devising agent-based computer models of this psychological study (Lenay, Stewart, Rohde & Ali Amar, 2011[*]), that has motivated some of the evolutionary robotics modeling work I did for my doctoral thesis (Froese & Di Paolo, 2010[*]; Froese & Di Paolo, 2011b[*]). We have been continuing this mutually informing dialogue in recent work, by exploring aspects of social imitation with variations of the perceptual crossing paradigm (Lenay & Stewart, 2012[*]; Froese, Lenay & Ikegami, 2012[*]).

John has also been instrumental in helping the enactive approach overcome its limited focus on individual agency. This is already indicated by his psychological studies in perceptual crossing, but he is also always the first to insist that such dyadic interaction is necessary but not sufficient for explaining human social life. In 2008 during a workshop on Enactive Approaches to Social Cognition, which I helped to organize, he got into some heated arguments with some of the other organizers and invited participants because of their, in his view, overly exclusive focus on dyadic interaction. He argued for a more prominent recognition of the fact that humans are always born into a pre-existing social world that is already saturated with its specific set of cultural structures, including artifacts, language, and norms of conduct (Steiner & Stewart, 2009). In response to this debate I tried to find a way of reconciling this heteronomy of culture with the relative autonomy of the inter-individual interaction process (Froese & Di Paolo, 2011a[*]; Torrance & Froese, 2011[*]).

Indeed, it was John who first piqued my interest in sociology, for example by giving a lecture on Durkheim's theory of the sacred during the Third CNRS Summer School on “Constructivism and Enaction: A new paradigm for cognitive science”, which was held in Cape Hornu in 2008. It is no wonder then that I jumped at the opportunity of inviting John to join me in visiting an exhibition of Teotihuacan culture, which had just opened in Paris during my exchange visit in 2009. Teotihuacan had been the first urban civilization in ancient Mesoamerica, and its ruins have fascinated me ever since I had first visited them a couple of years earlier. John's memorable comment at the end of our tour through the museum was the rather dry remark that the scientists did not seem to know very much about this culture. I agreed. And so now, just three years later, here I am in Mexico working together with archaeologists on a project trying to understand the origins of Teotihuacan society in terms of the complex systems theory of self-organization.

Thus, John was and still is an important part of my life in many respects. His unassuming yet encouraging presence at the many meetings we have attended together has given me the audacity to shake up the complacency of the status quo, and the heart to speak out loudly for a more rational and humane science in this all too troubled world. He has helped me to see how the great dichotomies confronting our intellectual understanding today – fact-value, mind-body, self-other, individual-social – can be overcome in concrete practice by developing a new kind of science that can be lived concretely and with dignity. As John has so eloquently argued in this book, heredity cannot simply be reduced to genetics, and I concur. Indeed, heredity is not just about tangible things. I am grateful for having inherited so much from John's way of life.

Mattéo Mossio

Comment 1

As John Stewart (JS) writes in the Introduction, “Qui aime bien châtie bien”. Just as he loves – and therefore chastises – the paradigm of Enaction (or more generally, as I see it, the theory of biological Autonomy), I'm making here a few critical comments on some key issues addressed in this book which is, in my view, a theoretically rich and stimulating one.


MM seems to prefer the term “biological Autonomy”; however as I see it, this term is more general, more vague, less clearly definable, and therefore less interesting than the concept of “Enaction”. Any paradigm having substantial ambitions does of course aim at having a broad scope; but the challenge is to combine this with some sharply defined cutting-edges. For this reason, I prefer not to dilute out the central focus of ambition, and thus to retain the term “Enaction”.

“Qui aime bien châtie bien” : a less literal translation of this French saying, which I thoroughly endorse, might be: “We are especially hard on those we love best”. My favourite illustration is Jane Austen's novel Emma, where the somber hero Rochester is especially hard on the heroine Emma, precisely because he loves her.

This may be a place to mention that I am deeply touched by Tom Froese's Epilogue – all the more so as I doubt whether I myself would have been able to write a text so unfailingly laudatory. A couple of months ago, I happened to mention this to Matteo Mossio; and he assured me that he would more than make up for any lack of criticism in his own comments. As will be seen, he has kept his promise – which will lead, I hope, to some invigorating exchanges.

Finally, I may also mention here that of the founders of the paradigm of Autopoiesis and Enaction, Francisco Varela is no longer with us, and Humberto Maturana is now 85 years old. Moreover, the early followers of this paradigm, such as myself, are also now getting on in years. It is therefore a good sign for the future that young researchers such as Tom and Matteo are weighing in with their contributions.

Comment 1a

Just to mention that in my view Autonomy, and not Enaction, is the pivotal concept of this family of approaches and models. Varela's famous book was actually entitled “Principles of Biological Autonomy”, and since then, substantial technical and theoretical work has been done for developing this concept, in my opinion with better results than those obtained for Enaction. However, although it would be useful to clarify the relations between the two concepts, I take this as being a side issue in the present discussion; accordingly, I'm not developing it here.

Comment 2

Let me start with a general issue, which concerns one of the fundamental tenets of the Paradigm of Enaction as presented by Stewart. As he recalls in the Introduction, the paradigm assumes the double equation “Cognition = Life = Autopoiesis”. Although the equation is indeed well known, I think that it does not make a useful job in the paradigm. Rather, it seems to me that it is quite counterproductive, and should be abandoned. The central reason is that the equation leans toward a form of reductionism, which might obscure – instead of capturing – some qualitative differences between Cognition, Life and Autopoiesis. Of course, there are profound links between these determinations; yet the richness and variety of living phenomena might call for finer-grained distinctions.


I readily admit that the form of an equation – “Cognition = Life = Autopoiesis” – is not ideal. I would certainly shy away from any suggestion of a straightforward identity between the three terms (as would be accentuated in an unfortunate formulation such as “Cognition = Life = Autopoiesis”, implying that the three terms are neither more nor less than synonyms). But I nevertheless maintain that these three terms – Cognition, Life and Autopoiesis – form a triad of key fundamental concepts in the sense that none of them can be properly understood in isolation, and thus that each of them calls for articulation with the two others. This gets us into considering the articulations (which implicitly includes specifying the differentiations) between these three terms – exactly the sort of discussion that MM gets into below. My proposal at this stage would be to replace the “equation” with the designation of a triad: i.e. “Cognition – Life – Autopoiesis”.

Now for the more detailed discussion.


arrow to header
  • Aftalion F. (1991)
    Aftalion F. (1991). A History of the International Chemical Industry. University of Pennsylvania Press, Philadelphia.
  • Atlan H. & Bousquet B. (1994)
    Atlan H. & Bousquet B. (1994). Questions de vie. Le Seuil, Paris.
  • Begley S. & Carey J. (1983)
    Begley S. & Carey J. (1983). Newsweek 64, Dec. 26th.
  • Berger P.L. & Luckmann T. (1966)
    Berger P.L. & Luckmann T. (1966). The Social Construction of Reality. Anchor Books, New York.
  • Berman M. (1981)
    Berman M. (1981). The reenchantment of the world. Cambridge University Press, Cambridge.
  • Bich L. & Arnellos A. (2013)
    Bich L. & Arnellos A. (2013). Autopoiesis, Autonomy, and Organizational Biology: Critical Remarks on "Life After Ashby". Cybernetics & Human Knowing, 19(4), 75-103.
  • Bourgine P. & Stewart J. (2004)
    Bourgine P. & Stewart J. (2004). Autopoiesis and Cognition. Artificial Life 10: 327–345.
  • Braitenberg V. (1984)
    Braitenberg V. (1984). Vehicles. Experiments in synthetic psychology. MIT Press, Boston.
  • Braverman H. (1974)
    Braverman H. (1974). Labor and Monopoly Capital: The Degradation of Work in the Twentieth Century. New York: Monthly Review Press.
  • Cairns-Smith, A. G. (1982)
    Cairns-Smith, A. G. (1982) Genetic Takeover and the Mineral Origins of Life, Cambridge University Press, New York.
  • Cairns-Smith A.G. (1985)
    Cairns-Smith A.G. (1985). Seven clues to the origin of life. Cambridge University Press, Cambridge.
  • Carneiro, J., Coutinho, A., Faro, J. & Stewart, J. (1996)
    Carneiro, J., Coutinho, A., Faro, J. & Stewart, J. (1996). A model of the immune network with B-T cell co-operation. I--Prototypical structures and dynamics. II--The simulation of ontogenesis J Theor Biol 182:513-530, 531-548.
  • Carneiro, J., K. Leon, et al. (2007)
    Carneiro, J., Leon, K., Caramalho, I., van den Dool, C., Gardner, R., Oliveira, V., Bergman, M.L., Sepúlveda, N., Paixão, T., Faro, J. and Demengeot, J. (2007). "When three is not a crowd: a crossregulation model of the dynamics and repertoire selection of regulatory CD4+ T cells." Immunol. Rev. 216: 48-68.
  • Dawkins R. (1986)
    Dawkins R. (1986). The Blind Watchmaker. Longman, London.
  • Derrida J. (1978)
    Derrida J. (1978). Edmund Husserl's Origin of Geometry. Nicolas Hays, Stony Brook NY.
  • Dumaret A. & Stewart J. (1989)
    Dumaret A. & Stewart J. (1989). L'hérédité de l'intelligence. Editions Dumas, Saint-Etienne.
  • Dupuy J.-P. (1992)
    Dupuy J.-P. (1992). Introduction aux Sciences Sociales. Ellipses, Paris.
  • Dupuy J.-P. (1994)
    Dupuy J.-P. (1994). Aux origines des sciences cognitives. Editions La Découverte, Paris.
  • Durkheim E. (1894)
    Durkheim E. (1894). Les Règles de la Méthode Sociologique. Presses Universitaires de France, Paris. English version : The Rules of the Sociological Method, (Ed. Steven Lukes; trans. W.D. Halls). New York: Free Press, 1982.
  • Durkheim E. (1897)
    Durkheim E. (1897). Le Suicide. Presses Universitaires de France, Paris.
  • Durkheim E. (1912)
    Durkheim E. (1912). Les Formes élémentaires de la vie religieuse. Le système totémique en Australie, Presses Universitaires de France, Paris. English version : The Elementary Forms of the Religious Life. (trans. J.W.Swain). The Free Press, London, 1965).
  • Eldredge, N., Gould, S.J. (1972)
    Eldredge, N., Gould, S.J. (1972). Punctuated Equilibria: an Alternative to Phyletic Gradualism. In Schopf, T.M. (ed.) 1972: Models in Palaeobiology. Freeman Cooper, pp. 82-115.
  • Fox-Keller E. (1983)
    Fox-Keller E. (1983). A feeling for the organism: the life and work of Barbara McClintock. Freeman, New York.
  • Frankel E. (1985)
    Frankel E. (1985). Corpuscular Optics and the Wave Theory of Light: The Science and Politics of Revolution in Physics. Social Studies of Science 6(2).
  • Franklin B. Advice to a Young Tradesman (1748)
    Franklin B. Advice to a Young Tradesman (1748) Philadelphia : Printed by Daniel Humphreys Garfinkel H. (1967). Studies in ethnomethodology . Englewood Cliffs, NJ: Prentice-Hall.
  • Froese T., Lenay C., & Ikegami T. (2012)
    Froese T., Lenay C., & Ikegami T. (2012). Imitation by social interaction? Analysis of a minimal agent-based model of the correspondence problem. Frontiers in Human Neuroscience, 6(202), doi: 10.3389/fnhum.2012.00202.
  • Froese T., McGann M., Bigge W., Spiers A., & Seth A. K. (2012)
    Froese T., McGann M., Bigge W., Spiers A., & Seth A. K. (2012). The Enactive Torch: A New Tool for the Science of Perception. IEEE Transactions on Haptics, 5(4), 365-375.
  • Froese T. (2012)
    Froese T. (2012). From adaptive behavior to human cognition: a review of Enaction. Adaptive Behavior, 20(3), 209-221.
  • Froese T. & Di Paolo E. A. (2010)
    Froese T., & Di Paolo E. A. (2010). Modeling social interaction as perceptual crossing: An investigation into the dynamics of the interaction process. Connection Science, 22(1), 43-68.
  • Froese T. & Di Paolo E. A. (2011a)
    Froese T. & Di Paolo E. A. (2011a). The Enactive Approach: Theoretical Sketches From Cell to Society. Pragmatics & Cognition, 19(1), 1-36.
  • Froese T. & Di Paolo E. A. (2011b)
    Froese T. & Di Paolo E. A. (2011b). Toward Minimally Social Behavior: Social Psychology Meets Evolutionary Robotics. In G. Kampis, I. Karsai & E. Szathmáry (Eds.), Advances in Artificial Life: Darwin Meets von Neumann. 10th European Conference, ECAL 2009 (pp. 426-433). Berlin, Germany: Springer-Verlag.
  • Froese T. & Spiers A. (2007)
    Froese T. & Spiers A. (2007). Toward a Phenomenological Pragmatics of Enactive Perception Enactive/07: Proceedings of the 4th International Conference on Enactive Interfaces (pp. 105-108). Grenoble, France: Association ACROE.
  • Froese T. & Stewart J. (2010)
    Froese T. & Stewart J. (2010). Life after Ashby: Ultrastability and the autopoietic foundations of biological individuality. Cybernetics & Human Knowing, 17(4), 83-106.
  • Froese T. & Stewart J. (2013)
    Froese T. & Stewart J. (2013). Enactive Cognitive Science and Biology of Cognition: A Response to Humberto Maturana. Cybernetics & Human Knowing, 19(4), 61-74.
  • Garfinkel (1967)
    Garfinkel (1967). Studies in ethnomethodology . Englewood Cliffs, NJ: Prentice-Hall.
  • Giddens A. (1984)
    Giddens A. (1984). The Constitution of Society. Polity Press, Cambridge.
  • Goody J.R. (1977)
    Goody J.R. (1977). The Domestication of the Savage Mind. Cambridge: Cambridge University Press.
  • Gould, Stephen Jay, 1989
    Gould, Stephen Jay (1989). Wonderful Life: The Burgess Shale and the Nature of History, W. W. Norton & Co., New York, NY.
  • Gould S.J. (1991)
    Gould S.J. (1991). Wonderful Life: the Burgess shale and the nature of history. Penguin, London. Traduction française: La vie est belle : les surprises de l'évolution. Le Seuil, Paris, 1993.
  • Guille-Escuret G. (1994)
    Guille-Escuret G. (1994). Le décalage humain. Le fait social dans l'évolution. Kimé, Paris.
  • Haldane 1947
    Haldane, J.B.S. (1947) What is Life?, Boni and Gaer, 1949 edition: Lindsay Drummond
  • Husbands P., Harvey I., Cliff D. & Miller G. (1994)
    Husbands P., Harvey I., Cliff D. & Miller G. (1994). The use of genetic algorithms for the development of sensorimotor control systems. In: P.Gaussier & J-D Nicoud (Eds), From Perception to Action, IEEE Computer Society Press, Los Alamitos.
  • Jacob F. (1970)
    Jacob F. (1970). La Logique du Vivant. Gallimard, Paris.
  • Jacob F. (1981)
    Jacob F. (1981). Le jeu des possibles. Nrf, Paris.
  • Jaynes, J. (1976)
    Jaynes, J. (1976). The origin of consciousness in the breakdown of the bicameral mind. Boston: Houghton Mifflin.
  • Kazatchkine MD & Morell A (1997)
    Kazatchkine MD & Morell A (Eds) (1997). Intravenous Immunoglobulin Research and Therapy. CRC Press.
  • Khatchatourov A., Stewart J. & Lenay C. (2007)
    Khatchatourov A., Stewart J. & Lenay C. (2007). Towards an Enactive Epistemology of Technics Proceedings of ENACTIVE/07: 4th International Conference on Enactive Interfaces (pp. 129-132). Grenoble, France: Association ACROE.
  • Lakatos I. (1970)
    Lakatos I. (1970). Falsification and the methodology or scientific research programmes. In: I. Lakatos & A. Musgrave Eds, Criticism and the growth of knowledge, Cambridge University Press, 1970, pp. 91-196.
  • Latour B. (1980)
    Latour B. (1980). The Three Little Dinosaurs or a Sociologist's Nightmare. Fundamenta Scientae 1, 79-85.
  • Latour B. (1987)
    Latour B. (1987). Science in Action. Harvard University Press, Cambridge.
  • Latour B. & Woolgar S. (1979)
    Latour B. & Woolgar S. (1979). Laboratory life: the social construction of scientific facts. Sage, Beverly Hills.
  • Lenay C., Stewart J., Rohde M., & Ali Amar A. (2011)
    Lenay C., Stewart J., Rohde M., & Ali Amar A. (2011). 'You never fail to surprise me': the hallmark of the Other: Experimental study and simulations of perceptual crossing. Interaction Studies, 12(3), 373-396.
  • Lenay C. & Stewart J. (2012)
    Lenay C. & Stewart J. (2012). Minimalist approach to perceptual interactions. Frontiers in Human Neuroscience, 6(98). doi: 10.3389/fnhum.2012.00098
  • Leroi-Gourhan A. (1964)
    Leroi-Gourhan A. (1964). Le Geste et la Parole. I. Technique et langage. II. La mémoire et les rythmes. Albin Michel, Paris.
  • Lestel D. (1998)
    Lestel D. (1998). L'innovation cognitive dans des communautés de partage de sens, d'intérêts et d'affects. Intellectica 26, 203-226.
  • Lord, A. B. (1960)
    Lord, A. B. (1960). The Singer of Tales. Cambridge, MA: Harvard University Press.
  • Luisi P.L.(2003)
    Luisi P.L.(2003). Autopoïesis: a review and a reappraisal. Naturwissenschaften 90, 49–59.
  • Maturana H. & Varela F.J. (1980)
    Maturana H. & Varela F.J. (1980). Autopoiesis and cognition: the realization of the living. Reidel, Boston.
  • Maturana H. R. (2011)
    Maturana H. R. (2011). Ultrastability ... Autopoiesis? Reflexive Response to Tom Froese and John Stewart. Cybernetics & Human Knowing, 18(1-2), 143-152.
  • Mauss M. (1925)
    Mauss M. (1925). Essai sur le don. Forme et raison de l'échange dans les sociétés archaïques. Introduction de Florence Weber, Quadrige/Presses universitaires de France, 2007.
  • McGregor, S. and Virgo, N. (2011)
    McGregor, S. and Virgo, N. (2011). Life and Its Close Relatives. In Kampis, G. Karsai, I. and Szathmary, E., editors, ECAL 2009, pp. 230-237, Springer Verlag, Berlin.
  • McMullin B. & Varela F . (1997)
    McMullin B. & Varela F . (1997). Rediscovering computational autopoiesis. In : Husbands P, Harley J (Eds), Proceedings of the 4th ECAL. MIT Press, Cambridge.
  • Merton R. K. (1942)
    Merton R. K. (1942). « The Normative Structure of Science » (1942) in Storer N.W. (ed.), The Sociology of Science, pp. 267-278. University of Chicago Press, Chicago.
  • O'Regan J.K. & Noë A. (2001)
    O'Regan J.K. & Noë A. (2001). A sensorimotor account of vision and visual consciousness. Behavioral and Brain Sciences 24(5), 939-1031.
  • Oparin, A. I. (1952)
    Oparin, A. I. (1952). The Origin of Life. New York: Dover. (Originally published in Russian, Moscow 1936; first English translation in 1938).
  • Oyama S. (1985)
    Oyama S. (1985). The ontogeny of information: developmental systems and evolution. Cambridge University Press, Cambridge.
  • Perelson AS & Oster G (1979)
    Perelson AS & Oster G (1979). Theoretical studies of clonal selection: minimal antibody repertoire size and reliability of self non-self discrimination. J Theor Biol 81: 645-670.
  • Prigogine A. & Stengers I. (1979)
    Prigogine A. & Stengers I. (1979). La nouvelle alliance, Gallimard, Paris.
  • Rohde M. & Stewart J. (2008)
    Rohde M. & Stewart J. (2008). Ascriptional and 'genuine' autonomy. BioSystems, 91(2), 424-433.
  • Salanskis J.M. (1992)
    Salanskis J.M. (1992). Le destin du modèle de Cantor-Dedekind , in Le Labyrinthe du Continu, J. -M. Salanskis & H. Sinaceur (Eds), Paris, Springer France, 190-212.
  • Saunders P.T. (1993)
    Saunders P.T. (1993). L'évolution des formes biologiques. Intellectica 16, 61-84.
  • Schumpeter J. (1934)
    Schumpeter J. (1934). The Theory of Economic Development: An Inquiry into Profits, Capital, Credit, Interest, and the Business Cycle. Cambridge Mass: Harvard University Press.
  • Segel LA & Perelson AS (1989)
    Segel LA & Perelson AS (1989). Shapespace: an approach to the evaluation of cross-reactivity effects, stability and controllability in the immune system. Immunology Letts 22: 91-100.
  • Simmel, G. (1900)
    Simmel, G. (1900). Philosophie des Geldes. Leipzig: Duncker & Humblot. English translation: (1990) The Philosophy of Money (Ed. D.P. Frisby). London: Routledge.
  • Simondon G. (1989)
    Simondon G. (1989). L'individuation psychique et collective. Aubier, Paris.
  • Simondon G. (1989b)
    Simondon G. (1989b) Du mode d'existence des objets techniques. Aubier, Paris.
  • Sims K. (1994)
    Sims K. (1994). Evolving Virtual Creatures. Computer Graphics Annual Conference Proceedings, New York, pp.15-22.
  • Smith A. (1776)
    Smith A. (1776). The Wealth of Nations. Strahan and Cadell, London.
  • Sohn-Rethel A. (1978)
    Sohn-Rethel A. (1978). Intellectual and Manual Labour. A Critique of Epistemology. Humanities Press, Atlantic Highlands N.J.
  • Stewart J., Gapenne O. & Di Paolo E. A. (2010)
    Stewart J., Gapenne O. & Di Paolo E. A. (2010). Enaction: Toward a New Paradigm for Cognitive Science. Cambridge, MA: MIT Press.
  • Stewart J., Khatchatourov A. & Lenay C. (2004)
    Stewart J., Khatchatourov A. & Lenay C. (2004). Enaction and Engineering. Interdisciplines,, Accessed: Feb. 2, 2010.
  • Stewart J. (1992)
    Stewart J. (1992). Life = Cognition: The Epistemological and Ontological Significance of Artificial Life. In F. J. Varela & P. Bourgine (Eds.), Toward a Practice of Autonomous Systems: Proceedings of the First European Conference on Artificial Life (pp. 475-483). Cambridge, MA: MIT Press.
  • Stewart J. (1996)
    Stewart J. (1996). Cognition = life: Implications for higher-level cognition. Behavioral Processes, 35, 311-326.
  • Stewart J. (2000)
    Stewart J. (2000). From Autopoiesis to Semantic Closure. Annals of the New York Academy of Sciences, 901, 155-162.
  • Stewart J. (2001)
    Stewart J. (2001). Radical Constructivism in Biology and Cognitive Science. Foundations of Science, 6(1-3), 99-124.
  • Stewart J. (2004)
    Stewart J. (2004). La vie existe-t-elle ? Réconcilier la génétique et la biologie. Vuibert, Paris.
  • Stewart J. (2010)
    Stewart J. (2010). Foundational Issues in Enaction as a Paradigm for Cognitive Science: From the Origin of Life to Consciousness and Writing. In J. Stewart, O. Gapenne & E. A. Di Paolo (Eds.), Enaction: Toward a New Paradigm for Cognitive Science (pp. 1-31). Cambridge, MA: The MIT Press.
  • Stiegler B. (1994)
    Stiegler B. (1994). La Technique et le temps, tome 1 : La Faute d'Épiméthée, 1994. Paris, Galilée.
  • Stiegler B. (1998)
    Stiegler B. (1998) Technics and Time, 1: The Fault of Epimetheus. Stanford University Press, Stanford.
  • Torrance S. & Froese T. (2011)
    Torrance S. & Froese T. (2011). An inter-enactive approach to agency: Participatory sense-making, dynamics, and sociality. Humana.Mente, 15, 21-53.
  • Varela F, Coutinho A, Dupire B & Vaz NM (1988)
    Varela F, Coutinho A, Dupire B & Vaz NM (1988). Cognitive networks: immune, neural and otherwise. In: Perelson AS (ed), Theoretical Immunology Vol II, New York: Addison Wesley, pp 357-374.
  • Varela F., Thompson E. & Rosch E. (1991)
    Varela F., Thompson E. & Rosch E. (1991). The Embodied Mind. MIT Press, Boston.Vidal-Nacquet P. (2000). Le monde d'Homère. Perrin, Paris.
  • Varela F.J. (1989).
    Varela F.J. (1989). Autonomie et connaissance : essai sur le vivant. Ed du Seuil, Paris.
  • Varela F.J. (1996)
    Varela F.J. (1996). Neurophenomenology: A Methodological remedy to the hard problem. J Consc Studies 3:330-350.
  • Varela F & Stewart J (1990)
    Varela F & Stewart J (1990). Dynamics of a class of immune networks. I. Global stability of idiotypic interactions. II. Oscillatory activity of cellular and humoural components. J Theor Biol 44: 93-102, 103-115.
  • Varela FJ & Coutinho A (1991)
    Varela FJ & Coutinho A (1991). Second generation immune networks. Immunology Today 12: 159-166.
  • Vaz NM & Varela F (1978).
    Vaz NM & Varela F, 1978, Self and non-sense: an organism-centered approach to immunology. Medical Hypothesis 4: 231-267.
  • Vidal-Nacquet P. (2000)
    Vidal-Nacquet P. (2000). Le monde d'Homère. Perrin, Paris.
  • Von Mises L. (1949)
    Von Mises L. (1949). Human Action: A Treatise on Economics. London: W. Hodge & Co.
  • Wächtershäuser, G (2006)
    Wächtershäuser, G (2006). From volcanic origins of chemoautotrophic life to Bacteria, Archaea and Eukarya. Philosophical Transactions of the Royal Society B: Biological Sciences 361 (1474): 1787–1808.
  • Wächtershäuser G. (1990)
    Wächtershäuser G. (1990). "Evolution of the First Metabolic Cycles", Proceedings of the National Academy of Sciences, Vol 87 200-204.
  • Wächtershäuser G. (1994)
    Wächtershäuser G. (1994). Life in a ligand sphere. PNAS 91, 4283-4287.
  • Wächtershäuser G. (1997)
    Wächtershäuser G. (1997) The origin of life and its methodological challenge. J Theor Biol 187:483-494.
  • Wächtershäuser G. (2000)
    Wächtershäuser G. (2000). Origin of Life: Life as We Don't Know It, Science, 289, 1307-1308.
  • Weber M. (1905)
    Weber M. (1905). The Protestant Ethic and "The Spirit of Capitalism" (Penguin Books, 2002) translated by Peter Baehr and Gordon C. Wells.


arrow to header
  1. Formally, there are additional premises : i) that life has not always been present from the very beginning of the Universe, and ii) that life did not have an extra-terrestrial origin and arrive on Earth from space. As to (i): the conditions when the planet Earth was formed were completely incompatible with any form of life as we know it; as to (ii), apart from its implausibility, it does not even solve the problem but just pushes it back to the place and time when life did first arise. Thus, I consider that the logical argument does go through.

  2. A version of this thesis that is less orthodoxically « scientific » in form, but which contains all the essential elements while being shorter and very readable, is Cairns-Smith (1985).

  3. Robots of this sort correspond, in the domain of “Artificial Life”, to what has been called « The Harvard Law of Animal Behavior » in experimental psychology: “Under the most carefully controlled experimental conditions, an animal does... whatever he damned likes”.

  4. « On n'interroge plus la vie dans les laboratoires ».

  5. In fact, as Lewontin (2003) has pointed out, there are few biological molecules that are as chemically inert as DNA. This is precisely the reason why it is possible to retrieve DNA from frozen mammoths, Egyptian mummies, or, as has been done more recently, from a few traces for the purposes of criminological investigation.

  6. Even if this text contained instructions about how to build a photocopier, a human engineer would still be required in order to understand the text, and, most importantly, to actually perform the material construction of the photocopier. Therefore, even in such a case, it is not correct to say that the text “reproduces itself.”

  7. Accepting for the moment that a “programme” really exists at all. We will come back to that question in the following section, in which we will examine in greater depth the question of ontogenesis.

  8. As in the joke where one of the characters says to the other: “Well, that old machine of yours, I already gave it back to you in a fine state, besides it was already broken when you passed it to me, and by the way you never lent it to me in the first place,” the speaker is piling up so many good reasons on top of each other (blithely ignoring the fact that they are mutually contradictory) that he is obviously trying to hide something.

  9. Actually, as Medawar (1957) has pointed out, few things are as un-natural as so-called “natural death” from old age. In natural populations, the vast majority of individuals die an “accidental” (but statistically normal) death well before attaining the age of senescence.

  10. The rate of “failed” embryos is probably higher than it seems, as relatively recent research has indicated that there are many miscarriages that serve to eliminate monstrous embryos. But, in addition to the fact that this spontaneous early abortion mechanism represents itself an adaptation, the success rate remains remarkably high.

  11. It may seem strange to say that the genes are “external” to the organism; but as we shall see more clearly when we go on to examine what “genetic information” really can do, even though genes are physically located inside nuclear chromosomes at the heart of each cell, it is actually quite correct to say that they are in fact epistemologically external to the somatic processes.

  12. That is, however, what Jacob conveys when he states that genetic information “are also the means of putting these plans into practice and of coordinating the system's activities.” However, it would probably be more consistent with the spirit of the concept of the “genetic programme” to say that this programme only gives instructions, in the same way as the architect gives instructions to the bricklayer. Note the implication: Once the architect has had his say, the bricklayer's work is taken for granted, and, in a certain way, “does not count.” This attitude of despising manual work is, perhaps, not unrelated to the fascination exerted by the idea of a “genetic programme.”

  13. In order to avoid any misunderstandings, let us clarify here that we are considering invariances; as soon as differences have to be explained, genetic information recovers all its prerogatives. So if we need to explain, for example, why a pig's offspring does not resemble a cat's, genetic differences become relevant again.

  14. This schema does not apply to insects.

  15. Why not a flat, two-dimensional sheet, or a one-dimensional string, or simply a collection of cells scattered around in the medium?

  16. More generally, the “phenotype” is not necessarily an organism's outside appearance. It can correspond to any measurable characteristic – even if it takes special observational techniques to observe it, whether physiological, biochemical, histological or other – that is “internal” to the organism.

  17. It may seem very hard to believe that anyone could ever imagine that the genes themselves ARE physically tall or short, and even less that the fact of BEING tall or short would be sufficient to make the corresponding organism tall or short. But sadly enough this is indeed the case – as the example of the “timid gene” cited in the last chapter illustrates. Of course, if pushed, even minimally sophisticated scientists will admit that this cannot be the case. Nevertheless, the lingering impression remains that genes somehow “carry” the characteristic, that they are in a sense “shortNESS” or “tallNESS”.

  18. In fact, this would already be the case if each parent had only two descendants (the strict minimum to ensure species survival): if the genes received from the parents were simply transmitted as such, without being copied, then after the transmission of one gene (for example, the paternal one) to the first descendant, only the other gene (the maternal one in our example) would remain to be transmitted to the second descendant. In that case, the ratios in generations F2, backcross, etc. would not be statistical, but exact – which is not the case in reality. We will not pursue any further these statistical considerations in this book, as they can become extremely subtle and complex.

  19. Mendel worked with seven differential “characteristics”: three characteristics were related to seeds (smooth or “wrinkled” form, brown or green teguments, yellow or green cotyledons at germination); two characteristics were related to the pods (with or without a narrowing, green or yellow color of the unripe pod); and two characteristics were related to the position of the flowers (axial or terminal) and to the length of the stems (long or short) (Mazliak, 2002).

  20. “Tessellation” is a regular tiling of polygons (two dimensions) or polyhedra (three dimensions). Here, the components C are tessellated in order to form a membrane.

  21. Technically, spatialized partial differential equations first need to be established for a(x,t) and b(x,t), i.e. A and B concentrations at each point of the volume inside the membrane [Bourgine & Stewart, 2004]. This makes it possible to describe mathematically the state of dynamic equilibrium in which δa/δt = δb/δt = 0; and then, by integration over the entire volume inside the membrane, to obtain ordinary differential equations describing the equilibrium condition. The process of hole formation also has to be modelled in order to define the frequency of holes of each dimension. This makes it possible to calculate the rate of loss of the B component through holes, and, therefore, to define the relationships between the three parameters kp, ks and a0 that correspond to a dynamic equilibrium.

  22. This is a relatively general result of cybernetics, in which a positive retro-action gives rise to bi-stability. This result is at the centre of the work of René Thomas (1998).

  23. From a mathematical point of view, it is also the membrane that makes possible the integration of all the spatialized differential equations and, therefore, makes it possible to formulate mathematically the relationship of parameters that is required for equilibrium. It is because the membrane enables the system to control its own boundary conditions that it becomes possible to re-model the system as a state-determined dynamic system.

  24. This statement could be softened by replacing the vivid expression “Life does not exist” with “Life is not (currently) a subject of study.” The second assertion leads to the same situation of hic et nunc, but leaves room for future developments. The prospects that this opens will be examined in the last part of the book "The enaction of social life".

  25. The weakness to be addressed here concerns the question of the “spontaneous generation” which is necessarily at the origin of life. Other weaknesses concern the absence of specific reference to the dynamics of sensori-motor coupling which are necessary for an autopoïetic organism to be clearly cognitive (see Bourgine & Stewart 2004); and the absence of a “genetic system”, necessary for autopoïetic organisms to be able to evolve towards forms of greater complexity: this is the theme of this chapter as a whole, to which we shall return in due course.

  26. A major objection to this, which rapidly comes to mind, is the phenomenon of ageing in multicellular organisms. Virtually all multicellular organisms die a “natural death” of old age (if they have not met with accidental death before), at the term of a lifespan that varies with the species (from weeks in the case of small insects to centuries in the case of sequoia trees) but is never infinite. This fascinating phenomenon – which is actually specific to multicellular organisms that arose much later in evolution, after the advent of a genetic system – is too complex to enter into here. The interested reader is invited to consult Stewart (2004) for a fuller discussion.

  27. From the point of view of the sociology of science, it is interesting to note that Wächtershäuser is not a professional biologist, but a patents lawyer interested in the theme of innovation. This is quite understandable: in the field of contemporary biology, “normal science” is dominated by DNA-centred molecular biology.

  28. Once again, we may note that the genetic information does not itself specify what it is coding for.

  29. Reference :

  30. It is as well to acknowledge that this is something of a rhetorical flourish : the point of view in question is not really that of the "immune system itself", but rather that of a human biologist acting as "spokesman" for the immune system. This being said, there remains a valid and important distinction between the two points of view - that of the classical immunologist, and that of the "spokesman".

  31. What was for some time taken as the best documented report, the case of two Indian girls raised by wolves, and which was notably cited by Maturana & Varela (1980), has recently been shown to be a hoax (Aroles 2007).

  32. “Radical constructivism” generalizes this “duality” to the whole of reality. Thus our knowledge of the natural world is clearly a human construction; however, as soon as this knowledge stabilizes, we spontaneously consider it as knowledge of an « independent reality » that is not constructed. We will explicitly examine this question below in terms of the sociology of scientific knowledge (see 4.4.1).

  33. I myself owe my appreciation of this dimension to the privilege of being associated with the Department of Human Sciences at the Technological University of Compiegne. Due to the vision of its founder, Guy Deniélou, the function of this Department is precisely to thematize the human and social meaning of technological systems. I will refer explicitly to the “Compiegne School” of thought on this question in 4.4.2.

  34. This is reminiscent of the ability of Cuvier to reconstruct an entire skeleton on the basis of a single bone.

  35. The extreme lucidity and political intelligence with which “primitive” societies can exert deliberate control over their own social structures is magnificently illustrated by the work of Pierre Clastres (1974).

  36. In French, even today the word for « money » is « argent » - i.e. quite literally “silver”.

  37. We may note that there is a definite logical incoherency here: since if one is irrevocably predestined anyway, how can one have any responsibility for one's situation?! But on the psychological level, this logical incoherency does not dissolve what is at stake – on the contrary, it raises it to a paroxysm.

  38. The brute fact of the matter is that even if he did not do so, out of humanitarian motives, he would be sidelined by less scrupulous competitors. We are here at the heart of the “social Darwinism” that we will come to in

  39. This is the explanation for the title of Sohn-Rethel's book, which may seem surprising given the sub-title which is A critique of epistemology. The title of the book is : Intellectual and Manual Labour

  40. This also explains how it is that the same function can be performed by simple pieces of paper... as long as they bear inscriptions which cannot be forged (or at least, which cannot be forged too easily) so that they carry the same guarantee. Two anecdotes provide a pleasant illustration of the striking contrast between first and second nature which comes into play with bank-notes. When I was seven years old, I inadvertently left a bank-note in my trouser pocket when it went to be washed. I was amazed to see my mother recuperate some damp fragments of the note, which still bore in barely legible form the number of the note, whereupon she took them to the bank and obtained in exchange... a brand-new bank-note! The second anecdote goes back to the 1970's, when some situationists spread confusion on a calm Sunday-morning market by proposing to buy and sell bank-notes at a price slightly off their nominal value. “Oyez Oyez! We will sell you these old, used bank-notes for 99 francs. But do you want one of these crisp, brand-new notes? – well, you will have to pay it 101 francs!” These utterances were obviously quite scandalous with respect to the postulates of the exchange-abstraction; just as obviously, the provocation was deliberate.

  41. Including, we may add, dyadic or other interactions with members of the same species ; Durkheim remarks elsewhere that if a man were reduced to having only empirical knowledge based on individual perceptions of this sort, « he would be indistinguishable from the beasts » (op. cit., p 487).

  42. In view of the social importance of tools, which we have examined in section 3.3, it is fascinating to see here that Durkheim himself spontaneously makes the association between conceptual categories, tools and social institutions.

  43. Many if not all tribes are organized in just two phratries, with half the clans belonging to each phratry.

  44. We have already seen, in 4.6.1, the fundamental importance of religion and “the gods” for the originary anthropological constitution. We recall here simply that, for Durkheim, the “gods” were nothing other than an archetypal expression of human society.

  45. Once again, the exceptions prove the rule. For example, in the old Soviet block, the Stalinist state imposed an official rate of exchange which was quite different from the “free” rate on the black market. But this did represent a hindrance (politically desired, in this case) to the development of market exchanges.

  46. As it stands, the term « random » is merely a cloak for ignorance. In order to have any positive meaning, the term “random” requires at least a framework, a specification of the set of categories which cover the variation in question, together with a specification of the relative probabilities for each category.

  47. “In the beginning God created the heaven and the earth... And God said, Let the earth bring forth the living creature after his kind, cattle, and creeping thing, and beast of the earth after his kind: and it was so. ... And God said, Let us make man in our image, after our likeness: and let them have dominion over the fish of the sea, and over the fowl of the air, and over the cattle, and over all the earth, and over every creeping thing that creepeth upon the earth. ...”. Book of Genesis, Chapter 1.

  48. Fisher showed mathematically that with random mating, 50% of the variance would be lost at each generation.

  49. Additional experiments involve observing the simultaneous inheritance of two characters, each controlled by a single gene; this gave rise to the phenomenon of « linkage groups », and eventually to a grand synthesis with the « chromosome theory ». I will not enter into these details here, because I hope I have said enough to convey the tenor of the Mendelian theory.

arrow to header
John Stewart


John Stewart entered Cambridge University with a Major Open Scholarship in Natural Sciences in 1959. After a first degree in physics he switched to genetics, and did research for ten years on genetic variation in kidney function in mice. He entered the CNRS in France in 1979, working successively in the fields of the sociology of science, theoretical immunology, cognitive science, and the philosophy of technology. He is the author of over a hundred scientific articles and several books, notably on the IQ heredity-environment debate, on genetic engineering, on the evolution of the immune system, on the relation between genetics and biology as a science of life, and most recently Co-Editor of a book Enaction: Toward a New Paradigm for Cognitive Science.. He is currently attached to the Technological University of Compiègne. He is the father of two children and has three grandchildren.

arrow to header
Tom Froese


Tom Froese received a Master's degree in Computer Science and Cybernetics from the University of Reading, UK (2004). He then obtained a Doctorate degree in Cognitive Science from the University of Sussex, UK (2010). He continued as a postdoctoral research fellow at the Neurodynamics and Consciousness Lab of the Sackler Centre for Consciousness Science, also at the University of Sussex. Froese then became a JSPS postdoctoral research fellow at the Ikegami Laboratory of the Department of General Systems Studies, University of Tokyo, Japan (2010–2012). From 2012 to 2014 he was a postdoctoral research fellow at the Self-Organizing Systems Lab of the Instituto de Investigaciones en Matemáticas Aplicadas y en Sistemas, Universidad Nacional Autónoma de México, Mexico. Starting in 2014 he is a faculty member of the Universidad Nacional Autónoma de México. Froese has published widely on enactive theories of the dynamics and phenomenology of life, mind, and sociality. Currently he is working on applying the concepts of enaction to gain a better understanding of the human mind and its complexities. His homepage can be found at:

Mattéo Mossio