Abstract
Little is known about the principles governing the extent to which nominal compounds are used. Adopting a contrastive perspective, this paper takes as its starting point the widely held assumption that German assigns a more important role to compounding than English. The notion of compound use is broken down into the three factors of (type and token) frequency, productivity and complexity. An analysis of newspaper and fictional language reveals an astonishing crosslinguistic difference in the interaction of frequency and complexity. The higher rate of German compared to English two-noun compounds decreases with increasing complexity to the point of overturning this ratio in favor of relatively more four-noun compounds in English than in German. This holds good for type and token frequency, expanding productivity as well as for both text genres. The consistency of the empirical results invites a structural explanation of the crosslinguistic difference. Two language-particular factors are invoked: general compound propensity and the range of typical compound types. The former explains the predominance of shorter compounds in German, the latter the predominance of longer compounds in English. German is argued to have a higher compound propensity than English while English allows a wider range of compound types than German. In particular, because English permits more nonlexicalized compounds and because the proportion of these items increases with compound size, the higher frequency of English compared to German overlong compounds is restricted to the upper end of the complexity scale.
Table 5 (p. 285) was originally printed with an error. Please find the corrected version of Table 5 in the Erratum in Volume 50, issue 3 (or click here 10.1515/ling-2012-err-v50i2)
©[2012] by Walter de Gruyter Berlin Boston
