Facts, theory and dogmas in historical linguistics: vowel quantity from Latin to Romance

The Latin-Romance continuum, with its two and a half millennia of documented history and with the great amount of scholarly work that has been carried out on it, still presents the historical linguist not only with unsolved problems, which is obvious, but also, less trivially, with the challenging opportunity to apply to these problems the method of historical linguistics in its fullest form. As Ferdinand de Saussure puts it, "la linguistique diachronique suppose à la fois une perspective prospective, qui suit le cours du temps, et une perspective rétrospective, qui le remonte". For any given fact, thus, a sound diachronic account must reconcile the evidence coming from reconstruction (perspective rétrospective) – which in turn consists of two operations, internal and comparative reconstruction – with that coming from philological inspection of extant relevant records (perspective prospective). This is well possible, in principle, within the Latin-Romance domain. In practice, however, much of the research that is currently being done on the diachronic development of the Romance languages (especially the kind of research one is likely to find in 'leading journals' in formal linguistics) does not exploit all the relevant evidence and as a result often produces accounts that demonstrably fail to hold water. Such accounts rest moreover on the underlying assumption that views work in (diachronic) linguistics as being primarily a contribution to a grander enterprise in which a general model of linguistic theory is being tested and developed. This same emphasis on theory, especially if contested, will duly verge on the dogmatic. Alternatively, historical linguistics can be conceived as an (autonomous) problem-solving discipline: in this vein, what really matters is the intrinsic merit of the specific analyses (in terms of economy and empirical scope) rather than their embedding within this or that general model. The point is illustrated with examples drawn from the study of sound change from Latin to Romance.