Relatively speaking can language affect perception Debi Roberson
Relatively speaking, can language affect perception? Debi Roberson
Collaborators: James Pak Rick Hanley Jules Davidoff Laura Shapiro Ian Davies Grev Corbett Marietta Vandervyver University of Essex Goldsmiths College Warwick University of Surrey University of Namibia
What do we gain from labeling our categories? Just being able to talk about them, or some other advantages, e. g. speed, efficiency, cognitive economy, flexibility, the ability to carry out complex cognitive tasks such as reasoning by analogy?
Investigations in the colour domain?
Different languages name the range of visible colours differently English naming Berinmo naming Himba naming
Two questions: 1. What would be acceptable evidence of language affecting cognition? 2. How different must two sets of descriptive terms be before there are observable differences in behavioural response from speakers of those languages?
All these populations show Categorical Perception of colors “A sharp peak in the relative discriminability of stimuli that cross a category boundary compared to the discriminability of items within categories such that continuous quantitative differences along a continuum are perceived as discrete qualitative changes at category boundaries”. (Harnad, 1987)
Same vs. different names Same name “Green” Different names Same name “Blue”
Blue-green stimuli in odd-oneout similarity judgements
Nol-wor stimuli in odd-one-out similarity judgements
Dumbu-burou stimuli in oddone-out similarity judgements
Percentage of judgements following predictions (based on the existence of a category boundary) for three languages
Impact of language on recognition memory
• Small differences in language category boundaries can lead to different patterns of memory for the same set of colors. • Cognitive organisation of color categories mirrors linguistic categories. Berinmo naming Himba naming
Singing the Russian Blues Winawer et al. , (2007) blue goluboy siniy
Russian speakers show a category effect in discrimination that disappears with verbal suppression
Fig. 1. Lexical categories influence perception in the RVF Gilbert, Aubrey L. et al. (2006) Proc. Natl. Acad. Sci. USA 103, 489 -494 Copyright © 2006 by the National Academy of Sciences
Korean and English speakers visual search across the boundary between Korean categories ‘yellow-green’ and green. (Roberson, Pak & Hanley, 2007) +
Grouping behaviour of English and Korean participants Mean RTs for English and Korean speakers to identify targets from distractors either across or within the categories yeondu and chorok.
English speakers sort these stimuli into essentially one group. They show no significant effect of category, no significant effect of visual field and no significant interaction. Korean speakers sort them into two groups. They show a significant effect of category and fast responders show a significant effect of visual field
Using eye-tracking - data for 7 English and 7 Korean participants who did not move their eyes from fixation. English speakers show no sig. category effects. Korean speakers show sig. category effects in RVF (p<. 02) but not in LVF.
Participants who moved their eyes English spekers show no category effect. Korean speakers show a category effect for both visual fields (both p<. 05)
If we accept that language can influence thought to some extent: When and how does it influence: Online: - linguistic structures/processes meddle in cognitive processes as they are happening - effects should go away/be diminished with verbal interference - effects might be stronger when subject is tested in language of interest Long-term / Underlying: - patterns in language shape the underlying cognitive processes/representations - effects should not go away/be diminished with verbal interference / loss of language. - effects might be seen even when subject is tested in other languages
Evidence to date is mixed as to the online / underlying influence of verbal coding on color cognition: Most research shows that effects go away/are diminished with verbal interference (Roberson & Davidoff, 2000; Winawer et al. , 2007; Gilbert et al. , 2006) AND task modification indicates a strategic element to reliance on verbal coding (Pilling et al. , 2003). BUT research with aphasic patients suggests otherwise (Roberson, Davidoff & Braisby, 1999; Haslam et al. , 2007) Developmental studies of young children show an increasing effect of categorisation with increasing color term knowledge (Roberson et al. , 2004; Daoutis et al. , 2006; Davidoff et al. , in prep) BUT this doesn’t address the issue of how deep the influence goes. Tasks involving memory probably maximise the influence of verbal coding. BUT perceptual tasks like visual search show very similar results.
Thanks for your attention!
Communication studies Ordered test array Random test array Easy to name targets Hard to name targets
Memory studies Ordered array The same set of test stimuli appeared in either an ordered or a random test array target Random array 5000 ms ISI 150 ms 1000 ms
Results Focals are recognised better in an ordered, but NOT a random array.
Why does the context affect recognition? In an ordered array, the observer can use verbal coding to ‘home in’ immediately on relevant area - reducing the number of stimuli to be searched. The name advantage is attenuated by the number of same-name (e. g. red) stimuli that must be searched (Guest and Van Laar, 2002) In random arrays subjects may strategically rely more on visual memory for stimuli - wiping the ease-of-naming advantage for focal stimuli.
Varied random arrays: : target in left half of array target in right half of array Target ÜAdvantage for all targets in left half of screen. ÜNo advantage for Focals in either half. ÜFocal items not more perceptually salient - advantage arises from reliance on naming. ÜParticipants rely less on name when test arrays are random. ÜHeavy reliance on verbal code may activate category prototype - at the expense of accurate visual coding.
- Slides: 37