Volume 49, Issue 4

This article presents several acceptability rating experiments concerned with crossing wh-movement in German multiple questions. Our results show that there is no general superiority effect in German, thus refuting claims to the contrary by Featherston (Linguistics 43: 667–711, 2005). However, acceptability is reduced when a wh-phrase crosses a wh-subject with which it agrees in animacy. We explain this finding in terms of the availability of different sorting keys for the answers to the multiple questions.

Most Northwest Solomonic (NWS) languages employ possessor-indexing or former possessor-indexing morphology to index subject on verbs. Although the phenomenon of dependent-marking languages using cognate morphology to case mark possessors and subjects is well known, head marking in the form of subject agreement using possessor-indexing morphology of the kind seen in NWS is less well understood. Moreover, most instances of cognate possessor-indexing and subject-indexing morphology typically result from independent pronouns in parallel processes of grammaticalization. In NWS the subject-indexing function arose from a well-established preexisting possessor-indexing function, demonstrating a diachronic functional shift from nominal marking to verb marking. Moreover, in several NWS languages two possessor-indexing hosts encoding distinct semantic categories of possession occur in verbal constructions marking different kinds of events, raising the question of how a semantic distinction between kinds of entities can be meaningfully applied to categories of events. This article surveys functional and formal characteristics of synchronic “possessive” subject-indexing morphology across NWS, and proposes an originating nominalization construction in Proto-NWS from which the synchronic phenomena arose. It reappraises Ross's (The development of the verb phrase in the Oceanic languages of the Bougainville region, Pacific Linguistics, 1982) hypothesis on the origin of the phenomenon, and the hypothesis that two distinct clause structures are involved in the presence or absence of this morphology. It then investigates several key problems not previously addressed, including why the order of subject-indexing morphology and verb it accompanies is the reverse of the order of possessor-indexing morphology and noun in nominal constructions, and how distinct nominal categories of possession encoded by the two possessor-indexing hosts could be reinterpreted as a verbal distinction of event type.

This article investigates the morphologically conditioned stress system of a previously undocumented Uto-Aztecan language, Choguita Rarámuri (Tarahumara). Stress distribution in Choguita Rarámuri results from a complex interaction between lexically prespecified stress, two systematic subpatterns (second and third syllable stress), a stress rule specific to noun incorporation constructions, and an initial three-syllable window, a highly unusual typological pattern. This article presents a cophonology analysis that captures the full range of stress alternations in this language through the association of morphological constructions with fully general phonological subgrammars, and through the association of morphologically conditioned phonological effects to the hierarchical structure of morphologically complex words. The cophonology analysis is contrasted with a Root Controlled Accent (RCA) analysis, an indexed constraint analysis argued for in Alderete (Natural Language and Linguistic Theory 19: 455–502, 2001a, Morphologically governed accent in optimality theory, Routledge, 2001b) to account for the stress patterns of Cupeño, a Uto-Aztecan language with stress alternations similar to those attested in Choguita Rarámuri. In an RCA analysis, stress alternations are derived through a single constraint ranking and faithfulness constraints indexed to specific morphological contexts. This article argues that an analysis that overlooks morphological constituency and which indexes faithfulness constraints to morphological environments undergenerates in the Choguita Rarámuri case.

Expressibility, namely the condition that whatever can be thought can be said, is for strong reasons considered as an essential property of natural languages. To avoid circularity, thought cannot be identified here as what language expresses. The present paper argues that completeness of language with regard to thought is a natural consequence of the fact that the language faculty is essentially the capacity to acquire and use combinatorial systems of symbols. In contrast to iconic signs, symbolic systems do not depend on similarity between signal and meaning, but are based on convention. This symbolic nature of language provides access to any domain of human experience, since no situational connection or similarity between signal and denotatum is required; the combinatorial character allows for any degree of detail, as it provides for expressions of arbitrary complexity. The symbolic and combinatorial nature of human languages implies their discrete and abstract character, by which they are limited to the expression of discrete meanings. Mental structures that are bound to similarity with the signal they rely on are therefore outside the range of language. Percepts of faces and the meaning of music are briefly discussed as mental representations that cannot be verbalized. The symbolic nature of language sets the limits of expressibility, but it also allows for metalanguage and definitions, which in turn are means to overcome local constraints on expressibility. Finally, expressibility is to be distinguished from codability, i.e., the preference for optimal expression and its consequences, which shape conventions and use of symbols.

Several languages have verbal tense, aspect and mood (TAM) and nonfiniteness markers that originate from spatial cases and adpositions. However, the evidence for the gradual loss of the original transparent and systematic correspondences is scarce. Due to the lack of analysis of rich and transparent verbal case systems in modern languages, the essence of the relationships between TAM and locative meanings is still a matter of debate, as in case of the progressive and the absentive. The Finnic dialect continuum provides unique data for clarifying the unresolved issues. Finnic parades an elaborate system of nonfinites with the properties of nominals and verbs. The nonfinites combine with a rich case system whose spatial meaning is bleached to different extent in the combinations. I coin a new term for the case phenomenon — cross-categorial case — and I analyze the Finnic rich cross-categorial case paradigms to provide evidence about the functioning of a transparent system of spatial forms and TAM meanings. A corpus study of the Estonian inessive nonfinites (the mas -construction) exemplifies how the correspondences in space and time as well as social cognition interact in the progressive and the absentive.