Function, Selection, and Innateness: The Emergence of Language Universals.
The central theme of the book relates to the question of why languages are similar on the one hand and show manifest differences on the other hand. Two answers seem possible, based on either the functionalist or the formalist approach. It is argued here that full explanatory adequacy in linguistic research can only be reached if the two approaches are unified. In the functionalist theory it is held that communication is responsible for the variation from language to language, whereas formal theory holds that it is the innateness of the language system that explains language universals and the internal structure of language. In order to prove the complementary roles of the two types a set of computational simulations is elaborated. What is aimed at, ultimately, is proof that language is an adaptive system in which language universals result from a historical process of functional pressure that is either biological or linguistic.
The first chapter, "A puzzle to fit," starts from the assumption that variation among languages is constrained in various ways. One aspect of this variation is investigated here, viz. so-called "fit" or "the appearance of design." After a lucid survey of different types of language universals and the constraints they map upon languages, a search for the fit of form to function is started. There are several types of functional explanation of language universals: a pragmatic one, an iconic one, and the appeal to the structure of discourse. Language use, so it seems, can have impact on universals; often this is a matter of pressure executed in the processing of language during an extended period. This is the functional explanation the author aims at, "a statement of fit between a given language universal and the pressures of processing" (p. 13). It is emphasized that this (functional) approach does not contrast with the Chomskyan paradigm. As regards the latter, the author firmly subscribes to an innate theory of syntax, a property of the brain that is biologically given. In the illustration of the latter the transformational theory is followed with its levels of adequacy and the process of language acquisition as a mere setting of parameters.
Although the innateness hypothesis is highly appreciated, it cannot be regarded successful in every sense; just like the functionalist approach, it does not succeed in solving the puzzle of fit. There remains a gap between the processing of language and the innate universals. To fill this gap, that is, to develop a mechanism of linking the two, a computer program is set up, which program must test the proposed solutions.
The rest of the book consists of two parts. Chapters 2 and 3 elaborate a program to solve the problem of linking, while the remaining chapters investigate the implication of the theory for functional explanation and innate hypotheses of language variation.
Chapter 2 on processing and word order explores how pressures on language use can give an explanation for language universals. To that end the performance theory of Hawkins (1994) is presented, in which word-order universals are treated. A solution is attempted to the problem of linkage, more particularly to the problem of how a difference in complex parsing results in a difference in cross-linguistic distribution.
It is then proved with several parsing examples that parsing complexity may result in different types of grammar. That languages correctly adapt to the design of the parsers is explained by Keller's so-called invisible-hand account (1994) of language change: changes are not intended by the language user but result from their actions. In the same sense universals are the result of human action across time and space -- a typically functionalist explanation, so it seems. Important in the changes is the ecological condition of the language users. Time and again the author accommodates the fit that emerges for some function in a broader framework, viz. that of biological evolution. Keller's invisible hand, then, explains how the influence of processing on language operates selectively. To prove this, a computational model is elaborated in which simulations of the adaptive process are construed. The simulations prove that the above performance theory of selection gives a good example of certain universals, such as those of word order. A mechanism for solving the problem of linkage thus comes nearer. However, Hawkins's performance theory, though important to the model developed here, is not accepted in full. To explain the adaptedness of languages, the hearer's rather than the speaker's selection seems essential (cf. also p. 135, referring to speaker's altruism defended by Newmeyer 1991).
The third chapter starts from the well-known Keenan-Comrie hypothesis (1977), an implicational universal on the structure of relative clauses and their pronouns. Again it is Hawkins (1994) who offers a correct explanation for the universality of' this complex rule. However, complexity of parsing cannot account for the hierarchical principle on its own. By means of a series of simulations it is proved that it is not only parsing complexity that exerts its influence but also morphological complexity: in general a speaker will prefer morphologically less complex forms. The situation becomes more complicated by so-called case coding: acknowledging case to the nominal element, the pronoun, in the relative clause. Case coding represents a trade-off between an increase in m-complexity and a decrease in p-complexity. This leads to the conclusion that "a gradient hierarchy of processing complexity cannot on its own give rise to the cross-linguistic implicational hierarchy of accessibility to relativization" (p. 84). After the treatment of some hierarchies (Why The team won the game they played and not ... it played), one question remains open: why don't all languages opt for minimum complexity? An answer is tried in chapter 4 on the limits of functional application. Since there are cases where processing asymmetry does not result in cross-linguistic asymmetry, the functional approach does not seem to give all the answers. It is especially unclear why parallel function has no cross-linguistic consequences, while accessibility has. This might be due to the fact that there is something in language acquisition that constrains the expected adaptation. An explanation or at least the beginning of an answer may be offered by a parallel in biology. There is a mismatch in nature in that it did not develop biological wheels; clearly there are structural constraints in animals barring the possibility of doing so. In relative clauses comparable innate constraints seem to limit some adaptations of form to fit a function, more specifically where grammatical operations work in different domains (wh-movement and relative-clause formation). Again, there are counter-examples to the hitter explanation (German free relatives, English post- and pre-genitive), but not only the LAD must be taken into consideration but also parsing principles. In other words, language universals originate as a result of both syntactic rules and processing: functional and innate influences thus seem to be mutually reinforcing. Whether this is indeed the case is answered in chapter 5, "Innateness and function in linguistics." First some competing theories on the origin of linguistic universals are examined. The question arising is to what extent Chomsky's universal grammar is proof of the adaptation to the communicative function. That autonomous syntax should not make use of functional pressures is certainly not correct: the whole of Kirby's book is proof to the contrary. Arguments that processing explains several universal principles of grammar are also taken from Newmeyer (1991).
In the end the author returns to the original question on the positions of the functionalist and the innatist. He argues that it is not by these positions that universals are explained; the status of the latter must be defined by solving the problem of linkage. A biological base for functional pressures cannot be denied. Here again Newmeyer's suggestions are followed: that, where problems for a hearer might arise, constraints are built in in the language faculty (1991:15). In modelling the evolution of language it is necessary to at least include functional pressures in order to explain universal constraints. As is most clearly demonstrated in Figure 5.3 (p. 133) two senses of adaptation must be discussed. On the one hand there is the glossogenetic selection (linguistic/historical) and on the other hand the phylogenetic selection (biological/natural). However, the two are related in that they form a link between functional and cross-linguistic distribution. [In the Index, reference to Phylogeny, etc., is missing.]
The conclusions of chapter 5 offer three clear and convincing claims:
1. functional pressures influence linguistic selection to give rise to observable language universals;
2. adaptation by linguistic selection operates within constraints imposed by universal grammar;
3. functional pressures influence natural selection, whose constraints result in universal grammar over a biological timescale.
Function, Selection and Innateness is a provocative and inspiring work, offering the link between functional pressure and the universal principles of grammar. Its computer simulations of the evolution of language convincingly sustain the theory that the grammaticalization of functional pressures has contributed to the development of the so-called language-acquisition device, Noam Chomsky's LAD.
Hawkins, John A. (1994). A Performance Theory of Order and Constituence. Cambridge: Cambridge University Press.
Keenan, Edward; and Comrie, Bernard (1977). Noun phrase accessibility and universal grammar. Linguistic Inquiry 8, 63-99.
Keller, Rudi (1994). On Language Change.' The Invisible Hand in Language. London: Routledge.
Newmeyer, Frederick J. (1991). Functional explanation in linguistics and the origins of language. Language and Communication 68, 3-28.
DROSTE, FLIP G.
University of Leuven
|Printer friendly Cite/link Email Feedback|
|Author:||DROSTE, FLIP G.|
|Publication:||Linguistics: an interdisciplinary journal of the language sciences|
|Article Type:||Book Review|
|Date:||Jul 1, 2000|
|Previous Article:||A response to Kanno's "The stability of UG principles in second-language acquisition: evidence from Japanese"(1).|
|Next Article:||Models of Inflection.|