SIL Electronic Book Reviews 2004-008

Introduction

The book, which constitutes a modified version of Kirby’s Ph.D. thesis, has received international acclaim as an innovative contribution to the discussion on the relationship between formal and functional linguistic approaches. Its origin lies within evolutionary linguistics, a considerably young subfield of the linguistic discipline. Applying innovative computational approaches to both formal and functional considerations of language evolution, Kirby comes to important conclusions in the debate on how language universals evolve.

The book assumes from its readers a general knowledge of functional theories and of how language universals operate, as well as some familiarity with the Principles and Parameters approach; Kirby himself refers to Chomsky (1981), among others.

Chapter-by-chapter content

Chapter 1, “A Puzzle of Fit,” sets the scene with a double observation: first, that constraints on language variation are not distributed randomly over the plain of possibility; and second, that language universals can be explained functionally through pressures of processing. Such functional explanations are then contrasted with syntactic theory according to which an innate knowledge of language is the reason for universal properties of all languages. The chapter closes with defining the primary research question of the book:

The problem of linkage. Given a set of observed constraints on cross-linguistic variation, and a corresponding pattern of functional preference, an explanation of this fit will solve the problem: how does the latter give rise to the former? (p. 20)

In chapter 2, “The Impact of Processing on Word Order,” Kirby combines Hawkins’ (1994) performance theory for measuring parsing complexity with Hurford’s (1987) ”arena of use,” suggesting that parsing complexity influences the variability of word orders in child language acquisition. This suggestion is then tested in a computer simulation system consisting of “speakers” producing “utterances” according to “grammars” and “acquirers” filtering these random utterances from the “arena of use” to form a trigger experience from which to build their own individual grammar, thereby becoming the new speakers of the next cycle. The trigger is computed from an equation involving both frequency of utterance and Hawkins’ formula for parsing complexity. Experiments with word order variables result in optimal orderings occurring after several cycles with almost 100 percent or, if optimal orderings are not available, the suboptimal wins over the worst. Similarly, cycles experimenting with the prepositional noun-modifier hierarchy (NRel > NGen > NAdj) result in the predicted order of change.

In chapter 3, “Hierarchies and Competing Motivations,” Kirby introduces the difference between p-complexity (for “parsing”), i.e., that more difficult variants are less likely to be included in a trigger experience, and m-complexity (for “markedness”), i.e., that less morphologically complex variants are more likely to be uttered by speakers. Furthermore, without a probability parameter potentially changing the values between both complexities at each cycle, the computer simulation community converged on only one type of certain hierarchies where several should be allowed. Kirby suggests that the explanation of implicational universals depends on these competing motivations; e.g., in gender marking:

p-complexity increases when gender is left unmarked, especially on second and third person expressions, but conversely m-complexity increases when gender is marked on any expression.” (p. 85)

Chapter 4, “The Limits of Functional Adaptation,” returns to the debate of formalism versus functionalism. Based on the observation that the functional approach fails to explain the co-dependence between forms of differing processing complexity in certain hierarchical universals, Kirby argues for the existence of the Language Acquisition Device (LAD) of the innateness hypothesis. Given the set of logically possible languages, only those occur which are also members of both the set of learnable languages (as determined by the innate LAD) and the set of languages which can be appropriately processed (as predicted by selection theory). In this, Kirby sees the interdependence of formalism and functionalism: “We are now at a point where the functional and formal (or innatist) perspectives are mutually reinforcing, rather than competing” (p. 122).

Chapter 5, “Innateness and Function in Linguistics,” reviews the recent literature examining the origin of the LAD, distinguishing an exaptationist explanation (i.e., that the form was reappropriated with little or no prior functional relevance) from an adaptationist approach (i.e., basically natural selection). This includes a discussion of five different views on the emergence of the subjacency condition which prohibits wh-movement across two bounding nodes, as well as an argument for integrating into functional explanation the syntactic autonomy assumption which previously has led generative linguists to reject any link between functional explanation and generative grammar. While rejecting the stance that universals like the tendency for consistent head-ordering have an innate basis, Kirby proposes the explanatory capability of both linguistic and natural selection in glossogenetic adaptation.

In conclusion, Kirby summarizes three claims which are based on the novel separation between m- and p-complexity and on the interdependence of functional and innate explanations, respectively:

1. “Functional pressures influence linguistic selection, which operates locally in the cycle of acquisition and use, to give rise, globally, to observable language universals, over a historical timescale” (p. 141).
2. “Adaptation by linguistic selection operates within constraints imposed by Universal Grammar” (p. 142).
3. “Functional pressures influence natural selection, which operates within physical and embryological constraints to give rise to an autonomous LAD, or UG, over a biological timescale” (p. 143).

The hitherto practised separation into “functionalist” and “formalist” camps in theoretical linguistics, however, is to be considered detrimental to perceiving the full picture.

Evaluation

Of course, the book has not been written specifically for descriptive linguists, and consequently, the average field linguist may find it too technical, the writing style too condensed, and the topic of not enough relevance for their everyday linguistic tasks. However, there are a number of gems in there which could be taken note of with benefit for descriptive linguistics. Kirby gives a concise overview of syntactic theory versus functional approaches, laying the theoretical foundation for integrating both perspectives in practice. His applications to specific language universals serve as an apt example of how practical observations should lead to generalizations in the bigger picture, in this case, language evolution. Moreover, in situations of language contact and change as they are often encountered by descriptive linguists in fieldwork, Kirby’s conclusions are applicable in investigations of which features are acquired and adapted from neighbouring languages, and why.

The caveats for the average descriptive linguist notwithstanding, the book is definitely recommended for anyone interested in linguistic theory, in language universals, and in the relationship between form and function. Be warned, however: light bedtime reading, this book certainly is not!