I find myself increasingly drawn to a whole host of questions in linguistics and cognitive science, instead of intonation alone which used to be my sole focus. While the topics are variable, I am beginning to converge on something like an approach that might be useful in studying a variety of topics. (Please count the qualifiers in that sentence.) In short, I am focusing more and more on how people accomplish certain tasks, almost procedurally, as opposed to what people can and cannot do. Of course, they are two sides of the same coin, but each focus does bring up slightly different questions. This interest in how we do something is tied up with what it means to have a causal model of some cognitive phenomenon. This idea isn't quite firm yet; however, I am approaching it increasingly with my working paper on the biological endowment for language. I hope to keep developing the idea in the dissertation, as well as future revisions of the endowment paper.

Here are some works in progress:

Acquisition / Learning Theory:

The biological endowment for language and the poverty of the stimulus. I assume without argument Chomsky's basic point from 1957 in his famous review of Skinner that to understand any complex behavior, we must understand the processes that occur in between input data and output behavior; i.e., we must understand the creature who is acting and understand their biology and cognitive endowment. I then am interested in whether or not the methodology of constructing poverty of the stimulus arguments actually gets us closer to this goal. The basic conclusion is 'no', as most linguistic theories are not truly causal models of cognition, while biology is certainly causal. The arguments here are not quite ready for prime-time, so if someone finds this and wishes to quote, please contact me so that I can at least say whether or not I still agree.

This is a follow-up Tuesday seminar presentation called Aristotle, Causality, and Poverty of the Stimulus. I just used a hand-out here (with typos), but you might be able to see where the causation thought has been heading. I've moved into reading about Bayes Nets and causal graphical models now. Once the causation reading is done, I will revisit the topic.

Dissertation Topic. I'm working on the full proposal now, but it should look at statistical learning in humans and its importance to word segmentation. More specifically, I hope to figure out more details on how statistical learning works, which should then give us interesting predictions about what things can and cannot be learned, as well as the time course of learning. It should be about 2/3 experimental and 1/3 modeling/theory.

Politeness, Face, and Korean Apology:

Concepts of face and apologies in Korean. This paper was developed with my collaborator Jeewon Hahn. It is still a draft, so do not cite without asking. The idea here is to reformulate Brown and Levinson's face theory based upon Korean apology use. The most exciting prospect from the paper, however, is the possible development of a subfield studying the growth and decay of face concepts in societies. I have a hunch that the misleading terms "collectivist" and "individualist" can be replaced by more accurate terms based upon face or identity theory.

Intonation:

An Acoustic Functional Grammar of English Intonation. This is a poster presentation from December 2006 at the Acoustic Society of America conference. It is an experiment in applying Paul Boersma's functional phonology to some basic facts of intonation. There is no paper to accompany the poster, so please contact me with any questions.

Joshua Steele. 1775. This paper published in the Proceedings of a local conference looks at Joshua Steele's attempts to describe English intionation in 1775. Besides a historical look, it examine his use of a modified music notation to transcribe speech. Here is the power point presentation of the paper.

Pitch Accents, Recursion, and the Construction of Local Discourse. This is a powerpoint of another department seminar where I look at the use of pitch height scaling across several sentences in political speech. Here's the abstract:

Previous work on the intonational structure of discourse has focused mostly on the phonological or phonetic expression of global discourse segment boundaries, as well as overall affective features of different styles of discourse. The current project looks at the intonational structure of local discourse, that is the relations between adjacent Intonational Phrases. Through the analysis of American English political speech, some new cues for the structure of discourse are identified. These cues include the repetition of intonational melody to ease processing, the use of deaccented phrases to close a grouping, and parallel and dominating pitch realization on nuclear pitch accents across Intonational Phrases. The pattern of nuclear pitch accent phonetic realization in particular can be used to build elaborate hierarchical discourse structures of a recursive nature. This recursiveness in intonational phonology would seem to be in opposition to the common theoretical assumption of strict layering, which, among other things, explicitly forbids phonological recursion. This will lead to a discussion of the relations between stress, rhythm, intonation, and grouping, and it is proposed that this recursiveness is not part of the normal metrical stress patterns of speech, out of which the Strict Layer Hypothesis grew, but instead a form of cohesion or generalized anaphora. The result is that even though the patterns identified are highly recursive, they are not necessarily in opposition to previous layering proposals.

I did an eye-tracking experiment in which I tested the impact of identical intonational phonology, according to the ToBI model, but differing phonetic pitch heights of pitch accents on discourse interpretation using the visual world paradigm. The pilot results were promising, but the stimuli need to be improved and then tested again with more participants. Yell if curious.

Intonational Stucture of Discourse. I completed a pilot project in Spring '05 looking at the intonational grouping tools that Malcolm X used in a political speech. This is the start of what turned into the Recursion lecture above. However, it's yet to come to full fruition.

Random Other:

I once lead UH's Cognitive Science Colloquium through a debate about Searle's Chinese room though experiment. My job was to defend Searle, while a compsci prof defended Turing. I wrote up these notes, which I just re-read and they are a little bit interesting. I mostly like them because apparently I was really punchy when writing them, writing helpful things like:

"So there. Did I say nah-nah-nah-naaah-nah? If not, then nah-nah-nah-naaah-nah." and...

"But the reason we were interested in strong AI – other than because we want to be sucked out an air lock while orbiting Jupiter, and who doesn't – is that it said it was going to tell us some really cool things about human psychology."