Copyright © 1999 Massachusetts Institute of Technology
As with any rapidly developing science, GENERATIVE GRAMMAR has given rise to a number of interesting philosophical puzzles and controversies. These controversies range from disputes about the object of study in linguistics, to issues about the relation between the language faculty and the external world, to questions about the legitimacy of appeal to rules and representations, to questions about proper empirical methodology.
One of the prominent philosophical debates in generative grammar has centered around the question of what sort of objects languages and grammars are. Katz (1985) distinguishes three general approaches to the question, roughly paralleling three traditional approaches to the nature of abstract objects: platonism, conceptualism, and nominalism. The platonist view would take the object of study in linguistics to be an abstract mathematical object outside of space and time, the conceptualist position would be a position like Chomsky's in which the object of study is a mental object of some form, and the nominalist view would hold that the object of study is a corpus of inscriptions or utterances.
The platonist view has been advanced most visibly by Katz (1981), although it may be that the position rests on a confusion. For example, Higginbotham (1983) has observed that even if grammars are abstract objects, there is still the empirical question of which grammar a particular agent is employing. George (1989) has further clarified the issue, holding that we need to distinguish between a grammar, which is the abstract object that we know, a psycho-grammar, which is the cognitive state that constitutes our knowledge of the grammar, and a physio-grammar, which is the physical manifestation of the psycho-grammar in the brain. If this picture is right, then the platonist/conceptualist dispute in linguistics may be trading on a failure to distinguish between grammars and psycho-grammars.
Perhaps more pressing is the dispute between the nominalist and conceptualist positions, a dispute that Chomsky (1986) has characterized as being between E-language and I-language conceptions of language. From the E-language perspective, a natural language is a kind of social object, the structure of which is purported to be established by convention (see Lewis 1975), and persons may acquire varying degrees of competence in their knowledge and use of that social object. On Chomsky's views, such objects would be of little scientific interest if they did exist (since they would not be "natural" objects), but in any case such objects don't exist. Alternatively, an I-language is not an external object but is rather a state of an internal system that is part of our biological endowment. An agent might have I-language representations of English sentences, but those internal representations are not to be confused with spoken or written English sentences. They are rather data structures in a kind of internal computational system.
Chomsky understands the I-language computational system to be individualistic (see INDIVIDUALISM). That means that the properties of the system can be specified independently of the environment that the agent is embedded in. Thus, it involves properties like the agent's rest mass and genetic make-up (and unlike relational properties like the agent's weight and IQ).
By itself, the dispute between I-language and E-language approaches has little philosophical traction; the actual direction of the field presumably settles the issue as to which is the object of study. Nevertheless, some normative claims have been offered. For example, Soames (1984) has suggested that if we attend to the leading questions of linguistics in the past, then linguistics has been (and ought to be) concerned with E-language. Of course, one might wonder why past investigations should restrict the direction (and leading questions) of current research. Chomsky (1993, 1995) not only disputes this historical story but has argued that E-languages are not suitable for naturalistic inquiry, because they constitute artifacts rather than natural objects. In other words, it is fine to talk about E-languages as long as one doesn't think one is doing science.
As we will see, the choice between these two general approaches to language is very rich in consequences. Both the claim that I-language is individualistic and the claim that it is computational have led to a number of philosophical skirmishes.
One of the immediate questions raised by the idea that I-language is individualistic has to do with the nature of SEMANTICS, and in particular referential semantics -- construed as theories of the relation between linguistic forms and aspects of the external world (see REFERENCE, THEORIES OF). In short, the worry is this: Putnam (1975) and many other philosophers have held that we need referential semantics to characterize linguistic MEANING -- that meanings "ain't in the head." But if this is right, then it is hard to see how semantics can be part of the language faculty, which is supposed to be individualistic (and hence "in the head").
This tension between I-language and referential semantics has led commentators such as Hornstein (1984) and Chomsky (1993, 1995) to be skeptical of the possibility of a referential semantics. However, Ludlow (forthcoming) has argued that the tension in these views is apparent only, because the connection between I-languages and referential semantics would parallel the connection between individualistic and relational sciences in other domains (for example, it would be similar to the connection that holds between the studies of primate physiology and primate ecology -- facts about physiology can shed light on the primate's relation to its environment, and vice versa). Inferences between individualistic and relational sciences are imperfect, but data from one domain can nevertheless be relevant to the other.
The idea that linguistic theory involves the investigation of RULES AND REPRESENTATIONS (or principles and parameters) of an internal computational system has also led to philosophical questions about the nature of these rules and representations. For example, Quine (1970) has argued that because many possible grammars may successfully describe an agent's linguistic behavior, there is no way in principle for us to determine which grammar an agent is using. For his part, Chomsky (1980) has argued that if we consider the explanatory adequacy of a grammar in addition to its descriptive adequacy, then the question of which grammar is correct is answerable in principle. That is, if we consider that a grammar must be consistent with the theory of LANGUAGE ACQUISITION, acquired language deficits, and more generally with cognitive psychology, then there are many constraints available to rule out competing grammatical theories.
Another set of worries about rule following have stemmed from Kripke's (1982) reconstruction of arguments in Wittgenstein (1953, 1956). The idea is that there can be no brute fact about what rules and representations a system is running apart from the intentions of the designer of the system. Because, when studying humans, we have no access to the intentions of the designer, there can be no fact of the matter about what rules and representations underlie our linguistic abilities. The conclusion drawn by Kripke is that "it would seem that the use of the idea of rules and of competence in linguistics needs serious reconsideration, even if these notions are not rendered meaningless." (1982: 31 fn 22)
Chomsky (1986) appears to argue that one can know certain facts about computers in isolation, but Chomsky's current position (1995) is that computers, unlike the human language faculty, are artifacts and hence the product of human intentions. The language faculty is a natural object and embedded within human biology, so the facts about its structure are no more grounded in human intentions than are facts about the structure of human biology.
If the language faculty is an internal computational/representational system, a number of questions arise about how to best go about investigating and describing it. For example, there has been considerable attention paid to the role of formal rigor in linguistic theory. On this score, a number of theorists (e.g., Gazdar, Klein, Pullum, and Sag 1985; Bresnan and Kaplan 1982; Pullum, 1989) have argued that the formal rigor of their approaches -- in particular, their use of well-defined recursive procedures -- counts in their favor. However, Ludlow (1992) has argued that this sort of approach to rigorization would be out of synch with the development of other sciences (and indeed, branches of mathematics) where formalization follows in the wake of the advancing theory.
A second methodological issue relates to the use of PARSIMONY AND SIMPLICITY in the choice between linguistic theories. Although tight definitions of simplicity within a linguistic theory seem to be possible (see Halle 1961; Chomsky and Halle 1968; Chomsky 1975), finding a notion of simplicity that allows us to chose between two competing theoretical frameworks is another matter. Some writers (e.g., Postal 1972; Hornstein 1995) have argued that generative semantics and the minimalist program (see MINIMALISM), respectively, are simpler than their immediate competitors because they admit fewer levels of representation. In response, Ludlow (1998) has maintained that there is no objective criterion for evaluating the relative amount of theoretical machinery across linguistic theories. Ludlow offers that the only plausible definition of simplicity would be one that appealed to "simplicity of use," suggesting that simplicity in linguistics may not be a feature of the object of study itself but rather our ability to easily grasp and utilize certain kinds of theories.
Finally, there is the matter of the nature of evidence available for investigating the language faculty. Evidence from a written or spoken corpus is at best twice removed from the actual object of investigation, and given the possibility of performance errors, is notoriously unreliable at that. Much of the evidence adduced in linguistic theory has therefore been from speakers' intuitions of acceptability, as well as intuitions about possible interpretations. This raises a number of interesting questions about the reliability of introspective data (see INTROSPECTION) and the kind of training required to have reliable judgements. There is also the question of why we should have introspective access to the language faculty at all. It is fair to say that these questions have not been adequately explored to date (except in a critical vein; see Devitt 1995; Devitt and Sterelny 1987).
Katz (1985: Introduction) offers that the philosophy of linguistics could soon emerge as a domain of inquiry in its own right, on the model of the philosophy of physics and the philosophy of biology. Given the number of interesting questions and disputes that have arisen in the interim, it is fair to say that Katz's prediction is coming true. The issues canvassed above provide a mere sketch of the current topics under discussion and point to a rich field of investigation in the years to come.
Bresnan, J., and R. Kaplan. (1982). Introduction: Grammars as mental representations of language. In J. Bresnan, Ed., The Mental Representation of Grammatical Relations. Cambridge, MA: MIT Press, pp. xvii - lii.
Chomsky, N. (1975). The Logical Structure of Linguistic Theory. New York: Plenum.
Chomsky, N. (1980). Rules and Representations. New York: Columbia University Press.
Chomsky, N. (1986). Knowledge of Language. New York: Praeger.
Chomsky, N. (1993). Explaining language use. In J. Tomberlin, Ed., Philosophical Topics 20. pp. 205-231.
Chomsky, N. (1995). Language and nature. Mind 104:1-61.
Chomsky, N., and M. Halle. (1968). The Sound Pattern of English. New York: Harper and Row.
Devitt, M. (1995). Coming to Our Senses: A Naturalistic Program for Semantic Localism. Cambridge: Cambridge University Press.
Devitt, M., and K. Sterelny. (1987). Language and Reality: An Introduction to the Philosophy of Language. Cambridge, MA: MIT Press.
Gazdar, G., E. Klein, G. Pullum, and I. Sag. (1985). Generalized Phrase Structure Grammar. Cambridge, MA: Harvard University Press.
George, A. (1989). How not to become confused about linguistics. In A. George, Ed., Reflections on Chomsky. Oxford: Blackwell, pp. 90-110.
Halle, M. (1961). On the role of simplicity in linguistic description. Proceedings of Symposia in Applied Mathematics 12 (Structure of Language and its Mathematical Aspects), pp. 89-94. Providence: American Mathematical Society.
Higginbotham, J. (1983). Is grammar psychological? In L. Cauman, I. Levi, C. Parsons, and R. Schwartz, Eds., How Many Questions: Essays In Honor of Sydney Morgenbesser. Indianapolis, IN: Hackett.
Hornstein, N. (1984). Logic as Grammar. Cambridge, MA: MIT Press.
Hornstein, N. (1995). Logical Form: From GB to Minimalism. Oxford: Blackwell.
Katz, J. (1981). Language and Other Abstract Objects. Totowa, NJ: Rowman and Littlefield.
Katz, J., Ed. (1985). The Philosophy of Linguistics. Oxford: Oxford University Press.
Kripke, S. (1982). Wittgenstein on Rules and Private Language. Cambridge: Harvard University Press.
Lewis, D. (1975). Language and languages. In K. Gunderson, Ed., Language, Mind, and Knowledge. Minneapolis: University of Minnesota Press, pp. 3-35.
Ludlow, P. (1992). Formal Rigor and Linguistic Theory. Natural Language and Linguistic Theory 10:335-344.
Ludlow, P. (1998). Simplicity and generative grammar. In R. Stainton and K. Murasugi, Eds., Philosophy and Linguistics. Boulder, CO: Westview Press.
Ludlow, P. (Forthcoming). Referential semantics for I-languages? In N. Hornstein and L. Antony, Eds., Chomsky and His Critics. Oxford: Blackwell.
Postal, P. (1972). The best theory. In S. Peters, Ed., Goals of Linguistic Theory. Englewood Cliffs, NJ: Prentice-Hall, pp. 131-179.
Pullum, G. (1989). Formal linguistics meets the boojum. Natural Language and Linguistic Theory 7:137-143.
Putnam, H. (1975). The meaning of meaning. In K. Gunderson, Ed., Language, Mind, and Knowledge. Minneapolis: University of Minnesota Press, pp. 131-193.
Quine, W. V. O. (1970). Methodological reflections on current linguistic theory. Synthese 21:368-398.
Soames, S. (1984). Linguistics and psychology. Linguistics and Philosophy 81:155-179.
Wittgenstein, L. (1953). Philosophical Investigations. Translated by G. E. M. Anscombe. New York: MacMillan.
Wittgenstein, L. (1956). Remarks on the Foundations of Mathematics. Translated by G. E. M. Anscombe. Cambridge, MA: MIT Press .
Copyright © 1999 Massachusetts Institute of Technology