Comments Welcome
George W. Grace
University of Hawaii
July 22, 1981
I pose the question as I have in the title because I assume that it is particularly difficult to understand that which is most familiar--that which we take for granted--and because I assume that the importance of having language to human affairs has never been properly understood. However, there is an increasing awareness of different aspects of this importance in such diverse fields as philosophy, anthropology, psychology, sociology, and political science. In fact, as any of the sciences concerned with man penetrates more deeply into its subject matter, it finds itself involved more deeply with language. However, there is perhaps insufficient appreciation of the fact that the aspects of language which are being encountered by these different disciplines are all aspects of the same thing. And particularly, there is very little allusion to any prospect that all of these human sciences (which are, therefore, necessarily sciences of language and its epiphenomena) could somehow make common cause with one another and, above all, with that science in whose custody language as a whole is understood to fall--linguistics.
There are many problems for which such a unified science of language--a science concerned with what it is like to have language- -could profitably come into play. It could be useful in assessing the experiments in teaching apes communication patterns based upon human languages. In what ways is what the apes have achieved like having language? Or again, it seems that it is in dealing with the consequences of having language that sociobiology proves most inadequate (cf. Sahlins 1976: 61 (1)). The unified science of language that I have just hinted at might play a very useful role in just such nature-nurture confrontations. Such a science might be particularly well suited to assist in defining the domain and the limits of sociobiology, and might also come to be its main complement in the study of the human species.
I have, in Grace 1981b and to some extent in 1981a, attempted to lay a few foundations on linguistics' side of the gap for an eventual bridge to achieve the unification. The exploratory proposals that make up the body of this note are directed at discovering possible emplacements for further foundations.
But before going further, I should say a few more words about the assumptions from which this note proceeds.
The principle assumption is that we learn what it is like to be us in large part by discovering where what it is like not to be us differs. I would like to enlarge briefly upon that point. We constantly, and inevitably, operate on the basis of assumptions. Furthermore, we do not really know much about the validity of most of these assumptions. In fact, we do not even know, except to a very limited extent, what they are. Our assumptions are largely inaccessible to introspection. One general characteristic of our unconscious assumptions is that we ordinarily assume that an unfamiliar thing will work in pretty much the same way as some apparently analogous familiar thing. That is, except where specific differences have been noted, assumptions tend to get transferred en bloc to the new system from the old one. But even that statement misrepresents the situation because it suggests that a great number of separate assumptions are bundled together and transferred. A more accurate representation would show what is transferred as one big, chunked assumption, in which it would be possible with suitable analytic procedures to separate out a great number of distinguishable different assumptions.
As I mentioned above, all of this has the paradoxical consequence that we learn what it is like to be us in large part by discovering where what it is like not to be us differs. It is thus that the different characteristics that are melded into our assumption chunk come to be separated out.
For example, in forming our assumptions of what we are, we have been inclined to attach particular importance to our conscious experience of ourselves. But now it is becoming increasingly evident that consciousness reveals very little about even those of .our functionings which we attribute to mind--e.g., awareness, thinking, learning, etc. In fact, it seems that consciousness is probably one of the epiphenomena of language (On this point see particularly Julian Jaynes 1976a: 66 and passim and Daniel C. Dennett 1978: esp. 30 and 169-71.
Another example might be the influence of our long tradition of prescriptivism toward language on our conception of the nature of language (cf. Grace 1981b .
What I will try to do in this note, then, is to attempt to throw some light on the question of what it is like to have language by attempting some speculations about what it was like for our pre-language ancestors not to have language as well as about the subsequent evolution of language. I will be concerned with these questions: What is it that we have added to the representational pattern inherited from our prehuman ancestors? When was it added? And in what increments?
In this section, I will attempt to find some indications of what the inherited pattern was like as a basis for the further attempt to specify what has been added to it. I will begin with Gregory Bateson's discussion (1972) of what Freud (cf. Freud 1938: 525ff ) called "primary process".
PRIMARY PROCESS. Sigmund Freud in his The interpretation of dreams proposed that dreams and the operations of the unconscious in general were structured by what he called "primary process" whereas conscious thoughts (especially verbalized thoughts) are expressed in "secondary process". In Freud's original conception, the role of secondary process was central (as is to be expected from the central role which conscious processes assume in our introspection). Primary process operated on that which the conscious mind found unacceptable and therefore rejected (repressed). As Bateson points out, we are now coming to realize that secondary process does not occupy the central position that Freud assumed it to have, but is in fact a limited epiphenomenon closely associated with language and verbal expression. It is primary process which (even though it is not directly accessible to introspection) represents the way the mind essentially works. Per contra, secondary process, I propose, is the way language works.
Primary process is said to be different in form and in subject matter from secondary process. As to form, "Primary process is characterized ... as lacking negatives, lacking tense, lacking in any identification of linguistic mood ... and metaphoric. These characterizations are based upon the experience of psychoanalysts, who must interpret dreams and the patterns of free association." (Bateson 1972: 129). [I find it notable that what primary process is said to lack corresponds rather exactly to what I have called the "specification of the condition of instantiation" (Grace 1981b: 102-3, 104-6)].
As to subject matter, "In primary process the things or persons are usually not identified, and the focus of the discourse is upon the relationships which are asserted to obtain between them. This is really only another way of saying that the discourse of primary process is metaphoric." (ibid.) [However, even in the time since Bateson wrote this there has been a notable increase in our awareness of how very basic metaphor is to language--to secondary process--as well (cf. for example, Lakoff and Johnson 1980, Jaynes 1976a:48-59)].
"The focus of "relationship" is, however, somewhat more narrow than would be indicated merely by saying that primary-process material is metaphoric and does not identify the specific relata. The subject matter of dream and other primary-process material is, in fact, relationship in the more narrow sense of relationship between self and other persons or between self and the environment." (Bateson 1972: 140).
"...the characteristics of primary process..are the inevitable characteristics of any communicational system between organisms who must use only iconic communication" (Bateson 1972: 140). They are characteristic not only of the dreamer, but also of artists and of prehuman animals or birds.
To review succinctly, there is a kind of representational process which is found in other animals as well as in man. In man it is found in dreams, in art (in varying degrees in different forms of art--especially of literature), and in unconscious processes generally. It is iconic in form, i.e., it represents abstract relationships. The relata are not specified; they are always the self and another individual or the environment.
Primary process corresponds rather well to the first (model- formulating) function proposed in Grace 1981b (101ff). The second function (specification of the condition of instantiation) corresponds closely to what has been said to be lacking in primary process. The third function (connecting the relata with the ken of the audience) is presumably unnecessary since the relata are presupposed and do not require specification.
It would appear then that we have continued a representational pattern inherited from our prehuman ancestors. As one approach to imagining the status quo ante we have considered the notion of primary process, which Bateson proposed as characteristic of non human animals as well as most human representational patterns. As a further approach, I would like to make a few comments about some characteristics of other animals, particularly Pongidae.
THE PONGIDAE. The Pongidae (great apes--orang utang, chimpanzee, and gorilla) are of particular interest on two counts. First, there have been numerous attempts to teach apes--usually chimpanzees--human language or some system of communication designed for experimental use but inspired by human language. Second, they are our closest living relatives by far and therefore presumably of particular significance in the task of reconstructing what our last non-language-bearing ancestor was like. The following rather miscellaneous observations may be of some relevance.
1. BRAIN LATERALIZATION. Some years ago it came to be supposed that language was located in the dominant hemisphere. Further research has revealed that the matter is more complicated than that, but it seems particularly true that for right handed adult males a much greater language deficit results from the unavailability of the dominant (left) hemisphere than from the unavailability of the minor (right) one. And it seems quite well established that there is clear specialization of the hemispheres and that the dominant hemisphere has a more important role as far as language is concerned. This role has presumably developed since our ancestors separated from those of the Pongidae.
In this connection it is very interesting to note that the performance of the "signing chimpanzees" has been compared with the linguistic performance manifested by the human minor hemisphere (e.g., cf. Hill 1974: 147). That, of course, suggests that what has been added to the prehuman representational pattern is essentially what is in the dominant hemisphere today. Jaynes (1976a) offers the interesting speculation that the development of the language capacities of the left hemisphere occurred very late, and that they were forced into the left hemisphere by a previous specialization of the right hemisphere. In his proposal the right hemisphere became the storage place of mnemonic and hortatory/admonitory formulae ("the voices of the gods") which served to guide complex behaviors. (Presumably these are to be interpreted as the forerunners of the maxims, adages, and proverbs which are known even today). This suggestion seems to accord well with Diana Van Lancker's (1975) proposal that "automatic" or "formulaic" speech is located in the right hemisphere while "propositional" speech is in the left.
2. PRIMATE SOCIAL ORGANIZATION. The conventional view has been that primate groups are organized on the basis of a dominance hierarchy. The basic principle, according to this assumption, was that of individual animals' knowing with respect to each other individual animal which of the two must prevail or yield in any encounter such as competition for food or sexual favors. These binary relationships were presumed to be maintained by aggressive displays on the part of the dominant animal. There was a dominance hierarchy among the males, with a particular dominant ("alpha") male at the apex of the entire group. There was often also a hierarchy among the females, with adult males generally dominant over females, etc. In short, the assumption was that what organization existed was based on each animal bullying whichever animals it was able to, or believed itself able to.
Recent research indicates that this picture has limited validity at best (cf. Rowell 1974), and that it probably arose as a projection of the researchers' expectations (based on their experience with Homo sapiens [or specifically Homo sapiens of the Free Enterprise persuasion?], it has been suggested). Patterns of dominance seem much less consistent and fixed than was assumed; one cannot predict with confidence which animal will prevail in an interaction. Female patterns seem particularly unstable, being significantly affected by presence of their young or by the stages of their reproductive cycles. Moreover, apes (in contrast to many monkeys, e.g., baboons) do not have frequent recourse to aggressive (agonistic) displays. They display relatively little threatening behavior. In fact, even among those species more inclined to agonistic behavior, researchers are now tending to shift the focus from dominance to subordinance (cf. Rowell 1967) that rather than looking to instances of competitive interaction to infer the rank order within a group, one will find them better manifested in the "attention structure"; that is, which animals are the recipients of the attention of which. Subordinate animals pay attention and thereby maintain a suitable distance, either following or giving way as appropriate, etc. This point is of interest here since it permits me to lead into the speculations of Jerome H. Barkow ( 1976) that, "...human beings pay attention not just to physically present others but to their internalized representations (and sometimes to representations which have never had corporeal existence)" (1976: 204).
To quote Barkow further, "By the Ramapithecus stage our ancestors' attention structure probably permitted them to internalize elaborate and long-lasting representations of close matrilineal kin and of the group's more dominant males" (ibid. 206).
"Certainly by the Homo erectus level these representations were sufficiently powerful to be attended to even in the absence of the originals themselves. For the young, this meant that mother in effect cautioned about danger even in her absence" (207).
"In a certain sense, our 'alpha-animal' is in our heads; Freud called him the 'superego'" (209).
These suggestions, especially the last, are quite reminisicent of Jaynes's "voices of the gods".
Barkow's speculations suggest an evolution in steps. Jaynes (1976a, 1976b) likewise attempts to reconstruct a sequence of evolutionary steps. I do not find the details of either proposal particularly convincing. Nevertheless, it seems clear to me that language must have evolved over time. I cannot construct any believable scenario wherein the development was instantaneous. It seems certain to me that there must have been long periods when our ancestors had something significantly more like language than any of the present-day apes do, but still something that we would be hesitant to call language. Moreover, I would imagine that there must have been a progression over time from something with only very limited language-like characteristics through something with increasingly great language-like characteristics to fully developed language.
So, part of my tentative answer to the "when" question is that the evolution must have occupied an extended period of time. Another part of the answer is that I cannot imagine that anything very similar to language as we know it could have existed before the Middle Paleolithic at the earliest (I am happy to see that Jaynes [ 1976a: 129-31, 1976b] agrees with me). The cultural progression before that time would be astoundingly slow if we ascribed it to creatures with language as we know it. It has been argued that the making of tools to a pattern presupposes language. I am not sure how it does so, although it is notable that no other animals make tools to a pattern over long periods of time non- instinctively, and I am willing to suppose that some sort of early stage of language-like development might be presupposed. However, to suppose that the peoples of the Lower Paleolithic and before had language as we know it seems to require the argument that technical progress is exponential (i.e., that the richer the technological base, the more rapid the rate of innovation from that base) to bear a heavier burden than it seems capable of. I therefore propose as a first hypothesis that the peoples of the Lower Paleolithic had something which we would not know whether to call language or not.
The idea that the evolution of language may have involved more than one innovation opens up the possibility that there were a sizeable number of different steps involved, and that the linguistic attainments of our ancestors may have been different at a number of different points in history. In fact, it seems virtually inevitable to me that if we could somehow (as with a time machine) contrive to observe evolving language at frequent intervals during its course of development, we would have great difficulty in deciding at what point we should assert that the people (animals?) had arrived at true language (2). I will conclude this note by attempting to distinguish different additions which have been made to the primate pattern of representation, although I will not try to arrange them in chronological order or to speculate about their dating.
I suggest that a number of kinds of additions to the primary process pattern can be identified. What follows is, of course, just a first attempt; certainly more distinctions can be made. On the other hand, it is clearly possible that some of the distinctions which I do make are false ones; i.e., some of what I treat as separate developments may, in fact, be no more than different perspectives on the same development. As I have indicated, I intend no claim that the order of presentation followed here bears any relation to the actual evolutionary sequence. The additions which I propose are:
1. Primary process representation is limited to two-place relations. There has been an expansion to more than two places.
2. Primary process representation permitted no "displacement", no distinction between the situation of speaking and that spoken of. Modern language permits the two to be separate, and requires that both be taken into account.
3. In primary process communication nothing is actually "said" in the strict sense. Primary process communicative behavior might be said to model a relationship, but nothing is actually said about it. What is added is a taking of responsibility by a sayer. It is what I have called a "specification of the condition of instantiation". Note that we can still disclaim responsibility for the meaning of our nonverbal behavior (or for what others claim it to mean) in a way that we cannot disclaim responsibility for (the meaning of) what we have said.
Note, however, that the kinds of things which are included under the heading "conditions of instantiation" seem to vary considerably in sophistication. They may very probably have been built up in a number of separate evolutionary steps.
4. In primary process representation, the identity of the two relata (which were necessarily immediately at hand) was self- evident. With the possibility of more than two relata and of absent relata, means of specifying their identity have become necessary. Thus, the noun phrase and noun phrase grammar must have been introduced at some stage (or stages--noun phrase grammar in modern language is sufficiently complex to suggest that it probably developed in more than one step).
5. Although this point and the next one follow less directly from what has been said about primary process, I think that the complexity of the relations/acts packed into verbs must have increased in what we may some day recognize as a fairly dramatic way. Some of this stage of development seems to be observable in the comparison of different contemporary languages (I interpret John Haiman's [1980] paper, "The iconicity of grammar" in this way [see fn. 2]).
6. All of the above makes for greater complexity of utterances, and this, in turn, requires the elaboration and implementation of more complex plans. Implementation of plans requires the seriatim execution of a prescribed sequence of steps; where the plan to be executed is complex, so is the sequence of steps (on the problem of explaining sequenced behavior, see Lashley 1951). It is worth noting here that sequenced behavior is frequently mentioned as a trait associated with the left cerebral hemisphere.
What is it like to have language? My first approximation of an answer to that question is, then, that it is to have the capacities involved in primary process plus the additional capacities given in the listing above (3). But that is no more than a first approximation; more careful and more expert analysis will surely change the list.
1. Sahlins writes (1976; 61), "If we were to disregard language, culture would differ from animal tradition only in degree. But precisely because of this "involvement with language"--a phrase hardly befitting serious scientific discourse--cultural social life differs from the animal in kind. It is not just the expression of an animal of another kind." Back up
2. It may even be the case that some steps in the evolutionary process might not have been reached by all contemporary languages (in this connection, cf. Haiman 1980. There is also a relevant paper by Paul Kay that I have not been able to locate in time). However, that is a very touchy subject. It is touchy first because we would probably assume a priori that if different contemporary languages have reached different evolutionary stages, the most advanced stage is that manifested by our own languages. And it is touchy for the more general reason that our society normally takes any suggestion that the characteristics or institutions of any group differ in any significant way from our own as a devastating putdown-- of the others! Back up
3. Charles Hockett must, of course, be recognized as the pioneer in the field with the design features which he has worked on for many years. Back up
Barkow, Jerome H. 1976. Attention structure and the evolution of human psychological characteristics. In Michael R. A. Chance and Ray R. Larsen (eds.). The social structure of attention. London, etc. : John Wiley and Sons, pp. 203-19. Back up
Bateson, Gregory. 1972. Style, grace and information in primitive art. Reprinted in: Steps to an ecology of mind. New York: Ballantine Books, pp. 128-52. Back up
Chance, M[ichael] R. A. 1967. Attention structure as the basis of primate rank orders. Man (n.s.) 2: 503-18. Back up
Dennett, Daniel C. 1978. Brainstorms: Philosophical essays on mind and psychology. Montgomery VT. Bradford Books. Back up
Freud, Sigmund. 1938. The basic writings of Sigmund Freud. Translated and edited by A.A. Brill. New York: The Modern Library. Back up
Grace, George W. 1981a. An essay on language. Columbia SC: Hornbeam Press. Back up
Grace, George W. 1981b. Ordinary language. ms. Back up
Haiman, John. 1980. The iconicity of grammar: isomorphism and motivation. Lg. 56: 515-40. Back up
Hill, Jane H. 1974. Possible continuity theories of language. Lg. 50: 134-50. Back up
Jaynes, Julian. 1976a. The origin of consciousness in the breakdown of the bicameral mind. Boston: Houghton Mifflin Company. Back up
Jaynes, Julian. 1976b. The evolution of language in the late Pleistocene. In Stevan R. Harnad, Horst D. Steklis, and Jane Lancaster (eds.). Origins and evolution of language and speech. Annals of the New York Academy of Sciences, Vol. 280, pp. 3l2-25. Back up
Lakoff, George and Mark Johnson. 1980. Metaphors we live by. Chicago: University of Chicago Press. Back up
Lashley, K. S. 1951. The problem of serial order in behavior. In L. A. Jeffress (ed.). Cerebral mechanisms in behavior. New York: John Wiley and Sons, pp. 112-36. Back up
Rowell, Thelma E. 1974. The concept of social dominance. Behavioral Biology 11: 131-54. Back up
Sahlins, Marshall. 1976. The use and abuse of biology: An anthropological critique of sociobiology. Ann Arbor: The University of Michigan Press. Back up
Van Lancker, Diana. 1975. Heterogeneity in language and speech: Neurolinguistic studies. UCLA Working Papers in Phonetics, 29. Back up
George W. Grace
August 3, 1981
I want to make two additions to my suggestion that the evolution of language must have involved several distinguishable steps.
First, although I proposed in the original note that the evolution of language had involved several steps, I tried to steer clear of the question of which steps preceded which. Obviously that is a difficult question to deal with responsibly--the evidence will necessarily be indirect, and it will doubtless take considerable time and effort even to begin to figure out what can count as evidence and how it is to be weighted. Nevertheless, I think that specific proposals are likely to be the best way to smoke out what ideas we do (consciously or unconsciously) have about what might count as evidence. And furthermore, I do have the feeling that certain developments seem likely to have preceded certain others. I will therefore propose a few hypotheses (speculations) about the sequence of steps in the evolution of language. I hope that they may stimulate some readers to feel that some of them seem right--in which case I hope those readers will be further stimulated to suggest what it is that makes us feel that that is so, i.e., what we are taking to be evidence. I hope further that those of them which seem to readers clearly to be wrong will stimulate those readers to try to specify in what way they seem wrong and what kind of evidence is applicable. With these remarks, I propose very tentatively the following hypotheses:
1. On the specification of relata: I suggest that the first stage may have consisted of just the addition of subject markers (or possibly absolutive [i.e., as opposed to ergative] or object markers) to verbs. At this stage, then, an utterance might have consisted of just a verb with such a marker, plus perhaps some specification of what I [alas, since the term is so awkward] call a condition of instantiation. (If it specified a condition of instantiation, I think I would be prepared to call it a sentence).
The addition of noun phrases would occur later (just possibly beginning with demonstratives and free pronouns?). Anyway, nouns would not appear before this stage. This proposal seems particularly interesting because it contradicts what seems to be a tendency in our culture to think of language as beginning with concrete nouns.
I would guess that modifiers came later and probably relativized clauses still later. [I have suggested elsewhere (Grace 1981b: 40) that nominalization of verbs in the frequency and complexity that we know it is a very recent development in our own cultivated language. Post-evolutionary?].
2. On the specification of the conditions of instantiation: I would suppose that the first stage must have involved the distinction of only a few such conditions. I would suggest that there might have been an early stage in which only, say, a realis, negatives, and questions were distinguished. Other tense, aspect and modal distinctions would gradually have been added at subsequent stages of the evolution.
After the first stages were past, I presume that evolution proceeded more or less simultaneously on various fronts. The early stages of evolution would, of course, include expansion of vocabulary (presumably primarily through metaphoric extensions) and the development of more subtle control of pronunciation as well as a gradual increase in the complexity of the speech plans being devised and executed (i.e, the utterances).
The second point which I want to make is that once we make the assumption that the evolution of language was not instantaneous, the distinction between the processes of that evolution and those of ordinary linguistic change becomes more problematic than it has been. Presumably linguistic differentiation, and therefore linguistic change in the traditional sense, has been going on ever since two critical conditions first appeared. The conditions are (1) the existence of arbitrary lexifications (i.e., arbitrary relations of signifier to signified) and (2) sufficiently extensive geographic distribution of the speakers of (pre-)language that they do not all participate in a single communication network. At that point I suppose that causes of change other than those responsible for the general evolutionary process must have come into play and that, as a consequence, a differentiation into independent linguistic traditions must have begun.
Does the beginning of this ordinary linguistic diachrony mean that the process of evolution of language has necessarily come to an end at that point? If we were to decide to say so, a corollary of that decision would be that we were thereby drawing the boundary between pre-language and language at that point. I.e., if we (1) say that the process of the evolution of language had ended at a particular point, and if we (2) agree that that process was carried through to completion--i.e., that it did culminate in language having evolved, then we must (3) conclude that fully- fledged language was in existence at the time when the evolutionary process came to an end. However, to make such a decision seems nothing more than a trick. There is no reason to suppose that whatever factors motivated the evolutionary sequence in the first place would necessarily have run their course just at the moment when the sufficient conditions for linguistic differentiation first appeared.
I assume that the evolutionary process must therefore have continued into the era of separate linguistic traditions (languages). By "the evolutionary process" I mean the same process (or at least processes which form a continuity with that process) as that involved in the pre-language stages of the evolution. I would speculate that the linguistic changes in the early period after linguistic differentiation began were mainly evolutionary changes (i.e., changes attributable to the same evolutionary motivations), and that the extent of parallellism in changes was much greater than it is today. I would guess that although the changes in different traditions might differ in details, the results would often turn out to be very similar because the same structural advances were being selected for.
But once we reach this point--once we accept the idea of general language-evolutionary forces operating alongside of other forces (e.g., internal structural pressures--what I call "drift") as causes of linguistic change in the earlier history of language (or pre-language)--what basis is there for assuming that those evolutionary forces no longer operate today? Is it not more reasonable to suppose that they do, in fact, still operate, but that the relative importance of their role in accounting for linguistic change has been reduced? In sum, I would suggest that from the point of view required here we can distinguish three kinds of causal factors in linguistic change. These are: (1) the same kind of selective factors which led to the development of increasingly complex pre-language to fully-fledged language (to, I would suggest, increasingly complex and sophisticated language) in the first place, (2) differential selective factors resulting from the development of highly sophisticated and highly differentiated cultures (these might be thought of as an adaptive radiation--in the biological sense--starting from the evolutionary factors), and (3) differential selective factors resulting from the highly sophisticated states of the languages themselves and the great differences in detail which distinguish them from one another (i.e, what I call "drift" factors).
| Home Page | The Ethnolinguistic Notes | The Ethnolinguistic Notes, Series 1 and 2 | Ethnolinguistic Notes, Series 3 | The Ethnolinguistic Notes, Series 4 | Reflections: Language Evolution |
| Reflections: Knowledge of Language | Personal Page | The Human Predicament | Why Write Unpublishable Things? | Modest Proposals | Odds and Ends | Pictures |
Last updated 5 January 1996