The influence of place and time on lexical behavior: A distributional analysis
- PDF / 1,112,497 Bytes
- 16 Pages / 595.276 x 790.866 pts Page_size
- 83 Downloads / 177 Views
The influence of place and time on lexical behavior: A distributional analysis Brendan T. Johns 1 & Randall K. Jamieson 2
# The Psychonomic Society, Inc. 2019
Abstract We measured and documented the influence of corpus effects on lexical behavior. Specifically, we used a corpus of over 26,000 fiction books to show that computational models of language trained on samples of language (i.e., subcorpora) representative of the language located in a particular place and time can track differences in people’s experimental language behavior. This conclusion was true across multiple tasks (lexical decision, category production, and word familiarity) and provided insight into the influence that language experience imposes on language processing and organization. We used the assembled corpus and methods to validate a new machine-learning approach for optimizing language models, entitled experiential optimization (Johns, Jones, & Mewhort in Psychonomic Bulletin & Review, 26, 103–126, 2019). Keywords Lexical organization . Lexical semantics . Distributional semantics . Big data . Machine learning
Bartlett (1928, 1932) conducted a classic experiment in cognitive psychology to examine how people remember. In the experiment, participants read an Indigenous American story entitled “The War of the Ghosts.” When prompted to recall the story, people inserted their own knowledge. For example, participants used the word boat (i.e., a word that they had experience with) in place of canoe (i.e., a word they likely did not know). Although Bartlett’s demonstration is remembered as a foundational example in the theory of reconstructive memory, it also makes another point. When people read or hear language, they comprehend that language through the lens of their own experience. For example, when asked to play a game of football, a person’s interpretation of that request might change depending on the side of the Atlantic where the person was raised. Similarly, if one is asked to play a game of “roque,” only people familiar with croquet variants from the late 1800s would understand the rules. But do the subtler differences in language experience exert a meaningful and distinguishable influence on people’s behavior and cognition? * Brendan T. Johns [email protected] 1
University at Buffalo, Buffalo, NY, USA
2
University of Manitoba, Winnipeg, Manitoba, Canada
By traditional methods of scholarship, a complete examination of language experience is intractable. However, recent advances in natural language processing, coupled with the availability of sizeable text corpora, have changed the game (e.g., Brysbaert, Mandera, & Keuleers, 2018; Chubala, Johns, Jamieson, & Mewhort, 2016; Green, Feinerer, & Burman, 2013, 2015; Johns, 2019; Johns, Jones, & Mewhort, 2019; Johns, Mewhort, & Jones, 2019; Johns & Jones, 2015; Jones, 2017; Jones, Dye, & Johns, 2017; Landauer & Dumais, 1997). For example, Johns and Jamieson (2018) applied theories of natural language processing to analyzing language use in published fiction. On the basis of the analysis,
Data Loading...
