Actualization: Linguistic Change in Progress (review)

The core of this collection often papers is Henning Andersen’s work on actualization and the role of markedness in language change. A’s theory is elaborated in a position paper (‘Markedness and the theory of linguistic change’, 21–57); presented at the Thirteenth International Conference on Historical Linguistics (Düsseldorf, 1997), it was the catalyst for a workshop on ‘Patterns of actualization in linguistic change’ at the Fourteenth International Conference on Historical Linguistics (Vancouver, 1999). The final paper (‘Actualization and the (uni)directionality of change’, 225–48) summarizes and advances the discussion in the earlier part of the book as well as some of A’s major proposals in the past three decades. These two papers serve as theoretical bookends for the volume, which is also introduced by A. Most other papers are particular studies that support or challenge his hypotheses. The book is dense. This is partly because it covers a very wide range of concepts and data. Partly it is because the key term ‘actualization’ is not uniquely defined (as the only ‘observable manifestation of grammar innovations in speech’) until the last paper (225); other terms are used in unfamiliar senses; for example, ‘actuation’ is defined as ‘real change in the system’ (79) rather than in the usual sense of ways in which change is initiated and why changes occur in one language at a particular time but not in another language or at another time (Weinreich et al. 1968:102).

As A points out at some length, actualization has not played a significant part in historical linguistics. In his view, inattention to actualization is largely due to two theoretical assumptions: (1) that change in use precedes structural change and that innovations in usage are chaotic and piecemeal (cf. Lightfoot 1999); this hypothesis can be summarized as ‘reanalysis is late’; (2) focus on covert change in grammars is considered more explanatory than focus on overt change (cf. Harris & Campbell 1995; but see Croft 2000 for a perspective more consonant with A’s).

In A’s view change is, as in the generative tradition, change in grammars, but the focus is significantly different. A privileges the hypothesis that ‘reanalysis is early’ (see also Kroch 1989). The frequency phenomena seen in S-curves are interpreted as the observable actualizations of the reanalysis (17), not the trigger. Furthermore, A seeks to account for the widely-attested fact that ‘change in usage progresses not in a chaotic, but in an orderly, grammatically conditioned step-by-step fashion’ (226) and involves the replacement over times of varying depth of old (O)-variants by innovative (I)-variants. This account requires three conceptual moves concerning (1) the nature of the relationship between grammar and use, (2) the definition of change, and (3) the intrinsic asymmetry of structural oppositions known as ‘markedness’.

For the first move, A invokes Timberlake’s (1977) discussion of reanalysis and actualization. Timberlake assumes a base system of content categories and syntactic relations embodying universal principles of grammar, together with a system of usage rules that link the base system to ‘more superficial categories’, including pragmatic and sociolinguistic rules that enable speakers to relate their system to that of the community (236).

Second, A distinguishes innovation (in the individual) from change (reanalyses between grammars, i.e. adoption). He conceptualizes changes as ‘the historical events in a linguistic tradition by which practices of speaking vary over time’ and which are ‘always manifested in synchronic variation’ (228). Those changes of prime interest to him are ‘evolutive’, brought about by learners (both children and older learners) in a speech community in which the innovated variants already exist (contrast borrowings) (234). Evolutive changes may be unconscious reanalyses or (partially) conscious coinages, hypercorrections, or extensions.

Third, A hypothesizes that directionality is explained by the fact that covariants ‘that are directly generated by a given base grammar are unmarked in relation to the covariants that are defined by its usage...