Bilinbgual Lexical Entrainment
What is lexical entrainment?
Lexical entrainment is a dialogue phenomenon that speakers adopt their interlocutors’ word choice [1]. As we see from the picture below, one speaker says “computer” and the other speaker says “laptop” while they both refer to the same object. They start with different terms, but the second speaker ends up using “laptop” as well. In this process, the second speaker is entrained to the first speaker.
Two types of entrainment
There are two mainstream explanations for the mechanism of entrainment, that lexical entrainment is an automatic or a meidated process.
- Automatic entrainment: Higher activation level leads to higher probability of adopting the word [2].
- Mediated entrainment: Features of the linguistic and social context affects speakers’ entrainment [4][5][6].
Does language proficiency mediate lexical entrainment?
We hypothesized that it does, and bilinguals’ proficiency in English is positively related with their entrainment rates. This is because less proficient bilinguals tend to have lower activation levels on the lexical items, especially for less frequent words.
Method
Language Experience and Proficiency Questionnaire (LEAP-Q)： this questionnaire collects participants’ language history, self-reported language proficiency, language exposure, accent, and other relevant factors in their first and second languages.
Picture Matching and Naming： this task measures entrainment behavior.
Participants were told that they were matched with someone picked from a dataset who had done this task before, and they would see object names provided by their paired partners in matching trials. In each experimental matching trial, participants would see two pictures with an object name given by their “partners”. They needed to choose one picture that matched the word.
In the experimental naming trial, they would see two objects, and they were instructed to type the name of the stimuli objects they had seen in the experimental matching trial. The target object in the experimental naming trial is the same one in the experimental matching trial. If the participant typed in the same word at the naming trial as the matching trial, this trial is marked as entrained.
There were in total 52 matching trials and 52 naming trials. The sequence of the trials was fixed, with an experimental matching trial first, two filler trials in between, and an experimental naming trial at last. This cycle repeated 26 times before the task ended.
Findings
The analysis supports the hypothesis that English proficiency mediates entrainment. However, the trend shown by the data is the opposite of our prediction. While we predicted that language proficiency is positively correlated with entrainment rate, the data suggested that the more proficient speakers tend to entrain less in the picture matching and naming task.
Entrainment in Mandarin
To further investigate this trend which is opposite to our original hypothesis, we conducted a replication analysis in Mandarin-English bilinguals. We hypothesized that for Mandarin-English bilingual participants, they would show similar entrainment behavior patterns in Mandarin: the more proficienct they are in Mandarin, the less they would entrain. The methods were identical to the English versions, but the materials were translated into Mandarin.The data showed no significant correlations between Mandarin proficiency and entrainment rate in Mandarin.
By running a t-test, we found that the entrainment rate in Mandarin is significantly higher than the entrainment rate in English condition, which supports the pattern we found earlier.
See the full report in our Cognitive Science Society conference proceedings!