Content uploaded by Scott Atran
Author content
All content in this area was uploaded by Scott Atran on Mar 17, 2014
Content may be subject to copyright.
Cultural mosaics and mental models of nature
Megan Bang*, Douglas L. Medin
†‡
, and Scott Atran
§¶㥋
*TERC, Cambridge, MA 02140;
†
Department of Psychology, Northwestern University, Evanston, IL 60208;
§
Centre National
de la Recherche Scientifique-CREA, 1 Rue Descartes, 75005 Paris, France;
¶
Institute for Social Research, University of
Michigan, Ann Arbor, MI 48106-1248; and
㛳
John Jay College of Criminal Justice, New York, NY 10029
This contribution is part of the special series of Inaugural Articles by members of the National Academy of Sciences elected on May 3, 2005.
Contributed by Douglas L. Medin, July 13, 2007 (sent for review May 21, 2007)
For much of their history, the relationship between anthropology
and psychology has been well captured by Robert Frost’s poem,
‘‘Mending Wall,’’ which ends with the ironic line, ‘‘good fences
make good neighbors.’’ The congenial fence was that anthropol-
ogy studied what people think and psychology studied how
people think. Recent research, however, shows that content and
process cannot be neatly segregated, because cultural differences
in what people think affect how people think. To achieve a deeper
understanding of the relation between process and content, re-
search must integrate the methodological insights from both
anthropology and psychology. We review previous research and
describe new studies in the domain of folk biology which examine
the cognitive consequences of different conceptualizations of
nature and the place of humans within it. The focus is on cultural
differences in framework theories (epistemological orientations)
among Native Americans (Menominee) and European American
children and adults living in close proximity in rural Wisconsin. Our
results show that epistemological orientations affect memory
organization, ecological reasoning, and the perceived role of
humans in nature. This research also demonstrates that cultural
differences in framework theories have implications for under-
standing intergroup conflict over natural resources and are rele-
vant to efforts to improve science learning, especially among
Native American children.
folkbiology 兩 mental models 兩 Native American 兩 science education
T
his review examines cognitive and behavioral consequences
of cultural differences in conceptions of nature and concep-
tions of the role of human beings in it. The evidence suggests that
these consequences are considerable, and as we piece them
together, they form something of a mosaic. We concentrate on
a set of comparative studies of a particular Native American and
European American culture population, but the results are
consistent with our previous findings and further suggest wider
implications for culture and learning, in general, and science
education, in particular. Our work falls at the interface between
cognitive psychology and anthropology. A few decades ago,
these two fields enjoyed a congenial division of labor, under
which cognitive psychology’s mission was to determine how
people think and anthropology’s focus was on what people think
(1). But recent empirical research by cultural psychologists,
cognitive anthropologists, and even philosophers (2–6) provides
convincing evidence that cultural processes affect cognitive
processes.
These new observations have led to corresponding theoretical
analyses. One approach to understanding cultural differences in
cognitive processing (7) appeals to a ‘‘cognitive toolkit’’ and
argues that cultures are associated with differences in the use and
accessibility of particular strategies drawn from a common set of
tools. Another approach (8) suggests that culture affects the
chronic accessibility of constructs for approaching the social
world (e.g., individualism versus collectivism) and that these
constructs also affect perceptual and cognitive processes asso-
ciated with the nonsocial world. Our orientation also draws on
the idea that abstract framework theories (or epistemological
orientations) constrain cognition but assumes that cultural
framework theories may be domain-specific (9).
Framework Theories, Cultural Models, and Meaning
Cultural framework theories provide individuals with skeletal
principles for meaning making, including beliefs about what sorts
of things are relevant, worthy of attention and in need of
explanation. Religion may be a good case in point (10). Here,
specific beliefs about the creation of the world, its ultimate cause
and function provide a lens that people may use to derive
meanings for particular contexts and behaviors. Religion may
provide a causal framework for important events surrounding
humans, such as illness, misfortune, and death.
Our work focuses on the domain of biology and the natural world.
We will describe our research with the two cultural groups living in
close proximity in north central Wisconsin, rural Native Americans
from the Menominee Nation and rural European Americans living
in an adjacent county in northcentral Wisconsin.**
Menominee. The Menominee are the oldest continuous residents
of Wisconsin. Historically, their lands covered much of Wiscon-
sin but were reduced, treaty by treaty, until the present 95,000
hectares was reached in 1856. There are 4,000–5,000 Menomi-
nee living on tribal lands in three small communities. Despite
economic incentives to the contrary, the Menominee have
preserved diversity and habitat types of their forest (13, 14),
which is managed by a tribal corporation. They enjoy an inter-
national reputation for their sustainable forestry practices (15).
European American. Just south of the reservation is Shawano
County (population ⬇40,000), the other focal area for our
research. The major sources of income in the town of Shawano
are light manufacturing, small-scale farming, and tourist recre-
ation, mainly in the form of hunting, fishing, boating, jet-skiing,
and snowmobiling. Shawano Lake is a major attraction, and
there are also several rivers and smaller lakes in the county.
Author contributions: M.B., D.L.M., and S.A. designed research; M.B. and D.L.M. performed
research; M.B., D.L.M., and S.A. analyzed data; and M.B. and D.L.M. wrote the paper.
The authors declare no conflict of interest.
‡
To whom correspondence should be addressed. E-mail: medin@northwestern.edu.
**An important aspect of this work is the evolution of the methodological and theoretical
frames our studies employ. Much of the work in cognitive science and subsequent
understandings of human cognition has been based on studies largely comprised of
white middle-upper class urban or suburban children and adults. The necessity to
diversify the representation of participants in empirical work in cognitive science cannot
be overstated; however, this raises at least two important issues that are not unrelated
to this work.
For many communities of color, especially American Indians, research often has been an
exploitive and an unwelcome enterprise. Indigenous scholars and communities have
been developing appropriate guidelines for conducting research with indigenous com-
munities and tribal institutional review board processes to ensure that previous research-
based assaults on tribal communities are not continued; these efforts are fundamentally
acts of sovereignty (11, 12). Second, research agendas need to meet needs of the
communities that are involved in the work. This means that much of the research
conducted becomes problem-driven. We believe that problem-driven research can pro-
vide rich contexts in which to do both applied and theoretical work.
© 2007 by The National Academy of Sciences of the USA
13868–13874
兩
PNAS
兩
August 28, 2007
兩
vol. 104
兩
no. 35 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0706627104
Outdoor recreation is important to many of the county residents,
and many adults have fished since they were young children.
Culture, Mental Models, and Children’s Biology
For at least several decades, there has been an enduring interest
in the development of children’s understanding of biology.
Perhaps the most influential work on this topic was pioneered by
Carey (16), who argued forcefully that development within the
domain of biological knowledge entails radical conceptual
change in children’s core concepts and theories. This conclusion
has been based largely on children’s and adults’ performance in
a category-based induction task (16). In this task, children
(ranging from 4–10 years of age) are introduced to a picture of
a base-item (a human, dog, or bee), and taught a new property
about it. For example, the interviewer might show a picture of
a dog, saying, ‘‘Have you ever heard of an omentum? Omenta are
very thin, yellowish things. Here is a picture of something that
has an omentum inside it.’’ Next, children are tested for their
tendency to generalize (or to ‘‘project’’) that property to other
objects (including other humans, dogs, aardvarks, dodos,
stinkoos, bees, worms, flowers, sun, clouds, harvesters, and garlic
presses).
Carey found that for young (4- to 5-year-old) children: (i)
projections from humans as a base were stronger than projec-
tions from either dog or bee; (ii) there were strong asymmetries
in projections to and from humans (e.g., inferences from ‘‘Hu-
man’’ to ‘‘Dog’’ were stronger than from Dog to Human); and
(iii) in some cases violations of similarity occurred (e.g., infer-
ences from Human to ‘‘Bug’’ were stronger than from ‘‘Bee’’ to
Bug). This pattern has been interpreted as evidence that young
children’s conceptions of the natural world are anthropocentric.
Research on children’s biology has been conducted almost
exclusively with individuals from North American, urban, tech-
nologically dependent populations. Thus, it is not clear which
aspects of children’s naı¨ve biology depend on cultural concep-
tions or conditions of learning or some combination of both. To
evaluate the role of cultural milieu and conditions of learning in
children’s inductive reasoning we initially studied three popu-
lations: urban (mostly European American) Boston children,
rural Wisconsin European American culture children, and rural
Menominee children of varying ages (17). The participants were
given Human, ‘‘Wolf,’’ Bee, and ‘‘Goldenrod’’ as bases and
several mammals, birds, reptiles, invertebrates, plants, natural
inanimates, and artifacts as targets. As in Carey’s studies,
properties were unfamiliar internal substances of the form ‘‘has
X inside.’’ The detailed procedures differed in a number of ways
from those used by Carey, so direct comparisons with her results
must be qualified with caution.
The most relevant data derive from the rural populations.
Both cultural groups showed the mature pattern of generalizing
in terms of biological affinity even at the youngest ages (4–6
years). This suggests that the anthropocentric pattern of gener-
alization noted by Carey may depend on a relative lack of
intimate experience with animals and plants (18). Rural children
generalized more from wolf to other mammals than from
humans to other mammals, a pattern consistent with humans
being seen as atypical animals.
But there were also two striking differences among the two
rural populations. Rural European American children of all ages
showed asymmetries in reasoning between humans and animals
and often justified a failure to extend a property from an animal
to humans on the grounds that ‘‘people are not animals.’’ In
contrast, Menominee children of all ages showed no reliable
human–animal asymmetries and were much less likely to say that
people are not animals. The Menominee creation story has
people coming from the bear, and even the youngest children are
familiar with the animal-based clan system. In short, there is an
explicit cultural support for a symmetrical relation between
humans and other animals.
The second main result concerns a strategy for reasoning that
we did not observe with urban children: reasoning in terms of
ecological relations. For example, children might justify gener-
alizing from bees to bears because a bee might sting a bear or a
bear might acquire the property by eating the bee’s honey.
Ecological reasoning was a common strategy among even the
youngest Menominee children but was only common for the
oldest rural majority culture children. This community differ-
ence was evident in both patterns of generalization (e.g., rela-
tively greater generalization from bees to bear and to plants) and
in justifications that children reported for their projections (18).
In summary, both culture and experience affect children’s
anthropocentrism and propensity for ecological reasoning.
Mental Models of Nature Among Fish Experts
To further explore the extent to which these findings permeated
these two communities, we studied models of fish and aquatic
ecology with local fish experts in the two communities (19, 20).
Standard sorting techniques and other probes were used to
explore each group’s categorization of local fish species and
perceived ecological relationships (fish–fish interactions). To
foreshadow our results, the data show that both groups of experts
shared a knowledge base concerning local freshwater fish and
their habitats, although that knowledge was organized differ-
ently across groups.
The participants were nonprofessional experts in freshwater
fish and fishing. We used peer nominations (snowball methods)
and then later confirmed expertise by probing for familiarity
with 46 species of local fish (20). Members of both groups
reported engaging in similar fishing activities, including fishing
both rivers and lakes in all seasons and using live bait, flies (that
they frequently tie themselves), and artificial lures to do so.
Participants of the two groups did not differ with respect to age
(mean, 46 years), fishing experience (mean, 41 years), or formal
education.
Knowledge Organization.
Measuring agreement.
To assess within and
across group agreement, we applied the cultural consensus
model, as developed by Romney et al. (21–23). The cultural
consensus model is a factor-analytic method for computing
agreement and disagreement in the structure and distribution of
information within and across populations. The model assumes
widely shared information is reflected in a high concordance, or
‘‘cultural consensus,’’ among individuals. Principal-components
analysis is used to determine whether a single underlying model
holds for all informants from a given population: a strong
consensus exists if (i) the ratio of the latent root of the first to
the second factor is high, (ii) the first eigenvalue accounts for a
large portion of variance, and (iii) all individual first factor scores
are positive and relatively high. In the case of a consensus, the
structure of the agreement can be explained by a single factor
solution, the ‘‘consensual model.’’ In this case, first factor scores
represent the agreement of an individual with the consensual
model. Less formally, consensus will emerge when there is a
modest to strong positive correlation of each informant’s re-
sponses with those of every other informant.
The cultural consensus model is also useful for analyzing within
and across group differences. These differences can be explored by
(i) comparing first and second factor scores of each individual and
(iii) analyzing patterns of residual agreement. Residual agreement
is calculated by subtracting predicted agreement from the observed
agreement (24–26). To the extent that within-group residual agree-
ment is larger than cross-group residual agreement, one has evi-
dence of reliable group differences.
Category organization.
Standard sorting techniques were used to
explore each group’s categorization of local fish species. On a
Bang et al. PNAS
兩
August 28, 2007
兩
vol. 104
兩
no. 35
兩
13869
PSYCHOLOGY INAUGURAL ARTICLE
spontaneous sorting task involving 44 local species of fish,
Menominee and European American fish experts showed overall
consensus but also reliable group differences. Further analyses
indicate that the Menominee consensus contains an ecological
component absent in the sorting of European American experts.
Multidimensional scaling yielded a dimension for Menominee
experts that correlates with fish habitat. In addition Menomi-
nee experts were reliably more likely than European American
experts to mention habitat in their explanations for the sorts they
created.
Do these differences will generalize beyond expert fishermen?
In related line work, we asked our experts to nominate other
people who fished a great deal but who were less expert (19) and
then interviewed the nominees and gave them the same sorting
task. Less-expert Menominee fishermen also tended to sort
ecologically and to provide ecological justifications. Less-expert
European American fishermen were more likely to sort in terms
of goals and less likely to sort taxonomically than more expert
fishermen.
The community-based difference in ecological orientation
found in these studies likely reflects preferences for organizing
categories rather than knowledge differences per se. In another
study, we specifically asked Menominee and European Ameri-
can fish experts to sort fish by habitat (20). In this case, we
observed a strong overall consensus and no group differences.
Ecological relations.
A subset of 21 fish species were selected to
probe directly for ecological relations. We presented all possible
pairs (210) to European American and Menominee fish experts
with a question. ‘‘Does fish A affect fish B, and/or does B affect
A?’’ If the answer was yes, the expert was asked to elaborate on
the relation. Again, we found a cross-group consensus and
reliable between group differences (20). For relations reported
by 70% or more of participants from either group we found: 85%
are reported by both groups, 14% by Menominee but not
European American experts, and 1% by European Americans
fishermen but not Menominee fishermen. Content analysis
revealed Menominee experts answer in terms of the entire life
cycle of fish (e.g., spawn, fry, adults); European American
experts generally answered in terms of adult fish. These results
suggest that European American experts organize their knowl-
edge around goals that target adult fish.
Again, we hypothesized that these were not differences in
knowledge but rather knowledge organization. In a follow-up
study, we again asked about fish-fish interactions, but reduced
the number of pairs from 210 to 34 and ran the task at a slower
pace (30 sec per pair rather than ⬇10 sec per pair). If European
American experts have the same knowledge but not necessarily
organized around ecological relations, then the group differ-
ences should disappear. They did. Using the relations reported
for these 34 pairs on the longer speeded task as a base, we found
that European American experts now reported reliably more
relations, including more relations involving immature fish.
Menominee experts showed no reliable changes across tasks; the
64% advantage found for Menominee fishermen on the longer
task was reduced to a nonsignificant 11%. This again suggests
that the community differences are in ‘‘habits of mind’’ or
knowledge organization, rather than knowledge per se.
Summary. Both Menominee fishermen and young Menominee
children show a greater ecological orientation than their major-
ity culture counterparts. These cultural differences among chil-
dren and fishermen immediately raise a series of questions about
the origin and distribution of an ecological orientation within the
Menominee community. Although knowledge of specific eco-
logical associations requires relevant observations and experi-
ence, might there be widespread orientations toward nature that
facilitate this way of organizing knowledge? How does the
community difference in the salience of ecological information
emerge across development? What role, if any, do values and
routine activities or practices play in shaping and organizing
knowledge and orientations? To address these questions we turn
to further studies with children and their parents.
Culture and Orientations Toward Nature
Values and Learning Goals.
Study 1.
The community differences in
knowledge organization found among Menominee and Euro-
pean American fish experts were paralleled by corresponding
community differences in approaches to nature among school
children and parents. To study these orientations we relied on
prior work by Kellert and associates (27) who have developed a
typology of values associated with the natural world (e.g.,
utilitarian, ecologistic-scientific, aesthetic, moralistic, dominion-
istic). We modified this coding scheme in several ways to adapt
it to our setting (see Table 1).
The relevant data come from interviews with rural Menomi-
nee and European American adults. In most cases the adults
were parents of children who participated in the studies de-
scribed earlier. The coding scheme was applied to two questions:
1. What are the five most important things for your child
(grandchild) to learn about the biological world? 2. What are
four things you would like your child (grandchild) to learn about
nature?
Table 1.Modified coding scheme based on Kellert (27)
Orientation Definition
Holistic A belief that everything in nature is interconnected; that there is harmony and balance in nature.
Spiritual A tendency to access or find spiritual meanings, lessons, importance in nature. Referring to Mother Earth, Creator
Traditional Referring to an activity in nature as something that has occurred throughout history and should continue for that reason.
Survival skills A belief that nature is something that can harm us if underestimated.
Moralistic Strong feelings of moral and ethical responsibility.
Ecologistic Precise Study and systematic inquiry of the natural world and belief that nature can be understood from empirical study
from a systems perspective. There is a tendency to relate species to other aspects of nature.
Abstract respect Respect for nature
Abstract liking Expressing a positive but abstract attitude towards nature (e.g., ‘‘I enjoy nature’’ vs. the more concrete, ‘‘I enjoy walks in
the woods’’)
Abstract
knowledge
Expressing learning goals abstractly (e.g., ‘‘I want to learn about nature’’ vs. the more concrete, ‘‘I want to understand how
beavers build dams’’)
Personal utility The physical benefits derived from nature as a fundamental basis for human sustenance, protection, and security. The
benefits are intended for the self or those in the immediate family or social network
Distant utility The physical benefits derived from nature as a fundamental basis for human sustenance, protection, and security. The
benefits are intended for those distant from the immediate family or social network
13870
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0706627104 Bang et al.
Results.
The dependent variable was the proportion of adults from
each community whose responses were consistent with a given
code on at least one occasion. We focus only on the differences
that are statistically reliable (see Table 2). The most salient result
was that Menominee adult responses were more likely to reflect,
spiritual, holistic, and traditional values. For example, they were
more likely to talk in terms of ‘‘Mother Earth,’’ say that they want
their children to understand that they are a part of nature, and
describe activities that are important to pass on to future
generations. Looking across these three codes, Nine of the 12
Menominee adults expressed one or more of these attitudes
compared with 2 of 12 European American adults (
2
⫽ 8.2, P ⬍
0.01). European American adults were more likely to mention
abstract liking than Menominee adults (9 of 12 vs. 3 of 12,
2
⫽
6.0, P ⬍ 0.01).
Almost all parents (and elders) expressed moral values and
beliefs about the need to respect nature. They differed, however,
in the perspective applied to these values. European American
adults mentioned the need to protect nature, implying a care-
taker relation with the natural world. Menominee adults, in
contrast, were more likely to stress that people are ‘‘a part of’’
nature. This distinction between being ‘‘apart from’’ versus ‘‘a
part of’’ nature reflects qualitatively different models of the
biological world and the position of human beings with respect
to it. We are currently exploring the idea that this difference in
subjective distance from nature affects environmental decision
making (see refs. 28–30 for evidence linking distancing with
judgment and decision making).The notion that every creature
has a role to play (implicitly or explicitly) reflects an ecological
framework. If this analysis is correct, we might also find this
difference in orientation toward nature in descriptions of prac-
tices associated with the outdoors.
Nature-Related Practices.
Study 2.
Much of the practice-related
work (31) focuses on discourse practices in and across historically
defined cultural communities. Our data come from an interview
with a group of Menominee and European American children
and adults. The interview was structured around a series of
questions about the interviewee’s practices and particular expe-
riences with respect to the outdoors. We began by asking about
33 different kinds of activities and left room for the participant
to volunteer additional ones. These practices were assigned to
one of three categories (fore-grounded, intermediate, back-
grounded), reflecting the position and role of aspects of the
natural world plays in the practice (see Table 3).
Results.
The survey data were examined for frequency and type
for individual practices and then condensed into the three
categories summarizing the role of nature in the practice (see
Fig. 1). The results indicate that European American partici-
pants were more likely to report practices in which nature serves
as background and Menominee participants were more likely to
report practices in which nature is fore-grounded. We treated the
three categories as providing a measure of subjective proximity
to nature and performed a 2 ⫻ 2 ⫻ 3 ANOVA with community,
age and proximity as factors. The interaction was highly reliable
[F (2, 42) ⫽ 5.92, P ⬍ 0.005]. There was no significant group
difference in total reported practices, although children reported
more total practices than adults (t ⫽ 2.66, P ⬍ 0.03).
Summary. The first study indicates that Menominee and Euro-
pean American adults differ in how they talk about nature and
the relationship of human beings to the natural world. The
second study shows that this difference in discourse is paralleled
by differences in outdoor activities or practices that both children
and adults engage in. Compared with rural European American
children and adults, Menominee children and adults spend
relative of their time engaged in outdoor practices in which the
natural world is fore-grounded and relatively less time engaged
in practices where nature is back-grounded. These findings
suggest that the Menominee ecological orientation in reasoning
strategies is paralleled in their framework orientation and prac-
tices. These findings carry over into intergroup conflict over
resources and may have implications for science education.
Table 2. Proportion of participants coded as reflecting various
value orientations at least once
Orientation Rural European American Menominee
Holistic 0.08 0.33
Spiritual 0.08 0.42
Traditional 0.08 0.50
Survival skills 0.75 0.75
Moralistic 0.92 0.83
Ecologistic 0.58 0.83
Abstract respect 0.58 0.58
Abstract liking 0.75 0.25
Abstract knowledge 0.58 0.75
Personal utility 0.42 0.75
Distant utility 0.42 0.17
The table only includes codes reflected in the responses of at least a third
of one of the groups.
Table 3. Outdoors practices and their categorical assignment
Fore-grounded Intermediate Back-grounded
Walks/hikes Yard work Laundry
Forest walks Landscaping Playing sports
Berry picking Powwow Basketball
Fishing Crafts Baseball
Hunting Firewood Volleyball
Gardening Ceremony/ Soccer
Trapping prayer Other
Collecting ginseng Swimming Dirtbiking
Wild ricing Biking Snowmobiling
Maple sugaring Boating
Sweat rocks Cooking
Harvest milkweed Work
Medicinal plants
Sitting outside
Fire circles
Camping
Fig. 1. Means of reported practices by category and community. Solid line,
Menominee. Dashed line, European American.
Bang et al. PNAS
兩
August 28, 2007
兩
vol. 104
兩
no. 35
兩
13871
PSYCHOLOGY INAUGURAL ARTICLE
Resource Conflict
Cultural frameworks and orientations toward nature have wide-
spread implications, perhaps most notably in the area of resource
management and intergroup conflict over resources. In our
studies with Menominee and European American fishermen, we
examined how the differences in orientation might be reflected
in values and attitudes toward different fishing practices (32, 33).
An additional focus was on how each group perceived the other.
We used several converging measures to examine values and
attitudes but report a single one here. Menominee and European
American fish experts were asked to rate 17 different fishing
practices on a 7-point disapproval, approval scale. These prac-
tices emerged from previous interviews.
Attitudes Toward Practices. Although we found a few cultural
differences in attitudes the most striking result was the overall
similarity of values. For example, both groups condemn selling
fish, keeping undersize fish, fishing on spawning beds, using
setpoles to catch trout, and culling out smaller fish to get the
largest possible limit.
Intergroup Perception. A follow-up study addressed the relation-
ship between the actual similarities and differences and the
perceived within and between group similarities and differences.
In this task we asked the same questions as before, but, rather
than exploring each individual’s goals and values, we asked each
participant to report what he thought would be the response of
members of his own group as well as members of the other group.
Specifically, we asked participants how they thought equally
expert members of their community or the other community
might answer the probes concerning values, goals, and attitudes.
First, participants were asked to answer the questions the way
they thought the typical fisherman from their community,
equally expert, would answer them. Next, they were asked to
repeat the task, this time answering from the perspective of an
equally expert fisherman from the other community (for Me-
nominee experts, how European American experts from this
area would answer and for European American experts how
Menominee expert fishermen would answer the questions).
The results on values and attitudes were strongly asymmetri-
cal. European American fishermen predicted that Menominees
endorse each of the practices that are roundly condemned by
both groups. In contrast, Menominee experts accurately pre-
dicted European American fishermen attitudes. In short, the
European American show striking misperception and stereotyp-
ing of Menominee values.
Predicting Within Group Variation. There is more to the story and
it underlines the importance of mental models. There are
individual differences in stereotyping and we have been able to
identify reliable correlates of these differences. The gist of our
findings is as follows: the more strongly a majority culture
fishermen’s mental model of fish and fishing departs from the
Menominee model (e.g., fishing for sport vs. for food), the more
stereotyping they show of Menominee values (33). This latter
result has two important implications. First, it shows that mental
models matter and second it underlines the point that cultural
groups are not uniform in their attitudes and one can gain
leverage by analyzing within group variation (see ref. 32 for
related ethnographic and historical data).
Science Education. Despite widespread efforts to reduce the
disparities, differences between minority and majority students’
achievement persists. Native Americans, specifically, drop out of
high school among the highest rates of all ethnic groups, and only
⬇6% receive bachelor’s degrees (34, 35). Nowhere is this
problem more apparent than in science learning. Over the past
10 years, Native American people received an average of 0.63%
of the bachelors degrees and an average of 0.48% of the
doctorates awarded in Science and Engineering (34). To make
these numbers more concrete, over the past 10 years, there has
been a total of 14 doctorates awarded in computer science, 10 in
physics, 5 in astronomy, 3 in ocean sciences, and 1 in atmospheric
sciences (34). In Biology there have been 108 doctorates
awarded; still, these numbers represent only 0.3% of the total
number awarded (34). This observation in and of itself is
something of a puzzle, given that Native American children
report greater self-efficacy for academic achievement in biology
than in other academic areas.
††
Our own research only deepens this puzzle. We have just
described evidence that rural Menominee children are preco-
cious with respect to an ecological orientation. In addition, on
standardized tests in the fourth grade, 50% of Menominee
children score at or above proficient levels, well above the
national average of 29% (see http://data.dpi.state.wi.us/data/
graphshell.asp?ORGLEVEL⫽DI&FULLKEY⫽08343403.
ZZZZ&DN⫽Menominee⫹Indian&SN⫽None⫹Chosen&
GRAPHFILE⫽GEDISA and http://nationsreportcard.gov/
science㛭2005/s0103.asp?printver). However, by eighth grade,
only 17% of Menominee students score at or above proficient
levels, well below the national average of 27%. In short, preco-
ciousness in biology does not lead to advanced or even average
academic classroom performance.
We are currently exploring the idea that conflict between the
practices associated with science instruction and Menominee
cultural practices is responsible for the disparity between the
knowledge of biology Menominee children bring to the class-
room and their performance in school science. This hypothesis
is consistent with other work exploring children’s repertoires of
practice, and the connection between children’s sense making in
these practices and the target disciplinary knowledge (36, 37).
This research perspective reframes the study of culture and
learning, because it implies that culture should not be under-
stood as individual traits, but rather as the constellation of ways
in which people think, act, and make sense of the world. By
studying culture in this way we can uncover potential leveraging
opportunities for improving the educational achievement of
those groups of children who have been placed at risk by current
educational practices (38). Our prediction is that student moti-
vation and (science) learning are greater when community
values, framework theories, activities, and practices are paral-
leled both in classroom orientation and curriculum-related
values, structures, and practices.
Most science curricula begin with ‘‘model species,’’ a partic-
ular biological kind that instantiates many or most salient
qualities of living things. Then curricula gradually build toward
a systems or ecological orientation. They rarely start with an
ecological orientation. Our findings suggest that, at a minimum,
this trajectory fails to use Native American students ‘‘habits of
mind.’’ Differences in framework theories need to be considered
in the motivating and structuring of learning environments. For
example, when ecosystems are introduced into curricula, they
rarely if ever include human beings. This neglect reflects cultural
assumptions about humans and nature, not the lack of influence
of humans on ecosystems. In addition, our research suggests that
using the intellectual resources Native students bring to class-
rooms may engage them in practices that more close resemble
expert practices and thinking.
††
Yamauchi, L. A., Greene, W. L., Annual Meeting of the American Educational Research
Association, March 24–28, 1997, Chicago, IL.
13872
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0706627104 Bang et al.
Conclusions
The studies we have reviewed and others (see refs. 39 and 40 for
parallel findings from cultural groups living in the lowland
rainforest of Guatemala linking mental models to agro-forestry
practices) suggest a complex interplay between what people
think and how people think. We subscribe to an epidemiological
approach to culture in which culture is studied as causally
distributed patterns of ideas, their public expressions and result-
ant practices and behavior in given ecological contexts. Note,
however, that it would be a misleading over-simplification to
equate ‘‘ideas’’ with independently transmissible elements or
propositions. Instead, we propose, framework theories, mental
models and epistemologies place ideas into a web of interrelated
beliefs that dictate plausible inferences and what is seen as
relevant (41).
Materials and Methods
Study 1: Worldviews.
Participants.
There were 12 adults from each
of the Menominee (seven female, five male) and European
American (six female, six male) communities involved in this
study.
Procedure.
The interview lasted ⬇1 hour and was structured
around a series of questions about (i) interviewee’s characteristic
outdoor activities and particular experiences, (ii) ways in which
children should and do learn about nature, and (iii) goals and
methods appropriate for science instruction. Our focus was on
two questions that came at the middle and toward the end of the
interview. The first was ‘‘What are the five most important things
for your child (or your grandchildren) to learn about the
biological world?’’ The second question was asked in the context
of discussing science education and it asked ‘‘What are four
things that you would like your child (grandchildren) to learn
about nature?’’
Responses to these two questions were used to code orienta-
tions toward the natural world. We found that it was much easier
to establish reliability on whether or not an orientation was
present or absent than to determine whether a statement should
be seen as a single instance with elaboration versus multiple
instances and our analyses are based only on the former.
Coding.
We expanded on Kellert’s original typology in several
ways: (i) a dimension was added corresponding to subjective
proximity to nature. For example, we created a distinction
between ‘‘distant utilitarian’’ (e.g., ‘‘the wood is used in con-
struction’’) and ‘‘personal utilitarian’’ (e.g., ‘‘we eat deer meat’’).
We also added categories corresponding to more abstract knowl-
edge, social value, respect, and liking (e.g., ‘‘I want my children
to enjoy nature’’). (ii) New categories were added corresponding
to categories that were either concrete (survival skills, as seen in
‘‘I want my child to be able to recognize poison ivy’’) or referred
to stances toward nature, such as traditional [referring to inter-
generational values and goals, spiritual, holistic (e.g., ‘‘every-
thing is related’’)]. Although we provide our full set of coding
categories in Table 1, we only report results for categories where
at least a third of the participants in one or more groups gave a
response consistent with the code.
Study 2: Nature Related Practices.
Participants.
The rural Menomi-
nee sample consisted of 16 adults and 11 children who were
residents on the Menominee reservation. The European Amer-
ican sample consisted of 15 adults and 7 children who lived in
Shawano just south of the Menominee reservation.
Procedure.
This work is based on interview and survey data
collected in each community. Participants were asked to fill out
a survey to identify the kinds of activities or practices individuals
engaged in that involved nature and to indicate the frequency
with which they engaged in them. The survey was developed
through brainstorming sessions held with focus groups from each
community to generate a list of practices. There were a total of
33 different practice types on the survey with space for four
additional practices to be added.
Coding.
The 33 practice types were collapsed into three main
categories indicating the place and role of nature, or some aspect
of nature, within the practice. Initially, we thought of practices
either fore-grounding (e.g., fishing) or back-grounding nature
(e.g., baseball). However, there were some practices that did not
fit in either category as neatly (e.g., swimming). These practices
have been assigned to an intermediate category. To ensure
interrater reliability we had four coders (two Native American
and two European American) assign each practice to one of the
three categories. We then mapped these categories onto dis-
tances (background ⫽ 1, intermediate ⫽ 2, and foreground ⫽ 3)
and calculated the interobserver correlations. These interrater
correlations ranged from ⫹0.73 to ⫹0.86. We then averaged the
ratings across observers. Practices with mean scores of ⱕ1.75
were assigned to the category of back-grounded, practices with
mean scores ⬎1.75 but ⬍2.75 were considered intermediate, and
mean scores ⱖ2.75 were assigned to the fore-grounded category.
For a complete list of practices and their categories, see Table 3.
This research was supported by National Institutes of Health Grant R01
HD41653 and National Science Foundation Grants BCS 0132469 and
REC-0529650.
1. Lave J (1988) Cognition in Practice (Cambridge Univ Press, Cambridge, UK).
2. Nisbett RE (2003) The Geography of Thought (Simon and Schuster, New
York).
3. Shweder RA (1991) Thinking Through Culture: Expeditions in Cultural Psychol-
ogy (Harvard Univ Press, Cambridge, MA).
4. Nichols S, Stich S, Weinberg J (2003) in The Skeptics: Contemporary Essays,ed
Luper S (Ashgate Publishing, Burlington, VT), pp 227–248.
5. Saxe GB, Esmonde I (2005) Mind Culture Activity 12(3 and 4):171–225.
6. Rogoff B (2003) The Cultural Nature of Human Development (Oxford Univ
Press, Oxford).
7. Hong Y, Morris M, Chiu C, Benet-Martinez V (2000) Am Psychol
55:709–820.
8. Nisbett RE, Peng K, Choi I, Norenzayan A (2001) Psychol Rev 108:291–310.
9. Medin D, Atran S (2004) Psychol Rev 111:960–983.
10. Atran S (2002) In Gods We Trust: The Evolutionary Landscape of Religion
(Oxford Univ Press, New York).
11. Smith LT (1999) Decolonizing Methodologies: Research and Indigenous Peoples
(Zed, London).
12. Swisher KG, Tippeconnic J (1999) Next Steps: Research and Practice to Advance
Indian Education (Eric Clearinghouse on Rural Education and Small Schools,
Charleston, WV).
13. Beck DR (2005) The Struggle for Self-Determination: History of the Menominee
Indians since 1854 (Univ of Nebraska Press, Lincoln, NE).
14. Hosmer BC (1999) American Indians in the Marketplace: Persistence and
Innovation among the Menominees and Metlakatlans, 1870-1920 (Univ Press of
Kansas, Lawrence, KS).
15. Hall P, Pecore M (1995) Case Study: Menominee Tribal Enterprises (Institute for
Environmental Studies and the Land Tenure Center, Madison, WI).
16. Carey S (1985) Conceptual Change in Childhood (Bradford, Cambridge, MA).
17. Ross N, Medin DL, Coley JD, Atran S (2003) Cognit Devel 18:25–47.
18. Atran S, Medin DL, Ross N (2004) J R Anthropol Inst 10(2):395–420.
19. Medin D, Ross NO, Atran S, Burnett RC, Blok S (2002) Psychol Lear n Motiv
41:1–41.
20. Medin DL, Ross N, Atran S, Cox D, Coley J, Proffitt J, Blok S (2006) Cognit
99(3):237–273.
21. Romney AK, Weller SC, Batchelder WH (1986) Am Anthropol 88:318–
338.
22. Atran S, Medin D, Ross N, Lynch E, Coley J, Ek’ EU, Vapnarsky V (1999) Proc
Natl Acad Sci USA 96:7598–7603.
23. Weller S (1987) Am Behav Sci 31:178–193.
24. Boster JS (1986) Am Anthropol 3:381–399.
25. Coley JD (1995) Child Dev 66:1856–1874.
26. Lo´pez A, Atran S, Coley JD, Medin DL, Smith EE (1997) Cog Psy 32:251–
295.
27. Kellert SR (1993) in The Biophilia Hypothesis, eds Kellert SR, Wilson EO
(Island, Washington, DC), pp 42–72.
Bang et al. PNAS
兩
August 28, 2007
兩
vol. 104
兩
no. 35
兩
13873
PSYCHOLOGY INAUGURAL ARTICLE
28. Trope Y, Liberman M (2003) Psych Rev 110:403–421.
29. Liberman N, Trope Y, Macrae S, Sherman S (2007) J Exp Soc Psychol
43(1):143–149.
30. Bar-Anan Y, Liberman M, Trope Y (2006) J Exp Psychol Gen 135:609–
622.
31. Lee CD (2005) in Black Education: A Transformative Research and Action
Agenda for the New Century, ed King J (Lawrence Erlbaum, Mahwah, NJ) pp
43–114.
32. Medin D, Ross NO, Cox D (2006) Culture and Resource Conflict: Why Meanings
Matter (Russell Sage Foundation, New York).
33. Medin D, Ross N, Cox D, Atran S (2007) Hum Ecol 35(3):315–329.
34. National Science Foundation (2007) Women, Minorities, and Persons with
Disabilities in Science and Engineering: 2007 (NSF Publication No. 07–315) (NSF
Division of Science Resources Statistics, Arlington, VA).
35. Pavel DM, Curtin TR, Whitener SD (1998) Equity Excell Educat 31:48–
54.
36. Lee CD (2001) Am Educ Res J 38:97–141.
37. Warren B, Ballenger C, Ogonowski M, Rosebery AS, Hudicourt-Barnes, J
(2001) J Res Sci Teach 38:529–552.
38. Gutierrez K, Rogoff B (1993) Educ Res 32(5):19–25.
39. Atran S, Medin D, Ross N, Lynch E, Vapnarsky V, Ek’ EU, Coley J, Timura,
C, Baran M (2002) Curr Anthropol 43:421–450.
40. Atran S, Medin DL The Native Mind and the Cultural Construction of Nature
(MIT Press, Cambridge, MA), in press.
41. D’Andrade R (1995) The Development of Cognitive Anthropology (Cambridge
Univ Press, New York).
13874
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0706627104 Bang et al.