Abstract
The task facing us at the end of the last chapter was this: we had to find a means of constraining the theory of generative phonology, cutting down on extrinsic ordering and the postulating of abstract segments, without abandoning the significant generalisations which that theory allowed us to express. One restriction on extrinsic rule ordering we have discussed is the Elsewhere Condition. We saw that this general condition on rule application reduced the amount of extrinsic ordering required, since the order of application of some rules will follow from that general principle, rather than having to be stipulated in each grammar. If, in addition to this, certain ordering relations followed as a consequence of the structure of the grammar, that would also constitute an advance. Let us now look at a revision of the organisation of the grammar which has just that consequence.
This is a preview of subscription content, log in via an institution.
Preview
Unable to display preview. Download preview PDF.
Further Reading
For an introduction to Lexical Phonology, see Kaisse and Shaw (1985). For several case studies of rules applying lexically and postlexically, see Kiparsky (1985). A major early paper is Kiparsky (1982); if you have followed the abstractness debate from Kiparsky (1968) to Kiparsky (1973), you can pick up the development of Kiparsky’s ideas here, then proceed to Kiparsky (1985) and then later material. Much of the literature on level ordering and the evolution of the notion ‘cyclic application’ may prove difficult if you are unfamiliar with the issues in morphological theory, but the works cited here should be tractable, if rather dense in places.
Investigations within the LP model, such as Mohanan (1986), have brought out interesting relationships between the notions ‘neutralisation’, ‘derived environment’, ‘cyclicity’, ‘underspecification’, ‘preservation of contrast’ and the Elsewhere Condition. It remains to be seen exactly how the relationships between these ideas are best represented. The idea behind the Revised Alternation Condition may be expressed in LP terms as a Derived Environment Constraint: lexical rules may apply in derived environments only. It has also frequently been embodied in what is referred to as the Strict Cyclicity Condition (SCC): cyclic rules may apply in derived environments only (Kaisse and Shaw 1985: 17). We have not investigated the SCC here because of difficulties in presenting a consensus account of what it is for a segmental rule to be cyclic. Mascaró (1976) defines cyclic rules as non-automatic neutralisation rules, but there is disagreement as to just how much of the lexical phonology of a language must be said to be subject to the SCC. For Halle and Mohanan (1985), the SCC may cease to apply after a given level (in English, after level 1). Booij and Rubach (1987) argue that there is a class of lexical rules which are postcyclic (and thus not constrained by the SCC), which are implicated in certain sorts of cliticisation process (see Section 9.5 for some discussion of cliticisation, albeit of a postlexical sort).
Some phonologists (e.g. Kiparsky 1982) suggest that the SCC need not be stated as an independent principle, since the effects attributed to the SCC follow from the Elsewhere Condition (EC). There is also disagreement as to exactly how strict cyclicity effects may be derived from the EC (see Giegerich 1988 for an alternative account which does not appeal to the notion of a ‘lexical identity rule’ appealed to by Kiparsky). Other phonologists have denied that the effects of the SCC can be derived from the EC at all. See the sceptical comments made by Anderson (1974: 141) on the applicability of the ‘cyclicity’ notion to rules other than those which assign word stress, and Hualde (1989) for a claim that there may be cyclic rules which do not obey the derived environment constraint. Some metrical phonologists (see Sections 9.5 and 9.6), such as Liberman and Prince (1977), have sought to show that the notion of the cycle is inapplicable even to word stress assignment. Kiparsky (1979) is a major defence of the cyclic nature of word stress assignment.
Kiparksy (1985) has suggested that Structure Preservation (SP) too may be switched off after a given level. It is clear that, since SP and the SCC are two principal constraining forces in Lexical Phonetics, then any diminution in the number of rules which are subject to those principles will constitute a diminution of the constraints which LP puts on the power of phonological rules.
The idea that there may be ‘loops’ from one level back to another has also been proposed in the LP literature, and this constitutes a severe compromise on the rationale for the model. Szpyra (1989) denies that the level-ordered approach can be sustained, and thus denies that Lexical Phonology has any intrinsic merit. Halle and Vergnaud (1987) adopt the idea of level-ordered phonology, but not the idea of level-ordered morphology. For them, the morphology forms a separate module in the grammar which precedes the phonology, so that all of the morphology is located in the lexicon prior to the operation of the lexical rules. Otherwise, they accept various parts of the LP approach, such as the lexical/postlexical and cyclic/non-cyclic distinctions. Just how much of the core of Lexical Phonology will survive remains to be seen.
Copyright information
© 1993 Philip Carr
About this chapter
Cite this chapter
Carr, P. (1993). The Role of the Lexicon. In: Phonology. Modern Linguistics Series. Palgrave, London. https://doi.org/10.1007/978-1-349-22849-2_9
Download citation
DOI: https://doi.org/10.1007/978-1-349-22849-2_9
Publisher Name: Palgrave, London
Print ISBN: 978-0-333-51908-0
Online ISBN: 978-1-349-22849-2
eBook Packages: Palgrave Social & Cultural Studies CollectionSocial Sciences (R0)