Linguistic Variation - Volume 15, Issue 1, 2015

The notion that phonological features are organized into contrastive hierarchies has been entertained at different times in the history of linguistics. My main focus will be on the motivation for such hierarchies: what principles govern the ordering of the features? I will show that they have been motivated by three different principles: a) Activity: to identify the contrastive features that are relevant to the phonological computation, b) Minimality: to minimize redundancy in phonological representations and to maximize the amount of information conveyed by each feature, and c) Universality: to express universal tendencies in the nature of phonological inventories and the order of acquisition of feature contrasts. These principles do not necessarily conflict in every case, but in practice situations arise where they lead in different directions. To some extent Jakobson and Halle (Jakobson & Halle 1956; Halle 1959) and Clements (2001; 2003; 2009) appeal to all these principles, though they do so with differing emphases: Jakobson began by appealing to Activity, Halle came to stress Minimality, and Clements focused on Universality. I will argue on behalf of the centrality of Activity, which I understand to be the original and most compelling motivation for feature hierarchies.

The Variable Hierarchy hypothesis of Dresher (2009) has a number of far-reaching consequences and applications – beyond the domain for which it was originally developed – including overspecification, kinship terminology, vowel reduction and whistled languages. On the other hand, the Contrastivity-Only Hypothesis of Dresher (2009) is, in its present form, empirically too limited once one looks at a full system of phonological processes within a language, and some reference to noncontrastive features seems inevitable. However, I outline a possible set of diagnostics and distinctions that might allow one to localize and limit the reference to noncontrastive features.

In this paper, we present syntactic and semantic arguments that, in at least one articleless language (Lithuanian), bare nouns are able to project either NP or DP structures (cf. Franks & Pereltsvaig 2004; Ajíbóyè 2006; Pereltsveig 2006). We show that, in some syntactic contexts, Lithuanian bare nouns are only able to receive definite interpretations; in one context, they are only interpreted indefinitely. We then tie these interpretations to the presence or absence of D. Further, we highlight problems with the view that there is a categorical difference between articleless (NP) and articleful (DP) languages (Bošković 2010; Despić 2011), by testing Bošković’s and Despić’s generalizations about NP vs. DP languages on Lithuanian. In both cases, Lithuanian ends up behaving somewhat like an NP language, and somewhat like a DP language. Lithuanian is therefore a counterexample to the NP/DP split posited by Bošković, and must receive a different analysis.

A sub-set of languages with non-clause bound (NCB) reflexives exhibits Blocking effects. In these languages, a NCB reflexive is “blocked” from corefering with a higher subject when a lower subject does not agree with a higher subject for person. Accounting for Blocking effects has been challenging. Recently, researchers have posited a functional head that encodes point of view or sentience (e.g. Huang & Liu 2001; Tenny 2006; Chou 2012). However, these analyses do not account for all available data. Here, I offer a small modification to the most recent analysis of this type that accounts for available data by conforming to Chomsky’s cyclic transfer system (2007, 2008). I also provide exploratory data that suggests that (i) NCB reflexives exist in American English and (ii) they unexpectedly exhibit Blocking effects. Finally, I illustrate that the modified analysis for Mandarin also accounts for the distribution of NCB reflexives in American English.