Comments welcome
George W. Grace
University of Hawaii

Ethnolinguistic Notes

Series 4, Number 24

COLLATERAL DAMAGE FROM LINGUISTICS?
4. WHAT KIND OF LANGUAGE DOES THE "LANGUAGE ACQUISITION DEVICE"* REALLY PREPARE US TO ACQUIRE?

On the users of earliest language
On the effects of isolation and of its absence on language structures
Analytic processing and outward orientation
That the “language faculty" can’t have changed much since “earliest language"
About the “complexity" of insider-oriented languages
About how complexity got there in the first place
Conclusions: What kind of language does the “language acquisition device" really equip us to acquire?

On the users of earliest language

“…if grammar is tailored to the needs and properties of language users (to whatever degree), and language users now are not what they used to be, then it follows that grammar is probably not what it used to be." (Newmeyer 2002: 369)

Talmy Givón has described the kind of society in which our prehuman and early human ancestors lived as a “society of intimates" (“where all generic information is shared"-- Givón 1979: 297) . He contrasts the circumstances of communication in such a society with those of the “society of strangers" toward which subsequent social evolution has led.

The biological evolution that led to the human capacity for language presumably extended over many millennia. During that period and for more millennia afterward our (1) ancestors presumably lived in small groups where everyone knew everyone else, where all were in frequent contact with all, and where there was no division of labor other than by age and sex.

In the early period the total human population was, of course, very small. In most cases groups were probably fairly isolated, only infrequently encountering any others. Their relative isolation may constitute an important clue as to the nature of these languages since studies of linguistic change have indicated that isolation leads to what has often been referred to as “complexity" of structure.

On the effects of isolation and of its absence on language structures

The studies that I have in mind distinguish circumstances that favor what are sometimes described as “simpler" language structures from (opposite) ones that favor greater “complexity". Contact with other languages (or dialects) has been the focus of a number of these studies and has been identified by some as the critical factor favoring the kind of change that is described as “simplifying". At the same time, languages without such contact are said to become more “complex" in structure. C-J. Bailey has even proposed different terms to distinguish the two patterns of change. “Connatural" developments are the kinds that occur "when languages are left alone, i.e., when they have no contact with other systems", whereas those that occur "as the result of contact with other systems" are called “abnatural" (Bailey 1982: 10).

The use of the descriptive labels “simple" and “complex" deserves some comment. It is interesting to note that no one seems to feel any uncertainty as to which kind of structural characteristics should go with which label. The characteristics associated with “simplicity" might be described as analyticity and transparency. A language is analytic to the extent that the meanings of its expressions can be determined from their composition. It is transparent to the extent that it approximates a one-one relation between forms and meanings--i.e., absence of allomorphy, particularly of morphologically-conditioned allomorphy.

The easy learnability of simpler languages is particularly emphasized. By contrast, as Bailey points out, development in isolation leads to languages that would generally be regarded as hard to learn. He speaks of unusual sounds and difficult sound combinations. But perhaps more important is what might be described as a general structural opaqueness. In his words, "Connatural developments include synthesis (the amalgamation of functional particles as flectional bundles); abnatural developments include analysis (the breaking down of inflections into nothing or, more specifically, into functor words)" (Bailey 1982: 67).

Trudgill 1989 generally supports Bailey's claim, noting further that the differential effects of isolation and contact may be found in dialects as well as distinct languages. However, Trudgill notes that child learners of a language appear to have no difficulty with what count as complications for adults, and proposes that the “abnatural" changes simplify the learning task only if the learner is post-adolescent in age. He suggests, therefore, that the language contact whose presence or absence is relevant here is not contact of just any kind, but only contact of a kind that results in adults in significant numbers being required to learn the language.

William Thurston has made a similar proposal in his report on nine languages spoken in Northwestern New Britain (Papua New Guinea). He focuses particularly on the contrast between languages that were used as lingua francas and those that were used only with other members of the same community. He uses the term “exoteric" to describe the New Britain languages that have important lingua franca roles. Exoteric languages are described as being of necessity relatively simple. As a consequence of their lingua franca role, they “must be easy for neophytes to acquire" (Thurston 1987: 38).

Contrasted with the exoteric languages are those that Thurston labels “esoteric"--languages that function primarily for communication among people of the same social group. An esoteric language, he reports, tends to become increasingly complex; it “gradually accumulates morphological irregularities, morphophonemic complexities, highly-specific lexical items, constraints on derivation leading to suppletion, and opaque idioms". (Thurston 1994: 580).

What these studies add up to seems to be that attempting to make oneself understood to non-speakers (or non-native speakers) leads to “simplification" of style (a fact that is reminiscent of the widespread occurrence of more of less conventionalized styles of “foreigner talk"). At the same time, of course, the attempts of these non-natives to speak the language will themselves provide examples of such simplification. In other words, communications in which at least one interlocutor is not a native speaker foster simplified styles, and these styles may have a permanent place in (or at least a permanent effect upon) the repertoire of the language.

In Grace 1989 I distinguished between "dictionary-type" (or "holistic") knowledge and "grammar-type" ("analytic") knowledge that can be used in understanding speech or in preparing to speak. These two types of linguistic knowledge afford different strategies for processing incoming and outgoing utterances. I referred to the use of dictionary-type knowledge as “recognition strategy” and that of grammar-type knowledge as “analysis strategy”. Here I will use the better-known terminology of Alison Wray, who in a number of publications (see especially Wray (1992, (2002a) , has referred to the strategy of relying on stored knowledge as “holistic processing" and that of computing--i.e., resorting to grammatical knowledge--as “analytic processing".

It’s easy to see that the main effect of the increased analyticity resulting from the “simplifications" is to better adapt the languages for an analytic processing strategy. No doubt this facilitation of analytic processing is particularly helpful to newcomers to the language. In fact, it's helpful in two ways. Most obviously, it makes it easier for someone who doesn’t know the most “nativelike" ((Pawley and Syder 1983) encoding for a particular idea to construct a workable encoding ad hoc. (And such encodings can, of course, subsequently be decoded by means of an equally straightforward analytic process).

But the increased availability of analytic processing offers another benefit: it makes it easier for these outsiders to encode ideas for which no nativelike encoding exists at all, ideas that may be quite exotic to the culture normally encoded in the particular language. These ideas might derive from the outsiders’ own cultures, of course, but they might have other sources or even be entirely new.

Furthermore, the native speakers also benefit. The increased availability of analytic processing is after all not limited to non-native speakers. It also makes it easier for native speakers to encode novel ideas of their own. In fact, the point that it is easier for speakers of simple languages to express novel ideas is made explicitly by Thurston in the following passage (Thurston 1987: 41):

The speaker of a simple language, however, is more likely to know his/her language completely, whereas the speaker of a complicated language requires a great deal of esoteric knowledge and experience to be able to use it elegantly. Whereas the simple-language speaker can easily put novel ideas into words that are immediately understood and acceptable to others, the speaker of a complex language is more constrained in expression by the degree to which s/he knows the language, because s/he cannot necessarily know whether a made-up construction is already reserved for another meaning, or whether the form s/he wants is derivationally suppletive.

In contrast as we noted above, the presumed isolation of the groups speaking the earliest languages suggests a strong insider orientation. Consequently, we would expect that most of them never experienced any significant pressure to adapt for analytic processing, and that they always relied to a very high degree on holistic processing. We would, therefore, expect that their structures were what would be described as “complex" (as that term is being used here).

Why is the attempt to determine the characteristics of earliest language of any particular interest? The reason is that they provide evidence on the “language faculty"--the capacities for language use that are “hard-wired" in the human brain. The claim here is that these capacities must not be very different from those required for the use of earliest language.

Of course, it is certainly true that the skills required for proper use of a Modern Standard Language (especially its “elaborated code") seem very different from those that would have been required for the use of earliest language as I have depicted it. Different enough, certainly, that one might expect them to become a strong selective force in evolution. And I don’t intend to suggest that the kind of evolutionary factors that would affect the innate language faculty would have suddenly ceased to operate. However, the fact is that Modern Standard Languages have only existed for a few generations. (In my opinion, it is probably also significant that only a small proportion of any breeding population have ever become truly skilled in using the kind of written language that exploits--and that was originally responsible for--the more elaborate grammatical resources of such languages, and, furthermore, that these skilled individuals have not consistently emphasized such skills in their choice of mates).

It is true, of course, that the world has changed a lot since the earliest languages. The human population has grown astronomically. Technology has advanced and led to rapidly increased division of labor and social segmentation. Nevertheless, we should remember that it hasn’t been long--certainly not more than seven or eight generations--since the great majority of people everywhere in the world still lived all of their lives within a few miles of the place of their birth. And one must suppose that most conversations took place between people who had known each other all of their lives, who did the same kind of work, participated in the same religious observances, and generally shared the same knowledge and values. The greatest part of the language use at that time must still have been of a kind more suited to a society of intimates than one of strangers.

In fact, the ancestors of some of the peoples living today have remained preliterate and have generally continued--virtually until the present--living in circumstances even more like those that must have existed when language first emerged. Nonetheless, there is no indication that I know of that there is any difference between the innate language faculties of such people and those with an ancestry of highly skilled users of Modern Standard Languages. As far as I know the children of such “primitives" appear to be innately as capable of learning to compose and interpret highly technical expository prose as anyone else.

These considerations seem sufficient to justify the conclusion that evolution of innate language faculty had for all intents and purposes been completed by the time full fledged language had appeared.

Therefore, it seems that our innate “language acquisition device" must be adapted to the characteristics of earliest language--that is, it must be designed to prepare the newborn to acquire an insider-oriented language that relies heavily on holistic processing.

About the “complexity" of insider-oriented languages

The conclusion that earliest language was probably insider oriented and that it relied heavily on holistic processing suggests that it was also at least somewhat structurally “complex” in the sense in which that term has been used here. Of course, the correlation of insider orientation and holistic processing with “complexity” are not perfect. The studies referred to above all describe “complexity” and “simplicity” as conditions that develop gradually in response to the way the language is used. Accordingly, a lag between the circumstances of use and the relative “simplicity” or “complexity” of the language is to be expected; a language that was insider oriented and relied heavily on holistic processing might still not be “complex”. However according to our present understanding, that would occur only if the language had previously relied more heavily on analytic processing and had not yet completed the process of complexification. There’s no reason to suppose that that might have been the case for earliest language.

It seems natural to wonder whether there can be any positive reason why isolation would lead to actually increased complexity. “Simplicity" sounds like a good thing. Simplicity makes languages easier to learn, more adaptable, and more easily (and economically) described. Complexity makes them hard to learn, very difficult to describe, and (it seems) hard to use for novel purposes. What selective factors could possibly be operating that would lead a language to abandon relative simplicity in order to evolve greater complexity?

This is certainly a very interesting question. However, it must not be allowed to distract us from what is the main point here: that these complex structures do exist and that their association with insider-oriented languages seems to be well established. I do have some rudimentary thoughts on likely selective factors that I hope may serve as a catalyst for further discussion. However, beyond noting that it does seem clear that the difference between analytic and holistic processing is the key consideration, I will leave them for an endnote. (2)

About how complexity got there in the first place

The conclusion that earliest language must have been at least somewhat “complex” in structure, leads inevitably to the question of the origin of language. How can earliest language have become complex in the first place?

At first blush it might seem that the studies by Bailey, Thurston, Trudgill, and the others point to the answer. They are all concerned with the process of becoming: of becoming either simpler or more complex. Their findings could be applicable if there were some basis for assuming that our hypothesized highly complex earliest language was preceded by a simpler, more analytic predecessor and had developed from that in the ways these authors describe.

However, such an assumption would introduce a major new complication. What might such a predecessor have been? If it was earlier than earliest language, it could not, by definition, have had all the properties of a full-fledged language. Of course, there are serious gaps in our knowledge of the evolutionary sequence through which language as we know it developed, but to the best of our current knowledge the last antecedents of language in our primate ancestors consisted mainly of unanalyzable vocalizations (i.e., only holistic processing can have been available). Any suggestion that earliest language had passed through a more analytic and transparent state in order to reach the complex state that we hypothesize for it would surely seem quite far fetched.

But the question of how earliest language developed its complexity is--while certainly intriguing--not our concern here. (3). All that concerns us here is the fact that earliest language was in all probability insider-oriented and complex in the sense described here.

As I tried to point out at the beginning of this “Collateral Damage" series (Grace 2002a-c), there are many lines of research that depend on some sort of assumptions about the human language faculty. I have been trying to show that many of the assumptions of the Post-Chomskyan paradigm about what language capacities are innate are wrong. These assumptions are based on the observation of abilities that are learned--abilities whose functions developed through cultural, rather than biological, evolution.

The purpose of this Note has been to consider what some of the characteristics of earliest language might have been. These characteristics are of interest because they presumably represent the end point of the evolution of the innate capacity for language of our species. Or very near to it. As we saw above, there must have been very little evolution of any hard-wired capacity since earliest language, and consequently we should not expect any linguistic abilities that have developed subsequently to have been specifically provided for in the human brain.

If the reasoning given here is right then, we should conclude that:
1. What the LAD equips the infant to “acquire" is something like the skills required by earliest language.
2. Earliest language must have been insider oriented.
3. The use of analytic processing plays a minimal role in insider-oriented languages; they rely primarily on holistic processing.
4. Holistic and analytic processing make different requirements on a language--give it different kinds of characteristics. Languages that have few pressures for outward orientation or analytic processing can appear quite exotic to those of us whose background is in Modern Standard Languages--they seem to march to a different drummer.

There is good reason to wonder how much we really understand about how these “complex" languages work. As I pointed out in Grace 2002c, we approach them with a conceptual apparatus built on the basis of the Modern Standard Language (and to a great extent, the Western tradition of prescriptive grammar). This obliges us, in the words of Thomas Kuhn (1970: 5) to “attempt to force nature into the conceptual boxes" of that apparatus.

If by “natural language” we mean that which is innately specified, then natural language is surely better thought of as a repertoire than as a calculus. Before we will ever be able to understand our innate language capacity, we need to find a way to escape our culture-centric presupposition that the prototypical function of language is to encode and decode propositions and to gain some insight into how insider-oriented languages that rely primarily on holistic processing work.

* It should be understood that I do not believe that there is any single "language acquisition device". (That's why I always put it in quotation marks). However, the term does provide a convenient way to refer collectively to all of the innate capacities that--in combination--permit (cause?) us and no other species to learn language.

1. The “our" of “our ancestors", is meant to include everyone living today (including those whose ancestors lived in circumstances much like those described virtually until the present).(Back up)

2. Although I believe that the difference between analytic and holistic processing is the key consideration, I should begin by mentioning a point made by Thurston. He has pointed out that it is sometimes advantageous to be able to speak in a way that is understandable to insiders and not to outsiders. In fact, I think such occasions may be more common than I had ever realized; certainly the whole matter is deserving of more attention than it ordinarily receives. However, I don’t think that can be the entire explanation for complexity or even the main factor involved.

We’ve already observed that what “simplification" accomplishes is to adapt a language for an analytic processing strategy. That is, it makes it easier to assemble linguistic expressions from scratch, piece by piece. It also means, of course, that novel expressions can often be interpreted by analysis--by analyzing them into their components.

On the other hand, it seems that complexity arises to the extent that holistic processing predominates over analytic. It arises where the need for analytic processing is less significant. However, there are surely values other than irregularity as such that are selected for. Among them must surely be fluency and a kind of economy--for instance, Thurston (1989: 38) speaks of “phonological efficiency at the expense of morphological transparency".

My suggestion is that the economy at work is rather parallel to that proposed by André Martinet for change in phonological systems. Each language tends toward the best balance for its particular circumstances between two requirements. The first requirement is for clarity, for providing information in a clear and unambiguous form. The opposed requirement is for fluency and economy of effort generally. The balance sought is between the burdensomeness of having to spell everything out (or at least of working out how much spelling out is required--what information the interlocutor may lack) and the inefficiency resulting from too much misunderstanding or need for repetition.

The particular circumstances of insider-oriented languages reduce the pressure for spelling out and leave unopposed (or, rather, more weakly opposed) the pressure for facilitating fluency and economy of effort. To facilitate fluency is to facilitate holistic processing. Economy of effort requires leaving out what is inessential. Presumably this would mean what is inessential to being understood (or, more accurately, to having a good enough probability of being understood). What would constitute a “good enough probability" would presumably vary from instance to instance according to such things as the importance of the message, time constraints, the potential consequences of misunderstanding, etc. Too much explicitness would become tedious, but so, too, would frequent misunderstandings.

Leaving out what is inessential would presumably result in contraction, and probably--to the extent that the message is not unexpected--ellipsis. The resulting “complexity" would, of course, constitute no problem to the extent that the processing is holistic.

I offer these speculations for whatever value they may have. But whatever their value or lack thereof, the fact remains that complex structures do exist and their association with insider-oriented languages seems to be well established. (Back up)

3. Although, once again I again can’t resist offering a few speculative suggestions. Attempting to imagine stages through which a complex, insider-oriented language might have evolved from unanalyzable vocalizations seems like a promising exercise. I’ll suggest one conceivable scheme here as an illustration of what can be done.

Like Alison Wray ((2000, (2002b), I see the development as proceeding from a system which used only holistic processing to one in which analytic processing played a major role. (For convenience, I’ll use the term utterance to refer just to the linguistic form uttered--i.e., what a linguist would transcribe. Furthermore, I’ll use it indiscriminately to apply to “utterances" in the predecessors of language, including primate vocalizations.)

My scheme proposes the following stages:
(Stage 1) There is just a repertoire of unanalyzable utterances (maybe they should be called “vocalizations").
(Stage 2) (Probably in response to expansion of the repertoire) a kind of phonaesthesia develops that can be used as a mnemonic aid. That is, there is a suggestive association of parts of forms with parts of meanings that permits the use of analogy as a reminder. Analogy would also provide an aid for language learning--i.e., for interpreting utterances that one hadn’t previously encountered.
(Stage 3) The analogical possibilities presented by the phonaesthesia provide a basis for the recombination of parts to create occasional neologisms. It is, of course, also possible to combine (i. e., utter in sequence) whole utterances. Most such neologisms are probably unconscious at first, but with conscious instances gradually increasing in frequency.
(Stage 4) The possibilities of such recombination come increasingly to be exploited. As this happens, parts suitable for recombination are increasingly isolated and given conventional recognition as potential elements for combination--i.e., approaching or achieving the status of morphemes.
(Stage 5) Finally, the elements available for combination and the conventions governing their combination reach the point that they could be called a grammar. Likewise, the frequency of innovation increases to the point that a large proportion of utterances have something ad hoc about them. At this point the analytic processing strategy has taken its place as a full-fledged partner of the holistic.

That’s my suggestion of a possible sequence. I would imagine no one would want to apply the term “language" to anything before the final stage. What I find it particularly hard to speculate about is how much of such a sequence would have been driven by biological evolution, and at what stage the evolution of the “language acquisition device" would have been essentially complete. (Back up)

Bailey, Charles-James N. 1982. On the yin and yang nature of language. Ann Arbor: Karoma Publishers.(Back up)

Bernstein, Basil. 1964. Elaborated and restricted codes: Their social origins and some consequences. In John J. Gumperz and Dell Hymes (eds). The Ethnography of Communication. American Anthropologist 66(6), Part 2. (Back up)

Givón, Talmy. 1979. On understanding grammar. New York, etc.: Academic Press.(Back up)

Grace, George W. 1989. Recognition strategy and analysis strategy in language use. Ethnolinguistic Notes, Series 3, Number 37. Printout. Also (1996) Internet WWW page at http://www2.hawaii.edu/~grace/eln37.html. (Back up)

Grace, George W. 2002a. Collateral damage from linguistics? 1. The Post-Chomskyan paradigm and its underlying assumptions. Ethnolinguistic Notes, Series 4, Number 21. Internet World Wide Web page at http://www2.hawaii.edu/~grace/elniv21.html. (Back up)

Grace, George W. 2002b. Collateral damage from linguistics? 2. The cultural evolution of language. Ethnolinguistic Notes, Series 4, Number 22. Internet World Wide Web page at http://www2.hawaii.edu/~grace/elniv22.html. (Back up)

Grace, George W. 2002c. Collateral damage from linguistics? 3. The role of culture-centrism. Ethnolinguistic Notes, Series 4, Number 23. Internet World Wide Web page at http://www2.hawaii.edu/~grace/elniv23.html. (Back up)

Kuhn, Thomas S. 1970. The structure of scientific revolutions, second edition, enlarged. International Encyclopedia of Unified Science, Volumes I and II: Foundations of the Unity of Science, Volume II, Number 2. Chicago: University of Chicago Press. (Back up)

Newmeyer, Frederick J. 2002. Uniformitarian assumptions and language evolution research. In Alison Wray (ed.). The transition to language. Oxford: Oxford University Press. Pp. 359-75. (Back up)

Pawley, Andrew and Frances H. Syder. 1983. Two puzzles for linguistic theory: Nativelike selection and nativelike fluency. In Jack C. Richards and Richard W. Schmidt (eds.). Language and communication. London: Longman, pp. 191-225.(Back up)

Thurston, William R. 1987. Processes of change in the languages of north-western New Britain. Pacific Linguistics B99, Canberra: The Australian National University. (Back up)

Thurston, William R. 1989. How exoteric languages build a lexicon: Esoterogeny in West New Britain. In Ray Harlow and Robin Cooper (eds.). VICAL 1: Papers from the Fifth International Conference on Austronesian Linguistics. Linguistic Society of New Zealand, pp. 555-79.(Back up)

Thurston, William R. 1994. Renovation and innovation in the languages of north-western New Britain. In Tom Dutton and Darrell T. Tryon (eds.). Language contact and change in the Austronesian world. Trends in Linguistics Studies and Monographs 77. Berlin/New York: Mouton de Gruyter, pp. 573-609.(Back up)

Trudgill, Peter. 1989. Contact and isolation in linguistic change. In Leiv Egil Breivik and Ernst Håkon Jahr (eds.) Language change: Contributions to the study of its causes. Trends in linguistics, studies and monographs, 43, pp. 227-37. (Back up)

Wray, Alison. 1992. The focusing hypothesis: The theory of left hemisphere lateralised language re-examined. Amsterdam/Philadelphia: John Benjamins Publishing Company.(Back up)

Wray, Alison. 2000. Holistic utterances in protolanguage: the link from primates to humans. In Chris Knight, Michael Studdert-Kennedy, and James R. Hurford (eds.) The evolutionary emergence of language: Social function and the origins of linguistic form. Cambridge: Cambridge University Press. (Back up)

Wray, Alison. 2002a. Formulaic language and the lexicon. Cambridge: Cambridge University Press.(Back up)

Wray, Alison. 2002b. Dual performance in protolanguage: Performance without competence. In Alison Wray (ed.). The transition to language. Oxford: Oxford University Press, pp. 113-37.(Back up)