The Nonverbal Inlay in
Linguistic Communication

JOHN N. DEELY
with tables based on drawings by Brooke Williams

The emergence of semiotics as the architectonic of communications study as such sheds new light on and, to some extent, calls for a reconceptualization of linguistics in the traditional, scientific sense. Of course, the difficulty of defining “language” in a satisfactory way is celebrated, and is as much as any single factor responsible for the currently faddish but philosophically unsophisticated claims of animal experimenters to have taught “language” to chimpanzees, etc. I would agree simply with Professor Sebeok in regarding this current fad as “counterproductive and misleading”; and I would further argue that these widely heralded cases of language usage by carefully cued chimps are each one instances of the “clever Hans fallacy,” that is, the fallacy of researchers finding in the animals’ behavior what they themselves have put there by various muscular or behavioral cues (cf. Sebeok 1979).

But that would form the subject of a paper in its own right. Here I have a related but different objective, namely, to contribute to a clarification of the relationship between linguistics and semiotics by outlining a framework or model wherein the function of language in relation to behavior and social structure appears in its proper light. In other words, I want to suggest that the impossibility so far of achieving a satisfactory definition of “what language is” may be the result of a failure to appreciate the proper junctional status of language in the context of social behavior; and that the rise of semiotics provides an opportunity to reconceptualize our understanding of language in a framework that makes this proper functional status for the first time generally apparent and directly accessible.

One of the first and broadest divisions of semiotics is into anthropo-semiotics, or sign-systems as they function in a uniquely human context, and zoosemiotics, or sign-systems common to human and nonhuman animals. I think that language, properly conceived, is the key to rightly understanding the rationale of this contrast. To show my reason for thinking this, I want to introduce two distinctions. First, I want to distinguish zoosemiotic sign systems against the linguistic system that is species specific and unique to man. Then I want to set both of these systems in contrast with what I am going to call post-linguistic structures or systems,¹ by which I mean systems that come into existence on the basis of language and can only be understood in what is proper to them on that basis, but are not themselves linguistic; and once they have come into existence, they redescend, so to say, into the purely perceptual to become assimilated in a behavioral way to the society of non-linguistic animals. Thus, post-linguistic systems are normally simultaneous with language, and even, from the point of view of the individual language-learner, in certain ways precede ‘language’; but they always depend upon language for their proper existence, which transcends the modalities of simple perception and zoosemiotic signalling. Post-linguistic structures exist beside, alongside, aside from, language—yet are based on and derivative from it. And they react upon language, by influencing the semiotic exchanges that transpire through language (just as sensation does in pre-linguistic ways). Thus we find in linguistic behavior, as it were, a twofold non-verbal inlay embedded in and influencing semiotic exchange, one from the pre-linguistic order of perception, and one from the users themselves of language, as they are products of and have been shaped—even in their eventual perceiving—by a world of post-linguistic experience.

Here, the expression, “post-linguistic,” must not be separated from the notion of “experience.” I want to say that “experience” can be either pre-linguistic or both pre- and post-linguistic (though not post-linguistic without being also pre-linguistic), and that this is the proper contrast between zoosemiotics and anthroposemiotics taken in their fullest amplitude.

To see what is involved in these distinctions, let me develop the discussion around a sequence of diagrams or tables, which illustrate the double dichotomy between language and pre-linguistic structures of experience, on the one hand, and language and post-linguistic structures of experience on the other hand.

Consider first of all pre-linguistic experience (where the first nonverbal inlay in linguistic communication comes from), in a purely zoo-semiotic sense. For ease of imagination, I am going to focus only on the peculiarly anthropoid structure of experience as it began to emerge, roughly from the time of the Australopithecines; because, of course, to take into account insect and other animal forms far removed from the human, a great many conceptual modifications irrelevant to our present purpose would have to be introduced into the scheme. I think we would not be far wrong in envisioning an experience structure for anthropoids somewhat along the following lines (Table 1):

The initial and sustained immediate level of cognitive contact between organism and extra-organismic world is through the so-called “external senses,” whence we encounter colors, shapes, tastes, textures, odors. But connected with these, inseparably, and with no temporal gap, is a great deal else besides: to see colors is to see shapes, places, a plurality of things—so, the whole series of notions in Table 1 below the term “synthesizing” (which takes place within the organism experiencing) comprises the first data of experience out of which, eventually, language itself (the interface, let us say, between pre- and post-linguistic structures of experience) will arise.

The situation here bears looking at more closely (Table 2):

In this Table, we have the same picture of anthropoid experience, but broken down more clearly into its structural components. Note particularly the relation between the whole of perception and its immediate sensory core. Precisely here semiotic analysis enables us to give, for the first time, a satisfactory resolution to the controversy over “sense data,” as that controversy traces back in empiricist philosophies from present-day positivism to its beginnings in John Locke with his “simple ideas” of sense. The empiricist tradition has always had a problem figuring out why the mind structures sensory objects in the ways that it does. The origin of this difficulty was twofold. First, the direct objects of each external sense—colors, sounds, textures, tastes, and odors—were considered to be wholly effects within the mind itself, perhaps caused by external objects, but certainly not properties of those external things as such. This was in contrast to an older, less critical “realist” tradition which simply considered color to be in the grass, etc.

What eluded our early modern forebears, as their successors, was yet a third possibility for these initial data of sense: they could be regarded as neither intrinsic properties of physical things nor mere modifications of our subjective faculties, but precisely as properties exhibiting how things are in their action here and now on an organism possessed of this determinate range of sensitivities. This third alternative would account for the experimentally demonstrated relativity of sensory qualities without making of them mere constructs of subjectivity—while preserving, that is to say, their status as here and now revelatory of “the way things are” so far as involves the experiencing organism. The initial contact between cognizing organism and environment, on this view, is indistinctly subjective and objective. The world appears thus (e.g., colored) only to a subject, but it really is that way given that totality of conditions.

Equally serious was the characteristic modern failure to see that the initial sensory data are not “atomic” in character, but strictly semiotic. Thus, the initial synthesis of sensations made by the mind is not arbitrary or habit-controlled, as the early moderns, notably Hume, opined, but naturally determined by semiotic means. Since a sign is anything functioning to bring something other than itself into an organism’s awareness, and since colors, sounds, textures, etc., immediately bring along with themselves an awareness of plurality, positions, shapes, movements, and so on, one has only to regard sensation semiotically in order to realize that we are already given within it an outline of objective structure that is relative and changing, true, but naturally determined nevertheless. “Sense data,” in other words, already comprise an objective structure (not atomic elements logically discrete but habitually combined) that is the same, in the above sense of naturally determined, for all organisms. Species variation, consequently—that is, subjective differences of sensory range and variety—are strictly background phenomena at the level of the sensory core; and that is why biological mechanisms of camouflage (such as protective coloration) and deception, though principally dependent on the “qualitative” appearances or “properties” of bodies, are so widespread and consistently useful in nature across species lines.² Species variation comes to the fore only in the interpretation of the “common” data and in the uses that they are put to in predation and social exchange.

Now, language, in the sense that linguistics studies it, and as it constitutes a species-specific system, is something that emerges, so to speak, “on top” of this common structure of anthropoid experience. To understand it in relation to that experience sufficient for present purposes, it is enough to indicate how the “natural” and the “conventional” relate in the separating off of specifically linguistic from purely behavioral sign systems in anthropoid interaction.³

Prior to the advent of cognitive organisms, relations are only physical, that is, obtaining between presently existing entities. Once cognition is introduced, the situation drastically changes. Now the physical relation of A to B (e.g., of clouds to rain, or of smoke to fire, etc.) acquires the possibility of a new, further dimension: A can represent B to C, a cognitive organism. What was formerly only a physical relation can now become also a recognized (a known) relation that makes A, besides being the cause of B (or B, besides being the effect of A), its sign. Nor is this all. Because the relationship as such has its proper being “between” A and B, while it has its reality independent of cognition from the existence of A as, say, cause of B (or from B as effect of A, depending on which relation is considered), it can now exist as “between” A and B in cognition (memory, say) even after A (or B) has gone out of existence. Essence and existence, in the case of relation, do not go together in a way that ties relation to a unique subjective ground. Thus the same relation existing independently of cognition can also exist dependently upon cognition, and can continue to exist in cognition after its physical ground no longer exists! Moreover, relations existing first in cognition can subsequently be introduced into the physical order by bringing about the proper conditions within that order; and relations can be established in cognition that can have no counterpart in the physical world, just as there are physical relations that have not yet or will not ever enter into the cognition of some organism.

Physical relations are the paradigm case of “natural” relations, but cognitive relations too are forms of “nature,” and many such are exhibited in the instinctive behavior of animal species. The important thing for semiotics is simply that these two orders of relations are functionally equivalent in perception wherever it is a question of inter- or intra-specific interaction between individual organisms or between organisms and relevant features of the Umwelt. Comparatively “real” and “unreal” from the standpoint of what exists independently of cognitive activity as such, physical and cognition-dependent relations are strictly on a par from the standpoint of what works within experience to achieve some determinate end. Both types of relation are essential to the constitution of experience, as is clear from the above remarks; but within experience, it is not essential that the two be explicitly distinguished according to type, and indeed it is not always possible to separate out in experience what is the contribution of the subject and what is the contribution of the object to the experimental structure of a given interaction situation.

In philosophical tradition, one little-realized way of grounding the putative distinction between understanding and perception (“intellect” and “sense”) was precisely in terms of relation: perception reveals objects as they are only relative to the dispositions, needs, and desires of the organism perceiving, whereas understanding reveals in these same objects the further dimension of existence in their own right independent of relations to the knower.⁴

It is this way of interpreting the difference between “sense” and “intellect,” it seems to me, that is decisive for semiotics.⁵ For within the interweave or mix in experience of cognition-dependent and cognition-independent relations, only those organisms possessing the capacity to understand in its distinction from the capacities to sense and perceive, in the way just described, will ever be able even on occasion to discriminate between real and unreal elements in semiosis, that is, in the process of communication through signs. Stipulation, as a distinctive semiotic process, presupposes exactly this ability; and it is only in relation to stipulative decisions and their consequences that language can be said to be “conventional.” But stipulations, when successful, pass into customs, and customs into nature. Thus, sign-systems arise out of nature in anthropoid experience, become partially “conventionalized” in the sphere of human understanding, and pass back again through customs into continuity with the natural world as it is experienced perceptually by human and nonhuman animals alike (Deely 1978:7ff.). “Language,” in short, in the sense that is species-specific to homo sapiens, is nothing else than the “unreal” component of semiosis explicitly segregated and seized upon in its unique signifying potential by the understanding in its distinction from perception and sense. As a result of the intervention of language, those organisms capable of seizing upon the difference between the “real” and the “unreal” elements of experience soon find themselves in an entirely different world. In prelinguistic experience, relations are not distinguished from the objects related. With language, it becomes possible to separate the two. The consequences of this simple feat are enormous, and without end—literally, for it is this that makes human experience an “open-ended” affair as a matter of principle (Table 3):

In this diagram, which looks very complicated at first glance, notice to begin with the same basic structure of anthropoid experiences—the zoosemiotic structure that is common to man and the higher anthropoids—that we examined in Table 2, retained in the lower half of the diagram. But in man, the processes eventuating that structure go up one step—but what a step!—higher. Passing through the layer of language, the social structures of anthropoid life “come out on the other side,” so to speak, in the form of such things as military establishments, civil governments, religious traditions—things which are not linguistic, are much more than the linguistic system, but which couldn’t come into being apart from language, enabling us as it does to systematically differentiate between relations (as offices and social roles, for example) and the things here and now related (the individuals concretely playing those roles in any given case). So, this diagram looks complicated, but when examined carefully, it appears that the upper portion is structured representatively by only three arrows (though indefinitely more, of course, could be added) each of which splits into two.

Follow, to begin with, the left-most arrow. As it splits toward the right, it leads toward a military establishment as it is experienced by linguistic animals belonging to the society of which that establishment is an expression. That same establishment, of course, has an impact within the experience of any non-linguistic animals having commerce with the society, and elements of it will be perceived and assimilated by semiotic processes, but now only insofar as these elements of the military establishment have “gone below” the linguistic layer. This is indicated by the left-most split of the leftward arrow, which leads back toward the sphere of simple perception and pre-linguistic experience. The domesticated dog, for example, will well perceive a relative social importance of individuals as they affect his own well-being; but without language, he will never perceive the four stars of the general, the bars of the lieutenant, or the stripes of the sergeant for what they are as properly military insignia. Post-linguistic structures, in what is proper to them, are forever hidden from the standpoint of pure perception and pre-linguistic experience.

Thus we have animal experience in the sense of zoosemiotics below the broken lines of language, and, both above and below the lines, and indeed “centered” by them, human experience which, for the reasons given, includes what the animals have, plus the post-linguistic institutions or structures which exist properly consequent upon language and owing to it. It is thus that human experience is much broader than that of even the highest of animal species in the total biological community without “understanding.”⁶ Post-linguistic structures exist properly above the linguistic interface; but they also exist in some way other than what is proper to them assimilated and accounted for in animal experience below that interface, and, as they are encountered perceptually, we must further distinguish the sense in which they are encountered by non-linguistic anthropoids from the sense in which they are encountered by post-linguistically formed but alien organisms (“outgroups”: the broken arrows to the right of the diagram leading back up through the interface).

In the taking account of things in experience by animals, one way of describing what results from post-linguistic structures would be as two levels of natural systems—a primary (or infralinguistic) level and a secondary (or supralinguistic) level (Table 4).

If we adopt in imagination the point of view of an alien—an alien in the sense of someone from outside the given culture, or even from another planetary system—observing for the first time life in a given human society, what would appear? A series of regularities, or “natural systems,” which, upon examination, would appear in one sense primarily as objects or patterns of perception, but in another sense, as objects of understanding, as derivative from the unique indifference of language to “reality” as perception alone reveals and grasps it. The so-called “primary” natural systems comprise real (physical, i.e., cognition-independent) relations of the human organisms to one another and to their Umwelt. Further observation would reveal, however, within this primary network of perceptually accessible interactions, and carried by them (conflated with them in social life), a secondary network of (comparatively) unreal relations conveying the historical experience of the group as something transcendent, not to individual and social reality, but to the perceptually accessible elements as such expressing the “primary” systems. The so-called “primary,” or perceptually available systems, thus, “feed into” the linguistic interface in and through social interaction; but what “comes out on the other side,” so to speak, is not language, but society enculturated, society as a potential cultural system.⁷

The main point here is that the system of language, to be best understood, should not be conceptualized as a whole unto itself, but as an interface, a perceptually diaphanous network of unreal relations intervening as such between the specifically human language users and the layered manifolds of experience they seek to understand. It is in such a light, I conclude by suggesting, that the relation of linguistics to semiotics might well be further examined.

NOTES

1. So far as I have been able to determine, this distinction I am proposing between linguistic and post-linguistic systems does not appear anywhere in previous semiotic literature. This strikes me as somewhat improbable, so I should not be surprised to later learn that it does have a counterpart in some previous book or essay as yet unknown to me. Nevertheless, I must report that the only distinction found so far even analogous to the one I am proposing is Jurij M. Lotman’s distinction between primary and secondary modeling systems, first called to my attention by Donna Jean Umiker-Sebeok in conversations preparatory to this paper: see Daniel P. Lucid (1977:7ff.), and J. Lotman (1977a:95-98; 1977b: passim); and that this distinction as it is employed in Soviet semiotics, though unquestionably related to the distinction being proposed here, is not the same distinction, and indeed is not yet clear in its foundation. Lengthy discussions with experts versed in the Soviet literature, notably Dr. David Danow of the University of South Carolina, served only to confirm the ambiguity as to the grounding of Lotman’s usage. My choice of terminology for expressing the proposed distinction was a difficult one. “Paralinguistic” would have been my preferred term for what I am here calling “post-linguistic,” but previous appropriation of this term by linguists (see Sebeok, Hayes, and Bateson 1964; Crystal 1974) made this choice likely to create endless and needless misunderstandings. The final choice of “post-linguistic” was influenced more than anything else by Charles Morris’ discussion of what he called “post-language symbols” (1971:122-25); but whereas Morris is naming a particular group and functioning of signs as such, the structures I am referring to incorporate and generate signs without being in themselves semiotic, even though they owe their origin to semiosis and are normally permeated by semiotic dimensions.

2. Julian Huxley, in his classic study (1942:414), went so far as to say that Cott “has shown that concealing and revealing coloration, when properly investigated, remain the paradigm of adaptive studies, and has thoroughly turned the tables on captious objectors.”

3. Elsewhere (Deely, 1978: especially pp. 7ff.), I have dwelt at length on the essential ambiguity of this traditional dichotomy between the “natural” and the “conventional.” That account provides the background concepts for most of what follows here.

4. The ‘sensory core’ we have distinguished within perception (Table 2 above) consists of the impressions produced in the organs of sense by the action here and now of the surrounding environment on the organisms. Never given as such in our experience, which is always of perceptual wholes, pure sensations are known only derivatively and by an analysis which proceeds from the realization that unless there were such first elements or data at the base and core of perception, we would find that all knowing in every respect entails an infinite regress, and so could have no point of origin (cf. Price 1950:3, 7, 149, 155).

Now sensations are always given with and by perceptions, that is to say, within an elaborate and detailed network of objectivity that is the work of memory and imagination as well as of the so-called external senses. Moreover, the perceptual field is not only determined by the individual experiences of an organism, but, even more profoundly, by the anthropoid history of the higher organisms (which alone concern us at present) as it has been built into their genetic constitution over the centuries by the complex processes of ‘natural’ or evolutionary selection. Thus, a given organism has a ‘natural’ perception of certain objects as friendly or hostile, alluring or repulsive, prey or predator, etc., and these naturally given determinations of perception—sometimes called “instinct”—precisely consist in the catching of a given element of experience within a net of “unreal” or cognition-dependent relations whereby the perceiver apprehends its objects not principally according to what they are “in themselves,” as it were, but rather according to what they are so far as the perceiving organism is involved. The same remarks apply to cognition-dependent relations attached to objects not by the a-prioris of organic constitution (the organism’s genetic and selective history), but by the simple learnings about things built up through experiences the organism undergoes.

Now outside the human species, and indeed often enough inside the human species, this difference between relations obtaining among objects so far as the actions or behavior of the individual organism is concerned and relations obtaining among objects prior to or independent of the self-interests of the perceiver in the perceived—this is never thematized, never disengaged as such as an explicit component of a categorial scheme. Yet, prior to such a thematization, in all our direct experiences of objects, the objects are given in an apparently unified way which in fact conceals the profound differences between an object as a thing in its own right and an object as an element of experience being accounted for by the perceiver in terms entirely born of its own needs and desires.

Thus, all objects of direct experience, from an independent viewpoint, are known to be an at best imperfectly discriminated amalgam of what is, in terms of cognition-independent being, being and nonbeing, that is, cognition-independent and cognition-dependent interwoven networks of objectified relations.

The discrimination of relation as such, as a mode of being distinct from and superordinate to related subjects, seems to take place only upon a comparative analysis—not always self-conscious and transparent to itself, be it said—which is also able, in principle at least, to further distinguish among relations so discriminated between those whose entire actuality is the work of the perceiving organisms (whether idiosyncratically, by custom, or by social institution) and those which obtain physically as well as through our experience. This relative discrimination of cognition-dependent elements in the objective structures of experience and (in further refinements) beyond the experimentally given—never wholly secure, to be sure, because never exhaustive (caught as we are in time which throws up new structures in direct apperception faster than we can reflexively disengage their pure elements)—is what underlies the possibility of a system of signs containing irreducibly stipulated components demonstrably understood as such by the controlled flexibility human beings display in imaginative discourse and, perhaps especially, fairy tales.

This peculiar capacity for thematizing and critically reassessing under various circumstances the line between reality and unreality, exhibited in the partially controllable indifference of our discourse to what is and what is not the case, is found nowhere else in the animal kingdom, certainly not in the recent “sentences” constructed by our neighbors in evolution, the chimpanzees. It is an activity unique to and in some ways definitive of human understanding. (See Deely 1975; and Notes following.)

5. In 1632, Poinsot, in his Tractatus de signis (642a24-29; 301a1-306b45; esp. 304b11-306a5; 747a33-b2), expressly regarded such an approach as an essential propaedeutic to the doctrine of signs. Among contemporary authors, only Jacques Maritain (1957:87-91), a student of Poinsot’s thought, has made any development of what is involved here, and that only in the brief passages which I cite:

“The birth of ideas and thus of intellectual life in us seems bound up with the discovery of the signifying value of signs. Animals make use of signs without perceiving the relation of signification. To perceive the relation of signification is to have an idea . . . .

“For the first stirring of an idea as distinct from images, the intervention of a sensible sign is necessary. Normally in the development of a child it is necessary that the idea be ‘enacted’ by the senses and lived through before it is born as an idea; it is necessary that the relationship of signification should first be actively exercised in a gesture, a cry, in a sensory sign bound up with the desire that is to be expressed. Knowing this relationship of signification will come later, and this will be to have the idea, even if it is merely implicit, of that which is signified. Animals and children make use of this signification; they do not perceive it. When the child begins to perceive it (then he exploits it, he toys with it, even in the absence of the real need to which it corresponds)—at that moment the idea has emerged.”

“The discovery of language, then, coincides with the discovery of the relation of signification, and this would explain why, as a matter of fact, the invention of language and the birth of ideas, the first release of the intellect’s power, probably took place at the same time.

“It is conceivable, I think, that a genuine language of natural sensory signs may have preceded language strictly so called (made up of conventional sensory signs), and that the latter may have developed out of the former. The ‘miracle’ would have happened at the moment when man, beyond the fact of using natural gestures to express hunger, anger, or fear, would also have grasped the notion that this gesture was possessed of the virtue of signifying. By the same stroke a field of infinite possibilities would have opened. Then, once the relation of signification was discovered, the process of arbitrarily selecting or inventing other gestures and of using them as conventional signs no doubt developed quite rapidly.”

“. . . as a philosopher I wish to . . . emphasize that what defines language is not precisely the use of words, or even of conventional signs; it is the use of any sign whatsoever as involving the knowledge or awareness oj the relation oj signification, and therefore a potential infinity; it is the use of signs in so jar as it manifests that the mind has grasped and brought out the relation of signification.” (Maritain’s italics.)

“. . . In any case, the invention of those particular conventional signs which are words, the creation of a system of signs made up of ‘phonemes’ and ‘morphemes’ was in itself a second ‘miracle,’ a further discovery of human intelligence, no less characteristic of man, but less essential than, and by nature not prior to, the discovery of the relation of signification.

“So far we have spoken of genuine language. Let us point out that the word ‘language,’ when referring to animals, is equivocal. Animals possess a variety of means of communication but no genuine language. I have observed that animals use signs. But, as I also pointed out, no animal knows the relation of signification or uses signs as involving and manifesting an awareness of this relation.

“The full import of this is best realized in connection with the use of conventional signs by animals”—e.g., in the case of bees (von Frisch 1950), or the more indirect but no less intriguing case of the balloon flies (Kessel 1955).

Non-linguistic animals “use signs—and they do not know that there are signs.”

The prospects of such an analysis suggest that the famous “deuxième article” of the Statuts of the Société de Linguistique de Paris, approved in March of 1866 (Mémoires . . . , 1868:3: “La Société n’admet aucune communication concernant . . . l’origine du langage . . . .”), may be principally an admission of the radical inadequacy of presemiotic approaches to the semiotic phenomenon par excellence, human discourse.

6. If one considers the difference between the post-linguistic and the pre-linguistic systems, and the fact that the post-linguistic structures, when they acquire a semiotic dimension, as they invariably do (if indeed they do not have it from the outset), can be understood in the way that is essential to them as post-linguistic only by linguistic animals, then it is also clear that as they redescend below the linguistic interface, as they are experienced by the non-linguistic animals, signs designating them function in a completely different way. The post-linguistic meaning never survives as common to the two levels. The materially same sign designating a post-linguistic structure in those of its elements that are accessible in pure perception pre-linguistically is formally diverse from its designating what is actually proper to those elements as post-linguistic.

Take the case of civil government. Consider the notion of a president—the President of this country. If I were the Premacks, I would teach my chimpanzee each time, say, Richard Nixon walked into the room, to put on the board plastic symbols indicating “Here comes the President.” Now, if I were the Premacks, I would forthwith claim that I had taught Sarah or Washoe or whichever chimp the meaning of the word “president.” But the fact of the matter is that the monkey doesn’t see a president. Indeed, a president, as such, never appears to the eyes. The monkey doesn’t have clue one as to what is referred to by the term “president,” as it designates something distinct and distinguishable from the given concrete individual. When the chimp associates the word “president” with Richard M. Nixon, and when a human being says, “Richard M. Nixon was the President of the United States,” the materially same sound or symbolic marker for “president” is functioning in a completely different sense, once above the linguistic interface as well as below it, the other time below it only. To stipulate the meaning of a word, to teach it to a chimpanzee bv standard associative techniques, and then to assert that the word is functioning in the chimpanzee’s sign system in precisely the stipulated sense is a particular version of “clever Hans” that I would call the fallacy of linguistic anthropomorphism.

7. This would suggest that the complete reduction of culture to social system, as characteristic of British anthropology and of American anthropology since the time of Radcliffe-Brown, is an oversimplification. Earlier American anthropology, with its concept of culture system as “superorganic” and wholly transcendent to social system, went too far in assigning autonomy to the cultural vis-à-vis the social. But the opposite view, that culture is nothing but the uniquely complex forms of social organization proper to man, also goes too far, as a result of an inadequate understanding of semiosis as it occurs in specifically human language. It is true that the unreal relational components of human experience only exist through the cognitive functioning of living individuals, and in this sense the cultural system does have actuality only in and from social interaction. But this “unreal” dimension of experience recognizable as such, and as providing “a substitute for experience which can be passed on [through language] ad infinitum in time and space” precisely because it is cognitively separable from this or that specific concrete individual or group of individuals with whose activity it is here and now—or was there and then—de facto identified, is in itself something distinct from even though immanent within social interaction and social system, and is the ground of the cumulative transmission of learning that makes human society as enculturated different in kind from the animal societies that cannot jump the links of individuals connecting the generations. This is one of the points of view, I would suggest, that reveals most sharply the revolutionary importance of semiotics for anthropology and for clarifying the foundations of the social sciences generally.

REFERENCES

Crystal, D. 1974. Paralinguistics. In Current Trends in Linguistics, vol. 12: Linguistics and adjacent arts and sciences, T. A. Sebeok, ed., 265-95. The Hague: Mouton.

Deely, J. N. 1975. Modern logic, animal psychology, and human discourse. Revue de l’université d’Ottawa 45:80-100.

——. 1978. Toward the origin of semiotic. In Sight, sound, and sense, T. A. Sebeok, ed., 1-30. Bloomington: Indiana University Press.

Huxley, J. S. 1942. Evolution: the modern synthesis. London: George Allen & Unwin.

Kessel, E. L. 1955. The mating activities of balloon flies. Systematic Zoology 4:96-104.

Lotman, J. M. 1977a. Primary and secondary communication-modeling systems. In Soviet semiotics: an anthology, D. P. Lucid, trans, and ed., 95-98. Baltimore: Johns Hopkins University Press.

——. 1977b. The structure of the artistic text, R. Vroon, trans. Ann Arbor: University of Michigan Press.

Lucid, D. P. 1977. Introduction. In Soviet semiotics: an anthology, D. P. Lucid, ed., 1-23. Baltimore: Johns Hopkins University Press.

Maritain, J. 1957. Language and the theory of sign. In Language: an enquiry into its meaning and function, R. N. Anshen, ed., 86-101. New York: Harper & Brothers.

Mémoires de la Société de Linguistique de Paris 1868, vol. 1. Paris: Librairie A. Franck.

Morris, C. 1971. Writings on the general theory of signs. The Hague: Mouton.

Poinsot, J. 1632. Tractatus de signis. In J. Poinsot, Ars logica, B. Reiser, ed. Turin: Marietti, 1930.

Price, H. H. 1950. Perception. 2d ed. London: Methuen.

Sebeok, T. A. 1978. “Talking” with animals: zoosemiotics explained. Animals, December 1978, pp. 20-23, 36.

——. 1979. Looking in the destination for what should have been sought in the source. In The sign and its masters, T. A. Sebeok, ed., 85-106. Austin: The University of Texas Press.

Sebeok, T. A.; Hayes, A. S.; and Bateson, M. C., eds. 1964. Approaches to semiotics. Transactions of the 1962 Indiana Conference on Paralinguistics and Kinesics. The Hague: Mouton.

von Frisch, K. 1950. Bees, their vision, chemical senses, and language. Ithaca, N.Y.: Cornell University Press.

Language and
the Genetic Code

ROBERT B. LEES

Molecular biologists frequently refer to certain features of the cellular nucleus using linguistic terms; they distinguish between familiar physical, chemical, and biological properties on the one hand, and, on the other, linguistic structure or behavior. Some have called attention to similarities between languages and the so-called genetic code. Are these references mere nonce allusions or trivial metaphors, parallel to, for example, the language of flowers?

A few studies have emphasized significant differences between the superficial structure of languages and that of the genetic code. Nevertheless, I shall argue that despite these obvious contrasts, there is after all a deep connection between them—a very abstract and distant connection, the study of which may eventually elucidate fundamental characters not only of our linguistic faculties but even of the mind itself, its origin, and its evolution.

To appreciate the nature of the analogy between language and the genetic code, or more properly, between the mind and the cell, we must extract the relevant details from the better understood of the two, the biological case. I shall begin, therefore, with a whirlwind sketch of the current favorite view of the origin and the evolution of life. I suggest that the mind arose and evolved in parallel fashion, aided by the invention of an internal representation system, language, in much the same way that biological evolution was aided by the invention of the genetic code.

The first significant insight into the nature of life occurred when the Moravian abbot Gregor Mendel, in the 1850s, developed a convincing explanation for certain remarkable cases of inheritance in which the proportion of individuals in a population of progeny bearing a certain trait inherited from their parents is expressible by a fraction with very small integral numerator and denominator. Perhaps Mendel’s most noted case is the one involving a crop of field peas, where each plant is a hybrid cross between a purebred yellow parent and a purebred green parent; all these children are yellow, but when they in turn are hybridized to yield a field of grandchildren peas, ¹/₄ of these descendants are green while ³/₄ are yellow. Mendel explained the appearance of elementary arithmetic in otherwise quite unintelligent peas by postulating that individuals do not inherit their color directly; rather, they inherit abstract “factors” from their parents, one from each. The yellow-pea factor (Y), he explained, is, moreover, “dominant,” and the green-pea factor (g) is “recessive.” Thus, since each hybrid child bears a Y and a g, it is yellow; and since random matings among gY children must yield ¹/₄ gg, ¹/₄ gY, ¹/₄ Yg, and ¹/₄ YY grandchildren, only ¹/₄ of the grandchildren will be green (gg).

This brilliant insight, though simple to us and actually applicable to only a few cases of inheritance, analyzed biological descent into deep structures (factors) and surface structures (traits), or in more modern terms, genes and traits, the so-called genotype and its reflected phenotype. Mendel’s identification of genes (he did not actually use that term) introduced the possibility of conceiving and representing inheritance in terms of concrete chemical entities in a germ cell, passed on from generation to generation.

Later, A. R. Wallace and Charles Darwin each painstakingly contributed to the development of the other cornerstone of biology, an explanation of the evolution of traits; and they did this without benefit of a viable theory of inheritance, for neither had heard of Mendel’s success. Charles Darwin explained the appearance of new species simply by noting that (1) if traits are, by and large, inherited by progeny, (2) if in every generation there are always (for whatever reason) individuals of various types, of greater or lesser degree of diversity, (3) if some types (for whatever reason) leave behind more progeny than others, and (4) if environmental conditions affect the reproductive success of a type, then, inevitably, as environmental conditions change slowly over geologic time, new types will emerge, their concentration in the population will increase at the expense of other reproductively less successful types, and some types which formerly interbred will no longer do so and hence will comprise new species. (Incidentally, to the extent that more complex and less complex types can coexist in slightly different habitats or niches, the more complex types must eventually appear; this, then, solves the puzzle of why evolution has produced increasingly complex organisms, as though it were preprogrammed or directed by an outside agency.)

Nearly everyone has by now heard of the 1953 discovery by J. D. Watson and F. H. C. Crick in London of the stereochemical structure of DNA (deoxyribonucleic acid), for which they received the Nobel Prize in 1962. Most of our children learn in school of the fundamental importance of DNA. Although it was first isolated among the various nucleic acids by Miescher in 1858, it was not until the middle of our century that it became clear to most biologists that DNA bears the secret of inheritance.

There was a very good reason for scientists’ reluctance to acknowledge DNA’s paramount role: it had to prove itself capable of unbelievably delicate discrimination. The operation of the immune system, by which every organism distinguishes between its own proteins and those of others, is an example that clearly illustrates how any substance that is responsible for specifying the constituency of biological types must itself bear a very high degree of specificity. Until recently, proteins were the only substances known to exist in such a profusion of subvarieties in a living organism that they might be the locus of this required specificity. Few researchers, if any, suspected that a substance as simple in gross structure as a nucleic acid could encode such a high degree of biological individuality.

Like proteins, DNA and the closely related RNA (ribonucleic acid), are large polymers, like starch, rubber, and plastics. A polymer is any substance whose molecule is an indefinitely long chain of monomers, each of which is a small molecule capable of attaching at one end to another monomer in a fixed way and at the opposite end to a third monomer in a fixed way. When the monomers are of more than a single species, the chain is a copolymer; when the linear order of the monomers in a mixed chain is not predictable by any fixed rule, the chain is irregular copolymer. DNA is an irregular copolymer of four different monomers, each of which is one of four organic bases attached each to a deoxyribose sugar ring that is itself attached to a phosphoric acid molecule. The bases are adenine (A), thymine (T), cytosine (C), and guanine (G). RNA differs from DNA only in having a molecule of ribose in each monomer as its sugar and in having uracyl (U) as one of its bases in place of T.

What Watson and Crick discovered is that DNA normally consists of two of these long strands intertwined helically, with the phosphates on the outside, the bases stacked along the inside of the tube, and the two helices attached to one another by hydrogen bonds from each of the bases on the one strand to a corresponding base on the other. The monomer bases are so matched in size and in structure as to form two invariant pairs: A and T can easily fit together across the space between the two strands, and C can match G. This is critical for the two functions performed by the DNA (or RNA) in every cell.

The individuality of the DNA borne by a given individual or type inheres in the particular choice and sequence of its bases along either strand. In human DNA, there may be a million bases along a strand. Thus, there are, for all practical purposes, infinitely many different varieties of DNA available to specify an individual—expressed exponentially, at least 4^1000000, or 10^6000000!

Because of the rigid pairing of the bases (A-T and C-G), the two strands of the double helix are exact complements. This is the basis of the surprisingly simple mechanism which underlies self-reproduction. At mitosis, an enzyme-catalyzed chemical reaction unzips the helix, and each of the strands picks up a complement mate from the surrounding soup of component parts and builds a new double helix of its own, each element identical to the original, barring accidental errors (one of the sources of mutation).

The second function of DNA and RNA in the cell is to preside over the synthesis of proteins. An organism consists mainly of proteins; differences among organisms are mainly the result of differences in the proteins that compose them.

A protein is an irregular copolymer of twenty different monomers, each an amino acid linked to the next in the chain by a peptide bond. For an organism to acquire the amino acids for the assembly of its proteins, it must either synthesize them itself or eat another organism that already contains them. Proteins taken in from another organism must usually be restructured by the ingesting organism to ensure that only its own proteins appear. Its cells consult their DNA to determine which strings of amino acids to put together.

This assembly process occurs on organelles in the extranuclear cytoplasm (ribosomes). Therefore, the precious information from the DNA must be transcribed and conveyed out of the nucleus to the ribosomes as needed. That task is performed by short, single strands of RNA. This messenger-RNA (^mRNA), reads a relevant section of DNA by complementation, A-U or C-G, passes outside to a ribosome, and picks up, again by complementation, small pieces of transfer-RNA (^tRNA), each of which carries a particular amino acid; then the ribosome assembles, monomer by monomer, a chain of these accordingly.

We can now define the genetic code as the translation schema for determining in each case which of the twenty amino acids is selected by the information that the ^mRNA has transcribed from the nuclear DNA strand. Since there are only four different bases used for encoding, but twenty different amino acids, clearly the code must involve some combinatory schema with at least three bases per translated acid (two bases can distinguish only 4² = 16; three can separate 4³ = 64). Despite several ingenious theories designed to match sixty-four to twenty, clever experiments eventually determined that each translation is indeed specified by a triplet of DNA monomers, and the code is therefore quite degenerate, although the resulting redundancy is, of course, a welcome antinoise resource. Three amino acids are each encoded by six different triplets, five by four triplets each, one by three triplets, nine by two triplets each, and two amino acids by a single triplet each. That makes sixty-one triplets for twenty amino acids; the extra three triplets serve as full-stop punctuation marks!

This superficial linguistic analogy can be carried still a bit further. A triplet of successive nucleotides (DNA monomers) on the DNA strand plays a role like that of a word. Each is one of a finite set of minimal functional units that bears a symbolic significance: a word (in most cases) has a meaning, a triplet specifies a particular amino acid. Thus, a nucleotide is analogous to a phoneme. The stretch of DNA transcribed by ^mRNA at any one reading is translated into a protein and is the analog of a sentence; there are an infinite number of them, and it requires higher-order information not connected to the transcription and translation mechanisms to determine which stretch is to be read. A sentence is chosen because of the function of its meaning in a setting, and a stretch of DNA is read because of the function of its protein in the cell. A triplet is called a codon; the stretch of DNA that is read is called a cistron. For a sizable family of cases, a cistron is an appropriate interpretation for a gene or for one of Mendel’s factors. Finally, the punctuation mark which flags the point along the DNA strand where a cistron begins functions like the capital letter that starts a sentence.

Recent research reveals that reading is initiated and terminated in a much more complicated manner than is implied by the simplicity of the local mechanisms. The complicated gene-expression system that determines when a certain gene is read is not itself a simple matter of reading off a message; rather, it involves the whole complex chemical control system of the cell and all of its external environmental influences. Similarly, the choice of a certain sentence, or even of a certain word, is hardly ever grammatically specified, but instead involves the whole mental life of the speaker.

It has also been known for some time that traits are, for the most part, not determined by single genes, just as the expression of a certain idea may require the formulation of several different sentences. It can even happen in some cases that a given stretch of nucleotides functions in the transcription of parts of two different genes. Analogously, a word or a string of words may be structurally ambiguous, that is, capable of functioning in two different constructions or sentence types. For example, in one reading of the sentence “He left out the back,” the word out is in the direction adverb, as in “He left by the back door” (departed thence); in a second reading, the word out is in the verb, as in “He left out one word” (omitted it). Thus, we have crude analogs of syntax, ambiguity, and homophony, all normal features of any linear encoding of information which functions in a complex system of higher order.

The genetic code is universal. Except for a few minor variants, every living creature on earth employs the code described above—not merely the overall schema, but even the translations of individual codons in the code book. So far as we can tell at present, the translations are largely unmotivated from a chemical, physical, or biological point of view, in exactly the sense that words are largely arbitrary from a semantic point of view; that is, one could have made oneself just as clear in English had the verb buy meant ‘sell’ and sell ‘buy’, but so too could our genetic code have fulfilled its function had the codons AAU and AAC translated each to the amino acid glutamine, while CAA and CAG translated each to asparagine, instead of the other way round. This remarkable fact bears witness unimpeachably to our kinship with every other type of organism on this planet.

The boundaries between life and death have become much clearer in this light; in fact, that puzzle no longer torments us as it once did, for nothing of great importance hangs any longer on our choice of exactly where to draw the boundary, except for the legal problem of death in humans. A virus is best viewed as a nonliving piece of genetic-message DNA (or RNA) with some associated proteins that requires a living-cell environment for self-reproduction. This requirement highlights the significant fact that the living cell, and hence every living creature, is a strictly self-referent system: the genetic message which specifies the constituency of a certain cell is read and translated only within that cell itself (though it could in principle, of course, be read and translated within any living cell).

A cell, then, is a bag of chemicals, mostly structural and control proteins. These are organized into various organelles and structures which interact with one another in an immensely complex way to absorb available nutrients and energy and to grow to the point of splitting into two daughter cells. The various reactions are, for the most part, mediated by control proteins, or enzymes, whose concentrations in different places wax and wane.

Enzymes are proteins, or polypeptides, seen from the point of view of their remarkable catalytic function. An enzyme is synthesized as a long strand of linearly linked amino acids; once it is synthesized, it forthwith coils up almost like a ball of string. The cross-links which fold it into its quasi-spherical shape are specific to each protein, and that protein folds in the same way on every occasion. The resulting ball has special regions on its surface, complex cavities among the twisted and intertwined strands, where certain smaller molecules can fit. Chemical reactions among these molecules are activated by the electrochemical effects of the surrounding coils of the enzyme substrate.

The strands of DNA in a cell are also accompanied by a certain protein cover, mainly histones, which, together with the more mobile enzymes, comprise a control system that opens up and closes down certain cistrons, or genes, at various times.

Thus, the cell is a complex cojitrol system of interacting elements, a dynamic microcosm of chemical structures. When these are interacting, we say the cell is alive (or, more obscurely, we attribute to the cell the reification ‘life’), just as we refer to an automobile engine as on or off. By obstructing a critical reaction, or a sufficiently large number of noncritical ones, we may bring the process to a halt; for cell or engine we may say that we have killed it, that it has died.

Of course, there is an interesting difference between cells and engines, but it has nothing to do with the character of life itself. Engines do not reproduce themselves. However, there is no good reason why engines could not be constructed so that they could propel automobiles and also occasionally assemble new engines from spare parts; the technological problem is indeed formidable, but it is not in principle unsolvable. At present, however, each engine must be assembled by a man or a factory, while cells are born of parent cells (i.e., assembled from raw materials by a parent of a sort).

This fact often confuses one into supposing that the first cell must have had its life inspirited by a prime blower. This misconception arises from several failures in judgment. First, it is not the cell now before us whose spontaneous assembly must be explained, but rather the quite primitive Ur-cell of long ago. Second, the nondegenerating assembly that became the Ur-cell arose by chance, the survivor of a billion or so years of chemical recombination in sterile seas. The cell we see today is the offspring of literally countless subsequent evolutionary interactions.

Third, there is an important difference between intelligence of self-consciousness in complex animals and life in cells. The elements of a higher nervous system are themselves cells, already very smart; the elements of a cell are chemical molecules, by comparison quite stupid. Cells are very well understood in principle, no matter how complicated an actual cell may be. The operation of the complex nervous control system that underlies self-consciousness or creativity in primates is barely understood at all. The role played by the genetic code in organizing and maintaining the integrity of a cell and its genetic line is well understood; the role played by the mind’s representational systems in organizing and maintaining the integrity of the personality or of an intellect is barely understood at all.

Yet the analogy between these two levels is unmistakable. On at least two separate occasions in the history of our corner of the universe, a new kind of complex control system of interacting elements arose spontaneously to generate a self-contained, homeostatic, evolving organism. The first, the biological world of life, arose on a substrate of chemical interactions, and in time it invented a genetic code. The second, the mental world of the intellect, arose on a substrate of nervous interactions in the brains of higher species, and in time it invented a linguistic code.

Others than myself have entertained such an analogy, notably the eminent contemporary philosopher Sir Karl R. Popper. He considered the traditional mind-body problem and, at least for the time being, accepted a trialist view of three hierarchically interconnected but, in a sense, independent worlds: first, the physical universe of molecules, cells, and brains; and parasitic on it a second domain of mental events—thoughts, intentions, and emotions; and generated from this, in turn, a third world of the products of the mind—the world of theorems, proofs, sciences, philosophies, and religions.

It is not implausible that mental events are just physical events, albeit complex ones, in the nervous system. Such a resolution of the traditional mind-body problem is not a simple-minded physicalist reduction of mental events to physical objects. An event in the nervous system may be a very much more sophisticated and abstract notion than that of an object or a substance; and it may be quite some time before we can specify or predict the nature or occurrence of such complex interactions among countless numbers of axons, synapses, or neural nets; but it no longer seems mysterious to us that a higher organism experiences mental events when, for example, one part of its hrain scans activities in another, just as a sense organ scans some external input of auditory or visual or tactual signals.

However, the independent world of products of the mind specified by Popper, if it is a viable notion, does seem quite puzzling. In the sense of the foregoing paragraph, it is not at all clear how we should conceptualize the relation between physical objects or mental events and the content of a thought. Once an idea has been formulated, and maybe also represented, it can be rendered free from its creator and developed, or used as the basis of deductions, or incorporated into a theory, or refuted, and so on. In any case, it seems incongruous to view an idea as a mental event, even if it happened to originate in a mental event, or even if every idea must originate in mental events.

It is not difficult now to see a language as a device for representing ideas. It differs from other representation schemata used by men and animals just by the fact that it permits the representation of propositions, and it is to these propositions that truth-values may be attached and from these that arguments and proofs may be constructed. Primitive instances of argument may be possible for a linguistically innocent organism, just as primitive arithmetic conceptualization is exhibited by creatures who possess no numerals, but the advantages of sharing ideas can be achieved only with propositions represented concretely, either auditorily or visually, in a common mode. Moreover, linguistic encoding of meanings serve as, perhaps, an indispensable memory aid. Complex meanings are compressed into single words, succinct compounds or short phrases, and these can be stored not only semantically, but also phonologically or visually, and thus also permanently.

Let me summarize briefly here. Life arose spontaneously when certain chemical control systems crossed some boundary of complexity to become nondegenerating; organisms then further improved their ability to survive changes in the environment by inventing a finely tuned evolution system using simple self-reproduction and growth mechanisms which were based, in part, on a genetic code. An individual is delicate, and can be instructed by its environment to change only within narrow limits; but a species is rugged, and changes over long periods of time not by instruction, but by selection among independently occurring variants.

I advocate that linguistic competence be viewed analogously to the genetic code as a mechanism invented by minds to serve as a scratch pad for logic and a repository of ideas. Each mind dies with its substrate brain, but ideas, once shared or recorded, live on and evolve to nourish future minds. We should seek to formulate the basic principles that govern the appearance and evolution of complex control systems such as cells and central nervous systems; we may yet uncover the origin of and the motivations for language.