SPEAKER: Jason Bishop (College of Staten Island CUNY)
Focus projection, prominence, and individual differences: Evidence from (lexical) processing
Theories of Focus Projection (e.g., Selkirk 1995) attempt to explain the ability of a nuclear accent on the internal argument of a verb to "project" focus to larger constituents. For example, a single nuclear pitch accent on an object in English SVO constructions is said to be equally appropriate when focus is "narrowly" on the object, or "broadly" on the entire VP. An important implication of such theories is that non-nuclear accents (e.g., a prenuclear accent on the verb) have little role to play in marking the focus contrast, and indeed many theorists regard them as completely optional with respect to sentence meaning (e.g., Buring 2008). To date, however, experimental evidence has been mostly limited to off-line felicity judgments (e.g., Birch & Clifton 1995, Welby 2003), and results have been mixed.
The present study attempted to bring evidence from online processing to bear on the question of the optionality of prenuclear accents. The cross-modal associative priming paradigm was used, as it has been shown sensitive to both information structural (Blutner & Sommer 1988) and prosodic (Norris et al. 2006, Braun & Tagliapietra 2010) expectations. In the experiments presented, English-speaking listeners made lexical decisions about visual targets (e.g., BRUNETTE) which followed auditorily-presented SVO sentences in which the object was the prime (e.g., He kissed a *blonde*). To test listeners' expectations about how prosody relates to this contrast, both the size of the focus constituent (VP-Foc vs. Object-Foc) and the prosody (presence versus absence of a prenuclear accent on the verb) of the SVO sentences were made to vary. It was assumed that the presence of priming in the lexical decision task would be sensitive to listeners' expectations about the prosodic well-formedness of the test sentences.
Results from priming patterns indicated the following: priming is not an automatic process, but is dependent on (a) the interpretation of the sentence's information structure, (b) the presence versus absence of prenuclear accents, and, crucially, (c) their interaction. Interestingly, the latter finding is also somewhat dependent on individual differences in listeners' "autistic"-like personality traits, which have recently been implicated in the processing of other linguistic contrasts (e.g., Yu 2010, Xiang, Grove & Giannakidou 2012). The details of these finding are discussed, as are the implications for two separate theories of Focus Projection (e.g., Selkirk 1995 and Gussenhoven 1999), which differ in their treatment of prenuclear accents.