Author: tdrobert

Join us this Tuesday, February 21st, for a colloquium talk by Sam Zukoff (MIT), at 1:30pm in Hum 2, room 259. His talk is entitled “Stress Restricts Reduplication” and the abstract is given below:

This paper considers the typology of reduplicant shape, and argues that a system with freelyrankable templatic constraints on reduplicant size/shape over-generates. A survey of Australian languages with quantity insensitive left-to-right alternating cyclic stress systems finds that monosyllabic prefixal reduplicants are not attested; all prefixal partial reduplication patterns in such languages are disyllabic. The disyllabic pattern allows for complete satisfaction of all otherwise undominated stress constraints, whereas any monosyllabic reduplicant would induce violation of one of these constraints. The typological absence of the monosyllabic pattern in these languages thus follows only if templatic constraints (“Reduplicant Size”) must be subordinated to otherwise undominated stress constraints (“Stress Requirements”). This is captured through a meta-ranking condition on the phonological grammar: StressReq >> RedSize (S>>R). The paper further explores how this meta-ranking is compatible with prosodically variable yet predictable reduplicant shape in Ponapean, and an apparently problematic case of monosyllabic reduplication in Ngan’gityemerri which turns out to be the exception that proves the rule.

Our second talk this week will be Thursday, February 23rd, given by Ryan Bennett (Yale), at 1:30pm in Hum 1, room 202. His talk is entitled “Idiosyncrasy and contextual variability in the prosody of functional morphemes” and the abstract is given below:

Dependent morphemes (affixes, clitics) may idiosyncratically select for prosodic properties of their hosts (Inkelas 1990, Zec 2005, etc.). For example, the English comparative suffix -er does not attach to stems greater than two syllables in size (smart-er vs. *intelligent-er). Violation of a prosodic subcategorization frame may lead to simple ungrammaticality (e.g. *pretentious-er). In other cases, subcategorization requirements are met by restructuring the prosody of a morpheme’s host. In this talk we consider several case studies in which functional morphemes idiosyncratically impose a particular prosodic structure on their hosts, sometimes with dramatic results.

In Macedonian, preverbal object clitics are typically unstressable ( ‘(s)he saw him’, *). But in the presence of wh-words or sentential negation, such clitics are parsed into the same prosodic word as the verb and may bear stress ( ‘Who saw him?’). In Kaqchikel, a variety of diagnostics indicate that absolutive agreement markers have a different prosodic parse depending on the presence or absence of outer aspect marking (e.g. [xin-b’e] ‘I went’ vs. [in=jwi’] ‘I am intelligent’). The puzzle here is understanding why the prosody of “inner” morphemes (e.g. object clitics) depends on the occurrence of a specific “outer” morpheme (e.g. wh-words).

We propose that these patterns arise from prosodic subcategorization: the “outer” morphemes in question have subcategorization requirements which force re-parsing of their hosts, including any dependent morphemes present in the same structure. We account for this behavior in a novel theory of subcategorization which makes extensive use of prosodic recursion, and which emphasizes the prosodic result of combining a dependent morpheme with its host. We then consider possible extensions of this framework to Chamorro and English, and conclude with the theoretical and methodological implications of our proposal.

This Thursday, February 16th, Elliott Moreton (UNC) will be giving a colloquium talk at 1:30pm in Hum 1, Room 210. His talk is entitled “Emergent positional privilege in blend formation”, and the abstract is included below:

In many languages, sounds in certain “privileged” positions preserve marked structure which is eliminated elsewhere. Does that happen because learners are predisposed towards grammars which protect those positions, or because sounds in those positions are less likely to suffer misperceptive sound change? To test directly for predisposition towards positional faithfulness, a lexical blending task was used to force a choice of which source word to be more unfaithful to. Given source words which could be blended in two ways (e.g., flamingo + mongoose -> “flamingoose” or “flamongoose”), participants preferentially matched the blend that preserved more second-word phonology (“flamongoose”) to a definition which made the second word privileged. The experiments tested for two kinds of faithfulness — to segments and to mainstress location — and in three privileged positions: morpho-semantic heads (vs. non-heads), nouns (vs. verbs), and proper nouns (vs. common nouns). Results from English and, in one experiment, Spanish speakers show both segmental and stress privilege effects in heads, proper nouns, and, to a lesser extent, nouns generally. We interpret these results as evidence for the universal availability of positional-privilege constraints protecting any phonological property of any privileged position.