What comes before Language?
by David Foss
There appears to be, in the greater pan-linguistic community of human kind, a good deal of evidence that deep in the primordial foundations of language are a series of biologically endowed conceptual devices, which surface as grammatical regularities, and make possible the rapid acquisition of language on the basis of rather limited experience. Noam Chomsky has, on the basis of such regularities, and the mystery of language acquisition, suggested that, as homo sapiens, we are genetically equipped with a neural structure which predetermines techniques of conceptualization, methodologies in the construction of certain logical grammars, and the architecture of discursive life generally. It is tempting to think that the sort of linguistic equipment we begin with is in some manner conceptual, or is ultimately made up of concepts. While this must be true to some extent (at least insofar as the operands of these devices are themselves concepts), such a tact can be dangerously misleading.
The danger of calling these innate mechanisms concepts can be roughly indicated if we suppose that one such concept is red. For the sake of clarifying the distinction I wish to make, it will be helpful to check the compatibility of such a concept with the general view of Wilfrid Sellars, especially as it is laid out in Empiricism and the Philosophy of Mind. From Sellars’ perspective, the notion that sensory information is pre-linguistic, in the sense that it does not presuppose the acquisition of language, is nonsense. On its surface, the concept red just postulated would be utterly unpalatable to his view. That is, if by the concept red we mean some innate conceptualization of the chromatic region identified by the use of the term ‘red’. Admittedly, as we shall see, there are other (more flexible) ways in which something can be said to be a ‘concept’, but this is the most natural.
This concept of red, as a pre-linguistic concept shares all the characteristics of a linguistic concept except that it is 1) given to us by our innate biological endowment, and 2) lacks the linguistic tag “red” (or any linguistic context). In all other respects, this concept is indistinguishable from full blown linguistic concepts. In this form, I think it is clear that Sellars would reject it, or at least severely caution us to not take such a thing too seriously when seeking a basic unit of sense experience. In Sellars’ view, for any concept to bear cognitive relevance (let alone epistemic relevance), it must already be part of a wide array of linguistic concepts:
... while the process of acquiring the concept of green may — indeed does — involve a long history of acquiring piecemeal habits of response to various objects in various circumstances, there is an important sense in which one has no concept pertaining to the observable properties of physical objects in Space and Time unless one has them all — and, indeed, as we shall see, a great deal more besides.
Language learning for Sellars may be seen to move from a cognitive environment in which the first moves involve the demarcation of a continuous sense experience into manageable bits, rather than the more traditional view that we begin by joining distinct units of experience into complexes of these particulars.
... the process of teaching a child to use a language is that of teaching it to discriminate elements within a logical space of particulars, universals, facts, etc., of which it is already undiscriminatingly aware, and to associate these discriminated elements with verbal symbols.
The tools we use to ‘cut’ the stream of impressions that confront us during consciousness are probably hardwired in us. Indeed, there seems to be no likely alternative if we are to acquire langauge at all from such a beginning. But the sort of things which are innate would probably not be anything like the concept red. In fact, I’m not at all certain Chomsky (let alone Sellars) would consider a particular chromatic category as innate. What might be much closer to the sort of thing that would be built into our pre-linguistic apprehension of the world would be a very general ability to make rather arbitrary, albeit pragmatic, chromatic categorizations. But even here the faculty attributed to the pre-linguistic individual may be too powerful. While it is clear that language is not necessary to exhibit behaviors as though we could make distinctions of shape, color, texture, etc., (else how might we make sense of the non-deterministic behaviors of any non-human animal?) it is not entirely clear that the distinctions made are based on cognitive categorizations the sort of which tags like red will eventually come to individuate.
Vocabulary, especially with respect to color, does not merely identify pre-linguistic categories, but actually participates in defining these categories. Admittedly, there would have to be some sort of innate conceptual mechanism the terms of this vocabulary make use of, but the precise chromatic boundaries of each term seem to be set by their linguistic application. Or to use a more Sellarsian turn of phrase: the precise chromatic boundaries under which the application (or utterance) of a particular color-term, like red, will be appropriate, is established ad hoc during the learning of the language. I do not suppose this analysis is uncontroversial, but it is one I believe both Sellars and Chomsky would be inclined to agree with.
There might be the temptation to conclude from this that there is some sense to suggest that language development does presuppose the awareness of certain possible categories of immediate experience. Someone might ask, “Does this mean that the individual has some pre-linguistic apprehension of precisely what color she is looking at? Is the ambiguity of language (the indeterminacy of term boundaries), which is partially responsible for our initial rejection of foundationalism, simply the result of a fundamental (though in principle surmountable) difficulty in matching terms with such apprehensions?” If this were the case, epistemic certitude might be just a matter of ensuring a perfect match between terms of the language and such an awareness. This would, I think, be entirely unacceptable to Sellars. He would clearly insist that our simple apprehension of color (in terms of color matching across situations) is at least in part the result of linguistic training. This is not to say that there are no neurophysiological mechanisms which give somewhat precise indications of color, and which act as a sort of limit on the degree of perceptual precision we may obtain in distinguishing chromatic variation; but that our cognition of color gradations (and, more obviously, our cognition of color synonymy) will be an intimately linguistic affair. If this were not the case, and Sellars were to agree with this model, he would be justifiably accused of reintroducing sense-datum in their most pervasive form.
The notion of a concept, for Sellars, is almost always presented in its linguistic manifestation. If something in our heads were to be pre-linguistic, it would certainly not be a concept (or even conceptual) in his view. Talk of concepts seems to be generated by the effort of a fairly advanced linguistic community to construct a theoretical language capable of distinguishing self ascription from normal discourse. This is seen as the problem of inner episodes:
[How can there] be inner episodes — episodes, that is, which somehow combine privacy, in that each of us has privileged access to his own, with intersubjectivity, in that each of us can, in principle, know about the other’s.
I do not suppose this is always what Sellars has in mind by concept. In fact, I would venture to say that his view is quite compatible with the recognition of some sorts of proto-linguistic concepts, or more probably pre-linguistic conceptual mechanisms. These concepts would not be strictly linguistic in the sense that they could not in isolation cause or determine any linguistic manifestation. But they would be proto-linguistic in the sense that they are in fact necessary for the construction of linguistic concepts. The mortar and bricks of language use itself, they place limits on the sorts of linguistic architectures which are possible, but are neither sufficient for, nor entirely determining of, the specific language system they facilitate. Such things as these, whether they be called proto-linguistic concepts, pre-linguistic conceptual mechanisms, or some even more obtuse verbal formulation, are ultimately not only compatible with the view Sellars puts forward, but absolutely necessary for any view which, like his, rejects the existence of linguistic entities which presuppose no language mastery, and make up the basic units of pre-linguistic experience.
W.V.Quine, from a position very much like Sellars’ in general recognizes something like this, and emphasizes the distinction between such mechanisms and the language itself:
Language aptitude is innate; language learning, on the other hand, in which that aptitude is put to work, turns on intersubjectively observable features of human behavior and its environing circumstances, there being no innate language and no telepathy.
Quine calls this aptitude an innate ‘quality space’. Chomsky, among others, takes issue with the specific notion of ‘quality space’, but not with the motivation for suggesting such a cognitive tool. Exactly what these pre-linguistic mechanisms are, or even at what level they operate, has been cause for a great deal of disagreement, and is not likely to be settled for quite some time.
I should mention that, as I have been using the adjective, pre-linguistic (whether applied to apprehensions, concepts, experiences, or a posited stage of language development) has been intended to indicate that realm of human cognition which presupposes no language use, whether it be prior to the full acquisition of a language, or is simply causally prior to it. It might appear that such a class of cognitive activity is not entirely welcome in Sellars’ account. Certainly, he makes the claim (in response to the danger of a regress in terms of conceptualization necessarily presupposing prior conceptualization):
... one can have direct (non-inferential) knowledge of a past fact which one did not or even (as in the case envisaged) could not conceptualize at the time it was present.
The difficulty with this response is that it is far from obvious that one can retain pre-conceptual information at all. If a ‘fact’ is not in some deep manner incorporated into the cognitive fabric of an individual it does not seem possible that it could ever be accessed by her. This difficulty may be the result of not seeing Sellars make the needed distinction between linguistic concepts, and what I have called very roughly proto-linguistic concepts. Clearly, in this particular context he is referring to a condition under which the concept in question is not simply one borne of unconceptualized past experience, but is moving from the simple apprehension of historical concepts to the use of them as evidence (a different sort of conceptual transition which nowhere involves pre-conceptualized information). But it does seem that a general thesis is being proposed, especially in the case of language acquisition, that historical pre-conceptualized sensory data can be appealed to during the construction of linguistic conceptual entities. My contention is that any raw ‘sensory data’ which is going to be retained for later recall, must already be conceptualized. In order to avoid the regress, this process of conceptualization must at some point not presuppose the prior acquisition of language. And such a class of processes would necessarily involve (at least) a proto-linguistic conceptual schema into which retrievable data would be placed for later reference.
It is interesting to note that while pre-linguistic mechanisms must exist innately in order that language emerge from the flood of sensory experience we are born into, it’s not so clear that proto-linguistic conceptual schemas also exist innately. They are, I would want to say, the natural and inevitable result of the confrontation of pre-linguistic mechanisms with ‘raw sensory data’ (however this is to be understood pre-linguistically). But I would not want to commit myself to the rather untenable position that these schema are unchanging in terms of their internal mechanics.
Another interesting feature of proto-linguistic conceptual schemas is that they are not at all the sort of things which can help us decide whether a particular linguistic concept is correctly derived from experience. As a dynamic and shifting base, these schemas would form a very poor epistemic foundation, and indeed this would be quite contrary to their purpose. The realm of proto-linguistic entities which I have proposed, is not designed to account for language per se, but to account for its acquisition. Why is this important? This model is not en effort to describe how language comes to fit the world of individual experience, but the other way around. It is a model, like Sellars’ general approach, which attempts to describe how the world of individual experience comes to fit language. Proto-linguistic concepts play the role of providing us with the conceptual material out of which we carve language.
Would Sellars except some form of proto-linguistic concepts? I think he must, for the reasons mentioned. Without them the acquisition of language would be a strange affair indeed. But the danger still exists that these would be interpreted as capable of arbitrating linguistic claims (along the lines of ‘immediate experience’). The temptation is one which rests on a picture of language acquisition which is fundamentally passive. If it is imagined that language is learned through simple observation we might suppose that the process is one of matching some pre-linguistic vocabulary and grammar to that of the language we are learning. But this is not what I take to be the case. Indeed, language acquisition seems to be a fundamentally active process, in which the subject aggressively seeks out the boundaries of expressive behaviors, whether they be the utterances of terms, the formulation of syntactical structures, or the use of intonation.
This is not to say that when we learn language, what we are simply learning is a set of certain expressive behaviors and their respectively appropriate use. Chomsky rightly points out:
... when we learn a language we are not “learning sentences” or acquiring a “behavioral repertoire” through training. Rather, we somehow develop certain principles (unconscious, or course) that determine the form and meaning of indefinitely many sentences.
What we acquire is probably something like a system of rules and terms; a complex of conceptual processes governed loosely by a set of active mechanisms which drive and modify the system in an effort to match the linguistic community. Whatever the contents of pre-linguistic mechanisms or proto-linguistic concepts are, it seems clear that they could not be something as linguistically complicated and precise as the concept red. In spite of this color’s status as ‘primary’ in our optical theory, it is far from clear that its boundaries are not learned.