Lexical phonology

Prosodic phonology

Chomsky's model of transformational grammar was based on the assumption that phonology was related to syntax and morphology in the following way: readjustment rales are applied to morphs in the underlying representations (D-structures of morphemes), and then the surface structures (S-structures) are subject to (1) the phonological component, which produces phonetic form (PF) and (2) the semantic component, which produces logical form (LF). In addition to rules spelling out grammatical forms Chomsky and Halle proposed readjustment rules that divide complex syntactic expressions into "phonological phrases".

The theory of prosodic phonology which is a later development of generative phonology claims that phonological rules do not operate on syntactic representations at all, but on phonological representations derived from syntactic structures but bearing little resemblance to the latter. Prosodic phonology is not restricted to breaking down phrases of great length into phonological phrases. In prosodic phonology, phonological phrases are part of a hierarchy of prosodic categories the smallest of which is the syllable and the largest, the phonological utterance (which may include more than one sentence). In some ways, prosodic phonology can be considered an extension of metrical phonology, since the prosodic hierarchy includes the syllable and the foot, two important concepts in metrical phonology. Phonological rules can be written to apply within prosodic categories (domain span rules), at the juncture of two categories of a given type (domain juncture rules) or at the end of a category (domain limit rules).

The generative phonological theory that contains powerful derivational devices also needs derivational constraints to avoid large-scale exceptions marking for rules which are not productive. For example: sane — sanity is assumed to be derived from the underlying form [san] which through Tense-Vowel Shift became tense [e] in sane, and in sanity turned into [a] through Trisyllabic Laxing. What the theory needs is an automatic derivational con­straint that blocks Trisyllabic Laxing in nightingale and ivory as well as rul­ing out the possibility of spurious derivation of the lax stressed vowels in camera and pedestal from underlying tense vowels. Here is such a constraint: Derived-Environment Condition. It runs as follows: "Feature-changing rales apply in derived environments only." It constrains Trisyllabic Laxing in such a way as to permit the rule to produce the lax vowels in sanity, serenity, etc., but to block its application in nightingale; moreover, the spurious derivation of the lax vowel in camera through Trisyllabic Laxing is ruled out. The 'trisyllabic' environment of the stressed vowel in san­ity, serenity is brought about by the attachment of the morpheme (-ity) to the morphemes (sane), (serene) etc. while in both nightingale and camera the trisyllabic environments are underlying rather than derived.

In a wider sense lexical phonology, which is concerned with the phono­logical regularities found within the recurrent morphological units, severely constrains the choice of possible underlying representations in every morpheme of the language. The lexical phonology is confined to specifying the pronunciation of individual words, but the postlexical phonology also includes aspects of the pronunciation of phrases and sentences (and longer utterances).

What are the implications of the dynamic models in generative phonology?

Creativity is the key concept here. Human language is creative in that each utterance, outside of a few formulaic expressions like "Good morning", is created anew in accordance with the grammar and phonology of the language.

The speaker's knowledge of general rules enables one to create new and new forms according to the same rules. Speaking a language is creative in the sense that any speaker may produce an infinite number of utterances with the lexicon of morphs and the phonological rules. The creation of new words to correspond to new ideas and technological advances is the norm in most societies. Thus, for instance, about 1500 new words have appeared in the latest edition of Collins dictionary. We need to know phonological rules to predict the pronunciation of new words. We can assume that the same rules play a role, even if only a passive one, in the pronunciation of familiar old words as well. We can see that phonology is a changing scene, it is developing and its rules are productive.

To conclude:

Generative phonology differs from phonemic theory in its objectives and methods.

The main difference between the two theories is that Generative Phonology "seeks to establish a single underlying representation for every morpheme, from which the speaker will then derive, or 'generate' the appropriate phonetic surface form in any given context. This objective is not shared by phonemic theory. The fact that long has the phonemic representation /tan/ in isolation but [lngg] in the morphologically complex form longer is not a problem for this theory: for a phonemicist whose primary interest is to establish the sound system of the language, this is a morphological problem which therefore lies outside his field. For a generativist, on the other hand, complex derivations such as the one deriving [д] from /ng/ are a matter of necessity rather than choice, arising from the overall objective of Generative Phonology. Similarly, vowel alternations such as that in divine — divinity fall into the domain of Generative Phonology: here, again, is a morpheme with two different phonetic forms which therefore must be derivationally related. For the phonemicist, on the other hand, a pair of words such as this one is without interest except that it instantiates the two phonemes, /ai/ and /i/." (Giegerich 1998:301)

Дата добавления: 2014-01-11