Sociolinguistic Typology: Social Determinants of Linguistic Complexity by Peter Trudgill (review)

Trudgill's argument is that "complexity" (a dubious term, see below) of a language Lx correlates, statistically though not absolutely, with five factors whose values also tend to cluster into two poles:

• • high versus low contact with other languages, measured by the extent of adult outsiders acquiring L_x as second language (pp. 16, 34, 99);

• • unstable versus stable social situation of the L_x-speaking community over time (pp. 3, 9);

• • large versus small L_x-speaking community (pp. 99-100);

• • weak versus strong (dense or multiplex) networks within the L_x-speaking community (p. 104);

• • low versus high extent of communally shared information (p. 127).

The definition of high versus low contact is carefully constructed, excluding various types of situations that for other purposes could be considered as high contact. By Trudgill's definition L_x is not in a high contact situation when its own speakers also speak a national language or lingua franca, even though this may lead to extensive borrowing. Nor is it in a high contact situation when there is recurrent intermarriage with speakers of another low-population language, so that child bilingualism is dominant. The only situation that meets his definition of high contact is when large numbers of adult outsiders learn to speak L_x imperfectly.

One polar type has high contact, an unstable social situation, a large community, weak networks, and low communally shared information. This syndrome, characteristic of most standard national languages but also of creoles, strongly favors (and perhaps compels) simplification or the maintenance of inherited simplicity. The other polar type has low contact, a stable situation, a small community, strong (dense) networks, and broad sharing of information. This is the small-scale language community, isolated by geography, by an absence of in-migration, or by both, and it allows (but need not compel) complexification or the maintenance of inherited complexity. I refer to these below as the "standard/creole" and "relic" social situations, for lack of better terms. At the end of the book, mixed types based on differing values for the first three criteria above are recognized (p. 147).

This is probably the most thoughtful and knowledgeable articulation of a growing movement reacting against the all-languages-are-the-same model ("equicomplexity" [p. 51]) that one associates not only with typologically uninformed bioprogrammers but also with many defenders of the dignity of minority languages or dialects, especially creolists (such as Michel DeGraff) and social dialectologists (such as Trudgill's mentor William Labov).

Trudgill sticks to morphology and phonology. The morphology chapters deal with four types of complexity and complexification (pp. 21-22, 62): categorial load (presence versus absence of nonuniversal categories), irregularity versus regularity, opacity versus transparency, and degree of redundancy. Categorial load (my term) corresponds to what Trudgill calls at first "paradigmatic redundancy" and later, more perspicuously, "addition of morphological categories" (pp. 22, 62). On the other hand, we might merge irregularity-regularity with opacity-transparency, which overlap considerably.

Leaving aside these terminological quibbles, the bottom line is that high load, irregularity, opacity, and extensive redundancy can develop in relic situations (e.g., the Northwest Coast of North America, the Caucasus, New Guinea, Australia). Standard/ creole situations strongly favor developments in the opposite direction.

The phonology chapters focus on phoneme inventories and to some extent on the presence or absence of unusual sound shifts. Aside from sheer phoneme count, Trudgill is also interested in the geometry of vowel systems: standard/creole situations require feature symmetry and maximal dispersion (acoustic clarity at the expense of articulatory effort), while relic situations allow tighter systems that may include typologically marked front rounded vowels, and that present reduced acoustic dispersion.

Correlations of social situation with phoneme count leads to a conundrum: standard/ creole situations require a middling range (enough for basic communication but not much more), while relic situations allow anything from luxuriant proliferations of phonemes (Caucasus, Northwest Coast) to extremely compressed, bare-bones systems (Hawaiian, to which we can now add Pirah...