-
Reading Certainty: Evidence from a Large Study on NLP and Witness Testimony
- Author(s):
- Ben Miller (see profile)
- Date:
- 2019
- Group(s):
- Digital Humanists
- Subject(s):
- Computational linguistics, Digital humanities, Natural language processing (Computer science), Oral history
- Item Type:
- Conference paper
- Conf. Title:
- DH2020 Conference
- Conf. Org.:
- The Alliance of Digital Humanities Organizations
- Conf. Loc.:
- online
- Conf. Date:
- July 20–24, 2020
- Tag(s):
- Cultural analytics, Natural language processing, Witness histories
- Permanent URL:
- http://dx.doi.org/10.17613/k860-bk55
- Abstract:
- Witness testimony provides the first draft of history, and requires a kind of reading that connects descriptions of events from many perspectives and sources. "Reading Certainty" examines one critical step in that process, namely how a group of approximately 230 readers decided whether a statement about an event is credible and factual. That examination supports an exploration of how readers of primary evidence think about factual and counterfactual statements, and how they interpret the certainty with which a witness makes their statements. This presentation argues that readers of collections of witness testimony were more likely to agree about event descriptions when those providing the description are certain, and that the ability of readers to accept gradations of certainty were better when a witness described factual, rather than counter-factual events. These findings lead to a suggestion for how researchers in linguistics and the humanities could better model the question of speaker certainty, at least when dealing with the kind of narrative non-fiction one finds in witness testimony.
- Metadata:
- xml
- Status:
- Published
- Last Updated:
- 3 years ago
- License:
- Attribution-NonCommercial-NoDerivatives
- Share this:
Downloads
Item Name: miller-dh2020-reading-certainty.pdf
Download View in browser Activity: Downloads: 82