Chapter 3
Symbolic Cognition
Having discussed the general nature and structure of information, I would now like to return to the processes of symbolic encoding that provide the interface between the phenomenal world and the semantic world of human beings. I will therefore expand on and provide more detail and context for certain concepts discussed in the preceding chapter. The purpose of this chapter is to establish as clearly as possible the specificity of human cognition in relation to animal cognition, with respect to both the processes of individual cognition and the emergent processes of collective cognition. Rather than a systematic presentation, this chapter provides a spiraling series of meditations in which the same themes are re-examined in increasing detail from different perspectives.
Section 3.1 delimits the field of symbolic cognition. Section 3.2 defines the type of reflexivity specific to human cognition. Section 3.3 discusses the power of human symbolic cognition, in particular its capacity to generate cultural phenomena. Section 3.4 focuses on the impossibility of separating the phenomenal and conceptual dimensions of symbolic cognition. Then section 3.5 discusses the openness of symbolic cognition, its creativity and the unlimited diversity of its manifestations. Section 3.6 completes the chapter with an inventory of the differences between human and animal collective intelligence. This final section provides a transition to the next chapter (“Creative conversation”), which deals with contemporary human collective intelligence as enhanced by the digital medium.
3.1. Delimitation of the field of symbolic cognition
3.1.1. Singularity
We know that biological evolution invented the eye (and the visual faculty) several times in the line of vertebrates and invertebrates: the eyes of octopuses, bees and apes are not all derived from the same initial eye. In contrast, symbolic cognition is a hapax of evolution: it emerged one time only, in the lineage of the primates who mastered fire. Unique in the history of evolution, symbolic cognition is indissociable from a reflexive, or self-referential, capacity of creation, exchange and transformation of the cognitive operators we call concepts. All cognition implies categorization. Only humans represent their categorization operators using symbolic systems and explicitly use symbols as objects of manipulation and contemplation.
3.1.2. Social and technical dimensions
As a general rule, symbolic systems are collective mechanisms produced and transformed at the level of cultures. Their holistic functioning and their coherence become apparent only when our intellectual lens is focused on a society or institution as a whole. The canonical example is always languages, but it is clear that musical, religious, political, legal, economic, technical, ludic and other symbolic systems belong to the same conventional, collective dimension of cognition. Just as human memory is embodied in a great many environmental, technical and institutional mechanisms1, symbolic systems can obviously incorporate technical and social elements that go beyond strictly personal cognition. For example, in the 21st Century, networked computers externalize many functions of syntactic manipulation and interconnection of units of meaning. These functions may be carried out by hardware and software modules shared by millions of people: tools for processing numbers, texts, images, sounds; search engines; online dictionaries and encyclopedias, etc. With respect to the social extension of cognitive processes, an example of a symbolic system that is not a language and that functions on a cultural scale is a country’s legal categories, rules and judicial procedures. The legal categories are the “dictionary”, while the rules and procedures provide the “grammar” of the legal symbolic system. Historical experience shows that this type of symbolic system is capable of organizing a very effective process of collective cognition. The conceptual thought of human beings is thus almost always structured by symbolic systems that pre-exist and transcend them. We can see individual cognitive systems as processors associated in a distributed calculation using shared symbolic systems and operating on the sociocultural scale of a mixed technobiological collectivity.
3.1.3. Symbolic manipulation goes far beyond linguistic competence and “reason”
The distinguishing feature of human beings is traditionally said to be that they are the animals endowed with language, reason or logos. But I want to stress here the fact that the faculty of explicit conceptualization that distinguishes us from other species cannot be reduced to the mere linguistic encoding of information. What medieval philosophers called the “intellective faculty” can manipulate all the forms of symbolic encoding characteristic of the human species, not only language. The symbols that are used in intellectual operations can be iconic, musical, choreographic, mathematical, technical, religious, political, economic2, legal, culinary3, vestimentary4, sexual5, erotic6, parental7, medical8, etc. This means that our capacity to explicitly manipulate categories is not only the condition that makes speech possible, but it also underlies all cultural institutions. These institutions presuppose: (i) systems for symbolic encoding of the objects of human experience; and (ii) the use of this symbolic encoding for the distributed techno-social processing of those objects. It is therefore clear that symbolism goes far beyond languages. Moreover, within linguistic encoding, symbolic cognition is not limited to the purely logical function – deductive, inductive, abductive – or the reasoning function in general. It includes all actual or possible uses of linguistic symbols, all kinds of “language games”9, whether they are practical10, poetic, rhetorical, ludic, affective or other.
3.2. The secondary reflexivity of symbolic cognition
3.2.1. The primary reflexivity of phenomenal consciousness
Some researchers11 see the beginning of cognitive processes in the biological functioning of the cell or of plant organisms because of the autopoietic and selfreferential nature of living things in general. It is generally agreed, however, that “sentient” or “conscious” cognition begins only with animals that have nervous systems. Nervous systems interpose complex computational circuits between the sensory reception of information (excitation of the sensory nerves of touch, receptors of the retina, taste buds, etc.) and the control of muscular movement. Heinz von Foerster noted that the nervous system works on its own results much more than on raw sensory data received by the sensory receptors12. Indeed, in the most advanced animal organisms there are many more neurons that receive their inputs from intermediate neurons than neurons that are fed directly by sensory receptors. Sensory qualities such as colors, shapes and odors are not received directly from the external world, but are actually calculated from the patterns of excitation of the sensory receptors. There would be no colors in a world without eyes, optic nerves or the complex biological computation machine of the brain, just as there would be no pleasure or pain without nerve impulses and the transmission and reception of various chemical messages in the interdependent ecosystemic networks in which animal organisms participate.
The main effect of neural calculations is the categorization of sensory data, first at the most basic level (pleasure or pain, salty or sweet, round or angular, blue or yellow) and then at the level of the construction of objects such as prey, predators, partners or indicators of them. And perception is always colored by an affect (starting with attraction and repulsion), whether this affect is conscious or not. Based on the reception and calculated production of chemical messages, emotions such as fear, aggression and sexual attraction come to infuse a range of subjective energies into the sensory world constructed by the apparatus of perception.
There are no raw phenomena, on one hand, and no categories that exist independently of phenomena, on the other hand. Rather, phenomenal experience and perceptual-affective categorization are two aspects of a single cognitive process. The phenomenal experience of animals is the subjective counterpart of the work of categorization and affective coloring carried out by their nervous systems. A flow of phenomenal experience emerging from the sensory-motor loop brings with it the perception of objects, qualities and poles of attraction and repulsion, which necessarily correspond to categories. There is no visual image, for example, without colors, light contrast or contours of some kind. In short, to perceive is to categorize. The process of categorization that occurs in the sensory-motor loop — and particularly in the neural computation that takes up most of that loop — in my view shows the reflexivity of animal cognition. The animal is “conscious”: that is, its phenomenal experience shines from within, lighted by the primordial glow of the perceptual: the (non-visual) light of categorizing experience with its affective tonality. This primary reflexivity of animal cognition is constitutive of the phenomenal world in general.
3.2.2. The secondary reflexivity of discursive consciousness
We come now to symbolic cognition and its secondary reflexivity13. The characteristic feature of symbolic cognition is its capacity to represent – and therefore to conceive – the organizing categories of experience, using classes of phenomena. The classes of phenomena (auditory, visual, etc.) representing the categories are signifiers and the categories themselves are signifieds. In the primary reflexivity of the animal sensorium, the categories are implicit: they are incorporated into the modus operandi of the neural circuits. In the secondary reflexivity of the human intellect, the world of categories becomes explicit; it goes from the wings of the neural circuits to the stage of phenomena. The activities of certain assemblies of neurons14, those that distinguish color and stabilize the visual category red, for example, are reflected, oddly, in a class of sounds (or in a series of visual characters): “red” in English. This class of phenomena is itself recognized by means of other dynamics of neural excitation, so that the brain becomes capable of designating its own activities using phenomenal images chosen (by the culture) as means of self-reference. The symbol is therefore based on an encoding of categories in two degrees, which involves not only the nervous system at the level of categorization of phenomena, but also a conventional correspondence — established by the collective intelligence of the culture — between signifiers and signifieds at the level of the projection of categories onto phenomena. This representation of the process of categorization in the phenomenal world is the essence of language. Through human language, the world of categories becomes an object of cognition and the activity of knowing can therefore reflect itself. It is this circular process that I call the secondary reflexivity of symbolic cognition. In short, given that at the most abstract level “the observer” is a system of categorization of a flow of data, symbolic cognition is intrinsically reflexive because it permits self-observation by the observer, i.e. the cognition of a categorization system by itself. This is only possible because the categorization system is projected onto the data flow that feeds it.
3.3. Symbolic power and its manifestations
At the origin of human cognitive reflexivity is a capacity for symbolic manipulation that is more general than language, and more basic than music, myths, rituals and techniques. Jacques Derrida15 speaks of a writing originating in thought that is in no way a transcription of speech, and whose marks are not the traces of any previous presence. This primordial writing can also be related to the basic intuitions at the origin of Chomsky’s “universal grammar”16or to the “language of thought” of the philosophical tradition17, but without limiting it to merely being the archetype of languages. As I conceive of it, this cognitive proto-writing is, rather, the abstract objective counterpart of our general capacity to arrange symbols on some kind of grid and to carry out symmetrical, reversible operations of reading and writing on those symbols. This universal abstract capacity for reading and writing is an innate symbolic potential that the hunter-gatherers of oral cultures realized long before the literate people of scribal civilizations. We find this same cognitive potential at the source of the three main types of games that characterize human beings: semiotic, social and technical.
We are distinguished, first, by our linguistic, narrative, musical and artistic abilities in general. We enjoy producing, modifying and transmitting messages. No other species on the planet plays with signs as we do18. There is no need to insist on this obvious fact.
Second, the complexity of our institutions and our social relationships goes far beyond that of other primate societies. Without going into detail on the intricate, complicated legal and political structures of the great civilizations, it suffices to consider the importance of rituals in the vast majority of cultures and in every realm. In ritual, it is people, their costumes, their attitudes, their words or songs, their deeds carried out collectively, with each one playing a role, that act as signifiers. Confucian teachings consider excellence in the practice of ritual (which includes not only special ceremonies, but also everyday behavior) as one of the goals in the education of a cultured person. Filial piety, familiarity with classic texts, elegance in writing, precision in language and ease in performing rituals are part of the same human virtue19.
Third, if humans are a species of technicians, it is precisely thanks to their capacity to process any material object as a meaningful occurrence of a system of categories in complex relationships. The complexity of a printed circuit or an aircraft engine is equivalent to that of a tragedy in verse, a classical symphony or the design of certain Persian carpets. The excellence of contemporary engineering equals that of the sacred architecture of ancient Egypt or India in producing intricate structures and systems of alternating symmetries. Was Dedalus, the Greek hero of technology, not also the architect of the labyrinth? Handling a bow requires as much manual dexterity as using a paintbrush, and the same direct Zen intuition of the target. È cosa mentale. In their apologias for painting, Leonardo da Vinci and Vasari were still struggling with the old hierarchy between the vulgar mechanical arts, concerned with things, and the noble liberal arts, concerned with signs20.This inequality was more a matter of social convention in a particular time and place than of anthropological universality. The movements of the saw and the violin bow are similar: while one produces musical sounds, the other carves a three-dimensional shape. The actions of the musician or the carpenter have meaning within long traditions of practices, which in turn are part of larger cultural wholes. In all cases, signifiers are produced or manipulated: phenomena that are datable and addressable in the spacetime continuum, but extend from rhizomes deep in the virtual worlds generated by symbolic power, intangible worlds where meanings move.
3.4. The reciprocal enveloping of the phenomenal world and semantic world
A symbol consists of two parts: a signifier and a signified. We always encounter occurrences of the signifier part of a symbol in a phenomenal or sensory-motor mode. I am talking here of occurrences of signifiers and not just signifiers because signifiers are themselves classes of phenomena, and not phenomena that are dated and situated in the space-time continuum. Let us think, for example, of words (signifiers), which only have a place in the virtual system of language, as opposed to the pronunciations of these words (occurrences of signifiers), which are very much part of the space-time continuum. The processes of symbolic cognition always ultimately involve classes of interactions perceived, remembered, imagined or dreamed with phenomenal appearances, since concepts have to be represented by signifiers. Phonemes of language, notes of music, characters of writing, icons of ritual, religious or artistic expressions are such classes of phenomena. During actual cognitive processes, however, they are occurrences of signifiers, which are not only perceived but are also produced, transformed or displaced through actions, corporal movements, possibly by means of tools such as pens, brushes or musical instruments. Since categories or ideas (which are by nature abstract) cannot be perceived, imagined and manipulated independently of their perceptible signifiers, we can say that, for human cognition, the intellectual world of categories is necessarily enveloped — but also veiled — in the phenomenal world.
I would now like to show, still from the point of view of human cognition, that the phenomenal world is symmetrically enveloped by the world of the intellect, where relationships exist among categories. Let us begin by noting that the phenomena we perceive, produce and act on are generally named or labeled by us in one or more symbolic systems, in particular languages. We categorize not only the beings and objects we are in contact with, but also their relationships, the dynamics of their relationships and the rules of the games these dynamics obey.
Once a phenomenon is named or categorized in some way, it can be processed as the occurrence of a signifier, i.e. manipulated according to the conventional syntactic, semantic and pragmatic rules that characterize human communities. Thus, a dynamic configuration of phenomena only becomes meaningful on a playing field and according to invisible rules, which belong to the symbolic order. This order is symbolic because it goes beyond the perceptible appearances of the phenomenal world. It involves, in addition, networks of categories invested with all manner of affective energies according to widely varied yardsticks of measurement and evaluation21. This is how our musical intelligence processes sounds. We combine and decode the phonemes of language in this spirit, based on meanings and relationships among meanings in practical situations. Our politeness refines the choreography of social relationships according to complex meaningful patterns rather than mechanical relationships among material bodies. Practically all our interactions with perceptual phenomena can be thought of in terms of the recognition and transformation of meaningful configurations.
I began by saying that symbolic cognition always has a physical, phenomenal, sensory-motor counterpart; we write with our eyes and hands, we speak with our whole bodies. At the same time it should be recognized that all human works, all activities that are part of a cultural framework, including those that seem purely physical or material, also manipulate symbols. We interact symbolically with the phenomenal world and we manipulate images to have access to the world of concepts.
That is why, for example, the secret dance performed by hunters in the forest, involving times and winds, weapons and movements, animals and their tracks, is also symbolic in nature, as projected on the wall of the caves of Lascaux or Altamira at the dawn of prehistory. The same is true of the complicated procedures followed by car mechanics to disassemble and assemble engines, or seamstresses to stitch, mend and embroider garments. In all cases, including what seems at first glance to be a series of utilitarian actions involving only the material world, humans are interacting with occurrences of signifiers, images or bodies that refer to complex arrangements of categories, ideal models, evaluation criteria, scales of emotional intensity, game rules — a whole symbolic universe. We almost always, I repeat, treat actual bodies — including our own bodies — as meaningful images. We do so independently of the senses (hearing, sight, touch, smell, etc.) to which these bodies present themselves or the sensory-motor dynamics from which our cognitive calculation of the bodies emerges. Human beings cannot have any (phenomenal) experience without giving it meaning. What happens is real or fictitious, true or false, insignificant or important, good or bad, safe or dangerous, sad or happy — and to what degree, against the backdrop of what horizon of meaning, practical expectations or desires? Interaction with and between phenomenal bodies is therefore inevitably projected into a world of variables, operations and possible relationships that is not the world of material things in three-dimensional space, but that of conventional symbolic systems. Human symbolic cognition is a strange operator that connects and reciprocally envelops ideas and phenomena. This loop also goes between individuals, since symbolic systems are organized by culture and the phenomena we experience daily are socially co-produced.
3.5. The open intelligence of culture
Animals’ system of categorization is hard-wired in the dynamics of neural circuits and programmed in metastable balances of hormone concentrations. This does not exclude a certain plasticity, as shown by their capacity to learn. The fact remains, however, that the categorizations carried out by the nervous systems of non-human organisms are first encoded genetically at the level of the species. Despite this, the symbolic encoding of categories is not decided at the level of the human species, but in the context of cultural communities that establish and share symbolic systems. It is the capacity of encoding and symbolic manipulation that has been decided once and for all at the level of the species: linguistic capacity is innate or natural in humans. The codes themselves are adopted by cultures: languages, for example, are conventional and variable.
To grasp the nature of symbolic cognition, two points should always be kept in mind. First, symbols (signifier-signified relationships) never exist in isolation: they belong to sets of symbols that form systems: languages, writing, religions, political constitutions, economic rules, etc. Second, the users of these systems of symbols never exist in isolation either: to be effective as symbolic systems, languages, writing, religions, political constitutions and economic rules must function at the level of communities or societies.
As animal organisms, we participate in the collective intelligence of primate societies, the human communities we belong to. As carriers of the logos, we participate in cultural cognitive systems that are much more complex than those of societies of bonobos or gorillas. With respect to symbolic cognition, talking bipeds do not represent autarkic cognitive systems, but rather interconnected processors that carry out — with a margin of real but limited autonomy — the cultural computations of emergent collective cognitive systems. These emergent collective intelligences produce the cultural fabric, first because they interface with and in some way connect many symbolic languages and rules, and second because in doing so they coordinate individuals’ symbolic processing activities. This is how human institutions, in the broadest sense of the term, can function.
Even though the higher animals are capable of learning, the presymbolic cognition of the members of any animal species is usually confined to the closed circle of the categories hard-wired in its nervous system. Symbolic cognition, on the other hand, opens up a general capacity to use almost any durable assemblage of systems of categorization. With its capacity to process symbols, the human brain operates a little like a “universal machine” that can interpret and carry out instructions from a great many systems of categorization, as shown by the huge diversity of languages, music systems, literary genres, religious rituals and technologies created in the history of our species. In addition, at the level of human societies, processing of data is carried out by collective intelligences equipped with recording, communication and calculation devices that enhance our strictly biological capacity to classify experience and manipulate symbols. This means that human cognition is not in principle limited either in the variety of categorization systems and rules for manipulating symbols it can use or in the power of memory and processing of its reflexive intelligence.
3.6. Differences between animal and human collective intelligence
The first scientists to study collective intelligence were ethologists, who observed and modeled the behavior of animals. They showed that although the cognitive capacities of individual ants or bees are quite limited, anthills and beehives, when considered as “wholes” or “superorganisms”, are capable of solving complex problems in a coordinated way22. The combination of many simple individual behaviors results in complex, refined social behavior that exceeds the understanding of the individuals. Collective intelligence exists not only in insect societies but also in schools of fish, flocks of birds, herds of herbivores, packs of wolves and troops of apes. In general, living in societies in which individuals communicate and cooperate is a competitive advantage for many animal species.
Humanity is a highly social species and, as such, it manifests properties of collective intelligence just as other species of social primates do. To end this chapter on symbolic cognition, I would like to sum up the main differences between human and animal collective intelligence23. The scientific question is as follows: is the use of models of animal collective intelligence sufficient to describe symbolic cognition in cultural contexts? I think not.
The root of the difference between the two forms of emergent cognition is the innate biological capacity of humans to manipulate symbols, whether these symbols are linguistic, iconic, musical or other. Once again: we need to distinguish clearly between communication and symbolic potential. For example, many mammals and birds of the same species are capable of communicating among themselves to draw each other’s attention to food sources or the arrival of predators. It is clear, too, that mammals in particular are skillful at communicating emotions such as aggression, fear, joy and desire to mate. Communication can even be elaborate enough to encompass metacommunication, for example, in play activities24. But this does not make animals manipulators of symbols or masters of language. As they do not possess the symbolic potential that distinguishes humans, animals cannot ask questions, tell stories or conduct dialogs. Although a few ethologists consider certain animal societies to have a “culture”, i.e. a certain capacity to transmit invented or learned behaviors, it is clear that nothing equivalent to the history of technology, music or political forms (for example) exists in the animal kingdom or in any particular species. In contrast with the cultural (therefore human) history of architecture, beavers have built their lodges in the same way for as long as there have been beavers.
From the perspective that concerns us here, there are thus two essential differences that distinguish animal and human collective intelligence.
The first difference is that humans not only have remarkable capacities for problem solving; above all, as we have seen, they have reflexive consciousness, which is imparted by discursive thought, whether the speech underlying their thought is internal or part of a dialog25. Animals are also “conscious”, in the sense that their organisms support subjective experience such as perceptions, sensations of pleasure and pain, emotions, etc., but they have no autonomous reflection on their own behaviors. They do not think discursively about their actions before, during or after them, for the simple reason that, lacking language26, they have no means of maintaining any kind of rational thought. They do not represent themselves to themselves in the mirror of their own discourse. We must not, therefore, view human collective intelligence as emerging from the interaction of unreflexive behaviors that lack the autonomy provided by discursive thought – as is the case for collective animal intelligence. In human beings, a threshold has been crossed, because human collective intelligence brings together, connects and organizes individual cognitive processes that are radically more complex and exceptional than those of collective animal intelligence, cognitive processes that are in a sense illuminated from within by discursive reason27.
The second difference is that human collective intelligence is applied from one generation to the next over the course of history or cultural evolution. For example, the history of the processes of material production and transformation over the long term shows an increase in the power of the human species over its environment. Once pottery and metallurgy were invented, these processes were transmitted and perfected, and were added to what had previously been acquired in the history of technology. The same is true for communication media and systems of signs, such as writing or currency. In general, inventions that increase the power of the societies that use them are preserved, whether they involve material processes or symbolic institutions. Unlike animal collective intelligence, human collective intelligence learns not only on the scale of the time of a generation or the space of a society, but also on the much broader scale of the space-time of the human species as a whole.
1 See Geoffrey Bowker, Memory Practices in the Sciences [BOW 2005].
2 For example: writing and accounting systems, currencies, prices, deeds, banknotes, financial operations.
3 Culinary symbolic systems may combine or alternate: raw and cooked; hot and cold; crisp and soft; bitter, sour, sweet and salty; etc. The Chinese, Korean and Japanese cuisines are particularly subtle in this regard.
4 See Roland Barthes, The Fashion System [BAR 1990].
5 In the sense of sexual acts that are permitted, recommended or prohibited according to the parental, social, legal and sexual status of the partners, as well as their state of ritual purity, the calendar, etc.
6 The classic example of a traditional codification of erotic symbols is the Kama Sutra.
7 See, for example, Claude Lévi-Strauss, The Elementary Structures of Kinship [LÉV 1969].
8 The World’s medicines are based on very different symbolic systems, which can lead to surprising differences even with regard to anatomy, as shown by Shigehisa Kuriyama in The Expressiveness of the Body and the Divergence of Greek and Chinese Medicine [KUR 1999].
9 The concept of language game is one of Ludwig Wittgenstein’s great discoveries. After having developed a philosophy aimed at standardizing the logical and descriptive uses of language at the beginning of his career (in the Tractatus Logico-Philosophicus [WIT 1921]), Wittgenstein arrived at an open exploration of the grammars of actual language games at the end of his life. The practical uses of language games and the relationships between heterogeneous language games are at the forefront of his Philosophical Investigations [WIT 1958]. From the early to the late Wittgenstein, one theme nevertheless remains constant: that of the limitations of the intellectual faculty associated with language, and particularly the limitations of its capacity for self-description. See, for example, the final aphorisms of the Tractatus and many aphorisms in the Investigations.
10 Similarly to Wittgenstein, Austin, in How to do Things with Words [AUS 1962], and after him, Searle, in Speech Acts [SEA 1969] and Intentionality [SEA 1983], clearly showed that factual description and logical reasoning were only one aspect of language use. The “pragmatic” dimension discussed by these authors is less concerned with truth than with the more or less constraining practical force that conventional rules give to acts of enunciation such as promises, commitments, judgments, etc. As François Rastier suggests (in “La triade sémiotique, le trivium et la sémantique linguistique” [RAS 1990]), reflection on the pragmatic uses of language, at least in the Western tradition, probably dates back to the rhetoric of antiquity.
11 In particular, the Chilean school of biological philosophy represented by Humberto Maturana and Francisco Varela; see Autopoiesis and Cognition [MAT 1980].
12 See Observing Systems [FOE 1981].
13 In Language and Human Behavior [BIC 1995], linguist Derek Bickerton provides an excellent analysis of reflexive human consciousness based on linguistic capacity.
14 See Jean-Pierre Changeux, Neuronal Man [CHA 1983].
15 In particular in Speech and Phenomena, Writing and Difference and Of Grammatology [DER 1973, DER 1978, DER 1976].
16 See Syntactic Structures [CHO 1957], which outlines the formal core of this “universal grammar”, and New Horizons in the Study of Language and Mind [CHO 2000], in which Chomsky summarizes his philosophical positions on this subject.
17 See Le Discours Intérieur. De Platon à Guillaume d’Occam, by Claude Panaccio [PAN 1999]. Particularly notable is the concept of the “inner word” developed by St. Augustine in On the Trinity [AUG 2002].
18 See what a certain school of “French thought” has produced as a variation on this theme. For example, Barthes in The Fashion System [BAR 1990], Baudrillard in For a Critique of the Political Economy of the Sign [BAU 1981] or Guattari in Chaosmosis [GUA 1995].
19 See Herbert Fingarette, Confucius, The Secular as Sacred [FIN 1972], and Anne Chang, Histoire de la Pensée Chinoise [CHA 1997].
20 They were trying to show that painting belonged to the liberal arts, to rhetoric, even though it did not use words.
21 In his Philosophical Investigations [WIT 1953], Wittgenstein observes that “language games” are not only linguistic phenomena, but forms of life.
22 See the classic works by Edward Wilson, The Insect Societies [WIL 1971] and Sociobiology: The New Synthesis [WIL 1975]. See also more recent works by Bonabeau and Théraulaz, such as Intelligence Collective [BON 1994] and Swarm Intelligence [BON 1999].
23 By animal, I mean here non-human animal, although, strictly speaking, humans are obviously also animals.
24 This point was made by Gregory Bateson in Steps to an Ecology of Mind [BAT 1972].
25 For the great Russian psychologist Lev Vygotsky, the development of internal discourse (and therefore thought) is the result of an internalization of dialog. See Thought and Language [VYG 1986].
26 This lack is obviously not a “flaw”. Animals, including their various cognitive styles, are perfect as they are.
27 This reason may be sick, perverse, saturated with unconscious impulses, conditioned by somatic or cultural structures that are beyond it, but none of this prevents reason from existing and remaining, in spite of everything, unique to humanity.