FIVE

In What Sense Is Language
a “Primary Modeling System”?

The expression “primary modeling system”—coupled, as a rule, with the contrastive concept “secondary modeling system,” which emphasizes its derivational character in relation to natural language—has been central to Soviet semiotics of the Moscow-Tartu school since 1962, when it was proposed by A. A. Zaliznjak, V. V. Ivanov, and V. N. Toporov (in English in Lucid 1977:47-58; see also Rudy 1986, and Shukman in Sebeok 1986b:I. 166-168, 558-560).

In 1974 (Sebeok 1985:23n38), I interpreted the inferred concept—having checked my provisional understanding, when I gave a lecture at the University of Tartu in August 1970, with Ivanov—as follows:

The notion of a secondary modeling system, in the broad sense, refers to an ideological model of the world where the environment stands in reciprocal relationship with some other system, such as an individual organism, a collectivity, a computer, or the like, and where its reflection functions as a control of this system’s total mode of communication. A model of the world thus constitutes a program for the behavior of the individual, the collectivity, the machine, etc., since it defines its choice of operations, as well as the rules and motivations underlying them. A model of the world can be actualized in the various forms of human behavior and its products, including linguistic texts—hence the emphasis on the verbal arts—social institutions, movements of civilization, and so forth.

Although Ivanov graciously acquiesced at the time in my ad hoc formulation, this, in retrospect, seems to me to require still further elucidation. Accordingly, the purpose of the remarks that follow is to provide amplification and clarification.

The canonical delineation of modeling systems was framed by Ju. M. Lotman in 1967 (rendered in English in Lucid 1977:7) as “a structure of elements and of rules for combining them that is in a state of fixed analogy to the entire sphere of an object of knowledge, insight or regulation. Therefore a modeling system can be regarded as a language. Systems that have a natural language as their basis and that acquire supplementary superstructures, thus creating languages of a second level, can appropriately be called secondary modeling systems.” Natural language, in brief, is thus posited as the primary, or basic, infrastructure for all other (human) sign systems; and the latter—such as myth or religion—are held to be resultant superstructures constructed upon the former. In 1971, Lotman and Uspensky (in English, 1978) elaborated their view of the semiotic study of culture, noting that, in their scheme, language is viewed as carrying out a specific communicative function by providing the collective with a presumption of communicability.

An underlying question concerns, more generally, the concept of “model”—which is essentially a reductive analogy, and therefore ultimately a kind of icon—and its applications, if any, as a technical term in semiotics of the nonverbal and of the verbal in particular. Certainly, it is a fashionable appellation in the literature and philosophy of science, where it has, however, acquired many different connotations. Some of the more important of these—notably in logic, mathematics, and physics (but not in biology, more of which below)—are stimulatingly discussed by Hesse (1967).

The only recorded discussion of models in linguistics that I am aware of took place at the 1960 International Congress for Logic, Methodology and Philosophy of Science, with the participation (among others) of Bar-Hillel and Chomsky. The proceedings include a highly useful, although neglected, paper by Yuen Ren Chao, who correctly notes that, while “the term ‘model’ is relatively new in linguistics . . . the use of what may reasonably be regarded as models is as old as the study of language” (Chao 1962:558; for later references, see Welte 1974:1.386-387, and Stewart 1976; also cf. Koch 1986). Chao claims that the earliest mention of models in (American?) linguistics was in 1944 by Z. S. Harris. However this may have been, the term was thereafter used with increasing frequency, yet in a bewildering variety of senses: Chao lists no less than thirty synonyms or more or less equivalent phrases of “model” for the fourteen years he had surveyed. But none of these seems to conform to, or possess the scope of, the uses of “model” in the Soviet tradition.

Some twentieth-century pre-Chao and post-Chao models of semiosis—or aspects of semiosis—are illustrated by the graphic displays shown here, a modest sample chosen almost at random from a far larger number (cf. Fiske 1982, Ch. 2, for Gerbner’s, Newcomb’s, and Westley and MacLean’s models, among others). It should also be noted that these models are all, more or less, linked intertextually among one another, namely, their framers were aware of earlier models and their interpretations of these models were repositioned in the light of each later model.

The “convenient diagram of Symbol, Reference and Referent” (Fig. 5.1) was contrived, in the 1920s, by Ogden and Richards (1938:11). In Europe, the “organon model” of language (Fig. 5.2) by Bühler (1965:28) became widely influential after the mid-1930s. Shannon and Weaver’s (1949:5) schematic flowchart (Fig. 5.3), representing a general communication system, has become a classic that keeps being copied, since the early 1950s, with all sorts of variations, since it is heuristically so valuable and because it suggests ways of expanding the theory embedded in it. In the early 1960s, Sebeok (1972:14) tried to depict by way of a Morley Triangle the relationships between Bühler’s model and Jakobson’s (orginally, 1960:253, 257) more comprehensive information-theoretical schema of six constitutive factors, each of which is posited to determine a different function of language; this was, in turn, actuated by the Shannon and Weaver model (Fig. 5.4).

Chao does not press his own views, but it is clear that, had he developed them, they would have mirrored common semiotic principles by changing their parity. What he does say (1962:564) is that, in his model of models, “there are things and models of things, the latter being also things but used in a special way.” One would now rather say that there are objects and signs of objects, the former also being signs but used in a special way (cf. Ch. 1, above).

Chao then gives this example: “If we take any two things, say cabbages and kings, and make, say, a cabbage the model of a king, there is not likely to be much that is true of one that is also true of the other, though usually not zero, e.g. both are living things or can be, etc., but the modelity of cabbages with respect to kings is fairly low” (1962:564).

This can be rephrased in standard semiotic idiom in this way: a cabbage (aliquid) stands for (stat pro) a king (aliquo). If it is likely that much of what is true of one (that is, of the sign “cabbage”) is also true of the other (that is, of the object “king”), then perhaps one might amplify, with Peirce (1935-1966:2.257), that the cabbage tends to be a Dicent Sinsign, involving both “an Iconic Sinsign to embody the information and a Rhematic Indexical Sinsign to indicate the Object to which the information refers.” However, if very little is true of one that is also true of the other (even though it isn’t entirely zero), one might say, again with Peirce (ibid.:2.261), that the cabbage tends to be a Rhematic Symbol or a Symbolic Rheme, such as a common noun. In Jakobson’s much simplified version of semiosis (1980:11, 22), a model M, a cabbage, could be said to function as a renvoi to the thing T, a king, and this referral could, by virtue of an effective similarity, be iconic (after all, as Morris [1971:273] taught us, “Iconicity is . . . a matter of degree”). On the other hand, by virtue of an imputed, conventional, habitual contiguity, the referral could be symbolic, much as, for the experimental dog in the Pavlovian paradigm, the sound of a metronome became an arbitrarily paired symbol (that is, a conditioned reflex) for dry food.

Fig. 5.1

Fig. 5.2

Fig. 5.3

Fig. 5.4

Soviet conceptions of models and modeling systems clearly owe much to Jakob von Uexküll’s theory of meaning (1982; cf. Gipper 1963, Ch. 6, and Sebeok 1979a, Ch. 10), developed, in Hamburg during the first four decades of this century, by this great biologist in a series of sagacious, if quirky, contributions to semiotics. Stepanov (1971:27-32), for instance, singles him out for extended mention in the course of his sketch of (then) current trends in modern (bio)semiotics.

Jakob von Uexküll’s highly original Umwelt-Forschung—which its creator viewed as a scientific theory anchored in Kant’s a priori intuitions—is truly a fundamental theory, as much of sign processes (or semiosis) as of vital functions. Moreover, his conception at once utilizes a pivotal model—the famous “functional cycle,” this simple, albeit not linear, diagram (Fig. 5.5) by which, as Lorenz (1971:274) noted, “a vast programme of research is implied”—and in itself constitutes a cybernetic theory of modeling so fundamental that the evolotion of language, as I have argued elsewhere (Sebeok 1985c, 1986c), cannot be grasped without it.

The term Umwelt has proved notoriously recalcitrant to translation, although “subjective universe,” “phenomenal world,” and “self-world” variously approximate Jakob von Uexküll’s intent. However, “model” renders it more incisively, especially in view of his credo (1982:87) that “every subject is the constructor of its Umwelt.”

As Jacob (1982:55) later explained with utmost clarity: “Every organism is so equipped as to obtain a certain perception of the outer world. Each species thus lives in its own unique sensory world, to which other species may be partially or totally blind. . . . What an organism detects in its environment is always but a part of what is around. And this part differs according to the organism.” The-world-as-perceived depends crucially on each organism’s total sensorium and on the way its brain integrates sensory with motor events. But the inclusive behavioral resources of any organism must be reasonably aligned with its model of “reality” (Natur), that is, the system of signs its nervous system is capable of assembling—or it will surely be doomed, by natural selection, to extinction.

Fig. 5.5

Schneirla’s biphasic approach-withdawal theory (1965) furnishes a minimal model that must have been crucial for the survival of all animal types, from protozoans to primates (including humans). Such a miniature model—or “modelita,” in Chao’s (1962:565) sobriquet—evidently requires much the same organs but is played out in two functionally opposed systems, one for the reaching of food and mates, the other for the evasion of noxious situations. A key postulate of this holistic oppositive A-W theory, allowing as it does for plasticity through experience, is that it cyclically relates every organism’s Innenwelt (or inner world, “comprising” as Lorenz explains [1971:275], “the whole of its bodily structures and/or functions”) to its charactistic habitat (Umgebung, or observer’s Umwelt; after Jakob von Uexküll 1909).

The Innenwelt of every animal comprises a model—whether of a minimal A-W type or of a more eleborate kind—that is made up of an elementary array of several types of nonverbal signs (variously elaborated by Jakob von Uexküll [1982:10-11] under such labels as Ordnungszeichen, Inhaltszeichen, Lokalzeichen, Richtungsqzeichen, Wirkzeichen, and the like, none of which can be discussed here). Solely in the genus Homo have verbal signs emerged. To put it in another way, only hominids possess two mutually sustaining repertoires of signs, the zoosemiotic nonverbal, plus, superimposed, the anthroposemiotic verbal. The latter is the modeling system the Soviet scholars call primary but which, in truth, is phylogenetically as well as ontogenetically secondary to the nonverbal; and, therefore, what they call “secondary” is actually a further, tertiary augmentation of the former. The congruity of this expanded paradigm with Karl R. Popper’s famous Worlds 1-2-3 model (Eccles 1979, Lecture 6; Sebeok 1979a:204-205) is unmistakable: his World 3 is the World of Culture; his World 2, “the other uniquely human world” (Eccles 1979:115-116), explicitly encompasses language and develops together with the former “in some kind of symbiotic interaction”; and his World 1 is the whole material world of the cosmos, both inorganic and organic, including machines all of biololgy.

The earliest known species in the genus Homo is the form Louis Leakey named habilis, first described in 1964 and now usually regarded as a short-lived, transitional African form of some two million years ago, ancestral to all later hominid species. With a brain capacity of 600-800 cubic centimeters (cc), this ancestral creature must have had a mute verbal modeling device lodged in its brain (Sebeok 1986c), but it could not yet encode it in articulate, linear speech. Language is, in fact, among its quintessential taxonomic markers (in conjunction with chipped pebbles and clusters of animal bone that evince deliberate cutting and breaking).

The evolutionary success of habilis is corrobrated by the very swift appearance, a mere half-million years later, of the succeeding species, H. erectus, with a brain volume of 800-1,200 cc; this speedy attainment is undoubtedly due to the species’ linguistic competence, also indirectly manifested by its possession of tool kits exhibiting standardized design, the use of fire, and its rapid global dispersion.

Starting about 300,000 years ago, an archaic form of H. sapiens evolved out of the erectus species, with a growth of skull capacity up to 1,400 cc, and many concurrent novelties. It is reasonable to conclude that this premodern human already had the capacity to encode language into speech and the concomitant ability to decode it at the other end of the communication loop. H. sapiens sapiens appeared a mere 40,000 years ago, with our brains averaging 1,500 cc.

The cardinal points in this brief scenario are twofold: language evolved as an adaptation; whereas speech developed out of language as a derivative exaptation over a succeeding period of approximately two million years. These twin propositions need to be made plain with reference to a suggestion by Gould and Vrba (1982). These authors emphasize the distinction between historical genesis and current utility, suggesting that characters that evolved for other usages (or none) may later come to be co-opted for their current role. The former operation is customarily called adaptation; for the latter, they propose a new designation, exaptation.

Accordingly, language—consisting of a set of features that promotes fitness—had best be thought of as having been built by selection for the cognitive function of modeling, and, as the philosopher Popper and the linguist Chomsky have likewise insisted (see also Ch. 7, below), not at all for the message-swapping function of communication. The latter was routinely carried on by nonverbal means, as in all animals, and as it continues to be in the context of most human interactions today.

Several million years later, language, however, came to be “exapted” for communication, first in the form of speech (and later of script, and so forth). This relatively brief elapsed time was required for a plausible mutual adjustment of the encoding with the decoding capacity, but, since absolute mutual comprehension remains a distant goal, the system continues to be fine-tuned and tinkered with still. Gould and Vrba give many interesting examples of comparable biological processes, stressing that current utility carries no automatic implication about historical origin, and concluding with the empirical observation that “most of what the brain now does to enhance our survival lies in the domain of exaptation” (1982:13). The common flaw in much evolutionary reasoning—“the inference of historical genesis from current utility”—has egregiously contaminated virtually all researches, in the nineteenth century and even quite recently, into the problem of the origin of language, which researches have therefore proved intractable to most probes based on such unbiological principles.

It is interesting that in the other universal domain of human modeling (which cannot be discussed here) where nonverbal—or, as Bullowa (1979:9-10) termed it, “extra-verbal”—communication clearly has exclusive primacy over language, to wit, in ontogenesis, the identical marring feature “has delayed the study of the earliest human communication,” namely, “our habit of thinking of communication consisting mainly of language.”

As Peirce (1935-1966:1.538) taught us, “Every thought is a sign,” but as he also wrote (ibid.:5.551), “Not only is thought in the organic world, but it develops there.” Every mental model is, of course, also a sign; and not only is modeling an indispensable characteristic of the human world, but also it permeates the entire organic world, where, indeed, it developed. The animals’ milieu extérieur and milieu intérieur, as well as the feedback links between them, are created and sustained by such models. A model in this general sense is a semiotic production with carefully stated assumptions and rules for biological and logical operations.

This is as true of bees (Peirce 1935-1966:5.551) as it is, on a far vaster scale, of Isaac Newton’s and Albert Einstein’s grand models of the universe. Einstein, for one, it will be recalled, constructed his model from nonverbal signs, “of visual and some of muscular type,” and labored long and hard “only in a secondary stage” to transmute this creation into “conventional words and other signs,” so that he could communicate it to others. “The words or the language, as they are written or spoken,” Einstein wrote in a letter to Hadamard (1945:142-143), “do not seem to play any role in my mechanism of thought. The psychical entities which seem to serve as elements in thought are certain signs and more or less clear images which can be ‘voluntarily’ reproduced and combined.”

The relatively simple, nonverbal models that animals live by and that normal human infants likewise employ are more or less pliable representations that, as we saw, must fit “reality” sufficiently to tend to secure their survival in their ecological niche (an ethological expression that in semiotic parlance, refers to the Umwelt as viewed by an observer of the subject under scrutiny). Such “top-down” modeling (to use a current jargon borrowed from the cognitive sciences) can persist, and become very sophisticated indeed in the adult life of exceptionally gifted individuals, as borne out by Einstein’s testimonial or by what we know about Mozart’s or Picasso’s ability to model intricate auditory or visual compositions in their heads in anticipation of transcribing this onto paper or canvas. This kind of nonverbal modeling is indeed primary, in both a phylogenetic and an ontogenetic sense.

Language itself is, properly speaking, a secondary modeling system, by virtue of the all-but-singular fact that it incorporates a syntactic component (for there are, so far as we know, no other in zoosemiotic systems, although this feature does abound in endosemiotic systems, such as the genetic code, the immune code, the metabolic code, and the neural code). Syntax makes it possible for hominids not only to represent immediate “reality” (in the sense discussed above) but also, uniquely among animals, to frame an indefinite number of possible worlds (in the sense of Leibniz).

Thus is the human being able to fabricate tertiary modeling systems of the sort Bonner (1980:186), for instance, calls “true culture,” requiring “a system of representing all the subtleties of language,” in contrast to “nonhuman culture,” and thereby produce what the Moscow-Tartu group has traditionally been calling a “secondary modeling system.” It is on this level, redefined now as tertiary, that nonverbal and verbal sign assemblages blend together in the most creative modeling that nature has thus far evolved.

________

This chapter was originally prepared for the Proceedings of the 25th Symposium of the Moscow-Tartu School of Semiotics (Imatra, Finland, July 27-29, 1987). It first appeared in the Proceedings volume, Semiotics of Culture, ed. Henri Broms and Rebecca Kaufmann (Helsinki: Arator, 1988). Versions later appeared in the People’s Republic of China (in Chinese), Hungary (in Hungarian), and Portugal (in French).