Can Linguistics Use a Shave from Ockham’s Razor?
University of Washington
Friday, May 18, 2012
3:30 – 5:00 p.m.
This presentation is a progress report of work done so far on a project to apply Ockham’s Razor to a slice of linguistic literature, studies on French phonology over the last half century plus. Conventional wisdom holds that Ockham’s Razor constitutes a principle of parsimony, Pluralitas non est ponenda sine necessitate, ‘A plurality is not to be posited without necessity’. This principle has been used as a guide in linguistic theory over the years, as well as in the sciences. However, no linguist has ever examined what Ockham himself actually did! In my research, I have tried to understand and summarize Ockham’s methodology, which, a few centuries after his death, was given the label ‘Razor’. I consider it invaluable to engage Ockham’s contribution to 14th century metaphysics, before applying his principle of parsimony to modern linguistics. In this encounter, we discover a subtle interplay between the Razor and, unknown to most, an ‘Anti-Razor’, of which Ockham held one version and a Franciscan colleague of his, Walter of Chatton, offered another interpretation. Crucially, the two principles are distinguished by markedness, as the Razor assumes the ABSENCE of an entity as normal, or unmarked, while the Anti-Razor considers the PRESENCE of an entity as unmarked. I will show how this markedness can be used for an understanding of Ockham’s nominalist reduction of the ten Aristotelian categories down to two. The Razor vs. Anti-Razor dichotomy also serves as the backbone of my overview of theories on French phonology.
I distinguish five stages, (1) American structuralism of the 1950’s, (2) Generative phonology in the 1960’s, (3) Concrete / Natural phonology of the 1970’s, (4) Non-linear phonology in the 1980’s, and (5) Optimality theory for the 1990’s and beyond. Tentatively, I propose an alternating evolution, in which (1) is best identified with the Razor, (2) with the Anti-Razor, (3) with the Razor, and (4) and (5) with the Anti-Razor. The only true dialectic relationship between Razor and Anti-Razor may be seen in the Generative phonology vs. Concrete phonology sequential interplay, the latter ‘shaving’ the former in much the same way that Ockham ‘razored’ the Aristotelian categories. At the end, however, I cannot conclude that Ockham’s Razor supplies a clear-cut and ‘universal’ metric of parsimony. Instead, I make the more modest proposal that the Razor creates for linguistic theory conditions in which such a metric becomes plausible and conceivable. In other words, Ockham’s Razor equates to a ‘claim of parsimony’, not to parsimony itself: It attempts to prevent not to be able to prevent the positing of unnecessary entities.