The Origin of Language Page 10

  In certain social contexts, this mediation may be taken to imply the existence of an asymmetric authority relation that transcends the immediate speech situation. But it is important to explain why this need not be the case. It is the situation of linguistic presence itself, the evocation of the scene of representation, that is the original source of the “authority” of the imperative. To be a participant in this situation, the hearer must defer his attention to worldly tasks in order to attend to the intentional model constructed by the speaker. In the imperative, the speaker takes advantage of this attention in order to extrapolate his linguistic intention into a worldly one aimed at the appropriation of its object. The ambiguity of the word “intention” here is not coincidental, nor is the instrumental nature of the imperative that exploits it. This identity of linguistic and practical intention resolves the paradox arrived at in the dialectic of the ostensive and thus produces a stable linguistic form, although, as we shall see, the use of this form will give rise to a paradoxical situation of a different sort.

  The originary ostensive would have been maintained until it was clear that no individual would attempt to appropriate the object, that is, that it was sacred to all. But in the profane use of the ostensive, the end of the utterance is the end of the linguistic scene, which would presumably give way to actions not dictated by the utterance itself. In contrast, the felicitous imperative’s extrapolation of linguistic into practical intentionality prolongs the interlocutors’ virtual presence on the linguistic scene until the assigned task is carried out, independently of other worldly claims. The deferral constitutive of linguistic presence now becomes the awaiting of an anticipated action.

  Grammatical Form

  Our discussion up to this point, concerned exclusively with intentional structure, has maintained the assumption that the primitive form of the imperative was substantially identical to that of the ostensive. As we know, however, although the nominal imperative continues to exist, the imperative in developed languages has a verbal form. We have seen that the temporality of the imperative, that is, its tense, is the prolongation of the linguistic scene in awaiting. The time of awaiting is both real, lived time standing outside the scene stricto sensu and a prolongation of the presence intended by the utterance. Thus the imperative includes within itself a model of a time other than that of its moment of utterance. We should contrast this with the simple identity of linguistic and real time in the ostensive, where the time of linguistic presence remains, as in the originary event, merely the time of deferral of action while attending to the speaker. The ostensive model has no temporal dimension; the word and its referent coexist in the same suspended present. The temporality of the imperative, although not yet a true tense independent of the scene of communication, like that of the declarative, is if not a temporal mapping of reality on language, already a mapping of language on reality. The hearer of the ostensive can immediately verify its informational content for himself, and so to speak discard the linguistic model that conveyed it; the hearer of the imperative must retain the model as a guide for his conduct, “verifying” it only upon the conclusion of his performance.

  Whether the imperative takes on a nominal or a verbal form, the anticipated result of the imperative is an action by the hearer, so that the verbal form provides a more explicit model of the action. This is not true merely in the trivial sense that “bring the hammer!” or “Give me the hammer!” is more explicit than “Hammer!” If we compare “Hammer!” (conceived as a nominal) with “Run!” (considered as a verbal substantive rather than a verb), the former requires the performance by the hearer of actions not explicitly stated, whereas the latter is a fully explicit instruction; the run, unlike the hammer, is wholly under the control of the hearer. This difference between nominals and verbals was not visible in the ostensive, where in either case the significant phenomenon was merely a thing/event present to the speaker and potentially to the hearer. In the imperative, the hearer can perform an action, but can only supply an object. This divergent relationship to nominals and verbals in effect inaugurates the governance of the former by the latter, although, as with tense, not yet in the fully realized model of the declarative, which presents the relationship of agent, verb, and object independently of any worldly action.

  It is nevertheless this action that permits us to explain governance from a generative perspective. It is perfectly conceivable that “double” or even “triple” ostensives may have existed, consisting of a verbal and one or more nominals; for example, on observing a flight of crows, “Fly! Crows!” or even something like “Burn! Fire! House!” But it would be an error to consider such utterances as true linguistic forms, and thereby as the direct ancestors of the declarative. The elements of such compound utterances would remain independent as potentially complete utterances in themselves, whereas the declarative sentence is not complete until all its positions have been filled.

  The fact is that no governance relation, even the inchoate one of the imperative, is conceivable in ostensive language. “Fly! Crows!” is no more a “sentence” than “Fly! Sky!” or “Sky! Crows!” or indeed any other combination of ostensives. Governance, which is a relation between linguistic elements of an utterance, as opposed to the intentional relationship (expressed in tense and person as well as in the reference of substantive words) of the utterance to reality, cannot be derived from the mere observation of relationships in nature, but only from the significant functioning of these relationships on the scene of representation. In the ostensive model, this interaction ends when it points out a significant phenomenon in the real world. The imperative model, however, includes not the mere presence of its object but the relationship to be assumed toward it by the hearer. If a compound ostensive like “Fly! Crows!” presents two independent observations, a compound imperative like “Bring [the, a] hammer!” or, if the notion of “bringing” be thought to beg the question, “Come! [with] hammer!” requires of its hearer not a contemplative analysis of the relation of its elements but a performance in which the referents of the elements are combined.

  We need not suppose the analytic counterpart of this practical operation to be present in the mind of the hearer. It suffices that his performance be more explicitly determined in its verbal than its nominal aspect, so that, in the example at hand, the hammer cannot be provided without the hearer’s coming to the speaker, although the action of coming can be performed without the hammer. And in general, the verbal element will be performed by the hearer as agent, accompanied in some way by the nominal element as object or instrument. Thus although a bystander could describe the performance in our example in an ostensive coming! hammer! analogous to fly! crows!, the performer of the act himself could not be unaware that his coming was “governing” the appearance of the hammer. The phenomena described by the ostensive, whether or not they involve human agents, are independent of the linguistic model that refers to them, whereas the imperative specifically requests a human performance.

  Similarly, the extension of the scene of representation created by the imperative to the awaiting of a requested performance adds the notion of tense to the verbal element of the imperative. As a result, this element acquires in its most rudimentary form the character of a verb. In the nominal imperative, the requested object is merely made present, but in the performance requested by the verbal imperative (Run!), the object created in order to put an end to the awaiting is defined within the temporality of the imperative speech-act. The same hammer may be requested today as yesterday, but not the same run.

  Here again, the contrast with the ostensive brings out the increased grammaticality of the imperative model: the ostensive, concerned only with the presence of its object, is indifferent to its temporal specificity. The “run” observed today is no more different from that observed yesterday than today’s hammer differs from yesterday’s hammer. The function of the ostensive is to designate phenomena of general significance, and significant objects are always “the same” objec
t because their appearance leads to functionally identical situations. “Fire!” always warns us of “the same” fire. But whereas the ostensive is unconcerned with the distinction between the identity of phenomena and their repetition, the imperative is not, because it requests its hearer to present the identical and/or reproduce the repeatable within linguistic time. Although the imperative does not yet permit the distinction among different tenses, it does respect the more fundamental distinction between the verbal, which possesses a tense, and the nominal, which does not.

  The genesis of the notion of person follows similar lines, although in contrast to that of tense, it can undergo internal differentiation in the context of the imperative model. The verbal imperative is personalized even in its basic “second-person” form because, again in contrast to the nominal, it requests an action to be performed, and thus made to exist, by the hearer. Just as we have seen that the “run” requested is a “run now,” so we may say that it is also a run-by-X, which is by no means identical to a run-by-Y. And as in the case of governance, the specificity of action on the part of the hearer of the verbal imperative may be presumed to be included in the intention of the speaker. Thus if several hearers are present and the speaker requests a hammer (Hammer!), the intentional model includes only the hammer. Even if one person is specifically addressed, this intentional structure is not violated if someone else brings the hammer, although the speaker’s expectations may be. But if he says “Come!” to one of the group, then the coming he is requesting could not normally be performed by any other.

  Now at this point “person” simply means second person, the contrast with the first person not having any basis in the intentional structure, the third being for the moment undefined. The speaker is normally at least the “dative” object of the imperative, and he may on occasion be its “accusative” object, as in a request for help or other personal services. But although personal “shifter” pronouns must have been among the first words, each individual being obliged to refer to himself or to the other by means of symmetrically “shifting” gestures, even as the accusative object of an imperative verb, the speaker is never in symmetry with the hearer. The performance requested of the hearer implies no contrast with one by the speaker.

  * * *

  Once the imperative acquires functionality in the profane world, it is easy enough to conceive how its third- and first-person-plural (Let’s) forms might have evolved. The use of language to assign tasks would naturally enough be extended to other parties, including the speaker, the principle being that an intentional structure that names an object on which work must be done, whether to make it present or to construct it from scratch, would (like Wittgenstein’s language games) not remain limited to the dialogue of two interlocutors. However, the paradigm in which all three persons play parallel roles as subjects of a verb cannot exist prior to the declarative.

  Chapter 9.

  Elementary Forms and Grammatical Structures

  The preceding discussion has shown that the requirement of human performance in the imperative is the source of the categories of tense, person, and governance that will become the touchstones of the grammaticality of the declarative sentence. By grammaticality we simply mean an utterance’s capacity for presenting a model of reality no longer dependent on the conditions obtaining during the communication situation, having corrected by means of “shifting” elements chosen from these categories for the specificity of time and speaker.

  But because we cannot discuss these categories in terms of a hypothetical imperative language, our presentation has unavoidably been oriented teleologically toward the declarative. Tense and person can already be associated with the verbal imperative, but they remain context-bound as they will not be in the declarative, without which they could not have emerged as fully grammatical concepts.

  The categories of person, tense, and governance (case) are precisely those that give rise to the most familiar paradigms of the declarative and its related modes, as a glance at a Latin or Greek grammar will show. Even in languages with few or no such paradigms, these categories must be marked in other ways. For they are constitutive of the notion of grammaticality itself, and the declarative sentence is inconceivable in their absence. Thus the intentional structure of the declarative not only admits of the possibility that oppositions between persons, tenses, and so on, will generate paradigmatic combinations in a given language, but, regardless of the morphological means employed, the declarative sentence presents a model that is situated with regard to the time of utterance and to the speaker and his audience, and where the relations among nominals and verbals are, within certain limits, specified. The constitution of this intentional structure, which we shall take up in the following chapter, is the result of a dialectical process generated by the internal contradictions of the imperative.

  Although if we would explain the genesis of the declarative from the imperative, it is heuristically useful to assume the inchoate existence of grammatical categories in imperative language, the intentional structure of the imperative does not require their specification. The imperative model minimally requires only the definition of the phenomenon whose presence is requested. The time of the action is not a true tense, but an extension of linguistic presence; similarly, the “governance” of the passive object by the active performer is a matter of practicality, not truly of grammaticality. It is senseless to speak of grammatically “correct” and “incorrect” imperative utterances, because the only relevant criterion, assuming that the addressee accepts his role in the intentional structure, is whether or not the task itself is well-defined in context. We may as well assume that the dominance in mature language of the more explicit verbal over the nominal form of the imperative was already present in imperative language. But the greater explicitness of the verbal form would not make it more correct.

  The category of person appears at first to constitute an exception because, in contrast with that of tense, it can be said to possess a true paradigm in the opposition between the second- and third-person imperatives. But for proof that this paradigm is not essential to the imperative, we need look no further than our own language, where the third- (and first-) person imperatives employ periphrastic constructions, in contrast to the second-person form, which simply uses the root form of the verb. (Even in ancient Greek, where the third-person imperative is classified as part of a paradigm, it contains a true ending [–(ε)τω] in contrast with the “zero” ending [–ε] of the second-person form.) But morphological evidence aside, the question of whether the basic form of the imperative is truly a “second-person” form involving an implicit opposition to third- and/or first-person forms must be answered in the negative. It is the real-world person, not the linguistic “person,” who is the true subject of the imperative.

  In a word, the elementary linguistic forms lack true grammatical structures because they are not yet fully emancipated from their origin in the original crisis. The scene of representation is still primarily a place of deferral of conflict, not contemplation of a model of reality. The imperative and ostensive are pragmatic, not theoretical, which is to say that the linguistic present internal to the utterances is not fully separated from the linguistic presence in which they are uttered. Thus although the imperative takes different forms, these can never be grasped as paradigms of possibilities inherent in the imperative intention; their use merely corresponds to different real-world situations.

  Language at this stage is anything but instinctual, but there is a sense in which behavioral models still apply: each word is still “associated” directly with the real or desired appearance of its object. Thus not only an ape but even a rat could be trained to “speak” in elementary language by pressing one lever when a cat appears and another when it is hungry, the two levers being connected to a mechanical voice which would produce, respectively, the ostensive “Cat!” and the imperative “Food!”

  Such models of human language, because they neglect the crucial element
of scenic presence, are etiologically inadequate to explain the elementary forms, and incapable of even conceiving an explanation for the higher forms. But so long as we confine our analysis to the practical functioning of the imperative, we will not touch on its contradictions and the forms generated in response to them. Our assertion that the proto-grammatical developments to which we have referred do not make use of true grammatical categories is in effect equivalent to saying that they could à la rigueur be described as accretions to the original ostensive arrived at under the pressure of “conditioned reflexes,” that is, by mere trial and error. Yet conditioned reflexes lead to insoluble pragmatic paradoxes that are incapable of provoking formal evolution.

  Whence the significance of the absence of a first-person form at this stage. If utterances are to be explained as resulting from “association” with the presence of their objects, then the self, being by definition always present, must either be spoken of constantly or not at all. The most elementary form of the recursivity that Chomsky sees, rightly if taken in a broad enough sense, as the mark of human language in opposition to pre-human signaling systems, is simply the speaker’s linguistic reference to himself.

  Dialectic of the Imperative (I)

  The intentional structure of the imperative has up to this point been presented as a structure in equilibrium: a verbal request establishes an awaiting of performance by its hearer, compliance with which abolishes the awaiting and terminates the prolonged presence that it maintained. This indeed constitutes the felicitous performance of the imperative. Because, however, this felicitousness requires the action of someone other than the speaker, it cannot be predicated of the utterance alone, whose constitutive expectations, however reasonable, may not be fulfilled by the addressee. The imperative utterance is not complete in itself. The asymmetrical positions of its speaker and hearer are not simply those of the speech situation; the hearer must, within the linguistic presence created by the speaker, not only hear but act. It is the contradiction implicit in this asymmetry that will lead to the creation of the “objective,” information-bearing declarative form.