Reviewed by Joseph F. Eska, Virginia Polytechnic Institute & State University

Variation is present in all natural languages, but it often does not find a place in diachronic linguistic reconstruction. This volume sets out to address specific problems in which the reconstruction of variation might enable further progress towards a solution.

After an introductory chapter in which Mary K. Niepokuj reviews the papers and discusses issues relating to the role of language variation in  linguistic reconstruction, Thomas D. Cravens examines cases in which Latin /p t k/ are not continued unaltered into Italian, but are unexpectedly voiced. He points out that underlying voiceless plosives are variably realized in contemporary Tuscan dialects and that such variation is also reflected in the orthography of medieval documents. He posits that such phonetic variation led speakers to abstract different underlying phonemes for each lexical item. Steven R. Geiger and Joseph C. Salmons look at the voice-onset time of coronal plosives in recordings of the Kölsch dialect of Wisconsin German from the 1940s, 1970, and 1999, and find that it has been decreasing dramatically for /t/ and less so for /d/ with the surprising result that it has diverged significantly from both standard German and American English. Emily L. Goss and Robert B. Howell examine variation in the writings of thirteen seventeenth-century inhabitants of The Hague with regard to dialect contact and find that they support the principles that marked regional forms are disfavored, while forms found in more than one dialect tend to be favored. Ray Harris-Northall demonstrates that grammars of Castilian Spanish from ca. 1500 do not reflect the standardization of the language, but aim to suppress certain variants. Brian D. Joseph deals with the difficulty of knowing when it is legitimate to project variation back to a proto-language to solve a problem in reconstruction. He treats several potential cases and concludes with the suggestion that variation might be responsible for linguistic drift. Cynthia L. Miller examines variation between the complementizers l’mr and lm, used to introduce direct speech, in fifth-century bce Achaemenid Aramaic legal documents and, after investigating a number of factors, finds that style is the most important one, with only l’mr occurring in the clause that gives the date of the document and names the parties involved. James Milroy argues that the role of the speaker is undervalued in the discourse of language change and that this must be corrected. He illustrates his position by examining variation in the realization of /t/ in Newcastle English, which he claims is the result of social factors. Martha Ratliff explains that variation in nominal prefixation in Hmong-Mien makes reconstruction almost impossible, but suggests that understanding the way that this variation works can allow scholars to ‘ignore the absorbed prefixes in our reconstructions of roots’ and ‘reconstruct a classifying prefix position for nouns’ (176). Paul T. Roberge addresses methodological issues concerning the reconstruction of variation in the early history of Afrikaans (ca. 1710–1840), concluding that the conscious manipulation of the language by speakers for reasons of inclusion or exclusion prohibits the reconstruction of variation in gender among Germanic languages and suggests that proto-Germanic should be reconstructed as having had only two genders, with the attested variation having arisen during the shift to a three-gender system. And Graham Thurgood argues that the major source of variation in languages of southeastern Asia is language contact, and, therefore, the reconstructed proto-languages should display a relatively small amount of internal variation.

Many of the papers in this volume show that, with due prudence, the reconstruction of variation in proto-languages can assist in the resolution of long-standing problems. Some go rather too far, however. Everyone interested in the reconstruction of proto-languages would do well to consult it.