The Structure of Experience:

1
The Structure of Experience

• Patterns of Intra and
• Inter-Cultural Variation
• James S. Boster
• University of Connecticut

2
Overview

• Outline the Information Economy Model
• Describe theme of the talk
• The disconnectionist architecture of
• distributed cognition
• Discuss examples
• Conclude

3
Main Question

• How does culture happen?
• How does a collective understanding of the world
  emerge out of individual cognition?

4
The Information Economy Model

• Just as economists study the production,
  distribution, and consumption of goods, cognitive
  anthropologists study the construction,
  distribution, and use of knowledge.

5
A definition of culture

• Cognitive anthropologists view culture as an
  information pool that emerges when members of a
  community attempt to make sense of the world and
  each other. Because individuals construct their
  conceptions of the world from their own
  experiences, their understandings vary from one
  another depending on the characteristics of the
  individuals, the nature of the domain learned,
  and the social situations in which learning takes
  place. (Boster, 1985)

6
Characteristics of culture

• Culture is learned knowledge.
• Individuals vary in their knowledge.
• An ethnography should describe and explain the
  social distribution of knowledge.

7
Problem

• This definition of culture is too Apollonian,
  as though all people want to do is reach an
  objective understanding of the world.
• We need a definition that is more Dionysian one
  that puts human passions and desires as part of
  what moves people to learn.

8
A better definition of culture

• Cognitive anthropologists view culture as an
  information pool that emerges when members of a
  community attempt to make sense of the world and
  each other as they struggle and collaborate with
  each other to get what they want (e.g., food,
  sex, power, and acceptance). Because individuals
  construct their conceptions of the world from
  their own experiences and for their own
  motivations, their understandings vary from one
  another depending on the characteristics of the
  individuals, the nature of the domain learned,
  and the social situations in which learning takes
  place.

9
Todays Theme

• Distributed cognition does not often have a
  connectionist architecture

10
Reigning metaphor of distributed cognition

• Connectionist model of distributed cognition
  (Hutchins)
• Distributed cognition as a neural net.
• Distributed cognition is not located in
  individual brains but distributed across networks
  of humans and artifacts.
• Can track representations as they are propagated
  and transformed through these networks.

11
Connectionist Network Diagram
12
Query

• Why should distributed cognition have the same
  connectionist architecture as individual
  cognition?

13
Alternate metaphor of distributed cognition

• Disconnectionist model of distributed cognition
  (Boster)
• Appearance of collective consciousness can emerge
  in the absence of network connections among the
  parts of the cognitive system.
• Distributed cognition as bird flock collective
  consciousness.

14
Bird Flock Taking Off (Still)
15

• Appearance of shared cognition is the result of
• Public information
• birds in flock are simultaneously visible to each
  other
• Common goals
• birds want to be in center of flock (to avoid
  predators and to cut wind resistance)
• birds want to avoid collision
• Individual variation
• birds vary in strength and flying ability

16
Centripetal and centrifugal forces in
distribution of knowledge
17
Examples

• Folk biology
• Similarity judgment
• Typicality judgment
• Clustering in free-recall
• Mating preferences
• Language acquisition
• Social Evaluation
• Social networks
• Stereotypes of national character

18
Cases of Universal Agreement

• Expect strong agreement within and between groups
  when information is freely available through
  direct observation, individuals have similar
  experiences and share motivation.
• In these cases, there is high agreement within
  and between groups but there is more
  within-cultural variation than cross-cultural
  variation.
• E.g., folk biological classification

19
Similarity Judgment Birds

• Diverse groups of humans (U.S. under-graduates,
  native South Americans, and scientific
  ornithologists) agree in their recognition of
  patterns of resemblance among a collection of
  South American bird specimens and agree also on
  the attributes of the birds to use in their
  similarity judgments (Boster, Berlin, and
  O'Neill, 1986 Boster, 1987 Boster and
  D'Andrade, 1989) .

20
Cases of Within-Cultural Agreement and
Cross-Cultural Differentiation

• Expect weaker agreement and more differentiation
  when part of the information is public, but
  individuals depend on each other for
  interpretation of that public information.
• In these cases, there is may be high agreement
  within groups but greater cross-cultural
  variation than within-cultural variation.
• E.g., child acquisition of language

21
Language Acquisition Spatial Relationships(with
M. Bowerman and S. Choi)

• Cross-cultural disagreement in naming of spatial
  relationships in English, Dutch, and Korean
• Language acquisition is a case of a developmental
  process that unfolds in a universally similar
  fashion but with culturally specific results
  Korean babies learn Korean while Dutch babies
  learn Dutch.

22
Languages vary in classification of spatial
relations

• English the picture is on the wall, the mug is
  on the table, the ring is on my finger.
• Dutch the picture is aan the wall, the mug is op
  the table, the ring is om my finger.
• Data identification of spatial relationship when
  modeled (tops on containers, hats on dolls, legos
  apart, etc.)

23
Results

• Some universal features attended to by all
  languages (e.g., separation vs. joining).
• Strong linguistic differentiation from earliest
  ages.
• Key to figures
• D Dutch, E English, K Korean
• Relative size of letter represents relative age

24
2.0-2.5
25
2.5-3.0
26
3.0-3.5
27
adults
28
aggregates
29
all
30
Cases of (Expected) Within-Cultural
Differentiation

• Expect greatest differentiation when knowledge is
  socially transmitted.
• Distributed cognition in this case should be the
  most connectionist.
• E.g., social networks, stereotypes of national
  character

31
Social Networks

• The social distribution of social knowledge
  should reflect structure of the social group.
• Individuals who are close to each other in the
  social network should agree more than more
  distant individuals.

32
Picture of south pole station
33
Winter-over crews in Antarctica (with J. Johnson
and L. Palinkas).

• Setting Amundsen-Scott South Pole Station
• The Bad Winter (18 men, 4 women)
• The Good Winter (19 men, 9 women)
• Data Social proximity of group members

34
Results

• Different groups have different structures.
• Bad winter fragmented into cliques
• Good winter organized into a core-periphery
  structure

35
Bad winter
36
Good winter
37
Correlation of Agreement with Group Structure
is neither strong nor robust.
38

• The relationship of the pattern of agreement to
  group structure is neither strong nor robust
• In fragmented social group, distribution of
  social knowledge has a slight but stable
  relationship with group structure.
• In cohesive (core-periphery) social group,
  relationship of distribution of social knowledge
  to group structure disappears when one corrects
  for overall agreement.

39
Conclusion

• Even a case in which we have a strong
  expectation of connectionist distributed
  cognition is mainly disconnectionist the pattern
  of private social ties is apparently a readily
  observable public fact.

40

• In heaven,
• all the policeman are English,
• all the car mechanics are German,
• all the cooks are French,
• all the hotel keepers are Swiss, and
• all the lovers are Italian.
• In hell,
• all the policeman are German,
• all the car mechanics are French,
• all the cooks are English,
• all the hotel keepers are Italian, and
• all the lovers are Swiss.

41
Stereotypes of National Character

• Method
• Send college students in California home to find
  out from their mothers what it is to be whatever
  ethnic identity they take for themselves (e.g.,
  Korean, Vietnamese, Chinese, Filipino, Latino,
  Anglo, etc.).
• Remove self-identity from propositions and
  replace with a blank.
• Remove 'Other people say that we are ...' to get
  negative propositions.
• Retain 30 belief frames.

42

• Method (continued)
• Translate from English into a number of the
  languages of Europe (Spanish, Italian, French,
  German, Dutch, Norwegian, Swedish, Finnish,
  Polish, Czech, Hungarian).
• Ask people of different nations to fill in the
  blank with as many nations as they think fit.
• 16 locations (Madrid, Barcelona, Vigo, Genoa,
  Naples, Paris, Cologne, Amsterdam, Stockholm,
  Oslo, Helsinki, Krakow, Prague, Budapest, London,
  and Storrs).
• Public settings (cafes, beer gardens, bars,
  railway stations, trains).
• 74 per location, 1184 total.

43
Results

• High agreement on national characteristics (mean
  r .67).

44
Attributes
45
Nations
46
Confidence intervals
47
Results

• Similarity of perceptions reflects location of
  perceived (r .42, p lt .0001).
• Similarity of perceptions reflects location of
  perceiver (r .17, p lt .05).

48
Factors 1 and 2
49
Factors 3 and 4
50
Results

• Disagreement makes sense
• most peoples think better of themselves than
  others do.
• neighbors differentiate more than peoples far
  away.
• invading your neighbors does not win their
  esteem.

51
What happens when you invade your neighbors
52
Conclusion

• Although stereotypes of national character show
  some local and regional chavinism, there is much
  of the pattern that is broadly agreed upon
  another case of distributed cognition that is
  more disconnectionist than connectionist.

53
Overall Conclusions

• Culture is often learned from public structure.
• Usually shared understandings arise from common
  experience of publicly available information.
• The architecture of distributed cognition is
  sometimes connectionist but not often.
• Disconnectionist distributed cognition is more
  typical.

54
Sunset
55
(No Transcript)
56
Caveat

• Agreement in folk biological similarity judgment
  is not inevitable.
• Differentiation can emerge when members of
  sub-groups differ in their interests in the
  domain and when there are alternate bases for
  judgment, some directly observable, others not.

57
Similarity Judgment Fish

• Cross-subcultural disagreement in fish
  similarities judgment (with J. Johnson).
• Data North Carolina undergraduates and expert
  fishermen judgments of the similarities among 43
  line drawings of fish.
• Results Experts sort on the basis of both
  morphological and behavioral/functional criteria
  while novices sort on the basis of morphological
  criteria alone.

58
CCA of fish
59
Novice MDS
60
Expert MDS
61
Why phenetics or traditional classification may
be better than a cladistic classification.