Gradient acoustic information induces long-lasting referential uncertainty in short discourses

ABSTRACT

Three experiments examined the influence of gradient acoustic information on referential interpretation during spoken language processing and how this influence persists over time. Acoustic continua varying between the pronouns “he” and “she” were created and validated in two offline experiments. A third experiment examined whether these acoustic differences influence online pronoun interpretation, and whether this influence persists across words in a discourse. Measures of eye gaze showed immediate sensitivity to graded acoustic information. Moreover, acoustically induced uncertainty persisted across a five-word delay: When listeners encountered a word that disambiguated the referent of the pronoun differently than it had originally been interpreted, the amount of time they took to recover from an initial misinterpretation was directly related to distance along the acoustic continuum between the pronoun and the endpoint corresponding to the correct referent. These findings show that fine-grained acoustic detail induces referential uncertainty that is maintained over extended periods of time.

KEYWORDS:

• Gradient
• acoustic
• eye-tracking
• reference
• discourse

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 It is common to describe /h/ phonetically as a diacritic to the adjacent vowel, rather than as a voiceless fricative. However, this distinction is not critical to the current study, insofar as listeners perceive our stimuli as /h/-like (i.e. interpreting the /h/ endpoint as he’s).

Additional information

Funding

This material is based on work supported by National Science Foundation Grants BCS 12-57029 and 15-56700 to Sarah Brown-Schmidt and a Beckman Institute for Advanced Science and Technology Postdoctoral Fellowship to JCT at the University of Illinois Urbana-Champaign.