Linguistics Seminar Series

Although the functionally unmarked word order in Hittite is robustly SOV, many other word orders are well attested. In addition to some that are syntactically licensed and bear various discourse structure functions, there are also a number of quite puzzling configurations that involve discontinuous constituents and appear unmotivated in terms of discourse structure. Violations of well-known syntactic constraints suggest that these orders are phonologically motivated. Building on previous evidence that Hittite has “phrasal stress”, I will argue that many if not all such orders reflect: (1) that the primary accent in all Hittite phonological phrases (which mostly match syntactic phrases) falls on the leftmost constituent; (2) that some prosodically weak constituents (e.g., indefinite adjectives) require a phonological word as their immediate leftward host, while others (e.g., relative adjectives) require only a phonological phrase; (3) that the last two rules are violable, resulting in some “exceptions” to the dominant patterns. Further study is needed regarding what determines the “prosodically weak” status of some elements.

This study discusses perceptions of variation across dialects of Arabic in the Arab world as revealed through a perceptual dialectology map task. On a map of the Arab world, female undergraduate students at Qatar University provided information about boundaries where people speak differently and labels for those boundaries. A correlation analysis of the boundaries showed that participants viewed Arabic dialects as constituting five major dialect groups: the Maghreb, Egypt and Sudan, the Levant, the Gulf, and Somalia. A closer analysis of the content of the labels revealed variation in terms of principal (Goffman 1981) on whom they draw in their judgments, the latter being either individual, regional (intermediate) or wide-scope generic. This analysis not only identifies more granularity in the concept of principal, it also quantifies the different kinds of principal and identifies statistical relationships between them, the labels, and the boundaries.