Children recognize distinct domains of value. This chapter argues that young children also recognize distinct domains of truth. Both value and truth judgments may be differentiated by age six or seven. However, research suggests that this parallel may not be present during the preschool years.
Charles W. Kalish
One of the main issues in the study of moral development has been children's value judgments. Researchers have been interested in how children evaluate events or propositions as good or bad. The focus has been on justifications: not just whether something is good or bad but why. Thus there is quite a bit of information regarding children's beliefs about the bases or warrants for value judgments. Another strand of developmental research has addressed children's developing conceptions of truth. Recently children's ideas about truth have been studied in the context of their developing theories of mind and of representation (Chandler, 1987, 1988; Montgomery, 1992). One important issue in this line of research is how children understand the justifications for judgments about what is true. Thus there seem to be considerable parallels between the research questions surrounding the development of truth and value judgments. The purpose of this chapter is to consider how research on judgments of value might inform our understanding of the development of judgments of fact.
Both truth and value can be understood as involving judgments or evaluations. An action or intention may be viewed as right or wrong. This kind of evaluation is a value judgment. We judge that someone was wrong to steal; the act is assigned a negative value. Similarly, a statement may be viewed as true or false. This kind of evaluation is a fact judgment. We judge that someone was wrong to claim that the sun revolves around the earth. While there are, perhaps, some differences in the entities that are the subjects of fact and value judgments (e.g., statements vs. actions) it is also possible to focus on their similarities as judgments. In particular, we believe there are warrants for both types of judgments. We can ask what makes a particular statement true, or why a particular action is good. There are some standards for the appropriateness or legitimacy of both fact and value judgments. These standards are that fact judgments should accord with what is true and value judgments should accord with what is good. Thus how people justify or defend fact and value judgments can tell us about their conceptions of truth and goodness, respectively.
Researchers have long been interested in whether there are developmental changes in children's conceptions of the objectivity of truth and value (goodness, Piaget, 1932). Thus, in the terms outlined above the question has been whether children see the warrants for judgments as objective or as subjective (Chandler, 1997, 1988; Kohlberg, 1969; Mansfield & Clinchy, 1993, 1997; Turiel, 1983). Is there something external to us that makes something true or good, or is the legitimacy of judgments of fact and value a human construction?
The primary suggestion of this chapter is that the perspective on justification developed for value judgments within the theory of social domains (Turiel, 1983) can inform thinking about the development of truth and fact judgments. In particular, the theory of social domains suggests that children have multiple, differentiated, views of the ways value judgments may be justified. Development is not a general progression from more subjective to more objective views (cf. Kohlberg, 1969), but rather even young children recognize both relatively objective and relatively subjective bases of value. By analogy, then, we may want to reexamine claims of general developments in children's thinking about truth (Chandler, 1988; Enright, Lapsley, Franklin & Steuck, 1984; Mansfield & Clinchy, 1997). I will discuss empirical evidence suggesting that young children do see different domains of factual judgments.
As adults we recognize that both judgments of truth and goodness may be of different types. In particular, both judgments of fact and value may be more or less objectively warranted. There are differences in modality. In some cases we feel the appropriateness of a judgment is a matter of physical or logical necessity. In other cases we see the judgment as more arbitrary, reflecting the influence of human decision and convention. This point has been well developed for value judgments in the literature on domains of social reasoning.
Turiel and colleagues (Turiel, 1983; Turiel & Davidson, 1986, see Tisak, 1995 for review) have argued that people see some value judgments as unavoidable and universal in place and time. These judgments, characterized as "moral," are thought to have force or warrant akin to logical necessity. People view some judgments, such as the judgment that murder is bad, as objectively correct or appropriate. This sense is demonstrated by intuitions that the value is universal and unalterable. People think that everyone should judge murder to be wrong and it is outside the scope of human action to change that (we can't make murder good). Other value judgments are contingent, dependent on particular decisions and social conditions. Some judgments, such as the judgment that it is bad to eat spaghetti with your fingers, are seen as conventional. This is demonstrated by intuitions that the judgment is local and alterable. Eating with fingers may be (properly) acceptable in some places and we could (legitimately) change our evaluation. Of course whether and when people treat value judgments as objective or as subjective are empirical questions. Note that Turiel also argues that conventional and moral evaluations are understood to have different justifications (for example, social harmony or justice, respectively). However, there is no necessary connection between forms of justification and beliefs about the objective or obligatory nature of evaluations (what Turiel refers to as "criterion judgments"). Thus it is criterion judgments that directly relate to intuitions about objectiveity and are most relevant to this discussion. From research on such judgments there is a considerable body of work that suggests that people, across a range of ages and cultural backgrounds, do seem to recognize both objective and conventional value judgments. What about the parallel case of fact judgments?
The most familiar cases of truth judgments have objective warrants. We are often concerned with judging the veracity of claims about reality such as "The earth is smaller than the sun." Clearly many people would accept that this statement is true or false independent of human decision or action. It is facts about the physical structure of reality that determine the truth or falsity of the statement. Thus we may treat the judgment as universal and unalterable: if it is true, it is true for everybody. However, the case of fact judgments raises a concern not relevant to the value judgments discussed above: Physical structure may change, perhaps even through the intentional intervention of people. Thus fact judgments (e.g., about locations of objects) are not universal and unalterable in the same way as are moral judgments.[1] However, it is still the case that fact judgments may be relatively more or less objectively warranted. In many instances it physical facts (however malleable) that determine truth. In other circumstances, truth does not depend on physical conditions. Examples here are conventional truths such as word-referent connections (e.g., "dog" means dog) and rules of games (e.g., a fly ball hit out of the park is a home-run). The truth of these propositions depend on people's intentions and decisions.
If the accounts presented above are accurate then there are some important parallels between judgments of facts and values. In both cases the warrants for judgments may be understood to be objective: the physical (or logical) structure of the world determines truth and goodness. However, both types of judgments may have a more subjective basis. Sometimes people decide or choose what is good and true. Thus, at least one developmental outcome is a differentiated conception of truth and value. However, this flexibility has often been thought to be hard won. Traditional theories of the development of judgments of fact and value suggest that children may be more limited in their understanding of the bases for judgments.
The development of reasoning about facts and values has often been approached from a broadly Piagetian perspective. In both cases reasoning has been said to proceed from simplistic conceptions to more elaborated and nuanced understandings. Often changes in thinking about truth and value have been thought to be reflections of general changes in the quality and sophistication of children's thinking. These theories generally suggest that children will recognize only a subset of the warrants for judgments recognized by adults. The younger the child, the more restricted the subset. However, stage theories of value judgment development have been challenged. Turiel (1983) has argued that even young children recognize different types of value judgments. It is less clear, in part because there has been less research in this area, whether children also recognize different types of fact judgments.
In Kohlberg's theory (Kohlberg, 1969), children progressed through different stages of thinking about value judgments. According to Kohlberg, children first understand the warrants for all value judgments to be based in personal preference or utility. They then come to recognize social norms and rules as a basis for judgment. Finally, people come to see some evaluations as grounded in universal (moral) principles. Thus there was a general developmental progression away from subjectivism and toward an understanding of more objective bases for value judgments.
In contrast to stage theories, there is now a considerable body of evidence that quite young children recognize different types of value judgments (Levy, Taylor & Gelman, 1995; Smetana, 1981; see Tisak, 1995, for review). They recognize that some evaluations may have a subjective basis while others are objective. In particular, researchers have argued that different types of content evoke different types of judgments. From an early age, people judge that actions involving harm or injustice are subject to objective evaluation. Turiel (1983) identifies judgments about issues of harm or justice as moral and argues that young children share with adults the sense that moral judgments are not contingent. In contrast, young children also treat some value judgments as matters of convention. The proper evaluation of some actions (such as forms of address or politeness) are relative to social norms and decisions. Other evaluations (such as color preferences) may be made entirely at the discretion of individuals (Nucci, 1981). Thus there are two points to Turiel's critique of stage theories: content matters, different kinds of evaluations may be thought to be appropriate for different actions; and there is no general progression, even quite young children recognize objective and subjective bases for value judgments.
Theories of the development of fact judgments have also been cast in terms analogous to stage theories of value judgments. Fact judgments, though, have been thought to show an opposite progression: Children start with an objective conception of truth and come to recognize more subjective influences. Several researchers have described increasing appreciation that internal, mental forces play important roles in judgments of fact (Chandler, 1988; Enright, et al., 1984; Mansfield & Clinchy, 1993). Thus children are characterized as acquiring more "constructivist" conceptions of truth.
It is important to distinguish two phases in thinking about the subjectivity of fact judgments. The first development, in early childhood, is the recognition of the possibility of variability. Young children, before age 4, seem to believe that the appreciation of truth is direct and unproblematic: They do not recognize that people can hold false beliefs (Wellman, 1990). Starting about age 4 children become aware of the possibility of error and gain increasing insight into the role that subjective states may play in leading to fact judgments (Montgomery, 1992; Wellman, 1990). This first phase represents children's increasing understanding that the mind plays an active role in the construction of belief (Carpendale & Chandler, 1996; Chandler, 1988). We can distinguish a second phase which involves appreciating the constructed nature of truth. In addition to recognizing the possibility of different judgments of truth, children come to accept the legitimacy of different judgments.
While five-year-olds know that people may hold different beliefs about the truth, they think that only one belief could really be true. Forguson and Gopnik (1988) argue that young children are "hyper-realists": They believe there is an objective reality against which all representations (judgments) may be tested. Two people may make competing judgments of fact but there will always be some correspondence in the physical world with would (in principle if not in practice) determine which judgment was "really" true. Chandler (1987, 1988) suggests that this level of understanding is characteristic of middle childhood. Not until adolescence are children thought to recognize the possibility that objective reality may provide only an equivocal basis for judgments of truth. In some cases there may be no single fact to the matter. Truth may be relative to a network of prior beliefs and background assumptions. If so, two people may legitimately hold different beliefs. A recognition of the possibility of divergent beliefs seems a necessary but not sufficient condition for this subjective notion of truth.
In making comparisons between the development of fact and value judgments it is the "second phase" or subjective notion of truth that seems most analogous to the developments hypothesized for children's moral reasoning. In both cases the question is how children establish the legitimacy of (conflicting) judgments. The first phase, understanding that people might come to discrepant judgments of truths, seems comparable to the understanding that they might make discrepant judgments of value. A discrepant judgment of value would be someone thinking that stealing is acceptable but someone else thinking it is not. It is the legitimacy of alternative judgments that is at issue. In the empirical literature, however, the distinction between questions about the possibility of alternative views and questions about the legitimacy of alternative views have not always been clearly distinguished (see below and Kalish, 1998 for discussion).
The above discussion reveals a substantial asymmetry between the development of fact and value judgments. While quite young children seem to recognize both objective and more subjective bases of value, a similar appreciation for fact judgments does not appear until adolescence. What could account for this difference? Below I will consider the possibility that stage-like accounts of conceptions of truth present an oversimplified picture. Perhaps, on analogy with Turiel's theory of domains of value judgments, fact judgments are differentiated by content. Children might hold less objective conceptions of truth in some cases than others. Thus the differences between fact and value judgments might be more apparent than real.
It is not always clear whether claims about the development of fact judgments are taken to characterize children's thinking about all knowledge or only knowledge of certain types of phenomena. Chandler (1987) does suggest that children may come to recognize the subjectivity of truth (what he terms "doubt") in some cases before others. He distinguishes between "case-specific" doubt, the recognition that there may be particular things we cannot know objectively, and "generic" doubt, the belief that subjectivity is a fundamental feature of all knowledge. It is this latter, more radical questioning of objective warrant for belief, that Chandler sees as precipitating the changes in adolescents' views of epistemology. However, the presence of case-specific doubts suggests that children may not be generally committed to the objectivity of truth. They may recognize different domains of fact judgments; in some cases truth might be seen as objectively determined, in others it might be seen as more subjective.
Most studies of children's conceptions of knowledge and truth have involved fairly complex situations. Children have been asked about claims mixing fact and value (e.g., whether and animal is a good pet, Mansfield & Clinchy, 1997), claims about ambiguous stimuli (Carpendale & Chandler, 1996), or claims about complicated situations (e.g., the causes of war, Leadbetter & Kuhn, 1989). Procedures typically ask children about conflicting judgments of fact: One person says an animal is a good pet, another says it is not. Older children recognize that claims about these issues cannot be verified on objective criteria. Rather beliefs can be assessed only with respect to some "background assumptions." Thus there is no independent, objective, vantage point from which to judge the truth of a statement. Younger children tend to see a right answer to these disputes. Believing that one person must be wrong is taken as evidence of an objectivist conception of truth. However, these data merely provide negative evidence. Perhaps young children would accept the subjectivity of other claims in other contexts.
Stimuli included in studies of epistemological development are not the only sorts of subjective truths. One might characterize these issues as ambiguously subjective. Children are supposed to realize that there is no single legitimate interpretation of facts. There are some physical conditions relevant to the claim or belief (e.g., physical qualities of the prospective pet), but those condidtions are not sufficient to settle on a single interpretation. Children must recognize both that the objective conditions are ambiguous and that conclusions from ambiguous data involve active mental construction. In this work, the normative interpretation is taken to involve treating claims such as "a juju is a good pet" (Mansfield & Clinchy, 1997) as neither true nor false. Commitments to the truth or falsity of the claim are understood to involve the imposition of prior beliefs or background assumptions on the data. Yet, it may be that the ideas of ambiguity and of the combination of fact and interpretation are difficult ones for children. It often seems that it is the appreciation of ambuguity, rather than mental construction, that is the focus of attention (cf. Carpendale & Chandler, 1996). Yet, understanding that the status of a truth claim may be ambiguous or uncertain seems different than recognizing that a truth claim may have a subjective warrant. Thus young children might accept that a claim could be true because of what people think or construct before they understand that evidence may be ambiguous. Conventional truths may be understood in this way.
There are a body of facts that are straighforwardly subjective. Prime examples are conventional truths. Consider questions such as whether Sunday or Monday is the first day of the week or whether a particular cut of meat is a "brisket" (cf. Burge, 1979). Other than a biblical literalist it is unlikely that anyone, any adult at least, takes the warrant for the belief that a day is the first day of the week to be some aspect of objective reality. To the extent there are disputes over these types of facts they will be disputes about the existence of consensual, inter-subjective, decisions and norms. These disputes will typically have a definite outcome: the truth values of claims about conventions are not ambiguous. Moreover the subjective nature of conventions may be quite explicit. Conventions are determined independent of physical facts, in contrast to the ambiguously subjective facts that partially depend on objective conditions. The subjective bases of conventions may be more obvious than the subjective bases of other sorts of facts.
Before considering whether young children understand that truth may be a matter of convention, it is important to clarify the claim that conventions are subjective. It may be objected that conventions are not subjective because it is possible to provide definite criteria for the truth or falsity of judgments of conventions. Further, those criteria will be external to the individual making the judgment. However, I would like to distinguish the idea of inter-subjective agreement from the idea of objectivity. That we agree on a decision or convention does not make it objective. This does represent a different usage of "subjective." For example, arguments from authority are typically regarded as evidence of an objective epistemology (e.g., Mansfield & Clinchy, 1997). But, in my terms, arbitrary authority is subjective, if we mean not determined by objective facts or the correspondence of beliefs and physical conditions. The key sense I would like to retain for "subjective" is that of "constructed." Conventions are cases of truths constructed through human mental/social activity. The standard direction of the relation between the world and mind in belief is that belief derives from, and must come to reflect, the physical state of the world (Searle, 1983). This is clearly not the case for conventions.
When do children come to recognize conventions as instances of constructed truths? One possibility is that such a recognition is a relatively late development, occurring as children give up their general objectivist views of the nature of truth. Alternatively, on analogy with accounts of distinct domains of value judgments, constructivist conceptions of convention may coexist with objectivist views of other types of facts.
A number of studies have explored young children's understanding of conventional facts. Conventions explored have included naming (e.g., the label for dogs), rules of games (who is "it" in a game of tag), and governmental laws (e.g., traffic regulations). These studies have used methods akin to those used in studies of social domains. For example, children have been asked to judge whether conventions may be changed and to judge the universality of the conventions. Conventional facts are typically contrasted with non-conventional, physical, facts such as the fact that rocks sink in water. Some studies show an objectivist bias. Young children tend to deny that conventions may be changed and expect them to hold universally (Lockhart, Abrahams, & Osherson, 1977). However, other studies suggest that children as young as six or seven treat intellectual conventions as more changeable and less universal than other sorts of facts (Komatsu & Galotti, 1986; Levy, Taylor, & Gelman, 1995; Nicholls & Thorkildsen, 1988). Differences between studies may be due to different procedures and the different types of conventions and physical facts probed (see Komatsu & Galotti, 1986).
While the existing literature provides support for the view that young children understand the conventional nature of some facts, methodological weaknesses limit the conclusions that may be drawn. In particular, the comparison between matters of intellectual convention and matters of (physical) fact is problematic. Questions about physical facts have asked whether it is possible for physical conditions to be different. For example, children have been asked whether there might be a place where rocks float in water (Komatsu & Galotti, 1986). What is at issue here is whether a particular physical event is possible. The assumption is that the truth of a proposition ("Rocks sink") is directly related to physical conditions: If a rock floats,"rocks sink" is false. However, when the same question is asked about a convention, the interpretation is not as clear. Asking about the variability or universality of a convention (e.g., "Could dogs ever be called 'rabs'?", Nicholls & Thorkildsen, 1988) has two senses; could such an event possibly occur? or could a new fact be true? The two senses of the question are the same for physical facts but need not coincide for conventions. For example, children may judge that the behaviors or events involving a convention could be different than they are: Someone could say "rab" when they mean dog. However, it does not follow from this that "rab" really means dog, that it is false that "dog" means dog. The behavior may change but the judgment about what is true remain constant. Thus it is important to distinguish children's intuitions about the possibility of alternative behaviors from their intuitions about alternative evaluations of those behaviors.
Possibility and Evaluation. Because the truth of conventions have some degree of independence from physical facts it is important to know how children are interpreting question of universality and alterability. Thus when children are asked whether a convention might be different in some other place they might answer "yes" meaning the behavior could be different without meaning to assert that the truth of the matter is different. Conventions could be seen as analogous to morals. Yes people may steal, but that doesn't make it right; Yes people might play that the second person tagged is it, but that doesn't make it true. The comparable disjunct for physical facts is not possible; Yes a rock may float, but that doesn't make it true? Some studies have asked children to evaluate the changes or differences in behavior (see Nicholls & Thorkildsen, 1988, see discussion below) but the asymmetry between physical facts and conventions suggests that comparisons between these two types of truths may not be the most apt. Note, however, that the appropriateness of the comparison depends on the purposes. If the goal is to explore children's intuitions about the regularity and consistency of events, rather than judgments of truth, then the methods described above may be more suitable.
The suggestion is that children's understanding of conventional truth may be more clearly examined in contexts where considerations of physical possibility do not arise. One such context is categorization. It has long been argued in the philosophical literature that some ways of categorizing the world are objectively correct while others are more arbitrary and conventional (e.g., Locke, 1707/1961). Objective categories are referred to a "natural" kinds, while more arbitrary, invented categories are "artifactual" kinds.[2] For example, the category of "weekend" is an artifactual kind; what counts as a weekend is a matter of convention. In contrast, "gold" is natural; we take membership in this category to be a matter of objective fact. Note the claim is not that it is impossible to form some categories, merely that it is incorrect. The distinction is not whether or not it is possible to categorize things in different ways (e.g., someone might call pyrite "gold" and might call Friday part of the weekend) but whether it might be legitimate (correct). Thus we can reason about the objectivity of warrants for category membership decisions independent of issues of physical possibility.
To assess whether young children recognize both objective and constructed categories, I have asked them to evaluate the legitimacy of non-normative categorization decisions (Kalish, in press). Four- to five-year-old children were introduced to an informant (a quasi-human puppet) who was said to come from a place far away where they "do a lot of things differently than we do." The informant answered a number of questions in ways that conflicted with children's own judgments. Children were asked to report when the informant was different but "ok" and when he was wrong. Of particular interest were evaluations sorting behavior. The informant was asked which two (of three) pictures were of the same kind. For example, one question asked whether a (picture of a) lion belonged with a cat or a dog.
The most interesting data came from sorting unfamiliar objects into superordinate level kinds. For animals, children judged that there was only one correct way to sort the items and the informant was wrong if his sorting did not conform to the "normative" system of kinds. There was only one correct way to categorize the lion, for example. However, children tended to be flexible in accepting the informant's categorizations of objects. It was "ok" for the puppet to sort however he wanted (and differently than the children would). For example, a bowl could be considered the same kind of thing as a basket or a can, either was acceptable. This was not simply an effect of being more or less familiar with the categories. Children showed considerable inter-subject agreement about how to categorize the pictures. In fact, in a control group the normative choice was clearer for objects than for animals. For example, children did not agree among themselves whether the lion belonged with the cat or the dog, but each child thought there was only one correct answer. Children's judgments were not absolute, categorization decisions for objects were not uniformly considered arbitrary. However, children did demonstrate that they would distinguish between more and less objective evaluations, independent of questions about physical possibility. Given that it was possible to sort all items in discrepant ways, children still judged some of those possibilities to be more or less legitimate.
The results from the categorization study tend to support earlier conclusions that young children understand there may be different kinds of warrants for judgments of fact. In this regard children's judgments of truth parallel their judgments of value. Rather than moving from generalized objectivism to subjectivism (or vice-versa) quite young children seem to make distinctions between more objective and more subjective bases for evaluations.
Distinguishing Fact and Value Judgments. However, before accepting the thesis that the development of fact and value judgments proceed in parallel there is one more option to consider. It may be that children think about fact and value judgments the same way and that is why they perform similarly on tasks assessing both kinds of judgments. Alternatively, it may be that tasks purporting to measure one kind of judgment are actually tapping the other. In particular, it is important to demonstrate that tasks assessing judgments of fact are not actually assessing children's judgments of value. Unfortunately, the studies described above cannot be clearly shown to tap intuitions about facts.
When children deny that a claim or behavior is appropriate they may do so because it is bad/naughty or because it is false. That is, they may see it as an inappropriate value judgment or an inappropriate fact judgment. In the studies of convention described above it is not clear whether responses were based on ideas about value or truth.[3] For example, in the categorization study described above (Kalish, in press), what made alternative sorts unacceptable? One possibility is that the alternatives were seen as erroneous and false. However, children could have made the same responses based on the judgments that the alternatives were bad and naughty. Simply asking children to evaluate the acceptability of some statement or action may not distinguish judgments of fact from judgments of value. Phrased in terms of Turiel's (1983) theory, the problem is that the same criterion judgments (intuitions about universality and alterability) may have different justifications.
One way to distinguish fact from value judgments may be to ask for justifications. This strategy has been pursued in the literature on domains of social judgments. Turiel and his colleagues (Turiel, 1983; see Tisak, 1995) have argued that children (and adults) give different types of justifications for universal (moral) and relative (conventional) social norms. For example, harm and justice are offered as justifications for objective, moral evaluations but social harmony may be the justification for relative judgments. Importantly, both these justification types seem to be based on considerations of goods and values rather than on considerations of truth and falsity. Thus, justification data might serve to identify whether children are making fact or value judgments (e.g., whether something is wrong because it is false or because it is naughty, cf. Kalish, 1998). However, an alternative course is examine the kinds of conditions which might lead children to change their judgments of fact.
The strategy pursued in the studies reviewed above has been to ask children whether discrepant judgments of fact and value might be acceptable and, from those acceptability judgments, infer what conditions they think may affect truth and value. Thus if it is acceptable for people from different social groups to make different judgments we can infer that the warrant for the judgment is seen as relative to social group. An alternative procedure is to see what conditions actually change children's judgments and infer that those conditions make the new judgments acceptable. Thus we can see that moving an item from one location to another induces children to change their judgments of fact (they now judge it is true that the object is in location X rather than location Y). We can conclude that children think a change in physical conditions may legitimate a change in a judgment of fact. Because children changed their judgments we can infer they believe it is appropriate to change their judgments. Similarly, we could show that children change their judgments of value after some group decision (e.g., after a vote they now say it is good to eat with your fingers). This change would suggest that children think it is appropriate to base judgments of value on group decisions. A similar strategy is appropriate in the case of fact judgments. A decision, for example stipulating a new convention, should induce people change their judgments of fact (change their beliefs). Will a decision about a convention change children's beliefs? Such an effect would seem to be a good indicator of a constructivist notion of truth.
Changing children's judgments of fact. Unfortunately, it is difficult to directly involve children in the establishment of conventions. The decisions establishing conventions tend to be fairly remote from children's experience. In the case of labels ("dog" means dog) it is probably only rarely that there is an explicit decision to set the convention. In other cases the decision-making process is restricted to a particular authority. However, some conventions may be established by an individual, even a young child. Thus some decisions about conventions may be quite familiar. Two example of such conventions are proper names and ownership. The name assigned to a pet (e.g., "Fido") or a doll is something an individual may decide. Similarly, we are quite familiar with individuals affecting changes in ownership (e.g., giving something to someone else). Young children participate in the establishment of these kinds of conventions. Thus the chosen, constructed, nature should be apparent. What remains to be seen is whether children recognize these conventions as establishing true facts (the dog's name is "Fido") at the same time they recognize the intentional source of the conventions.
In a series of studies Michelle Weissman, Debra Bernstein and I (Kalish, Weissman, & Bernstein, 1998) have explored children's understanding that decisions may establish a convention as true. In one study children were engaged by an experimenter in making a series of decisions about a doll. Decisions involved conventional attributes such as the name and ownership of the doll. In each case the experimenter first established an initial value of the attribute (e.g., This doll's name is "Sally." This doll belongs to me.). The experimenter then decided to change the attribute (e.g., rename the doll "Julie", give the doll to the child). Children were asked to make three judgments following each change. First they were asked a behavioral question, for example, "What should we say if someone asks us what the doll's name is?" A correct answer would be to say that the doll's new name should be reported. Two other questions assessed children's appreciation that the new attribute was now true of the doll. One question asked what was "really" the case. An example of a really question was: "Now that we've decided, what's the doll's name really?" A final question asked the child to judge whether a third party's assertion about the doll was correct or incorrect. An ignorant puppet (who didn't "know what we decided") either asserted the new or the old attribute (counter-balanced across participants). Children were asked whether the puppet's claims were "right" or "wrong."
In the context of the current discussion the most important result was a sharp difference in performance between five-year olds and six- to seven-year olds. Younger children answered the behavioral questions correctly (about what they should say) but had great difficulty with the other two questions probing judgments of truth. They were at chance in their assertions of what was "really" the case. Assessments of the puppet's assertions of the initial and changed values did not differ (younger children tended to judge the puppet wrong in both cases). In contrast, slightly older children performed near ceiling on all three questions. In particular they almost always asserted that the new attribute really was the case. Moreover they correctly judged that the puppet was right to assert the new value and wrong to assert the original. Older children, but not younger, seemed to appreciate that decisions could change what was really the case.
Research on children's understanding of conventions, though not always consistent, does allow some tentative conclusions. The first suggestion is that by the early elementary school years children seem to realize that some facts may be matters of convention. They recognize that conventional behaviors are relative: In different social circumstances different behaviors may occur. Further, it is not just the behaviors that are seen as relative but also the conventions themselves (the rules determining the propriety of behavior; Kalish, in press; Nicholls & Thorkildsen, 1988). Finally there is at least some initial indication that by six or seven children accept that decisions about conventions may change what is true (Kalish, et al., 1998). These results seem to indicate that the conceptions of fact and truth do parallel conceptions of value. As has been demonstrated in the literature on domains of social reasoning, young children neither treat all judgments as objective nor all as subjective. Of course, children's judgments of objectivity and subjectivity need not always coincide with adults' but this research suggests that at least some aspects of the basic distinctions are shared. Thus accounts of the development of children's conceptions of knowledge may need to be modified to reflect the possibility of a relatively early recognition of domains of epistemic judgments.
Although both fact and value judgments may involve domains, there may be some asymmetry in the development of the two kinds of judgments. In the study described above (Kalish, et al, 1998) six- to seven-year-olds seemed to understand that decisions may change truths, however preschoolers did not. Similarly, Flavell and colleagues (Flavell, Mumme, Green, & Flavell, 1992) have found that young children seem not to distinguish between more and less objective beliefs. For example, three-year-old children have just as much difficulty reporting that two characters might hold different value beliefs (e.g., about attractiveness) as reporting they might hold different factual belief (e.g., about locations of objects). This contrasts with the findings from the literature on social domains where even preschool-aged children make a distinction between universal and relative value judgments (Levy, Taylor & Gelman, 1995; Smetana, 1981). Young children may understand conventional value judgments but not conventional fact judgments. Moreover, Flavell's results suggest that young children may lose their differentiated sense of value when value judgments are presented as beliefs. Thus it may be appropriate to characterize preschoolers as generally objectivists with respect to epistemology and truth.
Of course, more research is needed to establish this asymmetry. However, there do seem to be good theoretical grounds for expecting that young children would not accept subjective warrants for fact judgments. One of the central findings from research on children's developing theories of mind is that desire is understood before belief (Wellman, 1990). Moses (1993) suggests a general pattern in which motivational states are understood before epistemic states. Children find it easier to reason about liking or attraction than to think about knowlege or belief. In particular, preschool-aged children are just coming to recognize the possibility that people may hold different views of the truth (Chandler, 1988, Wellman, 1990). Since recognizing the possibility of different judgments is a prerequisite to recognizing the legitimacy of different judgments, this delay constrains children's appreciation of the interpretive or constructed quality of truth.
In contrast, children's early understanding of desire may be related to their understanding of value judgments. Under at least some characterizations, desires are ascriptions of value. Desires involve judgments of what is good in the same way that beliefs involve judgments of what is truth (see de Sousa, 1974). Children very early recognize the representational diversity of desire. For example, 18-month-olds realize that someone else may like something that they themselves do not (Repacholi & Gopnik, 1997). An appreciation of representational diversity of desire may facilitate, or at least allow, the development of the distinction between objective and subjective value judgments observed during the preschool years. At the same time that children are struggling to understand that perception of truth is not direct and unmediated they already seem to know that assignment of value may be a social construction (Smetana, 1981).
The literature on children's developing theories of mind would seem to accord with the research findings discussed above. In particular, preschool-aged children recognize that some value judgments have objective warrants but others are based on social constructions and convention. In contrast, it may not be until the early elementary school years that children appreciate conventional as well as objective truths. Although the empirical evidence regarding preschooler's understanding of conventions (factual conventions) is quite limited, the above conclusions seem consistent with findings on children's developing understanding of belief and desire. To the extent that preschool-aged children understand conventions they seem to see them in motivational terms. People can choose what should and shouldn't be done. However, it is not until a few years later that children also appreciate that people can choose what is and is not true. An epistemic understanding of conventions may lag behind a motivational understanding.
The goal of this chapter has been to consider what light may be shed on questions of the development of conceptions of truth by taking as a perspective the development of value judgments. The development of value judgments has been conceptualized as involving several distinct domains of judgments. Children (and adults) do not view all value judgments in the same way but rather recognize different bases for different types of valuation. Taking this as a point of comparison it seemed reasonable to ask whether judgments of fact or truth might also show domain differences. Although studies are somewhat limited, the data do suggest that, by age 6 or 7, children recognize conventions as subjective or constructed truths yet see objective bases for other factual beliefs (Chandler, 1988). Thus not all truths are equally created. The fruitfulness of this comparison highlights the importance of continued explorations of the connections between research in children's developing notions of mental representation and studies of social domain judgments.
[1] This is because moral judgments are taken to be necessary (Turiel , 1983). The proper kind of truth judgment to use for comparison would be judgments of analytic truths such as "1+1=2." These truths hold in all possible worlds and so are not affected by changes in physical conditions: truly universal and unalterable. However, young children probably do not distinguish between physical and logical truths (cf. Komatsu & Galotti, 1988)
[2]See Keil, 1989 and Markman, 1990 for discussion of the development of children's natural kind and artifact concepts.
[3]Of course for adults the two kinds of judgments are not independent. In particular, it is bad, naughty, for a teacher to insist that students adopt a practice which is false.
Burge, T. "Individualism and the mental." Midwest Studies in Philosophy, 1979, 4, 73-121.
Carpendale, J. I., & Chandler, M. J. "On the distinction between false belief understanding and subscribing to an interpretive theory of mind." Child Development, 1996, 67, 1686-1706.
Chandler, M. "The Othello effect: Essay on the emergence and eclipse of skeptical doubt." Human Development, 1987, 30, 137-159.
Chandler, M. "Doubt and developing theories of mind." In J. W. Astington, P. L. Harris, & D. R. Olson (eds.), Developing theories of mind. New York: Cambridge University Press, 1988.
de Sousa, R. B. "The good and the true." Mind, 1974, 83, 534-551.
Enright, R. D., Lapsley, D. K., Franklin, C. C., and Steuck, K. "Longitudinal and cross-cultural validation of the belief-discrepancy construct." Developmental Psychology, 1984, 20, 143-149.
Flavell, J. H., Mumme, D. L., Green, F. L., and Flavell, E. R. "Young children's understanding of different types of beliefs." Child Development, 63, 1992, 960-977.
Forguson, L., and Gopnik, A. "The ontogeny of common sense." In J. W. Astington, P. L. Harris, & D. R. Olson (eds.), Developing theories of mind. New York: Cambridge University Press, 1988.
Kalish, C. W. Reasons and causes: Children's understanding of conformity to social rules and physical laws. Child Development, 1998, 69,
Kalish, C. W. "Natural and artificial kinds: Are children realists or relativists about categories? Developmental Psychology, (in press).
Kalish, C. W., Weissman, M. D., & Bernstein, D. "Children's understanding of conventional truths." Unpublished manuscript. Department of Educational Psychology, University of Wisconsin-Madison, 1998.
Keil, F. C. Concepts, kinds, and cognitive development. Cambridge, MA: MIT Press, 1989.
Kohlberg, L."Stage and sequence: The cognitive-developmental approach to socialization." In D. Goslin (ed.), Handbook of socialization theory and research. Chicago: Rand McNally, 1969.
Komatsu, L. K., & Galotti, K. M."Children's reasoning about social, physical and logical regularities: A look at two worlds." Child Development, 1988, 57, 413-420.
Leadbetter, B., & Kuhn, D."Interpreting discrepant narratives: Hermeneutics and adult cognition." In J. D. Sinnot (Ed.), Everyday problem solving: Theory and applications. New York: Praeger, 1989.
Levy, G. D., Taylor, M. G., & Gelman, S. A."Traditional and evaluative aspects of flexibility in gender roles, social conventions, moral rules, and physical laws." Child Development, 1995, 66, 515-531.
Locke, J. An essay concerning human understanding, 5th edition, Volume II. New York: E.P. Dutton. (Original work published 1707).
Lockhart, K. L., Abrahams, B., & Osherson, D. N. "Children's understanding of uniformity in the environment." Child Development, 1977, 48, 1521-1531.
Mansfield, A. F., & Clinchy, B, McV. "Continuity and change in children's epistemologies from ages 4 to 10." Paper presented at the Society for Research in Child Development, New Orleans, LA. March, 1993
Mansfield, A. F., & Clinchy, B, McV. "Toward the integration of objectivity and subjectivity: A longitudinal study of the epistemological development between the ages of 9 and 13." Paper presented at the Society for Research in Child Development, Washington, DC. April, 1997.
Markman, E. M. Categorization and naming in children. Cambridge, MA: MIT Press, 1990.
Montgomery, D. E. "Young children's theory of knowing: The development of a folk epistemology." Developmental Review, 1992, 12, 410-430.
Moses, L. J. "Young children's understanding of belief constraint on intention." Cogntive Development, 1993, 8, 1-25.
Nicholls, J. G., & Thorkildsen, T. A. "Children's distinctions among matters of intellectual convention, logic, fact, and personal preference." Child Development, 1988, 59, 939-949.
Nucci, L. "Conceptions of personal issues: A domain distinct from moral or societal concepts." Child Development, 1981, 52, 114-121.
Piaget, J. The moral development of the child. New York: Harcourt Brace, 1932.
Repacholi, B. M., & Gopnik, A."Early reasoning about desires: Evidence from 14- and 18-month-olds." Developmental Psychology, 1997, 33, 12-21.
Searle, J. R. Intentionality. Cambridge: Cambridge University Press, 1983.
Smetana, J. G. "Preschool children's conceptions of moral and social rules." Child Development, 1981, 52, 1333-1336.
Tisak, M. S. "Domains of social reasoning and beyond." Annals of Child Development, 1995, 11, 95-130.
Turiel, E. The development of social knowledge: Morality and convention. Cambridge: Cambridge University Press, 1983.
Turiel, E., & Davidson, P. "Heterogeneity, Inconsistency, and asynchrony in the development of cognitive structures." In I. Levin (Ed.), Stage and structure: Reopening the debate. Hillsdale, NJ: Ablex, 1986.
Wellman, H. M. The child's theory of mind. Cambridge, MA: MIT Press, 1990.
CHARLES W. KALISH is an assistant professor of educational psychology at the University of Wisconsin-Madison.