- Efficiency and complexity in grammars
John Hawkins has long been a key figure in the attempt to link together language change, psycholinguistics, and linguistic typology. His work over the past quarter-century has been an attempt to combine these fields in an innovative, yet data-oriented, way: his seminal work, Word order universals(Hawkins 1983), is often regarded as a classic. H’s ‘theme song’ has been the interconnectedness of performance and grammar, the notion that human linguistic processing has made languages the shape they are. In short, he argues that principles of performance can explain why languages look and behave the way they do. This work summarizes his latest thinking on the topic.
The title contains the book’s message, that even highly abstract and fundamental properties of syntax are derivable from simple principles of processing efficiency and complexity that are needed anyway in order to explain how language is used. The book contains nine chapters. Chs. 1–3 are, broadly, introductory; Chs. 4–8 provide more detailed information; and Ch. 9 summarizes the conclusions.
In Ch. 1, H explains his goal, which is to show that there are profound correspondences between performance and grammars. Grammars, he asserts, are frozen or fixed performance preferences. He labels this the performance-grammar correspondence hypothesis (PGCH),which he defines as follows: ‘Grammars have conventionalized syntactic structures in proportion to their degree of preference in performance, as evidenced by patterns of selection in corpora and by ease of processing in psycholinguistic experiments’ (3). This idea is not new. As he points out, these claims have formed the basis for proposals about word-order universals in his previous books. They have received support from a variety of scholars, though his ideas have never been recognized by Noam Chomsky and others who work on formal grammar. H states it as his personal view that work in formal syntax and semantics has now reached a point where it must take account of the type of processing constraint proposed here for two reasons: first, if there are systematic correspondences between grammars and performance, a theory that accounts only for grammars is not general enough and misses significant generalizations. Second, principles of performance offer a potential explanation for why grammars are the way they are, and they can help us to decide between competing formal models.
He then explains his general procedure as follows: first, he will find a language whose grammar generates a plurality of structural alternatives of a common type. Second, he will check for the distribution of these same structural patterns in grammatical patterns across languages. The PGCH predicts that when the grammar of one language eliminates one or more structural options that are permitted by the grammar of another, the restriction will be in accordance with performance preferences: the preferred structure will be retained and fixed as a grammatical convention, and the dispreferred structures will be removed, or retained only in a marginal form. He then briefly introduces his three general principles of efficiency: minimize domains (MiD), minimize forms (MiF), and maximize on-line processing (MaOP).
In Chs. 2–3, he discusses his background assumptions, then provides more detailed definitions of his efficiency principles and outlines their predictions. In Ch. 4, he elaborates and tests one of his principles in depth (minimize forms), and in Chs. 5–8, he assesses the individual and combined predictions of all three, before moving to the final summary chapter.
In many ways, those who have followed (and admired) H’s work over the years will find that they have heard much of this before, though here it is summarized more succinctly—and also more densely. Few of those who come to it fresh will find it easy to absorb, especially as he illustrates his ideas with examples from languages that are unlikely to be known by all of his readers, even though they are primarily European: he apologizes for the fact that he is forced by the lack of available materials to make untested predictions about non-Indo-European languages, which he hopes others will...