Output

Predictability, Anticipation, and Pre-activation effects on sentence processing (together with Evelina Fedorenko & Ted Gibson)

This builds on an effect we found in an earlier experiment (and which we since then have replicated). We found an environment in English in which verbs were processed faster the more material intervened between the verb and it's subject. We are currently preparing/runnning two more experiments that tease apart different interpretation of this effect. Is it merely anticipation (a la Konieczny 2000) or pre-activation of consituents that later have to retrieved from memory (as recently suggested by Vasishth, to appear), or predictability/surprisal (not that those are the same; see e.g. Levy, 2005).

More generally we are interested in what is anticipated/predicted (the presence of a verb? the type of verb? the verbsense?).

I'm helping with a ME study on quotative markers (designed by Isa Buchstaller)

In progress.

Four magnitude estimation experiments and two reading experiments (and going) on WH Questions (together with Inbal Arnon, Bruno Estegarribia, Philip Hofmeister, Jeanette Pettibone, Ivan Sag, and Neal Snider)

In this our WH Research Group we study certain order effects in English wh-questions. We found effects of the distance between the wh-filler and its intgration site (the gap) on acceptability, as well as effects of the complexity of the intervening material as well as the complexity of the extracted filler itself.

Several of our results were presented at CUNY Rethinking Superiority effects-a processing model, S-TREND, and SPLAT.

Existentials vs. Canonical subjects (together with David Beaver)

In progress.

This is part of a growing research group interested in the distribution of NP types between existentials and canonical subject constructions. David and I are using the Penn Treebank III corpora to confirm the trends earlier google searches revealed under more careful controls. More on this project later.

Relativizer Variation in non-subject-extracted relative clauses (together with Tom Wasow, David Orr; some also with Mary Ann Walter)

A series of studies on mostly the syntactically parsed Switchboard corpus of spoken English (but also the Wall Street Journal and the Brown Corpus of the Penn Treebank III). We have several presentations on this research (e.g., on production- vs. comprehension-oriented accounts of optional that, on accessibility effects, on predictability effects, etc. in non-subject-extracted relative clauses) and some write-ups have followed as well (e.g. for BLS, DGfS, CLS, DiSS'05).

We're planning to run further studies on the Parsed PPCME2 Corpus of Old English to investigate the diachronice development of relativizer variation.

I'm helping Gaetanelle Gilquin with an online study of sentence production, which uses written sentence elication. The purpose of this experiment was to test whether so-called prototypicality is the same as frequency of occurrence (two notions which are often conflated in the literature). In particular, the experiment investigated whether the distribution of verb senses (e.g. physical_transfer vs. abstract_transfer for the verb "give") obtained by free elicitation is the same as verb sense frequency in every-day language use. In the critical conditions participants saw a verb and had to construct a sentence containing it. The distribution of the elicited verb senses was taken to be an indicator of prototypicality (of the senses given the verb). Gaetanelle compared this distribution to the distribution of senses in both written and spoken corpora.

The results, which Gaetanelle presented at From Gram to Mind: Grammar as Cognition, argue that corpus frequency is not the same as proto-typicality (here: the distribution of senses in freely elicited data). The paper is not available online, but feel free to contact Gaetanelle Gilquin.

Differences between the production and comprehension of relative clauses (together with Evelina Fedorenko & Ted Gibson)

This experiment compares the effect of post-verbal complexity in relative clauses on comprehension difficulty with data on production complexity gathered in the Switchboard Corpus.

For results, see our CUNY presentation.

Phrasing in the post-nuclear domain (with Elisabeth Norcliffe)

Two small pilot reading experiments (4-5 subjects each) eliciting phrasing due to avoidance of syntactic ambiguity. The novel idea (thx to Edward Flemming) was to look at phrasing in the post-nuclear domain. Our data suggest that there are either intermediate phrase without recognizable pitch accents or there is prosodic phrasing (as evidence by phrase final lengthening) below the level of the intermediate phrase but above the level of the prosodic word.

For the results see our presentation at the 2005 LSA Meeting and the more recent proceedings paper for Interspeech 2005.

Processing of second occurrence focus (together with Daniel Richardson)

This experiments uses auditory presentation of short monologues containing focus and second occurrence focus expressions to compare the processing of both by means of eye-tracking.

Binding in Picture NPs

This experiment (20 participants) was conducted on-line using the WebExp Software package. The goal of the experiment was to test the effect of agentivity on binding of pronouns and anaphors in picture NPs.

For results see my LFG04 presentation Binding theory in LFG (and HSPG) revisited: The case of exemption or the paper.

Perception of non-standard focus-marking (with David Beaver, Brady Clark, and Edward Flemming)

A small experiment (14 participants) confirming that non-standard focus marking (in this case second occurrence focus-marking) is correctly perceived at above chance-level.

The materials of the experiment are available online.

For the results see our LSA presentation The perception of second occurrence focus or my 2nd QP.

Production of second occurrence focus (with David Beaver, Brady Clark, Edward Flemming, and Maria Wolters)

Reading experiment (20 participants) confirming that second occurrence focus is marked phonetically. We elicited material associated with focus-sensitive operators (e.g. only) in post-nuclear domains (i.e. following contrastive accent). In the absence of a pitch accent focus can be marked by increased intensity and length.

For the results see the LSA presentation on David Beaver's website or my 2nd QP.

Free Variation: Correlation between acceptability and production frequency (with Andrew Koontz-Garboden & Lev Blumenfeld)

This (still unfinished) experiment had four parts and was intended to examine possible correlations between production frequency of a person in free variation environments (e.g. certain binding environments where pronouns and anaphora seem to be equally acceptable) and the acceptablity judgments given by the same person.

Contrast vs. Focus - Categorical or Continuous?

A small pilot reading experiment eliciting a range of short dialogues. The contexts were manipulated in a way that caused different degrees of 'contrast/focus' from strong corrections ("No, it was JOHN ...") to mere answer provision of new information ("It was John ...").

I was interested in whether the phonetic realization of contrast and focus differed categorically or rather formed a continuum.

Wide vs. Narrow Focus (with Michael Wagner)

A small experiment with audio-visually presented contexts that were designed in a way that elicited answers with different focus structures (narrow or wide focus on the object or subject).

For some results see When Warriors mourn longer.