Matt Wagers

May 10 Why memory for syntax is so terrible (public lecture)

It is not too hard to remember what people say, but it can be quite difficult to remember precisely how they say it. In this talk, I will explore the science behind 'verbatim memory,' or memory for exact wording. Uncovering why verbatim memory is typically so fragile (but occasionally resilient) offers a window onto how rapid-acting short-term memory processes support the rich structure of language. This has implications not only for questions that interest linguists and psychologists, like how language acquired is by children; but also for ones with broader practical applications - like how much we should trust recollected speech in everyday contexts that matter, like the courtroom.

• Arnett & Wagers. 2017. Subject encodings and retrieval interference. Journal of Memory and Language. 93:22-54. https://doi.org/10.1016/j.jml.2016.07.005.
• Resumptive pronouns and the not-quite-verbatim memory of Twitter users.
• Bellik, Foley, Pizarro-Guevara & Roberts. 2017. Surface memory for prosody. Unpublished poster (Fall 2016 Psycholinguistics Seminar project).

May 11 Morphological informativity and grammatical licensing

In this seminar, I'd like to address two questions: (i) how do speakers integrate morphological information to guide the parsing and interpretation of filler-gap dependencies; and (ii) how does information signaled by morphology interact with information signaled by word order. The core data for this investigation will come from a series of experiments on the comprehension of constituent questions and relative clauses in Chamorro, an Austronesian language of the Mariana Islands; but we will also look at similar phenomena in Tagalog. In both languages, the inflection of verbs can convey information about the location of gaps - either through voice morphology or, in Chamorro, Wh-Agreement. Surprisingly this cue is not always leveraged immediately -– particularly when it points to a gap in object position. I will discuss the implications of this general finding for theories of how different sources of information contribute to parsing, and how syntactic descriptions incrementally feed semantic interpretation.

• Wagers, Borja & Chung. 2015. The real-time comprehension of wh-dependencies in a Wh-Agreement language. Language 91: 109-144. doi:10.1353/lan.2015.0001.
• Wagers, Borja & Chung. 2017. Grammatical licensing and relative clause parsing in a flexible word-order language. Submitted manuscript.
• Pizarro-Guevara & Wagers. 2017. The role of voice morphology in processing Tagalog A-bar dependencies.. In Nomoto, A., Miyauchi, T., & Shiohara, A., Eds., AFLA 23: The Proceedings of the 23rd Meeting of the Austronesian Formal Linguistics Association, pp. 228-242. Canberra: Asia-Pacific Linguistics.
• Chamorro Psycholinguistics na Project web site.

May 12 Competition among pronouns in the grammar and processing of Chamorro. Talk handout.

Much work on Binding Theory has implicitly assumed that reflexive anaphors are morphologically distinct from ordinary pronouns, and indeed in most of the world’s languages this seems to be so (Faltz 1977). Research on the comprehension of reflexive anaphors has likewise been concerned with linguistic systems in which reflexives and ordinary pronouns are morphologically distinct. Nonetheless, there are languages in which reflexive anaphors have the same morphological realization as ordinary pronouns. One such language is Chamorro. Reflexive anaphors in Chamorro look like ordinary overt pronouns. Although the language can use special morphology to mark a verb whose direct object is reflexive, this special verb morphology is optional.

In this seminar, I investigate the syntax and processing of pronouns in Chamorro. First, I'll show that Chamorro grammar treats reflexive anaphors differently from ordinary pronouns, even though the two types of pronouns generally have the same morphological realization. The details support a competition-based theory of anaphora most similar to that proposed by Safir 2014. Specifically: (a) reflexive anaphors are minimal pronouns which originate with just an index feature and acquire phi-features later in the derivation (Kratzer 2009), (b) ordinary pronouns, which originate with phi-features, always lose in the competition with a minimal pronoun with the same construal, and (c) any morphological differences between minimal pronouns and ordinary pronouns arise late, at spell-out.

Then, I turn to the question of how comprehenders confront the challenge of processing morphological pronoun forms in this language. I report on a picture-matching experiment on tablet computers conducted in the Mariana Islands, in which participants first saw a context screen accompanied by context-setting sentences, and then had to match a target sentence with one of two pictures, one depicting a reflexive event and the other, a disjoint event. Perhaps the most surprising result is that the parser prefers to construe overt pronoun forms as minimal pronouns even when a disjoint construal is allowed. We offer an account that derives this result from a competition-based theory of anaphora, together with the Chamorro-specific facts that (x) overt pronoun forms in this language can realize reflexive anaphors or ordinary pronouns and (y) object pronoun forms, which are second-position clitics, must precede the subject.