Learning How to Believe: Epistemic Development in Cultural Context

Abstract

Over the last decade, researchers have become increasingly interested in students' beliefs about the nature of knowledge and how these beliefs develop. Although initial psychological accounts portrayed epistemic development as a domain-independent process of cognitive maturation, recent studies have found trajectories of epistemic development to vary considerably across contexts. However, few studies have focused on cultural context. This article examines the role community values and practices play in fostering particular epistemological orientations by comparing the epistemological beliefs of 5th, 8th, and 12th graders (N = 200) from General and Religious schools in Israel regarding 2 controversies: belief in God and punishment of children. In both controversies, older participants were less likely than younger participants to consider the controversy rationally decidable. However, this shift emerged earlier in the God controversy than in the punishment controversy. In the God controversy, General pupils were less likely than Religious pupils to consider the question rationally decidable or their own beliefs infallible. But no such school differences were observed in the punishment controversy. Qualitative and quantitative analyses linked these differences to divergent discourse practices at General and Religious schools, suggesting that the relations between learning and epistemic development are more intricate than has been assumed hitherto.

Epistemology is an area of philosophy concerned with questions of what knowledge is and how it is justified. Although few people give these questions such detailed and sustained attention as professional philosophers, anyone attempting to acquire, produce, or evaluate knowledge relies, at least implicitly, on some set of epistemological beliefs. Such beliefs are of obvious interest to educators. To understand how students acquire, evaluate, and justify knowledge, we need to understand what they consider knowledge to be. And to help students become discerning consumers and responsible producers of knowledge, we need to understand how people learn to exercise reflective judgment in the face of competing claims.

The psychological study of epistemic development is undergoing something of a renaissance. Interest in this area can be traced back to Piaget (1970), Dewey (1933), and beyond (e.g., Plato's Theaetetus). But only recently have psychologists begun to draw together hitherto disparate strands of empirical inquiry to chart in detail the course of epistemic development from infancy to adulthood (see, e.g., Hallett, Chandler, & Krettenauer, 2002; Kuhn, Cheney, & Weinstock, 2000; Wainryb, Shaw, Langley, Cottam, & Lewis, 2004).

Initial research in this field was conducted largely within a neo-Piagetian paradigm. Epistemic development was characterized as a progression through discrete stages or levels of epistemological understanding, each following the other in invariant sequence and constituting a comprehensive transformation of the individual's conception of knowledge (see, e.g., Chandler, 1975; Kitchener & King, 1981; Perry, 1970). Researchers differed somewhat in their nomenclatures and in their criteria for distinguishing levels of epistemological sophistication, but they concurred broadly in positing at least two major shifts in epistemological understanding (for recent reviews, see Hofer & Pintrich, 1997, 2002). Initially, the individual subscribes to an objectivist conception of knowledge, in which every question is believed to have a single, correct answer that is knowable with absolute certainty. Subsequently, in a radical shift, objectivism is abandoned in favor of subjectivism, and the individual equates all knowledge claims with matters of personal taste or preference. Finally, a balance is achieved in which objective and subjective aspects of knowing are coordinated. At this “evaluativist” stage (cf. Kuhn, 1991), the individual sees knowledge as something that is constructed tentatively by evaluating the evidence for and against competing beliefs and points of view.

There was some debate about the precise relations of these stages to Piaget's stages of intellectual development (see, e.g., Boyes & Chandler, 1992). However, they were viewed by most researchers as a kind of “post-formal operations” taking up the formal operational attainments of hypothetical thinking and perspective-taking and applying them wholesale to knowledge claims as such, in a form of meta-metacognition (see, e.g., Kitchener, 1983; Kuhn, 2000; Moshman, 2003).

Empirical support for this model of epistemic development came from several parallel research programs. Overall, these studies provided substantial evidence of development in the hypothesized direction (Hallett et al., 2002; Hofer & Pintrich, 1997; King & Kitchener, 1994). However, the ages at which the key shifts in epistemological understanding were observed varied enormously from program to program. As Hallett et al. (2002) noted regarding the emergence of subjectivism, “one could read selectively from the literature and, with good reason, conclude that such abilities put in their appearance at either 4 or 6 or 8 or 12 or 16 or 20, or in receipt of a Ph.D.” (p. 289).

These vast discrepancies have led researchers to reexamine some of the assump-tions underlying their models and measures of epistemic development. In particular, researchers have begun to question the comprehensiveness of shifts in epistemological understanding and to attend more closely to the variety of contexts within which epistemic beliefs are held, employed, and articulated (cf.Elby&Ham-mer, 2001; Hammer&Elby, 2002, 2003). This attention to context has taken several forms. One line of research has sought to show that epistemic development is a some-what domain-dependent process that occurs with respect to some kinds of knowledge claim before others. For example, it has been proposed that subjectivism emerges in relation to aesthetic claims before it emerges in relation to claims about the physical world (e.g., Hallett et al., 2002; Kuhn et al., 2000; Wainryb et al., 2004). Another line of research has sought to show that epistemological beliefs are multi di-mensional rather than unitary. For example, it has been proposed that people's beliefs about the complexity of knowledge develop more or less independently of their beliefs about its certainty (e.g., Schommer, 1990, 1993). Although the accumulated empirical evidence is as yet suggestive rather than conclusive (see Elby&Hammer, 2001; Hallett et al., 2002, pp. 303–304; Kuhn et al., 2000, pp. 321–322), both research programs hold out the possibility of disentangling the course of epistemic de-velopment by identifying elements of epistemological understanding that emerge at different points in the lifespan(cf.Hallett et al.,2002, p. 290; Wainryb et al. 2004).

However, “domains” and “dimensions” are not the only contexts within which epistemic beliefs are situated. Just as people's epistemic beliefs may be affected by the particular content of the knowledge claims being evaluated, so too may they be affected by the particular procedures of knowledge evaluation practiced in, and sanctioned by, the communities in which they participate (cf. Hammer & Elby, 2003). Few studies have directly examined cross-cultural variation in epistemological beliefs. And those that have done so have tended to focus on comparisons between U.S. and East Asian college students (e.g., Chan & Elliott, 2002, 2004; Nisbett, Peng, Choi, & Norenzayan, 2001; Zhang, 1999). Nonetheless, the findings of such studies are highly suggestive. Especially striking is Zhang's finding that Chinese students' epistemological beliefs shifted over the college years in exactly the opposite direction to those of their U.S. counterparts. Specifically, rather than moving from objectivist to subjectivist conceptions of knowledge, similar to their peers at U.S. universities, Chinese students appeared to shift from more subjectivist conceptions of knowledge to more objectivist ones.

Such findings raise intriguing questions about the relations between culture and epistemic development, and about the relations between learning and epistemic development more generally. For a field of inquiry dominated by educational psychologists, remarkably little is known about these relations. As Hofer and Pintrich(1997) noted, “there is little empirical evidence for precisely what fosters epistemological development or how epistemological beliefs are altered” (p. 123). Research has shown that schooling makes a difference(e.g. Bell&Linn, 2002; King&Kitchener, 1994), but it has yielded little insight into how or why it makes a difference. Moreover, researchers have tended to equivocate between viewing epistemological beliefs as causes and effects of learning—as age-dependent constraints on instruction on one hand andasoutcomesof instructionon the other hand (cf. Kuhn, 1991, 2003; Schommer, 1990, 1993).

Cross-cultural studies of age trends in epistemological belief are one way to investigate systematically the relations between age, learning, and epistemic development. If the nature and timing of age-related shifts in epistemological belief vary cross-culturally, then this is presumably because participants in different cultures learn, through participation in practices particular to their respective communities, to treat knowledge claims in different ways. To the extent that researchers can pinpoint such practices and measure their impact on individuals' epistemological beliefs, we can begin to characterize more precisely the relations between learning and epistemic development.

However, in designing such studies, researchers must take great care to distinguish between reportage and editorial. “Development” is a value-laden term. Theoretical models of psychological development do more than describe a sequence; they attach values to different points along the sequence. Specifically, they define some psychological states or capacities as more mature, adequate, or sophisticated than others. When developmental studies are restricted to homogeneous cultural settings, within which there is little serious disagreement among experts about the relative adequacy of different psychological states or capacities, the risk of bias in diagnosing participants' levels of development is relatively slight. However, as the cultural heterogeneity of the sample increases, so too does the risk that the model on the basis of which participants' development is diagnosed is ethnocentrically biased against a portion of the sample. For example, beliefs that are defined as immature by the developmental model might be considered mature within one of the cultures from which the sample is drawn, or vice versa.

These dangers exist to some extent in all cross-cultural studies of psychological development (see Cole & Scribner, 1974; Greenfield & Bruner, 1966; Shweder, Mahapatra, & Miller, 1990). However, in cross-cultural studies of epistemic development the danger is even more clear and present. This is because there are at least four levels at which epistemological judgments can be made, and as one ascends from the first level to the fourth, the value assumptions implicit in these judgments become less visible to the naked eye but never quite disappear.

At the first, most basic level, there arejudgments about the epistemic status of par-ticular claims, such as whether a given assertion (e.g., “Eating people is wrong”) is known or merely believed (e.g., “John doesn't know that eating people is wrong; he merely believes that eating people is wrong”). At one level of abstraction up from such judgments, there are judgments about the epistemic status of general classes of claim, such as judgments about whether knowledge is possible about, say, matters of aesthetics or morality (e.g., “There are no such things as moral truths or moral knowledge; morality is a question of commitment—not of what is or is not the case”). At yet another level of abstraction up from such judgments, there are further judgments about the criteria by which epistemic states are distinguished one from another, such as judgments about what differentiates knowledge from belief or opin-ion (e.g., “Knowledge is justified true belief;if a belief is true but not justified, or jus-tified but not true, then it isn't known”). And beyond this third level, there is a fourth level of judgments about the relative adequacy of particular criteria for distinguishing between epistemic states, such as judgments about whether the subjectivist equation of knowledge with opinion is less adequate than the evaluativist's insistence that knowledge differs from opinion in being supported by evidence (e.g., “Subjectivism is inferior to evaluativism because it fails to account for our belief that some claims are better supported than others”).

According to this analysis, authors of contemporary models of epistemic development are themselves exercising a form of epistemological judgment in articulating their models. Specifically, they are operating at the fourth level (outlined previously), expressing their own beliefs about the relative adequacy of particular epistemological orientations. This is all well and good as long as the beliefs in question are compatible with the epistemological beliefs, values, and practices of the communities to which the model is applied. However, in cross-cultural studies of epistemic development, such compatibility cannot be assumed in advance of the investigation itself. Accordingly, if one wishes to avoid ethnocentrism, one cannot conduct a cross-cultural study of epistemic development without suspending, at least temporarily, one's hierarchical assumptions about the relative adequacy of particular epistemological orientations.

This is not to say that researchers are not entitled to opinions of their own about the relative adequacy of particular epistemological beliefs or that all talk of epistemic development is inherently ethnocentric. It is merely to point out that because standards of epistemological maturity may themselves vary across cultures, assumptions about the relative adequacy of particular epistemological beliefs must not be built into the design of cross-cultural studies. This study, therefore, departs from previous studies of epistemic development by dropping the assumption that some epistemological beliefs are inherently more adequate than others. By dropping this assumption, I do not affirm the contrary assumption that all epistemological beliefs are equally adequate. Rather, I adopt a form of methodological agnosticism to reduce the scope for ethnocentric bias in the study'sdesign. In the Discussion section, I return to the question of hierarchy and review the assumptions of contemporary models of epistemic development in light of my findings.

This study investigates the relations between age, learning, and epistemic devel-opment by comparing the epistemological beliefs of pupil sat Religious schools and General schools in Israel about two controversies: one religious and the other nonre-ligious. Specifically, this study asks three questions. First, how do beliefs about the nature of religious claims vary with age and school? Second, to what extent are these age and school differences attributable to underlying group differences in religious commitment and general epistemological sophistication? Third, how are variations in epistemological belief across ages, schools, and controversies related to educational practices at Religious and General schools, respectively?

ACKNOWLEDGMENTS

The data for this study were collected in the course of doctoral research at the Hebrew University of Jerusalem. The study was made possible by grants from the Israel Foundation Trustees, the Golda Meir Fund, the Mandel Foundation, the Memorial Foundation for Jewish Culture, and the Professor Eliezer Stern Institute for the Study and Advancement of Religious Education at Bar Ilan University.

Adam Afterman, Eyal Aviv, Tali Berko, and Yasmin Ron assisted in the construction and testing of the coding scheme. I thank Michael Chandler, Allan Collins, Rogers Hall, David Hammer, Paul Harris, Leslie Rupert Herrenkohl, Janet L. Kolodner, David Moshman, Mordecai Nisan, Sam Wineburg, and several anonymous reviewers for their comments on earlier drafts of this article.

Notes

¹In contemporary Jewish Israeli society, one's level of religiosity is a highly salient and well-defined social category. In this context, therefore, participants' evaluations of their parents'religiosity can be as-sumed to be made relative to the Jewish Israeli population as a whole, and not simply to adults the partici-pant knows. The successful use of this measure as a differential in previous studies among Israeli children and adolescents (e.g., those cited previously) and the high correlations reported in Table C1 between the parents'religiosity variable and other measures of religiosity provide further support for this assumption.

^aParents' religiosity items in the questionnaire were omitted by a few participants; accordingly, sample sizes for these analyses are 196 (rather than 200).

^bIn analyses of the effects of grade, 12th grade was the designated reference category; therefore, Grade-1 analyses evaluate group differences between 12th graders and 5th graders whereas Grade-2 analyses evaluate group differences between 12th graders and 8th graders; omnibus block chi-square statistics for entry of the grade variable into the model are provided at the end of the grade-1 row.

*p < .05.

**p < .01.

***p < .001.

^aParents' religiosity items in the questionnaire were omitted by a few participants; accordingly, sample sizes for these analyses are 196 (rather than 200).

^bIn analyses of the effects of grade, 12th grade was the designated reference category; therefore, Grade-1 analyses evaluate group differences between 12th graders and 5th graders whereas Grade-2 analyses evaluate group differences between 12th graders and 8th graders; omnibus block chi-square statistics for entry of the grade variable into the model are provided at the end of the Grade-1 row.

*p < .05.

**p < .01.

²As noted previously, the punishment controversy was chosen for its combination of structural similarity to the God controversy, on one hand, and its nonreligious content, on the other hand. However, the possibility remained that participants' religious backgrounds might nonetheless influence their beliefs about the punishment controversy. For example, participants from more traditional homes might be expected to be more pro-punishment. In fact, preliminary analyses reveal no significant school differences with respect to belief in punishment (see Table 2). Nevertheless, to rule out conclusively the possibility of such influences, parents'religiosity was included as a predictor in the regression analyses for the punishment.

^aAll t-test statistics are significant at p < .001.

^br = Pearson bivariate correlation with parents' religiosity.