Conceptualizing Dialects as Implicational Constellations rather than as Entities Bounded by Isoglossic Bundles
There are no grounds for thinking that the pursuit of dialectology with isoglossic approaches and in terms of the word geographies and similar productions that in the past have eventuated from this approach will ever tell us more than the little that they already have about human language, about the way languages vary, or finally--the ultimate goal--about human nature. The disproved assumptions and theoretically barren models and methods of the past have never surmounted the discrediting arguments that have been lodged against them (see further on). For such reasons, I will be advocating a change in approach during the coming years. Before saying briefly what sorts of change I will argue for, let me note that the data already gathered and to be gathered in dialectological research are a treasure which offers the theory of language and the theory of man far greater benefits than can be derived from cartological displays. The manner in which I think these data could be more profitably collated during the coming years of research is a purely linguistic one that befits our roles as linguists, as I will explain. Preliminary thrusts in this direction have, it should be noted, already contributed substantially to our understanding of the structure of language and its transpersonal use and provided analyses for problems in dialectology not amenable to other means of dealing with these data--including cartographic portrayals.
The first important finding of German and Romance dialectology (cf. e.g. discussions in Gauchat 1903; Hard 1966: 6-7, 29; Bach 1969: 56; Löffler 1974: 135; Coseriu 1975: 13) was that dialect boundaries do not exist: isoglosses do not bundle in a manner that would accommodate the definition of dialects as (mutually intelligible) subvarieties of a language bounded by isoglossic bundles- these bundles containing more or fewer isoglosses according to the greater or lesser importance of the boundary in question. Any bundlings that one postulates are set up at the cost of arbitrarily excluding certain features, fudging the apartness of isoglosses which are claimed to bundle, ignoring divergences among the isoglosses that may bundle in one locale but not in another, and excluding the