Chomsky has stated in a number of sources that there are three levels of “adequacy” upon which grammatical descriptions and linguistic theories can be evaluated: observational adequacy, descriptive adequacy, and explanatory adequacy.
 |
| sussy_Linguistics Theory |
If a theory or description achieves observational adequacy, it is able to describe which sentences in a language are grammatically well formed. Such a description would note that in English while a sentence such as He studied for the exam is grammatical, a sentence such as *studied for the exam is not. To achieve descriptive adequacy (a higher level of adequacy), the description or theory must not only describe whether individual sentences are well formed but in addition specify the abstract grammatical properties making the sentences well formed. Applied to the previous sentences, a description at this level would note that sentences in English require an explicit subject. Hence, *studied for the exam is ungrammatical and He studied for the exam is grammatical. The highest level of adequacy is explanatory adequacy, which is achieved when the description or theory not only reaches descriptive adequacy but does so using abstract principles which can be applied beyond the language being considered and become a part of “Universal Grammar.” At this level of adequacy, one would describe the inability of English to omit subject pronouns as a consequence of the fact that, unlike Spanish or Japanese, English is not a language which permits “pro-drop,” i.e. the omission of a subject pronoun that is recoverable from the context or deducible from inflections on the verb marking the case, gender, or number of the subject.
Within Chomsky’s theory of principles and parameters, pro-drop is a consequence of the “null-subject parameter” (Haegeman 1991: 17–20). This parameter is one of many which make up universal grammar, and as speakers acquire a language, the manner in which they set the parameters of universal grammar is determined by the norms of the language they are acquiring. Speakers acquiring English would set the null-subject parameter to negative, since English does not permit pro-drop; speakers of Italian, on the other hand, would set the parameter to positive, since Italian permits pro-drop (Haegeman 1991: 18).
Because generative grammar has placed so much emphasis on universal grammar, explanatory adequacy has always been a high priority in generative grammar, often at the expense of descriptive adequacy: there has never been much emphasis in generative grammar in ensuring that the data upon which analyses are based are representative of the language being discussed, and with the notion of the ideal speaker/hearer firmly entrenched in generative grammar, there has been little concern for variation in a language, which traditionally has been given no consideration in the construction of generative theories of language. This trend has become especially evident in the most recent theory of generative grammar: minimalist theory.
In minimalist theory, a distinction is made between those elements of a language that are part of the “core” and those that are part of the “periphery.” The core is comprised of “pure instantiations of UG” and the periphery “marked exceptions” that are a consequence of “historical accident, dialect mixture, personal idiosyncracies, and the like” (Chomsky 1995: 19–20). Because “variation is limited to nonsubstantive elements of the lexicon and general properties of lexical items” (Chomsky 1995: 170), those elements belonging to the periphery of a language are not considered in minimalist theory; only those elements that are part of the core are deemed relevant for purposes of theory construction. This idealized view of language is taken because the goal of minimalist theory is “a theory of the initial state,” that is, a theory of what humans know about language “in advance of experience” (Chomsky 1995: 4) before they encounter the real world of the language they are acquiring and the complexity of structure that it will undoubtedly exhibit.
This complexity of structure, however, is precisely what the corpus linguist is interested in studying. Unlike generative grammarians, corpus linguists see complexity and variation as inherent in language, and in their discussions of language, they place a very high priority on descriptive adequacy, not explanatory adequacy. Consequently, corpus linguists are very skeptical of the highly abstract and decontextualized discussions of language promoted by generative grammarians, largely because such discussions are too far removed from actual language usage. Chafe (1994: 21) sums up the disillusionment that corpus linguists have with purely formalist approaches to language study, noting that they “exclude observations rather than . . . embrace ever more of them” and that they rely too heavily on “notational devices designed to account for only those aspects of reality that fall within their purview, ignoring the remaining richness which also cries out for understanding.” The corpus linguist embraces complexity; the generative grammarian pushes it aside, seeking an ever more restrictive view of language.
Because the generative grammarian and corpus linguist have such very different views of what constitutes an adequate linguistic description, it is clear why these two groups of linguists have had such a difficult time communicating and valuing each other’s work. As Fillmore (1992: 35) jokes, when the corpus linguist asks the theoretician (or “armchair linguist”) “Why should I think that what you tell me is true?”, the generative grammarian replies back “Why should I think that what you tell me is interesting?” (emphasis added). Of primary concern to the corpus linguist is an accurate description of language; of importance to the generative grammarian is a theoretical discussion of language that advances our knowledge of universal grammar.
Even though the corpus linguist places a high priority on descriptive adequacy, it is a mistake to assume that the analysis of corpora has nothing to offer to generative theory in particular or to theorizing about language in general. The main argument against the use of corpora in generative grammar, Leech (1992) observes, is that the information they yield is biased more towards performance than competence and is overly descriptive rather than theoretical. However, Leech (1992: 108) argues that this characterization is overstated: the distinction between competence and performance is not as great as is often claimed, “since the latter is the product of the former.” Consequently, what one discovers in a corpus can be used as the basis for whatever theoretical issue one is exploring. In addition, all of the criteria applied to scientific endeavors can be satisfied in a corpus study, since corpora are excellent sources for verifying the falsifiability, completeness, simplicity, strength, and objectivity of any linguistic hypothesis (Leech 1992: 112–13).
Despite Leech’s claims, it is unlikely that corpora will ever be used very widely by generative grammarians, even though some generative discussions of language have been based on corpora and have demonstrated their potential for advancing generative theory. Working within the framework of government and binding theory (the theory of generative grammar preceding minimalist theory), Aarts (1992) used sections of the corpus housed at the Survey of English Usage at University College London to analyze “small clauses” in English, constructions like her happy in the sentence I wanted her happy that can be expanded into a clausal unit (She is happy). By using the London Corpus, Aarts (1992)was not only able to provide a complete description of small clauses in English but to resolve certain controversies regarding small clauses, such as establishing the fact that they are independent syntactic units rather than simply two phrases, the first functioning as direct object and the second as complement of the object.
Haegeman (1987) employed government and binding theory to analyze empty categories (i.e. positions in a clausewhere some element is missing) in a specific genre of English: recipe language. While Haegeman’s investigation is not based on data from any currently available corpus, her analysis uses the type of data quite commonly found in corpora. Haegeman (1987) makes the very interesting claim that parametric variation (such as whether or not a language exhibits pro-drop) does not simply distinguish individual languages from one another but can be used to characterize regional, social, or register variation within a particular language. She looks specifically at examples from the genre (or register) of recipe language that contain missing objects (marked by the letters [a], [b], etc. in the example below):
(1) Skin and bone chicken, and cut [a] into thin slices. Place [b] in bowl with mushrooms. Pur´ee remaining ingredients in blender, and pour [c] over chicken and mushrooms. Combine [d] and chill [e] well before serving. (Haegeman 1987: 236–7)
Government and binding theory, Haegeman (1987: 238) observes, recognizes four types of empty categories, and after analyzing a variety of different examples of recipe language, Haegeman concludes that this genre contains one type of empty category, wh-traces, not found in the core grammar of English (i.e. in other genres or regional and social varieties of English).
What distinguishes Haegeman’s (1987) study from most other work in generative grammar is that she demonstrates that theoretical insights into universal grammar can be obtained by investigating the periphery of a language as well as the core. And since many corpora contain samples of various genres within a language, they are very well suited to the type of analysis that Haegeman (1987) has conducted. Unfortunately, given the emphasis in generative grammar on investigations of the core of a language (especially as reflected in Chomsky’s recent workin minimalism), corpora will probably never have much of a role in generative grammar. For this reason, corpora are much better suited to functional analyses of language: analyses that are focused not simply on providing a formal description of language but on describing the use of language as a communicative tool. (P. 2 – 5)
REFERENCE
Charles F. Meyer. (2004). English Corpus Linguistics An Introduction. New York: Cambridge University Press.
