Maryland Mayfest 1999

So far, generative theories have accounted for regularities by extrinsic means, in the form of rules or constraints. I will claim that a significant source of regularities in the lexicon is also the nature of the representations themselves, based on the simple assumption in (1).

(1)
Mental representations of linguistic expressions form sets of entailments, each aspect of the representation entailing each of the others.

This means that two partially similar representations like (2a, b), each standing for a vector in multidimensional space (here, with 4 dimensions), will be under some pressure to be identical.

(2)
a. A B C D
b. A B C X

The reason is that, under (1), three of the entailments generated by (2a) (A=>D; B=>D; C=>D) will be violated by (2b), and vice-versa. Calculation of the effects of such entailments can be done in OT, where they can be treated like other constraints, themselves just types of entailments. I will argue that (1) is in fact the ultimate source of the OO-FAITH (output-to-output faithfulness, alias 'Anti-allomorphy/ Consistency/ Uniform exponence') constraints, which have been argued for in much recent work (Benua, Burzio, Kenstowicz, others). I will argue further that (1) successfully addresses the following three major problems that the introduction of OO-FAITH into the theory of the lexicon raises:

I. Conceptual redundance of OO-FAITH and traditional word-formation rules (WFRs), both yielding patterns of surface-to-surface similarity among words.
II. Modulation of OO-FAITH effects by the independent degree of similarity. For instance, cÛmparable is both semantically drifted and metrically unfaithful to comp·re, while comp·rable is neither, showing that evaluation of either semantic or metrical faithfulness needs to take account of the other.
III. Identifying the trigger of OO-FAITH. Recent work (Steriade, Burzio) shows that derived forms can in principle inherit structure from any co-member of their paradigm, not just from their 'morpho-syntactic bases'.

References

Benua, Laura (1997) Transderivational Identity: Phonological Relations between Words, PhD Dissertation, U. Mass. Amherst.
Burzio, L. (1994) Principles of English Stress, Cambridge University Press.
Burzio, L. (1998) 'Multiple Correspondence', Lingua 103, 79-109.
Burzio, L. (to appear) 'Cycles, Non-Derived-Environment Blocking, and Correspondence,' in Joost Dekkers, Frank van der Leeuw and Jeroen van de Weijer, eds. Optimality Theory: Syntax, Phonology, and Acquisition. Oxford University Press.
Kenstowicz, Michael (1996) 'Base Identity and Uniform exponence: Alternatives to Cyclicity,' in: Durand, Jacques & Bernard Laks(eds.) Current Trends in Phonology: Models and Methods. European Studies Research Institute, University of Salford Publications. (363-394).
Steriade, D. (1997) 'Lexical conservatism' in SICOL, 1997: Linguistics in the Morning Calm, Hanshin, Seoul, Korea.

Back to the previous page