| The comparative method in historical linguistics explains sound change as an aggregate of partial theories, to wit, the theory of phonetically-motivated regular sound change and theories about more sporadic diachronic processes, such as analogy and borrowing. From its inception, the efficacy of the comparative method, a well-tested research method refined over the last century by successively broader application to the Indo-European as well as other language families, has been challenged.; However, many of the recent challenges to the method are premature inasmuch as relatively few language families have been reconstructed in detail. Many properties, statistical and otherwise, of historical reconstructions have not been characterized with enough clarity and in enough detail to make a principled critique possible. Indeed, in many cases the data required to raise or answer criticism is not easily available. Until an exhaustive data set along with a (at least provisionally) completed analysis is provided, both critics and defenders are at a disadvantage.; 'Corpus-based' computational methods provide a means to create and test explicitly falsifiable hypotheses, to distribute the data and hypotheses to other researchers, and to study the properties of existing reconstructions.; Many obstacles present themselves to the creation of cross-linguistic corpora suitable for diachronic research. Well-known linguistic problems concerned with semantic and phonological comparability of disparate data sets must be solved. Also, computational barriers such as solution space size must also be addressed. These obstacles and possible solutions are discussed.; {dollar}{dollar}``Calomniez, calomniez, il en restera toujours quelque chose.''{dollar}{dollar}... |
