E. Margaret Evans
Abstract
This chapter notes that the use of intuitions to jump-start more sophisticated reasoning has been proposed for mathematics. The question addressed is whether core intuitions can also jump-start biological reasoning. Intuitive ideas can offer an immediate action plan that facilitates a rapid appraisal of the human mind or the natural world. Yet, there is a downside, such as a reliance on what may be inaccurate scientific judgments based on cognitive predispositions such as anthropomorphic or essentialist reasoning. Studies conducted with museum visitors are used to support the argument that specific cognitive predispositions can both help and hinder understand-ing. Further, this chapter claims that core intuitions can provide a series of stepping-stones, which, if navigated with care, may promote science learning.
The use of embodied intuitions to jump-start more sophisticated reasoning has been proposed for mathematics (Lakoff and Nunez). Can core intuitions also jump-start biological reasoning, in particular evolutionary reasoning? That is the question addressed in this chapter. Intuitive ideas can offer an immediate action plan (Kahneman), allowing us to make a rapid appraisal of the human mind or the natural world (e.g., a folk psychology or a folk biology: Wellman and Gelman). Yet, there is a downside, such as a reliance on what may be inaccurate
scientific judgments (Stanovich et al.) based on cognitive predispositions such as teleological or essentialist biases (E. Evans, “Emergence” and “Cognitive”).
Studies conducted with museum visitors will be used to support the argument that cognitive predispositions can both help and hinder biological reasoning, with a focus on evolutionary concepts. From the perspective of the science learner, it is important to identify when and how this happens.
Museum visitors provide an ideal population for such studies, as they are less likely than the general US population to reject evolution on ideological or religious grounds (E. Evans et al., “Conceptual Guide”). Further, it is possible to carry out a fairly rapid assessment of visitors’ reasoning before and after vis-its to exhibvis-its on evolution; as well, the inclusion of different age groups makes it relatively easy to compare age- and experience-related changes in reasoning, simultaneously. Results from such studies can provide the basis for developmen-tal learning trajectories. The main argument to be developed in this chapter is that to the extent that intuitive concepts provide a foundation for scientific reasoning (Duschl et al.), they should be incorporated into learning trajectories rather than treated as unwanted intrusions (e.g., Bishop and Anderson). Further, a developmental approach facilitates studies of the role that intuitive theories play in early science learning as well as their role as a default reasoning mode throughout the life span. A more nuanced understanding of intuitive reasoning modes would improve science learning more generally by providing a framework for closing the gap between intuitive and more reflective scientific reasoning (E.
Evans et al., “Encountering”)
The focus of this chapter is on evolutionary concepts because they provide a compelling test case. While the historical (e.g., Quammen) and contempo-rary (e.g., Miller et al.) evidence suggests that evolutionary ideas are rejected on ideological and religious grounds, many have argued that specific cognitive predispositions play a key role in this rejection. As demonstrated in historical (Blancke et al.) and contemporary populations, such biases render the human mind resistant to ideas of evolutionary change while also making it easier to accept the idea that the origin of life on earth was miraculous (see E. Evans,
“Conceptual Change” for a summary). Before describing those teleological (Kelemen; Shtulman and Calabi) and essentialist (Coley and Muratore; Gelman and Rhodes) biases theorized to make evolutionary ideas untenable (Mayr), I briefly review the extant literature on intuitive and reflective (scientific) reason-ing and consider how the two might compete or cooperate when students study counterintuitive topics, such as evolution.
In Two Minds: Intuitive and Reflective Reasoning
The claim that there are “two minds in one brain” (J. Evans, Thinking Twice), the one intuitive and the other reflective, has been advanced by Jonathan Evans and Keith Stanovich, among others, with convergent evidence drawn from a variety of fields including neuroimaging, to support the basic concept (J. Evans and Stanovich). Following Jonathan Evans’s Thinking Twice, in this chapter the terms intuitive and reflective will refer to what has been variously called system 1/
system 2, type 1/type 2, and implicit/explicit processing. Linking this variety of proposals is a common thread. The intuitive brain is thought to be evolutionarily old while the reflective mind is more recent and may (arguably) be exclusively human; the former is a rapid parallel processor of implicit belief-based informa-tion often at a preconscious level (sensainforma-tions and feelings), while the latter is a low capacity, effortful serial information processor of explicit information, often (but not always) at a conscious or meta-reflective level (J. Evans, Thinking Twice; J. Evans and Stanovich). Importantly, unlike type 2/reflective processing, type 1/intuitive processing is thought to be associative, operating independently of working memory and cognitive ability, while type 2 processing is rule based (J.
Evans, “Dual-Process”; Stanovich et al., “Complexity”).
Anyone familiar with the work on the emergence of children’s theory of mind and executive function will recognize that according to any of these defi-nitions the reflective mind is later developing, if assessed by children’s increasing ability to exercise effortful control over their thought processes and to read the minds of others (e.g., Wellman; Zelazo). Thus, a clear developmental prediction is that intuitive processes, particularly intuitive theories, are most likely to influ-ence the very young sciinflu-ence learner. However, intuitive theories also provide a default for any age group under circumstances when they have to process infor-mation rapidly or are not explicitly instructed that they should be mindful of the problem (J. Evans and Stanovich).
Do intuitive and reflective reasoning processes compete with intuitions impeding an understanding of the issue at hand? The classic demonstration of such competition has been the case of the three-premise logical argument or syl-logism (e.g., no A are B; some C are B; therefore, some A are not C) where the outcome is either logical but unbelievable or illogical and believable (J. Evans, Thinking Twice 114). In such studies, believable or unbelievable premises are substituted for A, B, or C. What tends to happen is that, with little reflection, participants will exhibit a belief bias and select the intuitively plausible but
logically incorrect solution (J. Evans, Thinking Twice). In this case the intuitive and reflective minds appear to compete, delivering different solutions depend-ing on the worddepend-ing of the problem and whether or not the participant exercises effortful control over his or her intuitive response. These kinds of studies also demonstrate, however, that the reflective mind can suppress the output of the intuitive mind, providing it is given the time to do so and the motivation (J.
Evans, Thinking Twice).
How might they cooperate? With the development of expertise in any field, it is the reflective mind that appears to oversee the initially effortful acquisition of novel skills; as the skills become routinized they are gradually subsumed under the direction of the intuitive mind. The classic example of this transfer of skill is driving. What is at first a long slow and mindful process gradually becomes a skilled activity that is performed speedily and intuitively with little access to the reflective mind (J. Evans, Thinking Twice). If, however, a novel problem arises, an icy road or an unexpected hazard, then the reflective mind swings into action and analyzes the problem, providing a new plan of action. What is important to note here is that in general the intuitive mind delivers a fast, effective, and accurate performance, particularly for routine tasks, even though it might pro-vide an incorrect predetermined response to a novel hazard. In decision making the intuitive mind draws on past experience, a lifetime of repeated associations between diverse activities, behaviors, and outcomes. In contrast, the reflective mind analyzes the immediate present or the possible future, running explicit simulations (what-ifs) of possible outcomes and planning actions based on these simulations (J. Evans and Stanovich).
This necessarily brief description of the two-mind theory raises a host of questions regarding the way in which an understanding of the joint actions of intuitive and reflective processes might be used to help improve the delivery and understanding of scientific knowledge. In terms of skill development, Sun et al. (“Interaction”) have suggested that implicit (procedural knowledge) and explicit (declarative) processes interact to produce learning; moreover, there are synergy effects. Although in the driving example just described the inter-action would appear to be top-down, with the reflective mind jump-starting the process, Sun et al. present evidence supporting the argument that there are likely to be a variety of ways in which the two kinds of learning might inter-act. Sometimes making the learning process explicit hampers learning; in other cases, implicit and explicit learning appear to be independent processes with different time courses. To model the possibilities, Sun et al. present a neural
network model with bottom-up intuitive processes as the driver, which accounts for different kinds of data, including cases in which learning occurs without conscious awareness. Two broad conclusions emphasized by these authors are the importance of modeling both implicit and explicit processes when studying learning and that bottom-up processes may be as important if not more so than top-down processes. Much of this research has focused on the learning of arti-ficial grammars or similar problem sets where the influence of prior knowledge is not necessarily an issue, but what about problem sets where belief biases are viewed as a key concern, as is often the case for science understanding (Bloom and Weisberg)?
Initially, it would seem, the acquisition of scientific knowledge is likely to be effortful and slow as it is overseen by the reflective mind, particularly if it involves the suppression of intuitions that are at odds with scientific ideas. But once sci-entific reasoning becomes routinized and effortless, what happens to these core intuitions? Are they supplanted? As the evidence indicates that the belief biases of the intuitive mind never disappear, even if they are temporarily suppressed (J. Evans and Stanovich), perhaps a better approach would be to investigate how they are implicated in the learning process.
For evolutionary reasoning, the basic claim advanced by many researchers is that belief biases or cognitive predispositions compete with and impede scien-tific understanding (e.g., Kelemen; Gelman and Rhodes). If this is the case, one would predict that measures of belief bias would be negatively correlated with evolutionary reasoning. In the following sections, I demonstrate that this is not always the case, first for natural selection, and second, for common descent. In fact, I go further and argue that often such cognitive predispositions may jump-start scientific reasoning in a bottom-up process.
The Role of a “Restricted Teleology” in an Understanding of Natural Selection
Even if a particular intuition appears early and is then supplanted by scientific concepts, this does not mean that the intuition impedes scientific understanding.
All this suggests is that it is a default reasoning mode brought to bear on a partic-ular topic when more reflective reasoning modes cannot be exercised for lack of knowledge, time, access, or processing capacity. Essentially, this is the argument that I pursue in my discussion of the effects of teleological and essentialist rea-soning (see Jarnefelt et al., and Jarnefelt, this volume, for a different perspective).
Of course, this does not mean that the same intuition may be helpful rather than a hindrance; such a claim requires a more nuanced argument.
One of the problems with research on intuitive reasoning is that the target intuitions are often underanalyzed or blended in such a way that it makes it dif-ficult to decipher the precise roles they actually play in student learning. This is particularly a problem with the extensive research on the role of teleological rea-soning in students’ understanding of natural selection (E. Evans, “Conceptual Change”). Teleological reasoning is problematic because it implies purpose, progression, and intentional design, whereas natural selection exhibits none of these features — in fact the inverse is true. However, in studies investigating these effects, purpose and intention are routinely conflated. Students often reason that out of necessity an organism needs a particular trait in order to survive, in which case the trait satisfies an intrinsic purpose, called need-based reasoning. But, even if students lack knowledge of the mechanism by which traits evolve, it does not necessarily follow that they believe the trait is acquired through the intentional efforts of the organism, called desire- or want-based reasoning (E. Evans et al.,
“Encountering”). Or, even that an external agent such as God or mother nature designs the organism for a purpose (E. Evans, “Conceptual Change”). Fine-grained analyses of these reasoning modes suggest the existence of a “restricted teleology,” whereby students may grasp the idea that an adaptation improves the survival of an organism without falling prey to anthropomorphic beliefs in which the organism or an external agent is able, like humans, to intentionally change the trait (E. Evans et al., “Encountering”).
In investigations of this line of reasoning, study participants should be given the opportunity to endorse or reject each of these reasoning modes in separate closed-ended questions. Ideally, to gain further insight into their underlying reasoning, they should also be asked open-ended questions. Using this range of measures, Evans and her colleagues have conducted a number of studies among museum visitors of different ages and found that with age and expertise visi-tors increasingly distinguish between want- and need-based reasoning. Typically, younger school children conflate want- and need-based reasoning, endorsing both; thus, for this age-group, these modes are often significantly correlated (Legare et al.), as might be predicted from the research on young children’s tele-ological reasoning (e.g., Kelemen). In contrast, older school children and adults often adopt a restricted teleological pattern, endorsing need-based reasoning while rejecting want-based reasoning (E. Evans et al., “Conceptual”; Spiegel et al.). For example, in a recent museum study (Horn et al.), 8- to 15-year-olds were presented with three diverse scenarios and asked to explain the adaptive changes
that occurred over time in the target species. Unlike their younger counterparts, the 11- to 15-year-olds rejected anthropomorphic explanations (e.g., the lizards changed over time because “they don’t like to get eaten”) and endorsed need-based reasoning (e.g., “because the different kinds [of anoles] need to adapt to their different environments”). Critically, it is the latter pattern that is positively correlated with natural selection reasoning (Horn et al.; Spiegel et al.).
This overall pattern of results suggests that a restricted teleology could jump-start natural selection understanding by drawing attention to the necessity of the adaptation for survival while increasing visitors’ sensitivity to natural mecha-nisms that do not involve the intentional actions of the organism. However, such studies do not provide evidence of a causal effect. Under what contexts might a restricted teleology be useful to the learner?
As the research team for a multicomponent exhibit on evolution, Evans and her colleagues investigated this phenomenon in more detail by incorporat-ing these types of scaffolds into the development of a narrative-based exhibition called “Charlie and Kiwi’s Evolutionary Adventure” (E. Evans et al., “Spiral”).
This exhibit focused on the adventures of a young boy, Charlie, as he traveled back in time to discover why kiwis (including his stuffed kiwi) lacked the abil-ity to fly and to find out what was so special about the ancestors of birds. In addition to a video theater experience, which conveyed the story of Charlie and Kiwi, the exhibition consisted of multiple components providing evidence of dinosaur-to-bird evolution. One of the challenges of an exhibition targeting school children was that the design team was inclined to use anthropomorphic language in the text, because it elicits interest and engages the young visitor. To assess how such language might affect understanding, the research team con-ducted several studies indicating that a restricted teleology might be a more effective cognitive tool.
In an initial qualitative assessment of children’s understanding of natural selection, children were asked to recall the story of the Galapagos finches’ sur-vival, after viewing an exhibit on that topic. The following examples of children’s language in their retelling provided converging evidence for the earlier findings (E. Evans et al., “Spiral” 49).
Interviewer: “Tell the story back to me.”
Six-year-old: “The finch wanted a bigger beak” [example of want-based reasoning]
Nine-year-old: “You don’t evolve because you want to . . . you evolve because you need to” [example of a restricted teleology]
As a follow-up to these qualitative studies, the research team ran an experiment in which children were randomly assigned to three conditions, in each of which children were told a story about bird evolution and then asked to recall the story.
In one condition the story was presented using the language of natural selec-tion, in a second condiselec-tion, want-based reasoning, and in the third, need-based reasoning (Legare et al., “Anthropomorphizing”). On a variety of measures, the 5- to 7-year-olds were the most likely to use anthropomorphic reasoning and, overall, the anthropomorphic stories elicited the fewest scientific explanations.
However, all children were more likely to use natural selection reasoning in their story recall if they had heard either the natural selection or the need-based sto-ries, and this was especially true for the 9- to 12-year-olds.
This kind of evidence prompted the exhibit development team to ensure that the distinctions between want and need, in the form of a restricted teleology, were called out in the exhibit text and particularly in the video experience. For example, in Charlie’s time travels he met up with his great-great-great-grandfather (who looked remarkably like Charles Darwin), and in one scene, while the two time-travelers viewed a nest of baby dinosaurs, some of which had feathers, the great-great-great-grandfather said: “Animals can’t just grow feathers when they want to. They have to inherit them from their parents.” Later, in another scene, when the old man was explaining why kiwis could not fly, he said: “No, but you must understand, Charlie: every bird has what it needs for where and how it lives.
It’s adapted to its environment” (E. Evans et al., “Spiral” figure 3.3). Note that in these examples, both intuitive and scientific explanations were incorporated into the text, thus merging top-down and bottom-up approaches into a single cognitive tool.
When the exhibition was complete, the research team ran a randomly assigned controlled study and found that the 5- to 7-year-olds who visited the target exhibit were much less likely than their peers who visited the control exhibit to endorse or use anthropomorphic language; additionally, the older children were more likely to use need-based reasoning and grasp the basics of natural selection than their peers who visited the control exhibit (E. Evans et al.,
“Spiral”). Such studies provide some evidence that when intuitive concepts are embedded in the narrative, be it in a curriculum or an exhibit, learners can tran-scend the intuitive and integrate more scientific modes of reasoning into their explanations. First, however, it is necessary for learners to distinguish between a restricted teleology, in which adaptive traits serve the intrinsic purpose of the organism, and a teleology that is intertwined with intentional reasoning.
The Role of an “Expanded Essentialism” in an Understanding of Common Descent
From the early days of research on cognitive biases that inhibit the understand-ing of biological concepts, essentialism has occupied a special niche in the minds
From the early days of research on cognitive biases that inhibit the understand-ing of biological concepts, essentialism has occupied a special niche in the minds