This week’s Chronicle is the continuation and conclusion of Chronicle 283, which critiques three attempts to explain the origin of language.

* * *

My fourth example is Bernard Comrie’s article “From potential to realization: an episode in the origin of language” (New Essays on the Origin of Language, edited by Jürgen Trabant and Sean Ward in the Trends in Linguistics series for Mouton de Gruyter [Berlin and New York, 2001]): 103-17. Less prudent than Bierwisch (see Chronicle 283), Comrie is ready to discuss the very originary moment in which our “language capacity” is actualized for the first time. He describes this moment as the manifestation of what he calls the “Dumbo factor.” At the outset Dumbo, Disney’s flying elephant, was able to fly, but was unaware of this ability. Then, as Comrie puts it, “when placed in a situation where his life depended on being able to fly, he made the attempt and flew” (103). We should note in passing the importance of life-threatening danger to actualizing Dumbo’s latent flying ability, which might be taken to suggest that Disney’s storywriter was closer than Comrie to an understanding of language origin. After discussing a number of cases where limited levels of input lead or do not lead to the emergence of language in children, Comrie makes his concluding argument:

We may now return to our basic question, namely: what level of input is necessary for language to arise? In the few cases where we can be reasonably certain that a normal child has been exposed to no input, language has not developed. Now, when language first originated this must have been the scenario, that is, the Dumbo Factor must have come into play: a creature that had a certain ability but was unaware that it had this ability must have become aware that it had the ability. We know of no clear modern instances of this happening, but this is not in itself evidence, since most children have the possibility of acquiring language under normal conditions and do so. At best we can say that we have no direct evidence of the Dumbo Factor coming into play, but circumstances make it unlikely that we would encounter such evidence. In one sense, then, our basic question of how humans came to know that they could communicate by means of language remains unanswered. (113-14)Without wishing to single out Comrie for ridicule, I think it is worth pointing out the inanity of this passage. Its fundamental argument is an obvious fallacy: (1) language use requires prior language capacity; (2) since capacity precedes use, there must be a time when one has the capacity without the use, and therefore without awareness of the capacity–the Dumbo Factor; (3) although normally individuals become aware of their capacity by contact with others already using it (i.e., speakers), there is no logical obstacle to actualizing it on one’s own when some external factor acts as a stimulus (“a situation where his life depended on being able to fly”); (4) clearly, in the first use of language, there were no prior users, so therefore (5) the origin of language can be explained by the Dumbo Factor. QED. It should be obvious to you, if not to the author, that this “argument” explains nothing at all, but is a pure artifact of the a priori distinction between capacity and use, potential and realization. Dumbo is a well chosen example, because real elephants can’t fly, and would have a great deal of difficulty in evolving the means to do so unless they were functionally employed in some intermediate state (for gliding, running faster, etc.), as wings are presumed to have evolved in birds. Exactly what it means for the Dumbo-protohuman to “become aware that it had the ability” to use language is left unspecified, since using language, unlike flying, is not something one generally does on one’s own. We would need at least two Dumbos together so that one might inexplicably begin to talk, and the other to understand and reply. Condillac proposed a far more convincing scenario of language origin between two children over 250 years ago.

The remainder of Comrie’s chapter falls back, in a similar if less self-aware manner than Bierwisch’s, on the idea that, since children will develop language once a vocabulary is supplied, we can get the lexicon of mature language from protolanguage, concluding with the following petitio principi:

. . . perhaps somewhat surprisingly, the main task in creating language seems to be providing the lexicon. Now, since protolanguage clearly has a lexicon, the early humans who had developed the ability to acquire human language but did not yet have a human language to acquire could in principle have simply taken off from whatever protolanguage they already knew and expanded it. If this scenario is correct, then while the provision of a lexicon is a task that does not in itself require the linguistic ability of humans, it is nonetheless a crucial catalyst for the realization of that ability. (114)In other words, “somewhat surprisingly,” there is no paradox, no problem at all. Protohumans acquire protolanguage, which subsequently evolves into human language, and presto, they have become human, as easily as elephants learn to fly.

My last pair of examples comes from a third recent collection, The Transition to Language, edited by Alison Wray at Oxford University Press in 2000.

Wray’s own article in this collection is entitled “Dual Processing in Protolanguage: Performance without Competence” (pp 113-37). She begins from the observation that language, particularly the language of everyday communication, is largely formulaic: “A striking proportion of formulaic expressions are used to manipulate others into physical, emotional, and perceptual reactions” (114). The next step in her reasoning is to remark that

the manipulative functions of our formulaic language correspond closely with those observed in the communicative behaviour of chimpanzees in the wild . . . On the basis of this parallel . . . I have proposed . . . that the holistic strategy for expressing manipulative messages in phonetic form may be considerably more ancient than the analytic strategy [i.e., the divisibility of the message into meaningful morphemes]. I have suggested that the holistic cries and gestures of our pre-human ancestors were transformed, over a long period of time, into a phonetically expressed set of holistic message strings, each with a manipulative function such as greeting, warning, commanding, threatening, requesting, appeasing, etc. (115)The reference to a “long period of time” should arouse our suspicion; it is used to glide over the passage from a fixed set of genetically determined calls with little learned variability to an unbounded set of unfixed “strings”; Wray’s strings have every appearance of being what Saussure calls arbitrary linguistic “signifiers.” There is not even an attempt at explaining how or why this passage from hard-wired cries and calls to protolinguistic software is accomplished.

In Wray’s hypothesis, these strings cannot be analyzed into component parts. But since the strings can multiply indefinitely, at a certain time (again determined without reference to any factors external to the individual protohuman brain) these cries/calls/strings stop being purely generic and begin to accommodate specific reference.

This [accommodation of individual denotation] is what I suggest came to pass at some point during the protolanguage period. The impetus could have been a small but significant increase in memory capacity, a cognitive change that made irresistible the expression of new connections between ideas, and/or environmental pressure to extend the scope of the protolanguage system. In all cases, the consequence would be a minor–but not a major–revolution in expression, opening the way to an analytic system, though not yet to syntax. (123)Thus everything that needs to be explained is attributed, first to a “long period of time,” next to “a small but significant increase,” producing a “minor–but not a major–revolution in expression.” One wonders what is wrong with a large increase and a major revolution; but the real intellectual operation being performed here, of which the author seems blissfully unaware, is to generate a structure of transcendence without anyone taking notice of it, and anything large would be likely to be noticed, if only by the reader.

Wray then goes on to propose a speculative hypothesis of how words came to be formed, that is, how the holistic strings of protospeech became analyzable into component parts. The following will give you the flavor of her speculations, which an uninitiated observer might call unscientific, or something worse:

In order to see how the first individual words could emerge, let us imagine that the frequency with which “fetch” messages for individuals are needed is sufficient to support individual designation, and that the system can handle it. The message inventory will include a separate message form for each of [what I say when I want you to] fetch N . . . In other words, there is a distinct phonetic form associated with each individual in the group, for the context of “fetch that person.” . . .Let us next imagine that a speaker wishes to get a hearer to give an apple to Mary, but, though the apple is in plain view, Mary is not present and so cannot be indicated by gesture. The speaker resolves the problem by saying first baku, “[what I say when I want you to] fetch Mary,” and then tebima, “[what I say when I want you to] give specified object (distant) to specified female person.” Since only one of the two variable referents (the apple) has been indicated by gesture, the hearer is seeking local information that will indicate who the recipient of the apple is. Provided baku is viewed as somehow relevant to this quest, the effect (whether on this first occasion or only after repeated usage) will be to interpret “[what I say when I want you to] fetch Mary” as a perceptual rather than physical manipulator (“conjure up an image of Mary”) or indeed, simply to understand baku as “what I say when I mean Mary,” . . . a straightforward referential name. (124)

I hasten to point out that baku and tebima are fantasy-words with no empirical basis; they are, as Pooh-Bah put it, “corroborative detail intended to give artistic verisimilitude to an otherwise bald and unconvincing narrative.” It would take me too long to point out everything that is wrong with this passage; let it suffice to say that this convoluted just-so story is designed to take the last step in the creation of the linguistic signifier, the arbitrary sign. Instead of beginning with the designation of something specific, a point of departure that was intuitively obvious to eighteenth-century writers such as Maupertuis or Adam Smith, Wray makes this her last and most difficult step, seemingly without realizing that all the rest of her hypothesis is a series of petitiones principii.

The reason why these perverse protohumans are said to create separate strings for “fetch Mary” and “fetch Rosalie” before creating the names “Mary” and “Rosalie” is that their language, by remaining “holistic,” remains on the other side of the Rubicon that separates the human from the animal. Into this pre- or protohuman “holistic” language, Wray smuggles the essential element of symbolic language, notably arbitrariness, without which the indefinite multiplication of “strings” would be impossible. Then we are told that a specific set of messages–those for “fetching”–become so frequent that a different one is created for each member of the group. It boggles the mind, or mine at any rate, to conceive that a community might have different calls for each member of a group without a concept of reference to that individual, and that the reference to the individual would be derived only from some combination of these calls.

Wray’s text is an extreme yet sadly typical example of the conceptual gradualism that complicates without in any way resolving the problem of the origin of human language by breaking it down into a series of arbitrary, speculative steps, each one of which presupposes precisely the symbolic relation between sign and referent that it purports to explain. The specialists, who not merely tolerate but actively encourage absurdities of the sort we have just witnessed, reject as unthinkable the notion that the community generates language by designating certain objects as sacred in order to preserve itself from the contagion of mimetic desire. The taboo against the sacred object is so powerful that even individual members of the group can be referenced only through the mediation of a call, “fetch Mary,” which we can somehow conceive before realizing that Mary is the object to be fetched. “Fetch Mary” is what I referred in to The Origin of Language (1981) as a nominal imperative, on the model of the surgeon calling “scalpel!” to his assistant. But before one can have an imperative, one must have an ostensive; before we can “fetch Mary,” we must have an ostensive association between this word/string and Mary herself. The “string” that we are using to fetch Mary must already designate or represent Mary; and for the community to represent Mary independently of the desire to remedy her absence, her presence must already be of particular significance. Reasoning along this line will lead us to the originary hypothesis, which confronts directly the origin of the symbolic sign and the transcendence it brings with it.

Baku and tebima aside, Wray’s theory is similar to Bierwisch’s; both rely on the growth of the lexicon to explain the evolution of linguistic structure, whether morphological (Bierwisch) or, in the present case, simply lexical in the strong sense, that is, referential. All the smoke and mirrors in Wray’s exposition serve only to hide the key difficulty, which Bierwisch is at least willing to admit he has not solved, of, reference aside, why or how there emerge arbitrary signifiers, not genetically linked to their function and capable of being multiplied indefinitely.

My final example is drawn from the same anthology: Derek Bickerton’s “Foraging versus Social Intelligence” (pp 207-25). Bickerton first proposed the idea of protolanguage in Roots of Language (1981) on the basis of his study of “pidgins,” simplified versions of languages spoken by non-native speakers that lack complex morphology or syntactic rules, and that evolve into “creoles” when they become the native languages of children brought up in these cultures. Bickerton’s argument (which has been contested by other specialists) is that creoles all share the same simple syntax, which cannot be derived from the more complex languages from which the creole is constructed, and may therefore be attributed ex hypothesi to an underlying “protolinguistic” mental substrate. The notion of protolanguage, the application of which is extended to such cases as apes who are taught versions of human language, appears to offer a bridge between pre-human modes of communication and language proper. It does not, however, as no intermediary concept can, answer the essential question of the origin of symbolic communication, the passage from animal signals to human signs.

In the article under discussion, Bickerton takes issue with what he refers to as a trend over the past few years of considering language to have arisen “as a direct result of increased and intensified social interaction”; the emergence of language would reflect the increased level of “social intelligence” required to organize an increasingly complex society. This hypothesis understands language as a means rather of social control than of amplifying practical cognition. To quote Bickerton:

[T]his chapter argues that the initial impetus for a [human] means of “information donation” quite distinct from means employed by other primates–that is, some form of protolanguage–arose directly from the requirements of group foraging, predator avoidance, and instruction of the young, rather than from specifically social interactions between individuals . . . (209)Bickerton’s main argument is this: given that human language is absolutely different from the communication systems of other primates, whereas the higher apes too have complex social relations, the uniqueness of language cannot be explained by the unique complexity of human social relations. Whatever selective pressures might have driven the development of language in humans, these same pressures, in lesser form, would have been present in our ape cousins; yet whereas our language is highly complex, “no other species has developed language at all” (209). I need not elaborate on how generative anthropology deals with this problem; what drives the emergence of language is not a mere increase in “complexity,” but an increase in mimetic capacity that makes animal forms of communication inadequate to prevent the breakdown of the social order in mimetic violence.

Bickerton goes on to make other points. He rejects Wray’s hypothesis examined above on the grounds that “it seems likelier that the original symbols consisted of single units with single meaning–meanings, moreover, that could somehow be ostensibly demonstrated from the immediate environment” (216). He also alludes to the “cheapness of tokens,” their ease of emission that makes cheating particularly easy. These are points highly congenial to the originary hypothesis: the ostensive use of the sign comes not from a proleptic need to show what it refers to, but from a need to represent-interdict a present object of desire; the cheapness of the originary token is not accompanied by “cheating” because the whole point of the sign is to deny the possibility of cheating by making the centrality and therefore the sacrality of the object of desire the first object of human consciousness.

Bickerton’s own explanation for the emergence of protolanguage is as follows:

. . . there is abundant evidence that hominid ecology did differ from that of other primates: our ancestors lived on open savannahs or in marginal woodlands rather than in deep forest, and they were primarily terrestrial rather than primarily arboreal. An even sharper distinction relates to an ecological niche occupied from the time of the Olduwan industry (approx. 2.3 million years ago) onwards. . . Hominids became principally scavengers and were able to compete with other scavengers and predators for the carcasses of large animals across a wide range of habitats. . . .

How did hominids succeed against fierce competition? Not, surely, by wandering round like a troop of baboons and eating what they happened to stumble on. They must have been able to (1) locate fresh carcasses with extreme certainty and rapidity, and (2) fight off competitors, probably with barrages of flung rocks. . . .

In such a context, the crudest beginnings of some form of language would have paid off from day one. Any hominid group capable of discriminating food sources (and perhaps also of indicating the relative dangers involved in their exploitation) would have enjoyed an advantage over other hominid groups. Note that the first linguistic communications need not have been monomodal, nor need their units have been arbitrary in the Saussurean sense. Directional gesturing with the hand, accompanied by the imitation of the noise made by a mammoth, could easily have been interpreted as meaning “Come this way, there’s a dead mammoth” . . . (218-19)

The key giveaway in this just-so story, which explains the origin of language by its cognitive-practical use, is found at the end of the first sentence of the third paragraph quoted above: language “would have paid off from day one.” Why is this expression there in the first place? Because the main difficulty of deriving language from its practical application is that language can have no practical application unless it already exists. “Day one” is a metaphorical way to posit a moment of origin without needing to define it. The originary hypothesis truly has something to say about “day one,” but Bickerton does not; there is no single moment when what he formerly called the “Rubicon” can be crossed, and his “ecology-based” theory is no advance whatever on the many earlier attempts to make the obvious usefulness of language the explanation for its emergence.

Texts like these, written and compiled by leading scholars in the field and published by respectable houses, are absolutely typical; I could have cited dozens more. In one sense, they make my task terribly easy, since their inadequacy as explanations of language origin is so patent that it can be exposed merely by reading them aloud. In another sense, however, these texts make my task very painful, since they bear witness, over forty years after Girard’s theory of mimetic desire and over twenty years after my hypothesis of the origin of language, to the fact that these ideas have not even brushed against the consciousness of the “serious” investigators of this issue.

I think I understand why this is the case; the idea that the origin of language, the very birth of the human, is something simple because cultural that some amateur like myself, inspired by another amateur like Girard, can hypothesize about, goes against the grain of the whole university-based structure of knowledge, the only alternative to which is the intellectual sub-stratum of cults and cranks. To put it differently, the origin of language is not the kind of problem that social scientists are used to solving, or able to solve. One might wish to call it an a priori ontological problem, but that would risk situating it in the pre-existent category of metaphysics. No, it is an anthropological problem, yet one of a sort that the science of anthropology has long abandoned. It is a trans-departmental problem that can only be approached by trans-departmental thinking.

* * *

We must dare to have faith in our ability to think the world and to experience its truths, so long as we are humble enough, and confident enough, to listen to the arguments of others. Non-specialist though I may be, I will continue to believe that the originary hypothesis is the minimal explanation of the human until someone shows me a reason not to. I hope you will agree that none of the examples discussed above offer much of a threat.

(Part II of a slightly modified version of a lecture delivered at Westminster College, Salt Lake City, in March 2003)