In Algona, IA, in 1957, children were funneled into two categories depending on whether they sang in the choir. “The kids that sang tended to be brighter, more middle-class,” Medin says. “Students who did not sing tended to come from the wrong side of the tracks.” He was lumped with the non-singers.
However, despite their poor grades, Medin found his rough friends intelligent and interesting. “There were a lot of things they were very curious about,” says Medin, who is now professor of psychology at Northwestern University (Evanston, IL). “And I just became intrigued by why these kids that seemed so motivated outside of school were failing in school.”
His interest in how individual differences affect learning styles led him to study psychology in college, and, later, to explore the interface between psychology and anthropology. In his career, he has striven to move beyond simplistic laboratory models of how the mind operates and to understand how our expertise and cultural background influence our mental picture of the world. His unique approach led the National Academy of Sciences (NAS) to induct him into its ranks in 2005. His Inaugural Article appeared in the August 28, 2007, issue of PNAS (1).
Space Monkeys
When Medin was a teenager, his family moved to Minnesota, where he attended Minnesota State University in Moorhead (then Moorhead State College). He began as a dual major in psychology and mathematics, but found the psychology teachers more interesting. He would later regret not having pursued more mathematics, but a brush with psychology research as part of a National Science Foundation summer research fellowship spent at the University of South Dakota (Vermillion, SD) got him hooked on studying animal behavior.
In fact, it intrigued him enough that, upon graduating from Moorhead, he committed to the Ph.D. program in psychology at the University of South Dakota. It was 1965, and NASA was racing the Soviets to the moon. Medin won a NASA fellowship in psychology to study the behavior of rhesus monkeys. In terms of sending humans into space, “there was the question of what we knew about perception, learning, memory in monkeys that might be relevant,” he says. “I was probably among the last of the NASA fellowships because they phased them out.”
Medin (Right) and his cousin share an early fascination with science.
Medin wrote his dissertation on how rhesus monkeys perceive shapes. In particular, he investigated whether “information theory” could predict which shapes monkeys could remember. According to information theory, symmetrical figures should require less effort to remember. In fact, some shape features did affect monkey memory, “but symmetry wasn't one of them,” Medin says. While carrying out these experiments, he had the idea that humans' language use might underlie major differences between the way monkey and human brains function. This thought—an interspecies version of the “Whorfian hypothesis” (i.e., that language influences thought)—spawned his interest in comparative psychology.
After completing his dissertation, Medin considered applying for a faculty position. His advisor, Roger Davis, had misgivings. “He told me, ‘Look, you're not going to get a good job from an unprestigious school like South Dakota; you really should do a post-doc,”’ Medin says. “What he didn't tell me is, you're unlikely to get a really good post-doc if you're not from a prestigious school.” But Davis had connections. He had studied under William Estes at Stanford University (Stanford, CA) and had subsequently sent him a star pupil, David Rumelhart, who was inducted into the NAS in 1991. “I think [Davis] wrote Estes and said, ‘I sent you Dave Rumelhart, do me a favor, I've got this green kid. Teach him some stuff,”’ says Medin.
Should Have Learned More Math
It turned out that, green as Medin was, he and Estes were a good match. In fact, Medin had done one of the experiments Estes had proposed in a grant application he had just submitted to the National Institutes of Health. “There were some coincidences,” Medin says. “I think that's what led him to take a chance on me.” At the time, Estes was just moving from Stanford to Rockefeller University (New York, NY), where Medin joined his group in 1968. Within a year, Estes had arranged a position for him as an assistant professor in large part “because my draft board didn't know what a post-doc was,” Medin says.
Estes was one of the founders of mathematical psychology, and around this time, Medin began to wish that he had completed that major in mathematics. “If I could have turned the clock back,” he says, “I would have ignored the yellowed cards that the calculus teacher was using from her notes that were 30 years old, and worked a little harder.”
His work—with monkeys and, for the first time, humans—hinged on constructing mathematical models to describe how animals and children discriminate or categorize. In a typical study, for example, researchers might examine how monkeys or children learn to categorize triangles. To do that, they might reward study participants each time they chose a triangle from a group of shapes. The dominant model at the time assumed that participants would compare new objects to an abstract, average idea of a triangle that they had developed.
But “when you average, you throw away a lot of information,” Medin says. So he developed a competing “exemplar” model, where people compare new examples, which act as “retrieval cues,” to specific objects they have already categorized. Consider how one might decide whether a certain animal is a mammal or a bird: “If it looks like a ground squirrel, it's probably a mammal, given that you know ground squirrels are mammals,” Medin says. He soon broadened his studies to include adults as well as children.
Natural Stimuli
Although Medin was making inroads with his exemplar model, he was beginning to question the kinds of categories he and his colleagues were using in their experiments. Most psychologists at the time used simple geometrical stimuli such as triangles or squares in their experiments. But Medin's ethologist colleagues at Rockefeller, such as Peter Marler (renowned for his elegant work on birdsong learning), had been pioneering the use of “natural” stimuli.
Discussions with Marler made Medin uneasy about using the more traditional artificial stimuli. Even in mathematical models, he found that conclusions drawn from abstract experiments often didn't apply to how the mind works in more complex models that better approximated real life.
By 1978, Medin had moved to the University of Illinois at Urbana–Champaign to lead his own laboratory, advise graduate students, and work with a broader range of colleagues than he could have at Rockefeller, where professors were concentrated in a more specialized wedge. He was also following Marler's lead: using more natural stimuli, such as trees and birds, which, he assumed, evolution had equipped the
“It's less important to try to cover a whole range of facts than it is to introduce students to a few big ideas.”
human mind to process. However, he was now working almost exclusively with humans. His work with non-human primates had dwindled because of his students' lack of interest.
The results of Medin's studies using natural stimuli led him down a new path altogether. Indeed, it soon became clear to him that people's expertise played a crucial role in how they categorize relevant objects and stimuli and how they use those categories to think about the world.
The Expert Effect
Medin uses a sample question to illustrate the effect of expertise: “Suppose there's one disease that white pine and weeping willow get and another disease that river birch and paper birch get. Which disease do you think is more likely to affect all trees?” The most obvious answer is that the disease afflicting pine and willow would be more likely to spread because they come from different taxonomic categories. That is the response from the standard undergraduate psychology study participants, who, Medin says, “can't tell the difference between a maple and an oak.” However, when Medin asked parks workers the same question, they replied that the disease afflicting birch trees would be more likely to spread.
“They said, first of all, that birches are incredibly susceptible to disease,” explains Medin. “So if one [birch] gets it, they're all going to get it. Secondly, they have a wide geographic range, both naturally and because they're planted as ornamentals, so there'd be plenty of opportunities for the disease to spread.” The parks workers, Medin explains, combine causal and ecological reasoning. The findings from this and other studies led him to conclude that most of the findings published about categorization and reasoning were only valid for “novices” with no specialized knowledge.
In 1989, Medin moved to the University of Michigan (Ann Arbor, MI), where he joined his colleague Ed Smith, with whom he had written a 1981 book on categorization (2). “The University of Illinois really maximized colleagues but didn't so much maximize [the] local environment,” says Medin. “It's flat farmland, and I decided I didn't really want to retire there.” But it was a short stay in Ann Arbor—in 1992, he decided to leave for Northwestern University because it held better professional opportunities for his wife, Linda Powers, a clinical psychologist.
The Rookie Anthropologist
Although Medin's stay in Michigan was short, he met several researchers there who have since proved to be valuable collaborators. One of these was anthropologist Scott Atran, a coauthor of Medin's Inaugural Article. Atran invited Medin to help him explore how culture influences people's mental models of their environment. In short, he was asking Medin to venture into anthropology. Atran had recently begun fieldwork in Peten, Guatemala, and Atran and Medin began to study how “folkecology” differed among three groups that shared the lowland rainforest: the Itzaj Maya, who have lived there for millennia; Spanish-speaking immigrant Ladinos; and Q'eqchi' Maya, recent immigrants from the highlands. All three groups rely on the same habitat for their livelihood. The researchers were trying to understand how culture might affect calculations of self-interest as the different cultures exploited common resources. They believe that a better understanding might enable the Guatemalan government and the United Nations to prevent a “tragedy of the commons” resulting in destruction of the rainforest.
“The government of Guatemala was just operating on the assumption that if you're Maya, you protect the forest,” Medin says. But he and his colleagues found that this assumption was incorrect (3). As one would expect, the native Itzaj Maya had the most sustainable forest practices, based on “a very rich ecological model of how the forest works,” in which species are valued according to a combination of their value to humans and how central a role they are perceived to play in the environment. However, the Q'eqchi' Maya, even though they knew the plants and animals well, were the most destructive group. The Q'eqchi' “had incredibly impoverished models of plant–animal interactions.” Surprisingly, the Ladinos, despite their Spanish heritage, had more sustainable practices than the Q'eqchi'.
From analyzing the social networks in the three groups, Medin and Atran discovered that the Ladinos were able to tap into Itzaj expertise. The two concluded that “ecologically sound behaviors” could be acquired by outsiders such as the Ladinos, whose culture initially had no relevant wisdom, as long as they could connect to experts. It takes time to learn, however, and the rate of rainforest destruction may be too rapid for such wisdom to reverse.
Cultural Prisms and Trophy Fish
Medin and Atran have continued to examine the effect that culture has on our perception of nature. In Medin's Inaugural Article (3), the two, along with a third colleague, Megan Bang, chose to ask how Native Americans—specifically, members of the Menominee tribe of Wisconsin—and majority-culture Americans, of equal degree of expertise as fishermen, categorized local fish species.
“Menominee experts tended to sort ecologically,” Medin says. “They could very rapidly come up with different kinds of ecological relations when we asked about fish–fish interactions. And the majority-culture guys were more likely to organize their knowledge either taxonomically or in terms of goals,” such as which game fish were more prestigious. Menominee experts also tend to consider the entire life cycle, whereas majority-culture experts think only of adult fish. There was an odd tangential finding also: The more majority-culture men focused on trophy fish, the more likely they were to misperceive or ethnically stereotype their Menominee counterparts.
Medin's childhood experiences remain relevant to his work. His Huckleberry Finn-like buddies in Iowa were fascinated by science. Underwater bombs were a favorite topic, he recalls, adding that “in those days, believe it or not, you could walk into a hardware store and buy dynamite fuse for five cents a foot.” The school system, however, did not recognize or reward their interest. Medin has found parallels to this in the way Native American children have difficulty learning science in a classroom setting.
Menominee children, whose language skills initially lag the national average by a year—and who come from families of low socioeconomic status—score above the national average the first time they write standardized science tests. “Four years later,” Medin says, “science has gone from being their best subject to being their worst subject. And that is a very striking phenomenon.”
Medin, Bang, and their colleagues are pursuing the idea that “framework theories” relevant in school and in students' home lives are clashing. The students would do better, he suggests, if teachers presented information in a manner similar to the way the students learn and practice community values. “It's less important to try to cover a whole range of facts than it is to introduce students to a few big ideas,” Medin says. One such concept might be evolution, but another could be a view of ecology through a system-level lens, parallel to the way Native American culture perceives the environment. To Medin, it is clear that, whatever your singing voice, your culture defines the way that you learn.
www.psychspace.com心理学空间网
