Previous work on vowel harmony (VH) in Palestinian Arabic (PA), particularly Kenstowicz (1981), has shown that this process is more adequately expressed in autosegmental than in linear terms, since the former permits the two otherwise uncollapsible regressive and progressive harmony processes to be stated in a single unified rule. The present study complements that autosegmental analysis. It shows that the autosegmental rule of VH in PA is constrained simultaneously by metrical and segmental boundaries. That is, high front vowels acquire the feature [+round] only if they are dominated by a foot marked [+round] and if they occur within the segmental domain of VH which comprises the stem and the preceding prefixes.

Sociolinguistic studies have shown that variation may occur on any of the accepted levels of linguistic analysis. Variation on the phonological level has been the most commonly treated in studies of English from Labov's (1966) New York study onwards. By contrast, it is rather rare that we find studies of morphological or lexical variation (the latter being the variation in the lexical form of words, for example, the Belfast alternation of /Λ/ and /Λ/ in the word foot (L. Milroy, 1980: 118)). However, morphological and lexical variation has been favoured in studies of some other languages, for instance Norwegian (Fintoft & Mjaavatn, 1980; Kerswill, 1985a), Swedish (Thelander, 1982) and Persian (Jahangiri & Hudson, 1982). On the other hand, studies of syntactic and prosodie variation have been altogether much less common (but see Cheshire, 1982; Local, 1982). In spite of differences in emphasis, the various methodologies used have all had the aim of discovering co-variation between linguistic and non-linguistic parameters.

Over the last decade or so, hierarchies such as the animacy hierarchy, the personal hierarchy, the case hierarchy, the definiticity and referentiality hierarchies and so forth have been identified as determinants of constituent order. It is my intention in this paper to review the set of hierarchies and rank them as determinants of NP sequencing in English. I shall from time to time compare and contrast the effect of a hierarchy in other languages with what we find in English; but it remains to be seen whether the ranking which obtains for English has a wider application: I make no a priori claim that it does. The hierarchies and families of hierarchies are ranked as follows, from the most powerful determinant to the least:

(1) the familiarity hierarchy;

(2) the topic < comment, given < new hierarchies;

(3) the universal sequencing conventions;

(4) the definiticity and referentiality hierarchies;

(5) the personal, social status, and role hierarchies;

(6) the dominant descriptor hierarchies;

(7) the formal hierarchies.

It is a standard assumption of government-binding theory that the relationship between a constituent displaced by the transformational rule schema Move α and its trace is subject to the locality condition known as subjacency, the central principle of the subtheory of universal grammar known as bounding theory (Chomsky, 1981, 1982, 1986). Subjacency requires that not more than one ‘barrier’ intervene between a moved constituent and its trace, but the definition of the relevant barriers has been, and remains, an issue of considerable controversy. In Chomsky (1977) it is suggested that NP and one of S or S¯ are the ‘bounding nodes’ for English, and many standard textbooks have since argued for NP and S (e.g. Radford, 1981: Ch. 7; van Riemsdijk & Williams, 1986: Ch. 4). Nevertheless, the possibility of cross-linguistic parametric variation may have to be allowed for, since Rizzi (1978) makes out a case for S¯ rather than S as the clausal bounding node for Italian in order to account for the freedom of extraction from so-called ‘wh-islands’ in that language. Chomsky (1980), however, puts forward the possibility that S¯ may be a bounding node universally, and that languages vary according to whether S is also. If it is, then there will be no long-distance movement (cf. standard German and Russian) unless individual verbs are specified in the lexicon as ‘bridges’ which nullify the barrierhood of S¯ (cf. the majority of verbs subcategorized by clausal complements in English). This view is revised and refined in Chomsky (1981: 307), where S¯ is taken to be a bounding node universally when it includes a complementizer or wh-phrase preceding a finite clause, in which case the finite clause S may also optionally be a barrier, and S is taken to be a bounding node when it is governed, as is the case after S¯-deletion in the complements of ‘raising’ predicates. Finally, Chomsky (1986) seeks to unite the definition of barrier for the purposes both of movement and government, assuming two barriers block movement and one barrier blocks government, by proposing that any ungoverned maximal projection is a barrier, and that any maximal projection immediately dominating such a barrier, whether lexically governed or not, is also a barrier by inheritance.

An interesting development in the last decade or so has been the increasing use that theoretical linguists have made of the notion ‘head’ – or rather, in order not to beg the question, of notions to which they have given the name ‘head’. The term has been around for a long time in linguistics, of course – for example Bloomfield uses it in relation to endocentric constructions (1933: 195), where the head is the daughter constituent which has the same distribution as the mother. Before that, Sweet had used ‘head-word’ to refer to any word to which another is subordinate (1891: 16, quoted in Matthews, 1981: 165). However, theoretical linguists made very little use of the term, or of the constellation of associated concepts, until quite recently. Its present status is due largely to work on X-bar syntax dating from Chomsky (1970), and especially to its recent manifestation in Generalised Phrase Structure Grammar (Gazdar & Pullum, 1981; Gazdar et al., 1985)–and even more so in the ‘head-driven’ variant of this (Pollard, 1985). But the improved status of ‘head’ is also due to some extent to the renewed interest in dependency grammar (Anderson, 1971, 1977; Matthews, 1981; Atkinson, et al, 1982; Hudson, 1984; Nichols, 1986). All these treatments agree not only in using the term ‘head’, but also in using it to refer to the element in some construction to which all the other parts of that construction are (in some sense) subordinate.

The label CLITIC has been applied to a wide variety of phenomena, from words that are prosodically dependent on neighbouring words (as are unaccented monosyllabic prepositions and personal pronouns in English) to words, or even individual morphemes, with idiosyncratic syntactic distributions (like the second-position pronominal and adverbial particles in many languages). I propose here to reserve the term for elements whose description requires more than the stipulation that they may or must be prosodically dependent.

The paradigm functions as a domain for certain processes of change. There can be changes in the number of grammatical categories or in the forms used to represent them. Thus, Benveniste (1968) contrasts change involving only the forms while the system remains the same (‘conservative’ change) with innovations in which there is change in the categories and the oppositions. In this paper, we will discuss both types of change in the Warlpiri pronominal system. Warlpiri is an Aboriginal language spoken in central Australia by about 3000 people. Our research has been carried out at Yuendumu, a remote community 300 km northwest of Alice Springs. While the majority of people in the community (about 750) have Warlpiri as their mother tongue, there are also about 65 non-Aboriginal people who speak English, and a few speakers of other Aboriginal languages closely related to Warlpiri. Although there is limited exposure to English, Warlpiri is the community language at Yuendumu.