The Division of Labour in Phonological Theory

Teamleader:
Patrik Bye

Other Team members:
bruce Morén
Tore Nesset
Sylvia Blaho

Time span:
Project starts:
26.05.2004

Project ends:
31.12.2007

Brief description and goals:
This project intends to explore the division of labour between the following explanatory principles in phonological theory.

Division of labour between representations and rules/constraints

Phonologists have long recognised a descriptive trade-off between representations and rules/constraints according to which the grammar computes the well-formedness of these representations. While all theories use representations of one kind or another, the explanatory burden assigned to rules/constraints differs according to the theory. Formal frameworks run almost the entire gamut from purely representational theories such as Firthian Prosodic Analysis to theories such as Optimality Theory in which the descriptive burden rests more or less entirely on the constraints. Other approaches such as derivational rule-based theory and Declarative Phonology occupy intermediate positions along the scale.

Another issue which falls under this rubric is the nature of representations given our assumptions about rules/constraints. Derivational rule-based theory and Optimality Theory alike assume that operations in phonology have transformational power. This assumption has made it possible to construe phonological primitives as having inherent phonetic content. Declarative Phonology, on the other hand, has had problems construing primitives in this way when confronted with natural language data. This prompts us to look anew at the nature of phonological representations to see to what extent our representational theories can be modified to bring monostratal theories into the fold of descriptive adequacy.

Universal/innate vs. emergence

A central question in modern phonological theory is the extent to which phonological rules/constraints and representational primitives should be seen as arising from an innate linguistic endowment, and also to what extent the formal properties of phonology are share with the formal properties syntax and other linguistic modules. The time is right to look at these question anew given recent directions arising out of the paper by Chomsky, Hauser and Fitch (2002), according to whom our innate linguistic endowment reduces to the ability to form recursive expressions from discrete categorical symbols. Is the substantive content of phonological categories informed by innate specifications or are they the product of convergent evolution under similar physiological constraints on articulation, perception, processing and memory? In addition to discrete combination, do rules/constraints make an independent contribution to generating the set of expressions of language L or can they be dispensed with in favour of richer theories of representation? Related to this issue is the extent to which phonological categories are modality specific.

Phonology vs. morphology/lexicon

Continuing in this minimalist line of enquiry, we wish to probe in the capabilities of the phonology more generally. Most approaches take for granted that phonology must account at least for phonotactic patterns and alternations. Alternations, however, may turn out to be treatable in terms of affixation, i.e. unification of the phonological descriptions of lexical entries with phonologically underspecified stems. Certain morphologized alternations, such as Initial Consonant Mutation in the Celtic languages, have been viewed in these terms. Variant alternations in the same morphosyntactic context would then entail a special type of allomorphy. Extending this line of analysis to alternations in general would bring morphologically and phonologically conditioned alternations under the same formal rubric.