Robert Kurzban University of Pennsylvania Perceptions of Race The Second CEFOM/21 International Symposium Culture, Norms, & Evolution Hokkaido University,

Robert Kurzban

University of Pennsylvania

Perceptions of Race

The Second CEFOM/21 International Symposium

Culture, Norms, & Evolution

Hokkaido University, Sapporo, Japan

August 8th

Claims

• “Automaticity” in categorization is an unlikely design feature (for “race”)

• Race is a proxy for the fundamental conceptual cognitive element, “coalition.”

• Providing an alternative coalitional dimension should be potent for changing categorization

Background: Social Categorization Literature

• Social Psychology: 3 social categories are automatically encoded: Age, Sex, & Race

• Efforts to attenuate categorization by race have been unsuccessful

Method: Memory Confusion

• Participants hear conversation & see photos

• In a recall task, participants are asked to recall who said what.

• Errors in recall index categorization

• Within-Between group errors calculated for each dimension (and statistically corrected)

• “Does anyone besides Rob have a question?”

• “Which professor from UBC was It?…”

Taylor et al., 1978

• Error rates at ~66%

• More within than between race and sex errors observed

• Categorization by race not changed by instructions to try to recall information.

There are certain categories that are highly accessible and difficult to suppress, in particular race and sex. Assuming that these categories are extremely salient and powerful, and are automatically encoded in the absence of any specific memory instructions, they may be rather insensitive to … contextual variations…  We believe that one of the contributions made by our research is to show how hard it is to interfere with strong, automatically activated categories…

Hewstone et al., 1991, p. 526See also:

Hamilton, Stroessner, & Driscoll, 1994

Fazio & Dunton, 1997

Wegner & Bargh, 1998 (Handbook chapter)

Additional Cites

• “…our theoretical framework and the latency data lead us to believe that the process began with attention being automatically drawn to the race of the Black targets.” Fazio & Dunton, 1997

• “Easily discriminable personal features—especially the “big three” of gender, race, and age—tend to activate preconsciously the categories or stereotypes associated with them.”

Wegner & Bargh, 1998 (Handbook chapter)

Theoretical Background: Race

• Adaptations are shaped by recurrent features of ancestral environment

• Sharp phenotypic gradients were not a recurrent feature of ancestral environments.

• It is unlikely that there are adaptations designed to encode race per se.

• Racial categorization is a byproduct. But of what?

What is racial categorization a by-product of? Color perception?• Stangor et al. expt 5: Colored sweatshirts

had no meaning; not used as a category by participants. Note. Attentional demands can’t explain this:

sex and race can simultaneously be strongly encoded.

• Perceptual similarity should affect encoding. It doesn’t (Stangor et al., expt 2)

What is racial categorization a by-product of? Reasoning about

natural kinds?• Various accounts posit humans essentialize

human groups.Rothbart & Taylor, Hirschfeld, Gil-White

• If so, this would explain the consistently strong results for encoding of race. But, these models do not (straightforwardly) predict that it should be possible to attenuate this process…

What is racial categorization a by-product of? Mechanisms designed to parse coalitions and

alliances?• “Races” were not a part of human ancestral

environments. Dynamic alliances and coalitions were.

• Categorization by race might be attenuated if race is not diagnostic of alliances, but another visual cue is.

• Additional predictions derived from this analysis:While sex could have evolved domain-specific

machinery, race could not.Categorization by sex should be stronger than by race,

everything else equal.

Experiment Details

• Racially mixed (4 White, 4 Black) basketball teams (all male)

• Competitive content in dialogue

• IV: Jersey color cue to team membership

• Prediction: Attenuate racial encoding

And you play like you’re in a zoo. Where you should be anyway

Sample Stimulus

Categorization by TEAM

0

0.5

1

1.5

2

2.5

Males (n=28) Females (n=27)

Participant Sex

Wit

hin-

Bet

wee

n E

rror

s

p < .05

Categorization by RACE

0

0.5

1

1.5

2

2.5

3

3.5

Males (n=28) Females (n=27)

Participant Sex

Wit

hin

-Bet

wee

n E

rror

s

Categorization by TEAM

0

1

2

3

4

5

6

7

Males (n=27) Females (n=25)

Participant Sex

Wit

hin-

Bet

wee

n E

rror

s

n.s.

Categorization by RACE

0

0.5

1

1.5

2

2.5

3

Males (n=27) Females (n=25)

Participant Sex

Wit

hin

-Bet

wee

n E

rror

s

n.s.

TEAM Accentuation

0

1

2

3

4

5

Males Females

Participant Sex

Wit

hin

-Bet

wee

n E

rror

s

RACE Attenuation (all s's)

0

0.5

1

1.5

2

2.5

Verbal Verbal & Color

Cue to Coalition Membership

With

in-B

etw

een

Err

ors

p < .05

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

verbal only + shared appearance

Cues to Coalition Membership

Eff

ect S

ize

race

coalition

Results

• Adding shirt color cue increases categorization by team (not surprising)

• Adding shirt color cue decreases categorization by race (surprising)

Experiment 2

• Same as above, except teams are all White, but of mixed SEX.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

verbal only + sharedappearance

verbal only + sharedappearance

Cues to Coalition Membership

Eff

ect S

ize

race

sex

team

Conclusion

• “Categorization” by “race” is not “automatic,” and can be attenuated, even after only relatively short exposure.

But does the attenuation of race replicate…?

Replication (N = 103)

0

0.5

1

1.5

2

2.5

3

Team Race

Wit

hin

-Bet

wee

n E

rror

s

p < .01

n.s. n.s.

Note: No sex differences

Research Program

Categorization

Cooperation

• More direct method: Reaction times to gauge categorization

• Additional control: cognitive load condition

• Additional contexts.