sum

Comments welcome
George W. Grace
University of Hawaii

Reflections on the evolution of human language:

2. The Emergence of Analytic Processing:

4. Summary of suggestions toward a hypothesis

I want to try here to pull together the suggestions I have made in this series of “Reflections on the evolution of human language” into something of an overall scheme. The scheme (that I’ll sometimes refer to as a “hypothesis” despite its preliminary nature) attempts to explain how the potential for analytic processing could gradually have emerged within a system of holistic signals.

I’m aware of two main assumptions that I should make explicit to begin with.

First, I’m assuming that at our starting point there were some kind of discrete, repeatable signals that were known and used by the ancestral population. More specifically, the assumption is that communication took (or could fairly be interpreted as taking) the form of discrete acts (in other words, antecedents of what we today call “speech acts”), and that these communicative acts depended on discrete signals. I’ll refer to such signals as the “vehicles” of the communicative acts in which they are used.

I don’t know of any basis at this point for guessing whether these signals were mainly oral, or gestural, or some combination of the two, or for that matter something else entirely. However, in spite of this uncertainty about the physical nature of the signals at this beginning stage, I’m going to use the term “utterance” to refer to signals appropriate for being used as the vehicle of a single communicative act. Thus, “utterance” here refers to the signal alone, not to the act as a whole. Furthermore, an utterance is repeatable--i.e., it is a type rather than a token.

Second, I’m assuming that what the early utterances evolve into mainly is utterances. In other words, that (as argued most notably in Wray 1998, 2000, 2002) these original holistic utterances evolved mainly into other utterances that were now analyzable rather than into units of a different type such as words and morphemes. The hypothesis is that the evolution proceeded by a gradual analysis of utterances into smaller units that are capable of recombination.

There is another characteristic of (contemporary) language that should be pointed out as having influenced this scheme: the fact that much non-compositional meaning--what we might call “semantic clues”--is encoded into signs that must nevertheless be learned and used holistically. In Grace 1987 I discussed the fact that when we create new lexical items, we almost always design them to suggest something about the nature of the referent. The primary objective of these “characterizations”, as I labeled them, may fall anywhere on a scale from providing a simple aid to understanding to outright manipulation of the intended audience.

It is important to remember that the fact that these new items characterize the referent does not mean that they are compositional. Knowing the etymology of computer is not sufficient to know what a computer is. Linguistics generally suggests a field of interest having something to do with language but nothing more specific. The designation philosophy is surely still less informative, and one might easily have guessed that right-to-lifers were, say, members of a movement opposing capital punishment. All of these can be described as “lexicalized”. A dictionary is still needed even though all of these designations have obviously been chosen for the light they throw on their referents (or the light they want to put them in). Although this fact hasn’t received a great deal of attention, language as we know it is permeated with such semantic hints or suggestions.

In Grace 2004a, b, and Grace 2005a I suggested that somewhat analogous semantic clues, although much more primitive ones, would probably have arisen quite accidentally early in the evolution of language--as soon as the inventory of utterances grew sufficiently large. And it was in those pages that I first introduced the elements of the scheme that I will present here as to how such clues might have served as a stepping stone along the path. This scheme pictures the development of real re-combinable morphemes as the result of a slow process in which holistic utterances were subjected to more and more penetrating analysis.

The scheme consists of what may only be described as preliminary suggestions about how such analysis might gradually have come about. The suggestions all come from phenomena that are observable in language today but appear not to be dependent on any conditions peculiar to full-fledged language.

At our starting point the system would have consisted just of holistic “utterances”. There are two possible kinds of preliminary development at this stage. They might have proceeded simultaneously or in either order:

1a. Gradual recognition of the possibility of utterances being motivated (i.e., form providing clues as to meaning--at this stage conceivably nothing more than holistic form to holistic meaning).

1b. The emergence of patterns in the form of utterances, patterns that probably developed independently of any associated meaning (or grammatical function). That is, there were instances where a part of one utterance was identical to a part of other utterances. Furthermore, the number of such instances grew as the inventory of utterances increased, and the mnemonic advantage provided by these recurring partials alone was sufficient to permit a significant expansion of the repertoire of utterances (for more detail, see Grace 2005a).

I’ve allowed for the possibility that two potentials of signs (in this case, the utterances of the pre-language) might have been discovered independently: (1) the potential of recurrent partials of form to be identified and exploited, and (2) the potential of the form of an utterance to reveal clues as to its meaning. In this scenario, the possibility of an association between partials of form and partials of meaning would have remained to be discovered. In any event, this discovery would have been required at some early stage.

That stage consists of gradual recognition by speakers of originally fortuitous associations of particular parts of utterances with particular parts of the meanings of the utterances as a whole. (This, of course, requires that they recognize that the conceptual situation/event to which an utterance refers may be in some part dissectible). I have thought of two present-day phenomena whose possible antecedents could easily have suggested such association I’m imagining that this association developed only gradually, and that the first step was probably no more than the dimmest sense that the form of an utterance might sometimes give some clue as to its meaning.

2a. The first of these phenomena is contamination or blending. The users of the early utterances might have made slips of the kind suggested by today’s slips of the tongue, with some of these slips producing blends analogous to those of today (Grace 2004b). That is, sometimes an individual confronted with a particular situation would be faced with a momentary uncertainly as to what to call it--as to which of two utterances would best represent it. And occasionally s/he would come out with a blend of the two.

Such blends might certainly have provided a hint of the possibility of utterances being motivated, and of the meaning of some utterances being partially--or even entirely--deducible from their form. In short, the concept of inferring meaning by analyzing form was suggested. Once the concepts of motivated utterances and analytic processing became sufficiently well established, the path to the idea that it should be possible to compose such analyzable utterances deliberately would have been begun.

2b. The second phenomenon is what has been called phonaesthesia (see Grace 2004a). The hypothesis is that as the inventory of utterances grew to the point of producing recurring partials, this would inevitably have produced cases somewhat analogous to the phonaesthemes that are increasingly being recognized in languages today. This again could serve as an indication that form might provide clues to meaning.

3. Where form-meaning analogies of either of these kinds were perceived, they might have played some role as mnemonic aids--as reminders of signals (utterances). In fact, the attachment of meaning to form would make for a much more adaptable mnemonic aid than the recurrence of forms alone. The analogies might also have provided an occasional aid for language learning--i.e., for suggesting interpretations for utterances that hadn’t previously been encountered.

The analogies might also have come to provide a feeling for the relative aptness of proposed neologisms when such occurred. I have no suggestion about how neologisms were created at this stage, but if the inventory of utterances did in fact expand, then they were being created from time to time. Anyway, at least to the extent that similar perceptions were shared by multiple speakers, these perceptions might have influenced the shape of subsequent vocabulary (=utterance) innovations.

4. At this stage, some individuals might undertake to compose occasional neologisms to refer to previously unnamed situations. I've suggested phonaesthemes as the most likely candidates for such recombination--as the most likely forerunners of morphemes (and words). However, it also seems possible that the success of some accidental blends might have suggested the idea of producing deliberate blends. In fact, it's possible to imagine an accumulation of blends being a source of some phonaesthemes; it's also possible to imagine conscious or unconscious awareness of phonaesthemes as constituting a favorable condition for the occurrence of accidental blends.

5. Whatever the precise derivation of the neologisms, occasional ones were, according to this hypothesis, subsequently picked up and used by others--some to such an extent that they became part of the recognized inventory.

6. To continue with the hypothesis, the success of some neologisms led to the subsequent production of more. As the production increased, particular partials (whether phonaesthemes or not) came to be recognized as particularly apt for recombination. There is gradual elaboration of neologism-coining to the point that its possibilities are recognized and sometimes exploited by at least some members of the community.

7. At the same time, some particular patterns of combination began to take shape--for example, certain elements came to be used in initial position in the combinations, certain others in final position. At this point, the recurring utterance parts (it’s probably fair to call them “morphemes”) would still have been quite tightly bound to very restricted sets of environments. Therefore, the concepts of morphological description would have been better suited to describing its internal structure than those of syntax.(1)

8. This would have been followed by something like the following sequence: as time passed the possibilities of such recombination came increasingly to be exploited. As that happened, parts suitable for recombination were increasingly isolated and given conventional recognition as potential elements for combination--i.e., approaching or achieving the status of morphemes.

9. Eventually, the elements available for combination and the conventions governing their combination reached the point that they could be called a grammar. Likewise, the frequency of innovation increased to the point that a significant proportion of utterances had something ad hoc about them. At this point the analytic processing strategy had taken its place as a full-fledged partner of the holistic.

10. Of course, eventually an increasing division between syntax and morphology had to occur. I have nothing to suggest about how this came about except to note that Alison Wray has made an interesting suggestion as to a possible first step. She (Wray 2002: 123ff) describes how a kind of topic-comment structure might have been improvised by combining complete utterances in sequence.

Concluding Thoughts

There is one other point that needs to be made. The preceding sketch of the hypothesis has entirely omitted any mention of the mini-systems proposed in Grace 2005b and how they would fit into the scheme. As far as I can see now, they would have had little effect on the sequence outlined above. I imagine the various steps described here as occurring in parallel in the different systems. By that I mean that more than one system would have been presented with pretty much the same opportunity at the same time. Presumably there would have been many individuals who were involved in the use of more than one mini-system. Once some of them had recognized and exploited such an opportunity in one system, they would have been likely to realize that the same insight might be applicable in others. In short, I imagine all to have developed pretty much in parallel, with a particular step forward occurring somewhat earlier in some than in others, but being subsequently diffused.

But maybe I should say something about how such a process might have played out as far as the interplay between biological and cultural evolution is concerned. I don’t pretend to be qualified to say much about such matters, but it does seem safe to assume interplays such as this:

1. Cultural evolution: at a given state of biological evolution, some individuals discovered an advantageous new communicative trick and exploited it in one or more mini-systems.

2. Biological evolution: this new trick proved advantageous enough as to produce a significant change in the way the population adapted to its environment--effectively modifying their ecological niche. This in turn provided a new selective factor affecting subsequent biological evolution. Thereafter biological changes that improved their capacities for exploiting the new niche were favored (the so-called “Baldwin effect”).

3. Cultural evolution: after any such biological change, individuals again gradually discovered new ways of exploiting the new capacities. Note that somewhat parallel new tricks would presumably have occurred in more than one mini-system.

Of course, cultural evolution occurs at a much faster rate than biological evolution. The above scheme is misleading in that there must have been long series of cultural innovations between biological innovations (nevertheless, the latter ultimately attained very significant proportions).

At our starting point, at least some of these systems of communication were already more advanced than those of other related species, and this advance bespoke better-adapted (for their purposes) cognitive capacities. Our hypothesis holds that the advantages afforded by increasingly effective exploitation of the environment led to selection for further advances in the relevant cognitive capacities. In return, most or all of the mini-systems tended to adapt to the new capacities, and consequently tended to evolve in a convergent direction. This feed-back cycle continued until the mini-systems could be seen as having met to achieve a kind of unified system. The mini-systems themselves could thenceforth be dismissed (by those who were so inclined) as nothing more than vocabulary sets.

Finally, it is worth emphasizing that the explanation proposed here would not be possible if we conceived of a language as only an inventory of lexical items and a system of rules for combining them. It seems particularly important to keep our explorations of the evolution of language free from the confines of what Hopper (1988) called the A Priori Grammar Postulate. He defined this postulate as making “the initial assumption that a grammar is a discrete set of rules which are logically and mentally presupposed by discourse” (1988: 118). This assumes that discourse implies composition, but even today normal language use requires an enormous amount of non-compositional knowledge, and if we go back far enough in time, there was no other kind of knowledge. One of my intentions here has been to encourage the thought that compositional knowledge is not only a latecomer, but that it may have been only gradually that it succeeded in making a place for itself at all.