Particularly Exciting Experiments in Psychology
October 8, 2015
Knowledge of and Exposure to One's Native Language Interferes With Acquisition of a New Language
All languages have regularities in the distribution or placement of sounds and words. However, there are disparities in how particular speech sounds are mapped to create phonologically meaningful units. For instance, in Hindi dental /d/ and retroflex /ɖ/ are two distinct sounds, but these are perceived by English speakers as variants of the single /d/ sound. Thus, language acquisition depends on perceptual learning of speech sounds, and statistical learning of structural regularities.
Earle and Myers (2015, Journal of Experimental Psychology: Human Perception and Performance) (PDF, 356KB) trained English-speaking participants to identify the dental vs. retroflex /d/ contrast in Hindi in one vowel context (e.g., dig vs. ɖig). Following training, they completed discrimination tests at three different time points, where they had to judge if two sequentially presented words that only differed in /d/ and /ɖ/ were the same or different. The words in each trial were always different spoken exemplars, so participants had to rely on recognition of the speech sound category rather than other acoustic cues.
Participants trained in the morning showed an initial improvement in discrimination performance, however, their performance decreased over time and was lower in a final test session the next day. In contrast, participants trained in the evening showed continual improvement across all test sessions.
In a subsequent experiment, the authors tested whether performance decline in the morning group could be attributed to interference from exposure to native speech sounds throughout the day, prior to sleep-related consolidation. All participants were trained in the evening followed by an interference session, where half the participants heard English /d/ sounds, which are highly similar to the Hindi contrast being learned, and the other half heard English /b/ sounds.
While both groups showed comparable gains in performance in the discrimination test immediately following the interference session, only the /b/ interference group showed additional gains in a second test session the next morning; the group that heard /d/ during the interference session showed no further improvement. These results suggest that the benefit of sleep for perceptual learning in second-language acquisition is mediated by the amount of exposure to similar native-language sounds between training and sleep.
Earle and Myers found that native language exposure interfered with perceptual learning for a new language. Finn and Hudson Kam (2015, Journal of Experimental Psychology: Learning, Memory, and Cognition) (PDF, 117KB) tested whether violating native language norms would interfere with statistical learning for a new language.
English-speaking participants listened to speech from either an experimental or control novel language for two hours. Initial consonant clusters for base-words that started with consonants violated English norms (e.g., thmahrey) in the experimental language. Both novel languages contained two prefixes that varied depending on whether the base-word began with a vowel or consonant (e.g., juh/juhb and woy/woys).
Learning these relationships requires participants to correctly segment a word into the prefix and base-word. One prefix was always presented immediately before the base-word (e.g., woy-thmahrey). For the other prefix, an infix (e.g., neyd) separated the prefix and base-word (juh-neyd-thmahrey). Following exposure, participants were asked to judge which of two sequentially presented words sounded more like it belonged to the language they had heard.
All participants were above chance for base-words that started with a vowel. In contrast, while control participants were also better for both prefixes for base-words that started with a consonant, experimental participants were only above chance for the prefix that was followed by an infix.
That is, participants failed to acquire the prefix-base-word relationship for words beginning with consonant clusters that violated English norms when segmentation of the prefix and base-word was not made salient by an infix. This suggests that prior linguistic knowledge of phoneme norms can interfere with statistical learning in a new language that relies on proper segmentation.
Together, these studies demonstrate how one's native language can interfere with perceptual and statistical learning processes involved in acquiring a new language.
Other Interesting Reading
- Herbort and Kunde (online first, Journal of Experimental Psychology: Human Perception and Performance) find that the spatial rules used to describe the interpretation of pointing gestures differ from the rules used to describe the production of these gestures.
- Tauzin and colleagues (online first, Journal of Comparative Psychology) show that dogs perceive human pointing as referring to spatial cues (e.g., location or direction), not target objects.
Particularly Exciting Experiments in Psychology™ (PeePs) is a free summary of ongoing research trends common to six APA journals that focus on experimental psychology.