An early step on the road to GA was taken in 1969-70, when Alain Cohen, a UCLA graduate who still teaches at UCSD, introduced me to The Pragmatics of Human Communication by Paul Watzlawick et al. (Norton, 1967), a product of Gregory Bateson’s psycho-communications “project” at Stanford. The notion of pragmatic paradox struck me as the key to the esthetic–so much so, that during a visiting professorship at Johns Hopkins in 1978, I offered a graduate seminar on le paradoxe esthétique with no predefined syllabus. The consequence of this decision, although professionally disastrous, was intellectually rewarding, since it was at Hopkins that I formulated the originary hypothesis.

Rereading Bateson after nearly forty years–and nearly thirty of GA–recalls my early affinities with his thought yet also reveals a profound divide. As suggested by the title of his best known collection of articles, Steps to an Ecology of Mind(Jason Aronson, 1972; all page references below are to this text), Bateson’s awareness of the interactive nature of thinking and consequent refusal to be confined within the individual psyche puts him in the line of thought that leads from Durkheim to GA. If nothing else, Bateson’s theory of the “double bind,” so to speak an Anglo-Saxon version of Girard’s “French triangle” (the quote is from James Joyce), guarantees his continued relevance. Yet he remains, as Derrida might have put it, behind la clôture de la métaphysique. Because Bateson analyzes communication in logical rather than semiotic terms, notably through Bertrand Russell’s “theory of types”–which creates a logical hierarchy that averts paradox by forbidding words to refer to others of their own ontological level or “type”–he lacks the conceptual means to grasp the qualitative difference between human and animal language. As Terrence Deacon (The Symbolic Species [Norton, 1997]) is still one of the few students of human language to recognize, the types of signs that C. S. Peirce called indexical and symbolic are discontinuous, ontologically and neurologically. Both humans and animals communicate through indexical signs, expressions of emotions that provide information to conspecifics, such as a cry of fear that serves as a warning signal. In contrast, the category of symbolic or in Saussurean terms, arbitrary signs, belongs only to humans, and Bateson fails to appreciate the impossibility of a logic-based “evolution” from one to the other.

Today when zoosemiosis is all the rage, it is considered bad form to insist on what anyone but a university intellectual can see is the qualitative superiority of human language over that of dolphins or bonobos. Indeed, Bateson’s “ecological” view of nature, in which not only humans and animals but their natural environment as a whole are all part of a single system, makes him a key ancestor of modern environmentalism, in its intelligent as well as its mindlessly victimary aspects. Bateson’s observations in his 1967 essay, “Style, Grace and Information in Primitive Art,” that “mere purposive rationality unaided by such phenomena as art, religion, dream, and the like is necessarily pathogenic and destructive of life,” that “unaided consciousness must always involve man in the sort of stupidity of which evolution was guilty when she urged upon the dinosaurs the common-sense values of an armaments race,” and finally, that “unaided consciousness must always tend toward hate” (146) are prime illustrations of the sixties critique of “capitalism” and the Cold War whose roots lie in the postwar repudiation of the Fascist exaltation of the principle of dominance that culminated in the Holocaust.


Bateson begins the 1955 essay “A Theory of Play and Fantasy” by describing his quest to understand the evolution of communication, “a very important stage [of which] occurs when the organism gradually ceases to respond quite ‘automatically’ to the mood-signs of another and becomes able to recognize the sign as a signal” (178). This stage is subsequently defined as “the drama precipitated when organisms, having eaten of the fruit of the Tree of Knowledge, discover that their signals are signals. Not only the characteristically human invention of language can then follow, but also all the complexities of empathy, identification, projection, and so on” (179). In this passage, despite the affirmation of the uniqueness of human language implied by the biblical reference, Bateson displays his unfortunate tendency to conflate language with forms of emotional interaction already in evidence in the animal world.

A visit to a zoo revealed to Bateson that playing animals “were capable of some degree of metacommunication, i.e., of exchanging signals which would carry the message ‘this is play’” (179). This gave him the idea of using play as his primary model for a higher level of consciousness than simple (indexical) “mood signs”; in order to play, an animal must be aware that its own actions and those of its fellows are not “real.” Bateson describes play in logical terms: “‘these actions [of play] . . . do not denote what would be denoted by those actions which the actions denote.’ The playful nip denotes the bite, but it does not denote what would be denoted by the bite” (180). After pointing out that this use of “denote” on two different levels violates Russell’s theory of types, he concludes that “the evolution of play may have been an important step in the evolution of communication” (181), presumably leading to human language.

Thus Bateson extends the term “denote,” which in its literal sense can only be accomplished by a signifier denoting its signified, to refer to the indexical association of one act with another. “The nip denotes the bite” may be interpreted as saying that in the universe of dog play, the nip means the bite, corresponds to what would be a bite in the “real world.” As for what the bite itself “denotes,” this cannot be defined in the play world, where the bite cannot exist; thus the paradoxical thought of its denotation constitutes a type violation.

Yet among dogs as among humans, an aggressive act does not normally “denote” anything. If we play cops and robbers, in Bateson’s terms, saying “bang bang” “denotes” shooting. If, however, I really shoot you, what am I “denoting”? Shooting might be considered an index of aggression, but what we refer to as an animal signal–a warning cry, for example–has a presumably adaptive communicative function, whereas a bite, like a gunshot, normally “communicates” nothing more than its practical consequences. Bateson’s metaphoric use of logical terminology beclouds the question of language origin that he originally posed in the unambiguously human terms of the Tree of Knowledge.

However pregnant with future evolutionary possibilities, animal play is not comparable to human symbolic representation, in which words and things belong to different universes rather than one being a “non-serious” form of the other. The formulation of the connection between nipping and biting demanded by Occam’s razor is not that the first denotes the second but that the first is an inhibited version of the second. Without going into the details of dog behavior, where play is largely subordinate to displays of submission and dominance, it suffices to say that a dog nips rather than bites because at the level of aggression at which nipping occurs, it has proved adaptive to avoid bloodshed by discharging this aggression while at the same time displaying an absence of intention to attack. The original source of the dog’s inhibition, we may assume, is fear of being bitten back, and eventually dogs come to “know” they are playing, but only an uncritical anthropomorphism would extend this knowledge to awareness that their nips “denote” real bites.


Bateson goes on to develop a second model of “denotation,” Korzybski’s model of map and territory, which would seem to make more explicit than play the symbolic rather than merely indexical nature of the relationship between its terms. A map is not a reduced version of a territory, but a separate two-dimensional diagram. Yet Bateson uses the map-territory model rather to illustrate the paradoxical effect of ritual, where “the discrimination between map and territory is always liable to break down, and ritual blows of peace-making are always liable to be mistaken for the “real” blows of combat” (182). Thus even after quoting Korzybski’s line, “The word ‘cat’ cannot scratch us” (180), Bateson describes, as he did with play, ritual blows as “mapping” the real blows of combat.

These formulations reflect a genuine anthropological intuition struggling to express itself in an inadequate vocabulary. Bateson’s fascination with the ritual confusion between map and territory ultimately leads him to a pregnant quasi-Durkheimian reflection on religion:

Finally, in the dim region where art, magic, and religion meet and overlap, human beings have evolved the “metaphor that is meant,” the flag which men will die to save, and the sacrament that is felt to be more than “an outward and visible sign, given unto us.” Here we can recognize an attempt to deny the difference between map and territory, and to get back to the absolute innocence of communication by means of pure mood signs. (183)

The “metaphor that is meant” corresponds nicely to the sacred as both signified and referent of the originary sign; the originary “name-of-God” is meant in both senses of signified and taken seriously. And the soldier defending the flag is Durkheim’s own example of the secular sacrality that links modern France to the stone-age tribes of central Australia. But Bateson never conceives that this “attempt to deny the difference between map and territory” might refer, through the mediation of sacred history, to an originary moment in which the uniquely human difference of map and territory, sign and thing, first emerged.

The “absolute innocence” sought by religious sacrament is in fact wholly different from the indexical “mood signs” we share with animals. The “innocence” of the latter is simply their pre-human dependence on biological “moods”; animals are innocent whether nipping in play or slashing away in earnest. In contrast, the “innocence” of the ritual conflation of map and territory derives from its evocation of a sacred being that not merely reduces violence to play but suspends action altogether, so to speak transforming a real-world cat into Korzybski’s ‘“cat’ that doesn’t scratch.” This sacred guarantee of communal peace must precede the ritual pseudo-violence that culminates in sacrifice–and that, as Bateson emphasizes, sometimes breaks down in real violence. Maintaining peace in the human community is a difficult and not always successful undertaking, but the point to retain is that the “absolute innocence of communication” that is the aim of ritual and of all other cultural phenomena depends on our ability to transcend the violence of our worldly “moods” in the shared contemplation of a scene of communal significance.

Bateson concludes the essay with the reasonable but insufficient observation that “paradoxes of abstraction [i.e., that confuse different logical types] must make their appearance in all communication more complex than that of mood-signals, and . . . without these paradoxes the evolution of communication would be at an end. Life would then be an endless interchange of stylized messages, a game with rigid rules, unrelieved by change or humor” (193). Although this less than perspicuous formulation makes the important connection between paradox and (the origin of) human language, it displays no awareness that linguistic signs are neither “mood-signals” nor meta-communications about mood-signals on the model of “this is play.” Indeed, given that animal play must be understood as helping maintain social organization by deferring intraspecific violence, human language can emerge only when play of this sort and the pecking-order dominance hierarchies it supports have proved inadequate to maintain order. The stability of the human community is founded not on a one-on-one pecking order but on the symmetrical interaction of all its members around a central sacred object that defers mimetic violence not by “subliming” it into play but by interdicting its appropriation as an object of worldly desire.

Before logical or linguistic categorization can take place, the community must establish through its exchange of signs a transcendental universe of representation, the world of Derrida’s différance. The sacred object, by being designated as inaccessible, as present on the scene of language rather than in the real world, is withdrawn as a potential source of mimetic violence. The fundamental paradox that underlies all logic and permits the subsequent establishment of “logical types” is that the sign designates its referent as “always already” part of this transcendent universe. What we point to is God, both/either because pointing makes it God, and/or because pointing recognizes its divinity.

Unlike the world of mathematics, the world of logic requires an object-universe to which its propositions refer–and reference to which risks generating paradoxical “objects,” such as Russell’s barber who shaves everyone who doesn’t shave himself. But this object universe can only be constituted in the first place by the act of faith that accepts that the object the sign designates as significant is significant independently of the sign. What bears out this faith is the ability of signification, of representation-as-sacred, to defer violence within the community. The social order rests on the equivalent of a Prisoner’s Dilemma enforced by the community as a result of its universal acceptance. The theory of types can only get off the ground once the first “type,” that of the sign as ontologically distinct  from the thing it refers to, has been saved from the violence attendant on confusing “map” with “territory,” the “aborted gesture of appropriation” with a real attempt to appropriate the object, by the communal faith that the signifying gesture designates a sacred being inaccessible in the real world.


In a later essay, “Redundancy and Coding,” which first appeared in a 1968 collection on animal communication published by semiologist Thomas Sebeok, Bateson makes a more focused attempt to distinguish between human and animal communication. He begins by rejecting the common-sense idea that “language replaced the cruder systems of the other animals” (417), not realizing that by denying this “replacement” he is in effect admitting that human language isdiscontinuous with animal communication and consequently cannot be derived from it.

After a good deal of inconclusive speculation on how our “digital” language might have evolved out of the “iconic” (that is, indexical) language of animals, Bateson settles on the emergence of the negative “not” as the simplest criterion of this transition. Bateson sees an important difference between “don’t” as a negative imperative that animals can communicate indexically, for example, by threat, and “not” as a simple negative referring in the case at hand to the animal’s own actions: “I will not bite you” (430), which can only be communicated symbolically.

[E]volution of a simple negative would be a decisive step toward language as we know it. This step would immediately endow the signals–be they verbal or iconic–with a degree of separateness from their referents, which would justify us in referring to the signals as “names.” The same step would make possible the use of negative aspects of classification: those items which are not members of an identified class would become identifiable as nonmembers. And, lastly, simple affirmative indicative statements would become possible. (431)

This absolute “degree of separateness” between sign and referent is what defines symbolic as opposed to indexical signification. Although Bateson does not make clear whether the negative is the source of this separateness or a sign of it, we can accept his principle that human language originates with the “negative,” provided that we interpret the latter not as linguistic negation but as cultural interdiction. This allows us to understand Bateson’s “don’t” vs “not” as opposing, in the first case, the animal’s one-on-one threat, and in the second, the communal scene of human language, where the reciprocal exchange of the “aborted gesture of appropriation” demonstrates that all members of the group have renounced the intention of appropriating the central object. If for no other reason, this seemingly adventurous extrapolation from Bateson’s thought is justified by the historical fact that Bateson’s “double bind” and Watzlawick’s “pragmatic paradox” provided the conceptual tools for the first step on the way from Girard’s theory of mimetic desire to generative anthropology.