Counter-Examples in Linguistics (Science): The Case of Circassian as a Split Anaphor Language by John Colarusso

Linguistica Atlantica Vol 25 (2004)

John Colarusso
McMaster University


Linguists often resist data that undermines the dominant paradigm to which they adhere. This paper examines split anaphors in Circassian, a language of the Caucasus, as a case study of such rejection.

A typology of counterexamples is devised and contrastively applied to physics and to linguistics, with etTortsmade to cite examples from each field. The split anaphor case is presented as an error in prediction and hence as a refutation of the Government and Binding paradigm. Its treatment is contrasted
with that of the orbit of Mercury, a comparable error in prediction of Newtonian mechanics.

A symmetry-breaking approach is taken to the problem of split anaphor (in which reflexivesare ergative while reciprocals are anti-ergative).A new explanation for ergativity is offered. This explanation predicts that only ergative languages with a particular rule coupling will exhibit split ergativity.

As anyone who has worked in linguistics can attest counter-examples have had a distinctly weak role in the recent history of the field. Followers of Chomsky have maintained that for a counter-example to have force, it must come with its own full counter-theory. Curiously, Chomsky himself has always been more open to examining data on its own merits, but such distortions are not unusual in the history of any field. In effect, this position taken by many of his students rules out virtually all counter-examples of any interest. I shall argue that this position, never articulated as far as I know in any formal document, is based upon an erroneous concept of the scientific enterprise.

Download the full-text document in PDF format