CULTURE AS METAPHOR: A NEW LOOK AT LANGUAGE AND CULTURE IN THE PLEISTOCENE Mary LeCron Foster The stones and bones of archaeological research into the ways of early hominids throw little light on causation. Despite increased insight into living habits and technology, the w of nascent culture remains unexplained. We must tap new sources if our understanding of this revolutionarily crucial period is to increase. Near the turn of the century, Durkheim and Mauss (1963) pro- posed that culture was a system of classification based on social forms. Although they were concerned with primitive classification, they recognized no break in continuity between that and the scien- tific classifications of civilized man: both were based on hierarchical systematization. Soon after the period in which they wrote, new analytical methods came into use in linguistics. These shed new light on the classificational structures of language. Several decades later, generative-transformational theory began to expand the scope of linguistic analyses and unexpected degrees and means of hierarchical layering were uncovered. Meanwhile, guided to considerable extent by the structural explorations of language, anthropologists began to expand our understanding of cultural classification. A similar methodology has been common to these approaches, from Durkheim and Mauss onward. While it is usually called structuralism, comparative structuralism might be a better term, since it involves either inter- or intra-cultural comparison in the search for the analogical bases of classification. Some- times, but not always, this involves a time dimension. The very fact of classification implies a recognition of similarity of items that are grouped together. Higher (or deeper) levels of classification imply still further degrees of likeness, not just between single items, but between structures themselves. Cultural classification is almost never explicit and those who manipulate cultural classes are rarely consciously aware of their existence. Metaphor is a useful cover term for the unspecified analogical equations of culture, since the analogies of metaphor are not specified. Those who have called culture "arbitrary," such as Geertz (1964) or Holloway (1969), or Saussure (1959), who spoke of the arbitrariness of the linguistic sign, seem to be arguing that culture (or language) is iconically liberated, which is, of course, far from the case. Nor is it arbitrary in a historical sense, since 2 it is inherited rather than invented, and is modified more or less accidentally, through use, rather than deliberately, through design. Causality is usually provided with a catalyst such as economic pressure, or the desire to increase procreative advantage. It is hard to see how these rational concerns could lead early man to devise hierarchical classification. If it has turned out to be adaptive, it surely was not so perceived at its inception. Durkheim and Mauss ascribed causality to social groupings and, in the Intro- duction to his translation of their essay, Needham points out flaws in their reasoning. However flawed their causal argument may be, I believe that the causes of culture were, indeed, social and, through social interac- tion, emotional. Durkheim and Mauss were astute enough to recognize emotion as a dominant characteristic of classification, although, as Needham also points out, they failed to integrate this important notion with the rest of their argument. It is precisely social and emotional imperatives that serve to tie culture back to nature from which it emerged. However, the model that we must devise to bring us back to this point must build on analogy, since analogy is the name of the cultural game. Analogic models for cultural analysis can be dangerous if not provided with sufficient controls; for, in the last analysis, it is probably possible to relate anything analogically to anything else, as critics of structuralism sometimes suggest. The best safeguard is that of redundancy, for cultural analogies tend to spread across modalities, replicating themselves in an open-ended fashion. A familiar example is Kuhn's (1962) scientific paradigm, which holds intra-culturally for any kind of science. On another hierarchical level is Dorothy Lee's (1959) model of lineality in American cul- ture. To borrow Jim Deetz' term, this provides a "mental template" that molds every aspect of American life (1979). The degree of analogical replicability depends on the hierarchical level of the analogical equation. Lee's is at a higher (or "deeper") level than Kuhn's. Culture as metaphor must have provided emotional gratification for its originators and must continue to provide that gratification today, since it has persisted. Natural sources of gratification derive in infancy from interaction with caretakers who provide for basic needs, such as food, tactile stimulation including eye con- tact, protection from harm, and social interchanges of many kinds. In adulthood, sexual gratification is added to these. Freud's developmental theories show that during maturation, gratification becomes integrated in accordance with individual life experiences. Dreams and fantasies provide a means for metaphoric recapitulation of experience. A further affective source of gratification is play. Primate play involves tactile experimentation with the environment and its resources, and pretense. Young apes, monkeys, human children and 3 many other mammals engage in mock battles. In adulthood this may be transformed into threats that are not carried through, or a variety of other deceptional strategies. Deception is not uniquely human, but in human beings it takes a novel form: self-deception, or fantasy. By about 18 months to two years of age, human children are actively fantasizing and acting out their fantasies in play. In contrast, as ape children grow older they become increasingly pragmatic: their play leads to practical ends. It becomes goal oriented in response to the exigencies of the immediate situation. For human beings, play continues through adulthood as fantasy production of many kinds, and becomes an end in itself. Culture, as metaphoric classification, could not exist without the human ability to fantasize. What is uniquely human is not rationality--as con- scious goal orientation or secondary process thinking, but the metaphoric capacity of primary process thought. Freud coined the primary-secondary process distinction, but the division of thought into two categories--one of logic, the other of metaphor--has a long history. Logic tends to be valued, metaphor is either devalued or suspect. Freud and his followers believe that primary process thinking is the more archaic of the two. As stated by Bateson (1972:140), secondary process . . . talks about things or persons, and attaches predicates to the specific things or persons which have been mentioned. In primary process, the things or persons are usually not identified, and the focus of the discourse is upon the rela- tionships which are asserted to obtain between them. This is really only another way of saying that the discourse of primary process is metaphoric. A metaphor retains unchanged the relationships which it 'illustrates' while substituting other things or persons for the relata. In a simile, the fact that a metaphor is being used is marked by the insertion of the words 'as if' or 'like.' In pri- mary process (as in art) there are no markers to indicate to the conscious mind that the message material is metaphoric. Both Bateson and Freud stress the pathological aspects of this thought mode, but recognize that it is also, to at least some extent, characteristic of normal individuals. The gestural iconi- city exhibited by primates indicates to Bateson that primary process thoughts and their communication are, "in an evolutionary sense, more archaic than the more conscious operations of language, etc." (Ibid., 141). Bateson fails to recognize that gestural iconicity is something new in the mammalian behavioral repertoire, that it becomes greatly expanded in human behavior, that language is far from a conscious process, and that man is the only creature whose secondary process operations are thoroughly subject to primary process conditioning. 4 For Hewes (1973) and others, the iconicity of ape gesture has become the basis of the argument in favor of a gestural origin of language. Despite the fact that Premack taught his chimp, Sarah, to communicate by manipulating non-iconic signs, he, too, argues for an iconic origin for phonemes (Premack 1972). One facet of his argu- ment is that chimps must be doing a feature analysis in associating together, and representing by a single sign vehicle, objects that are like. This likeness is quite concretely referential in the sense that objects represented by the sign for 'apple' are very similar one to the next. Chimps can recognize qualities such as 'brown' and this is more abstract, but still one assumes that one brown object with which they are presented is of a very similar shade to another. Premack's experiments with dogs indicate that they can classify in much the same way. Yet dogs exhibit little behavioral iconicity. Like apes, they are goal oriented and use their classificational talents only in the pursuit of short-term goals. Unlike Hewes and Premack, and despite his awareness of the iconicity of primate gesture, Bateson does not believe that it gave rise to language. He says that, if it had, it would have replaced gesture, whereas, in fact, gesture has persisted and even increased as one form of human communication. This is an effective argument but can be quite easily countered. When oral gesture became iconi- cally referential as incipient language, it separated from bodily gesture, which continued to signal emotional reaction to messages and events including messages delivered by the speaker himself. Intonation and stress, as paralinguistic phenomena, also serve this purpose and are so separable from the language that they accompany that a speaker may deliver quite different messages through each channel. Manual gesture also took two paths: one the earlier, signalling mode, the other a creation of metaphoric systems through technology. Language and technology thus became complementary modes in the interpretation of reality and the delivery of culture. Both language and technology are spatial modes, although lan- guage is less obviously so, since, until the invention of writing, and later sound recording, its spatially produced units vanished as soon as they were uttered. Since both language and technology stem from gesture, but part company from it, they seem more 'rational' and less iconic than they, in fact, are. To examine their early beginnings, we turn now to the Plio- Pleistocene, where stone tool production and, presumably, language made their early, crude beginnings over two million years ago. Isaac (1978:321) provides us with an incipiently cultural, socio- economic design for living from this period. Because stone tools are found in association with concentrations of animal bones, he assumes that these early hominids operated from some sort of "home base" from which they departed to forage for food, and to which they returned to share the products of this search with other members of their small group. Stone tools imply hunting, which, in turn, implies a fairly wide operational range. Extrapolating from the 5 evidence of modern hunter-gatherers, the assumption is easily made that males were the wider-ranging hunters and females the more narrowly ranging gatherers, restricted to some extent by the need to care for small children. This socioeconomic pattern is binary, both as to gender and to division of labor with each economic mode necessitating its own tool type. We have become aware, from the work of Levi-Strauss and others too numerous to mention, that binary organization is charac- teristic of cultural models. In Isaac's scheme, at one polar extreme are adult males, stone tools, animals for food and greater geographical distancing. At the other we have adult females, some sort of (assumed) perishable containing tools, vegetable food, children and narrower geographic range. One would assume that this narrower range associates females more closely with "home base"--an association that we find universally to this day, just as we also have women in close association with containers and men with killing and large cutting tools. At an earlier period than the one discussed by Isaac, stone tools were not created through flaking; they are only recognizable as tools because of signs of use. At the period under discussion stone tools were of two types: core tools, deliberately created by pounding to strike off flakes, and the flakes, which were not deliberately shaped, but were used for cutting. The binary oppositions which this model presents make it look very much like a system of classification (Figure 1). Hunting and Scavenging . . Economic Activity . . Gathering Animal Food . . . . . . . Food Sharing . . . Vegetable Food Men .Home Base . . . . . . Women + Children Flaked Core Tools . . . Flake Tools . . . . . Containing Tools? Stone Tool-Making . . . - <- . . . . Child Bearing Figure 1. Plio-Pleistocene Socio-Economic Structure (Adapted from Isaac, 1978, p. 321) Analogical justification for the model is not difficult to find. Biologically, adult males are thrusters and penetrators, while adult females are receptors and containers. By becoming a stone tool- using hunter, man transferred his thrusting and penetrating activity and its target from human female to hunted animal. By becoming a container-using gatherer, woman transferred her receptivity and containment functions from human male and human child to vegetable 6 material. The tools of either sex thus became analogical counter- parts of the gender-appropriate parts of their anatomy. In extending the analogy further we discover a kind of circu- larity. In a metaphoric sense, animals and women become equated as receivers of the male thrust, while vegetable materials to be contained become metaphoric equivalents of both men and children. We might even assume that overt and elongated vegetable food, such as seed pods, grain heads or leafy materials were metaphorically male, while roots that had to be extracted from the ground were metaphorically children. In order to hunt or collect effectively, each sex needed one arm for carrying and the other for grasping and thrusting. Since neurologists tell us that handedness is reflected in human brain structure, it would seem that the left hand became conventionalized early as the more passive containing or carrying hand, and the right as the more active grasping or thrusting hand. This would provide an early start for the classificational chains that Hertz and Needham (1973) and others have found to be universally associated with the right and left hands. It will be remembered that there is a near universal association of the left hand with females and the right hand with males. Further, the left is considered to be bad or inferior, the right good and superior. Since animal killing is justified because they stand in an inferior position to man, it is highly likely that the metaphoric identification of animals with women served to warp a value judgment into the classificational chains. While economic causality does not seem a necessary condition for classification, emotional causality would seem to be essential, for without affective attachment no classification could exist or be continued. If we start with an emotional commitment to play, which stems from an affectivity established in infancy by nurturant social bonds and tactile stimulation, and couple it with the affectivity engendered by the sexual drive, we can see how an iconic model that links sex and reproduction, food production and sharing with social roles having a biological base has a strong limbic motivational source, with the potential of indefinite maintenance even if the iconic links become so abstract as to be seemingly arbitrary. In order to fit language to this model, we will need to replace archaeological with ontogenetic evidence. Jakobson and Halle (1962) were struck with the fact that the earliest meaning-bearing sound sequences typically acquired by human infants were the same as those lost last by aphasics. Using these sounds they constructed what they called a "primordial triangle," consisting of E, t and a. In passing, they comment that m and n appear roughly at the same time as p and t in infant speech. This suggests that there was not one primordial triangle, as they hypothesize, but two, or a grouping of five sounds, classified by shared features. Thus, m and n share nasality, k and m share bilabiality, p and t are stopped (i.e., characterized by abrupt release), and t and n share tongue to tooth 4.) ( -% a) Oc) co = ce. . 4. (4-) Cl O a _ -H 44 4p ) ot'UCD $4 8 IH (a *r- ci) ai) U) d o 0 cn (1) R H r - 4 r - d 4i) .d d e 4 - H H A I o U) 4J (1 a) 0 04 ci) a Ea b-c 0- c i . U)4.) . t 4 ) .. = ' @ 4O 4-) o. 'H 4 - a) Ar tOn 4. _1 -r U0 ci) 4-) 0 H 0 0 Ez z I I ci) tP H 0 04 $4 ci) rx. 0 0 04 H 0 0 4.) ci) H ' 0 M . 0 0 ' d (.4) ci) 4 4.) *H 0 cd 44 I0 a) '-.O4. 41.) e a) r44 7 8 contact. A involves no closure or contact whatsoever, so it stands in a mediating capacity (Figure 2). In connection with their triangle, the authors do not discuss meaning. Stopped Nasal Bilabial p m a Lingual- t -n Dental Figure 2. Plio-Pleistocene Linguistic Units (Adapted from Jakobson and Halle, 1962, pp. 492-493) Jakobson (1962), in an article entitled "Why Mama and Papa?", discussed Murdock's (1959) analysis of cross-cultural data in which he demonstrated statistically an overwhelming correlation between sequences of either mama or nana with the meaning of "mother" and papa or tata with the meaning of "father" (deviations to dada, daddy, baba, etc., were permitted, and vowels could be any low vowel and not just a). To explain the universality of these terms, Jakobson spoke of the nasal murmur created when a nursing infant is suckling, contrasted with an open mouth associated with the father, a less intimately familiar figure. It does not seem to have struck him that the sounds with which he was dealing here are the same as those that he had postulated for the primordial stage. Jakobson's argument for the exigencies of feeding as a causal origin of these sounds may have some merit in explaining m, but is weak for the other sounds. An analogical argument can make a better case. M represents closure and p opening of the major facial ori- fice. In addition, m functions as contact and k as outward thrust of lips and expulsion of air. T, like p, involves abrupt opening, but also an extrusion of something inside: the tongue. N, like t, involves that inner something, but in contrast with t, n may be protracted while the breath is nasally expelled with no facial change. Thus, m and n are contained, while k and t are open and forward- or outward-thrusting. As human factors in the Plio-Pleistocene social metaphor, papa or tata serve neatly as icons for the adult male, or father figure, while mama or nana are analogically appropriate for the child- nurturing adult female. The analogical bases for semantic appropriateness suggest that objects other than human actors could as readily be designated by these syllables, or the single conso- nants that serve as onset. Thus, p(a) might appropriately represent a core tool and t(a) a flake tool. M(a) might represent any 9 container, such as mouth, breast, vegetable container, or hand, or the recipient of a male thrust, the hunted animal. By the same token, n(a) might represent whatever was contained: child, vegeta- ble material, breast milk, semen, etc. In other words, the abstract quality of the spatial function provides for polysemy rather than single reference, and reference is functional and spatial rather than substantive. Figure 3 unites the socioeconomic with the linguistic model and suggests some polysemic possibilities. It should be noted that there is no necessary direction of significa- tion. Any object or phoneme might easily serve as a sign for another within the same domain. Possibly language only gradually took over the major signifying function. A preliminary examination of the meanings of simple morphemes of the shapes discussed here in five languages with no known genetic interconnections and with wide geographic spread seems to bear out the hypothesis. There is no space to present these data here. It is interesting, however, that Clement Scott, who compiled the Nyanja dictionary (a Bantu language of Africa) that I used in the cross- language comparison, was greatly struck by the transparent sound-meaning analogies in that language. He calls attention to the tendency of the language to use sounds "in continual remembrance of their primary ideas in 'poetical' endeavor to convey thoughts in words" (Scott and Hetherwick 1951:254). For example, In Nyanja, m signifies 'collectivization' and n 'unification,' while p means 'incisiveness' and 'openness' and t 'open, stretching, extruding.' These are very similar to the meanings of these sounds postulated here. Nyanja a has a meaning of third person as well as objective case. These can be subsumed under the semantic rubric, 'other,' as can the meaning of Dravidian a, 'the remoter person or thing,' and the Proto-Indo-European neutral vowel, *e, which means 'to, toward, near, away from' and the like--in other words, 'other' as goal or target. If we think of the exterior-interior aspects of these sets, we find language beginning with manipulation at the front of the mouth. If we should postulate a simple, binary set for the Australopithe- cines, rather than the more complex, later hominid model that we have been describing, we might with some reason assume that the first opposition was between E 'outwardness' and m 'inwardness,' corresponding to male:female sexual roles. With home-based, small, social groupings involved in food sharing, the model expanded somewhat. This corresponds to Australopithecine unworked tools, consisting of selected stones unworked by flaking. The linguistic advent of separate tongue activity in production of t and n would then correspond to the striking off and using of flakes. During the Lower Paleolithic core tools began to move toward symmetricality, becoming perfected in the Acheulian hand axe. Later, as the genus Homo moved toward and into the Mousterian, flakes, as blade tools, began to be perfected through Levallois techniques, while core tools became less important. 10 The course of language must have shown similar perfection of tongue operations in contrast with the grosser, facial movement used to produce bilabial consonants. Jakobson and Halle (1962) hypothe- size the second primordial acquisitions to be i, u and k; i involves raising and flattening the front surface of the tongue, and u and k the back. The evolution of oral manipulation thus progresses from front to back of the mouth. These developments perhaps character- ized the early Mousterian and developed with progress in pressure flaking of stone blade tools. During the Upper Paleolithic, tools, tongue movements and stylistic variability seemed to reach nearly the symbolic complexity known today. It was probably during that period that linguistic iconicity lost its semantic transparency. J. David Sapir (1977) provides a model for metaphor that admits of two basic types: internal and external. In internal metaphor two terms from different semantic domains are linked in an equation by means of an unstated semantic interface appropriate to both terms. In external metaphor, these linkages expand to create a double linkage of the type: a is to b as c is to d. These external metaphors constitute system in language and culture. According to Sapir, "Once an external metaphor is set out as an opposition, fur- ther metaphors can be developed from the original simply by adding to or slightly altering the criteria of opposition" (Ibid., 28). These cultural metaphors are never arbitrary, but may become quite abstract. As Sapir says, "The more remote the terms are from each other . . . the greater are the possibilities of making a variety of non-arbitrary connections" (Ibid., 31). It is the possibility of exploiting external metaphor that created the Pleistocene revolution that saw the birth of culture. An interesting feature is that the mediating, semantic interface of each metaphoric equation was abstract enough to allow for consider- able polysemy, so that reference was always multiple rather than singular. It was the semantic abstractness of the interface, or the remoteness of the terms from one another that ultimately brought about the lack of iconicity between sound and meaning, or the seem- ing arbitrariness that we find today in both language and culture. The iconicity of ape and dog classification, to which Premack called attention, is at the most concrete end of the iconic scale, while cultural metaphor spans the continuum from concrete to abstract. In addition, sub-human classifications are closely attached to immediate, pragmatic goals, while the goals of cultural classification, like the classificational systems themselves, are often so abstract, remote and unconscious as to seem virtually non-existent. It is probably for this reason that man has tended to take goal-attainment out of his own hands and place it in the hand of God, or gods, who presumably are capable of comprehending the whole of the cultural model, which has left man himself, including anthropologists, in a state of bafflement. 11 REFERENCES CITED Bateson, Gregory 1972 Steps to an Ecology of Mind. New York: Ballantine Books. Deetz, James F. 1979 Towards a Structural Archaeology. Paper presented at the Annual Meeting of the Kroeber Anthropological Society. Berkeley, California. May 12. Durkheim, Emile and Marcel Mauss 1963 Primitive Classification. Chicago: University of Chicago Press. Geertz, Clifford 1964 The Transition to Anthropology. In Horizons of Anthropol- ogy. Sol Tax, ed. Pp. 37-48. Chicago: Aldine. Hewes, Gordon W. 1973 Primate Communication and the Gestural Origin of Language. Current Anthropology 14:5-29. Holloway, Ralph L., Jr. 1969 Culture: A Human Domain. Current Anthropology 10:395-426. Isaac, Glynn Ll. 1978 Food Sharing and Human Evolution: Archaeological Evidence from the Plio-Pleistocene of East Africa. Journal of Anthropological Research 34:311-325. Jakobson, Roman 1962 Selected Writings. 'S-Gravenhage: Mouton. Jakobson, Roman and M. Halle 1962 Phonology and Phonetics. In Roman Jakobson, Selected Writings. 'S-Gravenhage: Mouton. Kuhn, Thomas S. 1962 The Structure of Scientific Revolutions. Chicago and London: University of Chicago Press. Lee, Dorothy 1959 Freedom and Culture. Engelwood Cliffs, N.J.: Spectrum. Murdock, George P. 1959 Cross-Language Parallels in Parental Kin Terms. Anthro- pological Linguistics, Vol. 1, No. 9:1-5. Needham, Rodney, ed. 1973 Right and Left: Essays on Dual Symbolic Classification. Chicago and London: University of Chicago Press. 12 Premack, David 1972 Two Problems in Cognition: Symbolization, and From Icon to Phoneme. In Communication and Affect: A Comparative Approach. Thomas Alloway, Lester Krames, and Patricia Pliner, eds. New York: Academic Press. Sapir, J. David 1977 The Anatomy of Metaphor. In The Social Use of Metaphor: Essays on the Anthropology of Rhetoric. J. David Sapir and J. Christopher Crocker, eds. Pp. 3-32. Philadelphia: University of Pennsylvania Press. Saussure, Ferdinand de 1959 Course in General Linguistics. New York: Philosophical Library. Scott, David Clement and Alexander Hetherwick 1929 Dictionary of the Nyanja Language. London: Lutterworth Press.