language as a cultural emergent property
The empirical, the actual and the real
The real
In this kind of ontology, 'real' phenomena are generative mechanisms underlying empirically observable phenomena. Thus Chomsky is often identified as a linguistic realist because of his concern with the hidden operational code which makes instances of language behavior possible. There are, of course, a number of challenges to this approach, some of them presented by other realists. Carr (1990), for example, holds to a fairly traditional notion of what it is that linguistics is about. Its evidential base'... is something other than a collection of observed events' (p.30), and what linguists do is attempt'... to describe the mechanisms in the underlying linguistic reality which allow us to characterize these expressions [sets of sentences] as well-formed' (ibid.: 33). Carr, however, takes issue with the insistence in the Chomskyan tradition that 'linguistic objects and states of affairs are speaker-internal and ... theoretical linguistics is in some sense a branch of cognitive psychology' (ibid.: 3). This is because of his commitment to the intersubjective nature of linguistic reality. Drawing heavily on Popper's conception of intersubjectivity, Carr maintains that '[u]nder this ontological interpretation of the generative enterprise, we may say that a language, constituted by its rules, is a public object' (ibid.: 42).
The ontological status of linguistic 'rules' is a matter of continuing debate, and some linguistic realists part company with the Chomskyan tradition over a perceived conflation between two different kinds of 'rule', leading to different positions about the ontological status of unobserved 'realities'. Devitt and Sterelny (1999), for example, are skeptical of the claim that speakers must be possessed of propositional knowledge, represented in the mind and therefore psychologically real, as grammatical rules which are isomorphic with the rules formulated by linguists. Devitt and Sterelny are concerned not to dismiss the empirically observable reality of language, and to accommodate linguistic symbols as observable outputs into their description of language, as 'datable, placeable parts of the physical world: sounds in the air, marks on the page, and so on. They are not mental entities at all' (p.169). The detail of their argument about the ' ... two sorts of rules, the "structure-rules" governing the products of a competence, and the "processing-rules" governing the production of those products, rules governing the exercise of the competence' (ibid.: 171) is not our main concern. However, their position does illustrate that contemporary interpretations of realist philosophy in linguistics allow a greater role for 'objects or events in the physical world "outside the head" , (ibid.: 169)than is associated with the Chomskyan position. Thus, although Devitt and Sterelny take a somewhat different line from Carr (to whose work they do not refer), they too conceptualize linguistic realities as having an existence which is external to, and separate from, individual psychological realities.
One of the differences between Carr's position and that of Devitt and Sterelny is that, despite the commitment of both to the inter subjective properties of linguistic reality, the former is much less ready to concede the relevance of spatia-temporal events to the study of language. A realism which is so theory-driven that it effectively precludes the description of data is not one we would want to endorse. A critic of generative linguistics who makes the case well is Beaugrande:
A paradoxical scenario arose: reality being claimed for a technical construction while rejecting the reality of human discourse inclusive theory with exclusive practice ... . And instead of becoming a genuine normal science, generativist linguistics could only be a normalizing quasi-science that cannot describe authentic data but only the data it has expressly normalized, thereby rendering them empirically undecidable. (Beaugrande 1997: 35)
Carr maintains a commitment- which, significantly, we do not share to falsifiable hypotheses as a basic criterion of science. It is crucial for him, then, that linguists have a methodological tool which is at least consistent, if not identical, with the methods used to test hypotheses in other branches of science. Carr's approach adopts grammaticality judgements as its data, or evidential base, and adheres to a belief in the primacy of sentences ('abstract objects which do not exist in a context', 1990: 43) over utterances (spatia-temporal events). Utterances, what speakers actually say, with their notorious 'imperfections', exist, as we have explained, at the level of the empirical. Carr's position here is analogous to that of the realist social theorists Keat and Urry (19S2), whose argument, we feel, runs the risk of placing so much emphasis on the unobservable layer of reality that the empirical is demoted out of the picture. Carr defends his position by an appeal to 'the knowledge which allows us to make this crucial distinction'- that is, that between well-formed and ill-formed expressions. On the basis of that knowledge, he claims, 'we are able to dispense with much of what is given in a corpus; it is clear that corpora play little or no part in strictly theoretical linguistics' (Carr 1990: 99). Devitt and Sterelny, by contrast, give much greater weight to speakers' meanings. It is these, they say, that '... create the conventional written and spoken forms of the language. But it is because we have learned those conventions that we are able to have the rich variety of thoughts, and hence produce the rich variety of speaker meanings, that we do' (Devitt and Sterelny 1999: 155).
The domain of the real, then, in relation to language, includes both the mechanisms internal to the human organism which make language processing possible and those properties and powers of language which enable intersubjective communication. The significance of this second property of language will be considered more extensively below. At this point, however, we simply want to emphasize the claim that what is real about language includes its intersubjective as well as its intrasubjective properties.
