Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

On the distinction between false belief understanding and the acquisition of an interpretive theory of… Carpendale, Jeremy Ian Maxwell 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1995-059294.pdf [ 2.24MB ]
Metadata
JSON: 831-1.0088295.json
JSON-LD: 831-1.0088295-ld.json
RDF/XML (Pretty): 831-1.0088295-rdf.xml
RDF/JSON: 831-1.0088295-rdf.json
Turtle: 831-1.0088295-turtle.txt
N-Triples: 831-1.0088295-rdf-ntriples.txt
Original Record: 831-1.0088295-source.json
Full Text
831-1.0088295-fulltext.txt
Citation
831-1.0088295.ris

Full Text

ON THE DISTINCTION BETWEEN FALSE BELIEF UNDERSTANDINGAND THEACQUISITION OF AN INTERPRETIVETHEORY OFMINDbyJEREMY IANMAXWELLCARPENDALEB.A. Honours, Simon Fraser University, 1988M.A., Simon Fraser University, 1990A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOROF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDepartment of PsychologyWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJULY 1995© Jeremy Ian Maxwell CarpendaleIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)_______________________________Department of PS7hiO)/The University of British ColumbiaVancouver, CanadaDate II L,J(AJY 19?5DE-6 (2/88)11ABSTRACTTwo groups of 5- to 8-year-olds, and a comparison sample of adults, were examined in an effortto explore the developing relationships between false belief understanding and an awareness of theindividualized nature of personal taste, on the one hand, and, on the other, a maturing grasp of theinterpretive character of the knowing process. In Study 1,20 children between 5 and 8, and inStudy Two, a group of 15 adults, all behaved in accordance with hypotheses by proving to beindistinguishable in their good grasp of the possibility of false beliefs, and in their commonassumption that differences of opinion concerning matters of taste are legitimate expressions ofpersonal preferences. By contrast, only the 7- and 8-year-old children and adults gave evidenceof recognizing that ambiguous stimuli allow for warrantable differences of interpretation. Study 3replicated and extended these findings with a group of 48 5- to 8-year-old subjects, again showingthat while 5-year-olds easily pass a standard test of false belief understanding, only children of 7or 8 ordinarily evidence an appreciation of the interpretative character of the knowing process.111TABLE OF CONTENTSABSTRACT iiTABLE OF CONTENTS iiiLIST OF FIGURES vACKNOWLEDGMENTS viPREFACE viiINTRODUCTION 1An Approach to the Problem 2Evidence for an Early Understanding of Interpretation 7Evidence for a Slow Track Toward an Interpretive Theory of Mind 11Assessing an Interpretive Theory of Mind 15STUDY 1 19Method 20False-belief test 21Problems of Interpretation 22Lexical Ambiguity 22Ambiguous referential communication 22Ambiguous Figures 22Matters of taste 22Order of the Tasks 23Procedure 23Explanation Questions 23Prediction Questions 23Deviant interpretations 24Scoring the False Belief test 24Scoring the Problems of Interpretation 24Explanation questions 24Prediction questions 25Deviant interpretation questions 25Scoring the Matters of Taste 26Explanation questions 26Prediction question 26Reliability 26Results 27Discussion 29STUDY 2 32Method 32Subjects 32Procedure 32Results 33Discussion 34ivSTUDY3 .36Method 39Subjects 39Procedure 40Reliability 40Results and Discussion 40CONCLUSIONS 47Limitations and Suggestions for Future Research 49REFERENCES 54FIGURES 63APPENDICES 69APPENDIX AConceptual and Theoretical Background, and Empirical Evidence for theStudy of Children’s Understanding of Interpretation 70Interpretation and Representation 70The Concept of Interpretation 73“The Interpretive Turn”Interpretation in Modem Philosophy 74Deviant interpretations 78Theories of Mind and Social Perspective-Taking 80Evidence Regarding Children’s Understanding of Interpretation 85Interpretation and Visual Perspective-Taking 86Interpretation and Visual Ambiguity 87Epistemological Development in Adolescence 96APPENDIX BProtocol for Study 1 98APPENDIX CProtocol for Study 2 107APPENDIX DProtocol for Study 3 109VLIST OF FIGURESFIGURE 1.The “Duck-Rabbit” 64FIGURE 2.The “Rat-Man” 65FIGURE 3.Study 1. Proportion Correct Responses to the tests of Taste andInterpretations by Age-Group 66FIGURE 4. Study 3.Proportion Correct Responses by Age and Order of Presentation 67FIGURE 5. Study 3.Proportion Correct Responses by Age and Question Type 68viACKNOWLEDGMENTSI would like to take this opportunity to thank all those who have directly contributed to thesuccessful completion of this dissertation as well as those who have indirectly helped andsupported me throughout this project. First I wish to thank Michael Chandler, my supervisor, forhis guidance, contributions, and support in all aspects of the completion of this dissertation andthroughout my graduate career. His breadth of intellectual interests has contributed in importantways to my education. I also wish to thank the other members of my dissertation committee,Janet Werker and Charlotte Johnston, for their helpful suggestions at important points in thedevelopment of this project. Special thanks are due to Janet Werker for her support and for hercritical questions throughout my time at UBC. In addition, I would like to thank my colleagues,Chris Lalonde and Suzanne Hala, for their contributions and support. Thanks are due to CarmenSwanson for help with data collection for Study 1. The faculty and graduate students in the UBCdevelopmental group are also due thanks for their support and comments at various points in thisproject. In addition, thanks are due to the children, staff, and parents of the following day-caresand after-school-cares: Little Raven Day-Care, Pacific Spirit Day-Care, Jules Quesnel YMCAAfter-School-Care, the Kid’s Club, Kid’s Club Junior, and Queen Elizabeth After-School-Care. Iwould also like to express my respect and appreciation for Michael Chapman. Although he didnot contribute directly to this project, his ideas influence my thinking in many ways, and I expectwill continue to influence my work throughout my career. Finally, I wish to thank Caroline forher tolerance and support throughout this project, and I wish to express special appreciation forour children, Hannah and Max, the source of many of my ideas and the reason that everything isworthwhile.Financial support for this dissertation was provided by a doctoral fellowship from theSocial Sciences and Humanities Research Council of Canada, a University of British ColumbiaGraduate Fellowship, and in part by a research grant held by Michael Chandler from the NaturalSciences and Engineering Research Council of Canada.viiPREFACEA few comments are required to explain the organization of this thesis. The main line ofargument and presentation of evidence is written in a form that would be suitable for archivaltreatment in a journal. A thesis, however, should cover a topic in both more depth and morebreadth than is permitted by space limitations in ajournal. To accomplish this, I havesupplemented the main body of the text with additional sections, footnotes and appendices. Themain text can stand by itself, but at many points more detail is provided in footnotes, and the firstappendix goes into historical and conceptual issues in a broader manner than would be allowed ina standard journal article.1INTRODUCTIONStudies of children’s so-called developing “theories of mind” have so far concentratedattention primarily on 3-, 4- and 5-year-olds, focusing almost exclusively on their ability tounderstand the possibility of false beliefs.’ One justification for this restricted focus has been thecommon, but I believe mistaken, assumption that in evidencing a grasp of false beliefs suchpreschoolers also automatically demonstrate at least a fledgling understanding of the constructiveor “interpretive” character of mental life (e.g., Perner, 1991). All such legitimating claims aremistaken, I want to argue, for the reason that standard measures of false belief understandingprovide no evidence that is relevant to deciding anything whatsoever about the actual “interpretive”character of knowledge. Rather, as I will seek to show, such “standard” measures of false beliefunderstanding are only about the recognition that people who are more or less poorly informed areentitled to their own more or less ignorant beliefs. That is, tests of false belief understandingsimply fail, on close inspection, to concern themselves one way or another with actual matters ofinterpretation, which are ordinarily understood to turn upon the emerging realization that personspossessing precisely the same information still often come to sharply different views of, or beliefsabout, one and the same thing.If standard tests of false belief do not directly address questions about the interpretivecharacter of knowledge, then all real progress toward some empirical test of this matter must await1 Recently, there has been a great deal of interest within developmental psychology in how young children come tounderstand the talk and action of other people. An especially popular explanation for this developing ability issignaled in the choice of children’s ‘theories of mind” as the name for this new area (e.g., Astington, Hams &Olson, 1988; Butterworth, Hams, Leslie & Weilman, 1991; Frye & Moore, 1991; Whiten, 1991). According tothis “theory theory” (Morton, 1980) it is thought that children’s ability to understand others is the result of theiracquiring a theory-like understanding of their own and other minds--an understanding that they go on to employ inmaking predictions of human actions and constructing explanations of such actions. A critical moment in theemergence of this so-called “theory theory” occurred in 1983 whenWimmer and Pemer published details of aprocedure meant to assess an important component of children’s theories ofmind: the understanding that people mayhold to and act upon beliefs that are false. As these authors have convincingly reminded us, the concern with falsebelief understanding is important because an adequate understanding of beliefmust include an appreciation thatbeliefs can be mistaken. Wimmer and Perner also persuasively argued that demonstrated grasp of the possibility offalse beliefs is important methodologically for the reason that if children were only asked about their own and otherstrue beliefs then it would not be possible to distinguish between what reality is and is taken to be.2the introduction of assessment procedures that do in fact inquire directly into actual matters ofinterpretation. Acquiring real knowledge about the respective ages of onset of these conceptuallyand perhaps developmentally distinct abilities necessarily requires not only results from somestandard measure of false belief understanding, but also from some procedurally distinct methodof assessing children’s appreciation of the fact that one and the same stimulus event can supporttwo or more distinct interpretations. The research to be reported here introduced such measures,and provides a direct test of my own orienting hypothesis that children regularly achieve anunderstanding of false beliefs several years before they eventually go on to develop a grasp of themore demanding notion of interpretation.An Approach to the ProblemA subsequent step two in this effort to re-examine and perhaps rewrite the received historyof children’s emerging understanding of the interpretative nature of the knowing process has beento offer what has proved to be a serious challenge to the usual assumption that the so-called“standard” tests of false belief understanding actually provide unimpeachable evidence for agenuinely interpretive theory of mind. A prior step one, however, requires starting somewhatfurther back by carefully documenting the claim that, in the views of many key contributors to thisliterature, passing a standard false belief test is ordinarily thought to provide a sufficientdemonstration that mental life is already regarded as interpretive in character. Weliman (1990),for example, does explicitly state that at about the age of four to five years (when they ordinarilybegin to pass all standard appearance-reality and false belief measures) children also can be said tohave acquired “an interpretational or constructive understanding of representations” (p. 244).Similarly, Flavell, at least in some of his earlier writings, (e.g., 1988, p. 247) has suggested thatsuccess on a range of appearance-reality, as well as false belief tasks depends on the developmentof what he has termed Level 2 understanding--an ability that he claims ordinarily allows childrenof roughly 4 to already appreciate that “the same object can be represented in different, seemingly3contradictory ways.” Interestingly, in some of their more recent writings Ravel! and hiscolleagues (e.g., Ravel!, 1995) appear to have reread their own earlier evidence as actually bettersupporting the view that 4-year-olds have more of a “bottom up” than a “top down” or interpretiveview of mental life. As he and his colleagues now reason (Ravell, Green & Ravell, 1993, 1995;Ravell, Miller, & Miller, 1993), the insight that a representation can be false with respect to a realobject or event, still leaves room for the acquisition of the more complex concept that one and thesame object or event can sometimes afford multiple meanings that are equally legitimate.Perner (1991) most explicitly claims that success on usual tests of false beliefunderstanding necessarily implies a grasp of the interpretative nature of the knowing process. Thepresumed equivalence between false belief understanding and an interpretive theory of mind isjustified, in Perner’s view, because a necessary and sufficient explanation for both is thought tolie in their supposed common dependence upon an emerging “representational” view of mentallife. As he puts it, “around 4 years children begin to understand knowledge as representation,with all its essential characteristics. One such characteristic is interpretation” (Perner, 1991, p.275, italics in original). Ruffman, Olson and Astington (1991) come to a similar conclusion,explicitly dismissing the possibility that there might be a second, more interpretative stage inchildren’s understanding of mind. Instead, they read the results from their own research on visualambiguity as suggesting the “possibility that there is only one stage and that ambiguity and falsebelief tasks tap the same underlying development” (Ruffman, Olson & Astington, 1991, p. 101).Although all the theorists referenced above qualify their strong claims in various places byacknowledging that young preschool children actually may have no more than a “nascentinterpretive theory of mind” (Perner & Davies, 1991), whatever differences they do judge todivide the 4-year-olds and the adult’s understanding of mind continue to be taken as primarilyquantitative and skill driven, rather than qualitative. Wellman and Ridding (1994), for example,have recently suggested that it is not until after their preschool years that children older than 5 or 6first acquire a “conception of the mind itself as an independent, active entity” (Weilman &Hiclding, 1994, p. 1565). At the same time, however, these authors maintain that even4preschoolers see mental contents as being constructed “actively by the person, on the basis ofinference and subject to biases, misrepresentations, and active interpretation” (p. 1578).In contrast to all such “one miracle” views of epistemic development, I and my colleagues(e.g., Carpendale, 1995; Chandler, 1988; Chandler, Carpendale & Lalonde, 1995) haveconsistently promoted a different reading of the evidence, according to which the development ofchildren’s understanding of mind is not thought to begin and end with a grasp of the possibility offalse beliefs. In the place of all such either/or accounts is substituted our own better differentiatedclaim that, while the simple appreciation that ignorance necessarily promotes misunderstanding is,as advertised, an accomplishment of even the very early preschool years, the altogether morecomplicated task of appreciating the essentially interpretative nature of the knowing process doesnot occur until substantially later, and is probably best understood as only beginning to get underway by 6 or 7 or 8. That is, it has proven customary within the theory of mind literature toassume that passing a false belief test also counts as an early demonstration that mental life isinherently interpretive in character. I, however, mean to argue the opposite case, beginning with aclear demonstration that succeeding on standard false belief measures actually requires no morethan the altogether simpler understanding that persons with different experiential historiesregularly end up with different beliefs.The first part of the argument relies upon the fact that, on close inspection, false belieftests reveal themselves to be all about matters of relative ignorance, and not at all about matters ofinterpretation. That is, there is nothing about standard false belief tests that could possiblydemonstrate an understanding that two persons could differently interpret one and the same thing,because the events portrayed in such tasks are explicitly orchestrated to ensure that the subject andthe story protagonist, whose beliefs are inquired into, are actually exposed to differentinformation. Using Wimmer and Perner’s (1983) now classic story of Maxi and the missingchocolate as a case in point, it is evident on inspection that the story is carefully and self-consciously stage-managed in such a way as to ensure that Maxi is out of the room, and so isignorant of the fact that the location of his candy has been changed in his absence. This lack of5crucial information automatically results in Maxi’s holding to a now out-dated, and, consequentlyfalse, belief that is different from that subscribed to by anyone whose knowledge is somehowmore up to date. Although following this story and successfully predicting Maxi’s future actionswould clearly require an understanding of the possibility that beliefs can be false, nothing aboutgetting this story straight demands any understanding whatsoever of the everyday adult truismthat, even after experiencing one and the same event, people may, and regularly do, come awaywith differing interpretations (Chandler, 1992).22 This research is, in the broadest sense, an investigation of the development of children’s understanding of theconcept of interpretation. A good place to begin looking in the investigation of any concept is the dictionary. Thisis particularly the case when, as in the present research, the investigation concerns children’s development of whatamounts to the common sense understanding of a concept. According to The Oxford English Dictionary, themeaning of interpret is: “1. To expound the meaning of (something abstruse or mysterious); to render (words,writings, an author, etc.) clear or explicit: to elucidate; to explain.... 2. To give a particular explanation of; toexpound or take in a specific manner. Also, to construe (motives, actions, etc.) favorably or adversely” (p. 1131).On close inspection, there can be seen in these definitions two interrelated meanings both of which continue toresurface in different ways throughout the extensive debates over the meaning of interpretation. First, interpretationcan and has been taken to involve clarifying something that is otherwise obscure. The second and related meaningis that this process of clarification or explication always involves a construal. That is, understanding alwaysinvolves taking things in a particular way. This second use of interpretation clearly implies that, at least in somecases, there is room for more than one legitimate interpretation for the same object or message. Any suchunderstanding of mind as interpretive in this second sense should be seen as different from the insight, assessed byfalse-belief tests, that people who are differently informed are entitled to different conclusions.The following example may help to illustrate the difference between an understanding of false-belief andholding to an interpretive theory of mind. Imagine that two people watch the same movie but one of them goes tobuy popcorn at a critical moment in the film. A child with no more than an understanding of false-beliefs will beable to appreciate that these two people may well reach different conclusions regarding the movie because one ofthem lacks some key information due to being out of the room at a critical moment. Any such understanding offalse-belief would not, in and of itself, be adequate, however, to account for the fact that even if the two people bothwatch the entire movie they might still arrive at differing conclusions. To explain a case in which two peopledifferently interpret the same object or event requires something like what is referred to here as an “interpretive”theory of mind.A further common thread between interpretation and false belief understanding is that, in different ways,both can be used to account for what would otherwise be anomalous events, and thus, shore up faults that mightthreaten children’s theories of the world and the mind. For example, the idea of false-belief can be used torationalize seemingly anomalous behavior by suggesting that two people may react differently because one of themhas an out-thted belief that has been rendered false by lack of information. For example, a young child who had notyet achieved an understanding of false-beliefs would be at a loss in attempting to explain why, when returning fromplaying outside, Wimmer and Pemer’s Maxi character would look in cabinet A for his chocolate, when the childknows that Maxi’s mother had moved the chocolate to cabinet B. However, a child who had achieved theunderstanding that people may act on beliefs that are false would have no difficulty in explaining Maxi’s otherwiseapparently irrational action. In other words, we can account for some one’s action that is not consistent with ourown expectations by invoking a false-belief. Thus, the notion of false-belief can be employed in situations inwhich two people differ in the amount of information they possess because differential access to information can beassumed. When, by contrast, two people experience the same event yet come away with differing conclusions, theconcept of false-belief is no longer adequate for constructing an appropriate explanation, and the concept ofinterpretation must be invoked. In other words, the notion of interpretation could serve a function similar to thatascribed earlier to the developing notion of false belief and do so with the added advantage that interpretation can beinvoked in cases were it is not possible to claim that the two parties to a disputed account of reality actually hadaccess to different information. The capacity of interpretation to play such a role varies with the ambiguity of the6In contrast to mere false belief understanding, a more mature “interpretiv&’ theory of mindmust also be understood to include, I argue, the insight that the “mind11 influences how the“world” is experienced. This image of a two-way street connecting the mind and world, whichPiaget meant to convey with his metaphors of assimilation and accommodation, is perhaps bettercaptured here using Searle’s (1983) more contemporary talk of “direction of fit” .3 Under thisdescription, young preschool children are best characterized as beginning their epistemic careerswith an exclusive commitment to the unidirectional view that minds do all the changing bybeginning to “fit” or conform more and more perfectly to a world over which they have nointerpretive control. Although partially correct, this half truth leaves no room for lines ofinfluence running in the opposite direction--what Searle dubbed a “world to mind direction of fit.”This shortcoming limits young preschool children to what has previously been termed an“assimilation side” or “copy theory” of knowledge (Chandler & Boyes, 1982; Wellman, 1990).According to this view, objects are assumed to “transmit, in a direct-line-of-sight fashion, faintcopies of themselves, which actively assault and impress themselves upon anyone who happensin the path of such ‘objective’ knowledge” (Chandler & Boyes, 1982, p. 391). Consistent withsuch a “primitive copy theory,” knowledge is first seen “to reside in objective events whichtelegraph this information to any observer who gets in harm’s way” (p. 393), thereby ruling outof court the very possibility that knowledge acquisition necessarily involves a process of activestimulus. The less ambiguous the event, the more difficult it may be to apply the concept of interpretation, and, insome cases, we may have to resort to explanations which invoke some notion of psychopathology.3 Searle originally introduced the notion of “direction of fit” in his theory of speech acts (1969). By this accountso-called “assertive speech acts” (e.g., statements) are said to “in some way to match an independently existingworld” (Searle, 1983, p. 7), whereas, by contrast, “directive speech acts” (e.g., requests) are seen “to bring aboutchanges in the world so that the world matches the propositional content of the speech act” (Searle, 1983, p. 7).Searle (1983) went on to use this notion of “direction of fit” more generally as a way of characterizing the differentrelationship between the mind and the world for beliefs as compared to desires. Beliefs, according to Searle, can betrue or false, and thus they could be said to have a “mind-to-world” direction of fit (i.e., the mind should fit theworld). “Desires and intentions, on the other hand, cannot be true or false, but can be complied with, fulfilled, orcarried out, and we might say that they have the ‘world-to-mind’ direction of fit” (Searle, 1983, p. 8).Searle (1983) acknowledges that fitting is a symmetrical relationship, and, therefore, he clarifies his use ofthe notion of “direction of fit” by stating that of the two things meant to fit together, one or the other can be takenas given. For example, when Cinderella goes shoe shopping she takes her foot size as given and seeks a shoe to fither foot (“shoe-to-foot direction of fit”). “But when the prince seeks the owner of the shoe, he takes the shoe asgiven and seeks a foot to fit the shoe (foot-to-shoe direction of fit)” (Searle, 1983, p. 8).7interpretation. In the present view, something like such an initial “copy theory” of knowledgeboth characterizes the tacit epistemologies of young persons throughout their preschool years, andadequately accounts for any and all of the evidence so far brought out about preschoolersemerging beliefs about mental life.On this more fully “developmental” view of the stepwise process involved in graduallycoming to an adult-like theory of mind, it typically is not until their early school years that childrenare thought to initiate anything like a clear “shift from an object-centered or copy theory ofknowledge to a subject-oriented or constructivistic epistemology” (Chandler & Boyes, 1982, p.393). That is, it is only at this relatively late juncture, and some years after first grasping thepossibility of false beliefs, that children are hypothesized to begin gradually consolidating acapacity to appreciate, not only that different persons may have access to different information,but also that different persons can and often do attach different meanings to one and the samething.Where all this leaves matters is that, in contrast to the view that false belief understandingis equivalent to an interpretive view of mental life, I mean to evaluate the hypothesis that theacquisition of a constructive or interpretive theory of mind is a different and later occurringachievement. The research presented here aims to help arbitrate this controversy by introducingwhat are argued to be better reasoned measures of interpretive knowledge, and by showing thatpassing false belief measures is a necessary, but not sufficient condition for success on such testsof an interpretive theory of mind. Before bringing out this new line of evidence, however, it willprove useful to first review a short list of other claims, and other data, that together informed themethods and procedures of the present research.Evidencefor an Early Understanding ofInterpretationQuite apart from the question of whether false belief understanding is or is not the samething as a genuinely interpretive theory of mind, it still remains the case that there are other lines of8evidence commonly regarded as speaking to the question of whether preschoolers hold to somesuch interpretive view of mental life. Among the lines of evidence most often read as endorsingthe view that 4- and 5-year-olds already subscribe to such an interpretive theory of mind are thosedata meant to show that preschool children typically solve certain visual perspective taking andappearance-reality problems, or otherwise reveal a beginning appreciation of the diversity ofpersonal tastes, either before or along with their developing ability to pass various tests of falsebelief. In tallying up lines of evidence meant to support his own early-onset views, Perner(1991), for example, cites approvingly the earlier work of Flavell and his colleagues (Flavell,Everett, Croft, & Flavell, 1981) concerned with “level 2 visual perspective taking.” This work,which usefully establishes that even children younger than 4 are ordinarily already well aware thatone’s angle of regard can dictate “how” a display will look, is, however, perhaps better seen as auseful reminder that questions such as “how do you see that” and “how do you understcmd orinterpret that” ought to be regarded as equivalent only when otherwise simple talk about “see” and“looks like” is understood in especially metaphorical ways. That is, children who answerRavell’s classic question, “how do you see the turtle?” by answering, “standing on his feet,” or“lying on his back” did not, in all likelihood, intend, and should not now be understood after thefact as having offered up some juvenile epistemic commentary about how one and the same thingmight prove to have different meanings for different people. Rather, all such talk of differentangles of regard is most conservatively or carefully heard as talk about whether the visualinformation about turtles sliding into your retina arrives feet first or back to front. Perner is notalone, of course, in what may be a common but perhaps unacceptably profligate reading of this“level 2 perspective taking” data. Regardless of the company kept, however, any and all suchclaims about how one person’s line of sight is distinct from that of another is perhaps bestregarded as no more than a precursor to, and ought to be seen as importantly different from, thealtogether more demanding insight that two people can find different interpretative meanings inone and the same thing.9A similar discrimination needs to be drawn between fully fledged matters of interpretation,in the larger epistemic sense being sought after here, and those useful, but different insights thatare revealed when children first appreciate that appearances can sometimes be deceiving. Hereagain Perner (1991), and still earlier, Flavell and his colleagues (Flavell, Green, & Flavell, 1986),suggest that “knowledge about the appearance-reality distinction is but one instance of our moregeneral knowledge that the selfsame object or event can be ‘represented’ (apprehended,experienced, etc.) in different ways by the same person and by different people” (Flavell, Green,& Flavell, 1986, p. 2). The ability to recognize that one object can have both a mere appearanceand a different reality may initially appear to be equivalent to an understanding that some thingsafford multiple interpretations. Alternatively, however, and as hypothesized here, when childrenare asked to comment on both the appearance of an object as well as its underlying reality they donot need to relinquish the assumption that it is really oniy one thing, although it may appear to besomething else. Similarly, with false belief tasks it is a question of one reality and anothermistaken view of this reality. In contrast, tests of interpretation are seen here to require the abilityto recognize that there may be two or more equally valid interpretations of the same message orobject. In other words, passing an appearance-reality task requires only the insight that one viewof an object is right and another is wrong, whereas dealing with issues of interpretation requiresthe insight that multiple views of reality may be equally legitimate.4This series of studies is also related to an ongoing philosophical debate over the best way to view the nature ofthe mind and knowledge. According to traditional epistemology, and this is perhaps the dominant view ofknowledge in cognitive science, knowledge consists of forming accurate mental representations that correspond to apregiven world with properties that can be described prior to any cognitive activity. Since Pemer (1991) beginswith this assumption that the mind is representational in this strong sense (i.e., beyond the uncontroversial claimthat thought involves representing the world), it necessarily follows that “interpretation,” according to this view ofknowledge, involves a breakdown in the representing process which results in misrepresentation. However, thisrepresentational view of the mind is being increasingly questioned (e.g., Overton, 1994a, 1994b; Varela,Thompson, & Rosch, 1991; Furth, 1987; von Glasersfeld, 1979, 1982, 1984, 1988; Lakoff, 1987; Putnam, 1981,1987, 1988, 1990), and in its place is proposed an anti-representational, embodied, or constructivistic view ofknowledge, according to which interpretation is a necessary feature of the knowing process. Although this researchis in some sense independent of the debate between these two views of knowledge and the empirical researchproposed here will certainly not settle the controversy, this philosophical discussion is relevant because the view ofinterpretation proposed in this thesis is most comfortably included within a constructivistic position on knowledge,and, thus, would add one more reason for favoring such an epistemology.10A third and final line of evidence that might be brought out in support of the claim that 4-year-olds ordinarily possess an interpretative theory of mind is provided by the data of Flavell,Flavell, Green, and Moses (1990) showing that even 3-year-olds recognize that differentindividuals are characterized by different “tastes” or preferences. As the work of Ravell and hiscolleagues clearly shows, children of 3 to 4 seem to appreciate that they and the family cat havesharply different sentiments about just how tasty one and the same bowl of cat food will actuallyprove to be. Different measures of children’s understanding of taste, however, seem to assessdifferent competencies that develop between the ages of 3 to 10 (Mansfield & Clinchy, 1985;Zahn-Waxler, Radke-Yarrow, & Brady-Smith, 1977). Here it simply will not do to object, aswas done in the case of false belief and appearance-reality tasks, that the question of differentreactions to one and the same thing is simply never brought up. This recognition that twoonlookers may differ in their tastes may at first appear to reflect an early insight into the concept ofinterpretation.Although the appreciation of differences in taste seems superficially similar to the conceptof interpretation, I argue that differences in taste involve personal affective reactions that requireno external justification and are orthogonal to alternative interpretations that need to be based ongrounded reasons. In other words, an interpretation of something such as a novel can be offeredthat is independent of the liking versus disliking dimension. What is not clear here, however, justas it remains unclear across several centuries of philosophical debate (Gadamer, 1982), is whethersuch matters of taste do or do not actually contain any real epistemological content. In the face ofthese unanswered questions, it seemed appropriate to include in the present study measures ofchildren’s understanding of the diversity of taste, all in order to determine if children’s responsesto such esthetic matters are quantitatively and qualitatively different from their responses to tests ofinterpretation.11Evidencefor a Slow Track Toward an Interpretive Theory ofMindJust as there are lines of collateral evidence supporting the view that an interpretive theoryof mind is an accomplishment of the preschool years, there are also programs of research lendingevidence to the contrary view that any genuine appreciation of the interpretive character of theknowing process occurs more slowly, and usually not before the early school years. Far andaway the largest among these is that half a century’s worth of research devoted to the subjects ofchildhood egocentrism and the gradual emergence of so-called role- or perspective-taking skills(for reviews of this enormous literature see: Chandler, 1977; Chandler & Boyes, 1982; Shantz,1983). The existence of certain lethal conceptual and methodological genes lying near the heart ofthese studies ended up costing them most of their reproductive capacity, and eventually turnedthem into the “sports” of the theory of mind enterprise (Perner & Astington, 1992). It seems,however, that the poor performance of whole armies of 5- and 6- and 7-year-olds on standardrole-taking tests should not be taken too lightly. Any attempt to recount all that went wrong withthis once thriving program of social-cognitive research is too big an undertaking to be picked uphere. For present purposes it will be enough to point out that many of that literature’s mostcommonly relied upon measures of childhood egocentrism now appear, by present lights, to haverelied upon an under specified admixture of probes and procedures, some of which aimed to get atknowledge of false beliefs while others better served to illuminate more interpretive matters.Chandler and Helm’s (1984) so-called “droodle” procedure--a methodology usefully extended andclarified by Taylor (1988)--can serve as a representative case in point. Here subjects wererequired to set aside what they knew of a larger stimulus drawing in order to successfully adoptthe perspective of someone else who saw only a small cryptic part of this same stimulus. Clearly,as Perner and Davies (1991) have pointed out, the experimental manipulation employed in thisdroodles procedure shows much in common with standard false belief tasks that also rely uponthe genymandering of evidence such that subjects are ordinarily better informed than those whoselikely beliefs they are expected to comment on. What divided the original Chandler and Helm12procedure from the stripped down multiple-choice version substituted by Perner and Davies(which 4-year-olds succeeded at nicely) is that, in the original, subjects were required not only tokeep what they knew separate from what was known by others less well informed thanthemselves, but also to imaginatively put themselves in the place of such a partially ignorantbystander in order to successfully guess at how the limited information made available to themmight be reasonably interpreted. Burdened with such interpretive demands, it is perhaps not allthat surprising that Chandler and Helm’s 5- and 6- and even 7-year-olds often failed where Pernerand Davies’ 4-year-olds succeeded. While none of this is meant as a real substitute for that othermuch needed attempt to re-read all of the role-taking literature of the 60s and 70s in light of themore painstaking theories of mind literature of the 80s and 90s, it is meant to show that perhapsonly some, but by no means all of the many hundreds of those earlier studies of the emergingrole-taking competencies of young school-age children can be safely set aside. Rather, many ofthe 5-, 6-, and 7-year-olds who failed the procedurally complex role-taking tasks often laid out inthese studies did so, not because they were blocked from displaying their already secureunderstanding of the possibility of false beliefs, but rather for the reason that they were still enroute toward acquiring a genuinely interpretive theory of mind.Over and above the broad, but now somewhat moribund, literature on social role-takingthere also exists a small number of other distinct research enterprises aimed at also showing thatthere is more to the development of a mature theory of mind than the accomplishment of simple(i.e., first order) false belief understanding. Almost paradoxically, one of these lines of researchwas introduced by Perner and Wimmer (1985). A research effort that turned on the commonintuition that recognizing someone else’s belief to be mistaken is somehow less demanding thanthe still more recursive, twice removed, insight that is needed to appreciate that someone else’sbeliefs about someone else’s beliefs may also be mistaken. While there would seem to be noreason to doubt that keeping track of two recursions is naturally a bit more complicated thankeeping only one in mind, there is also no reason to automatically suppose that the future ofincreasingly adult like theories of mind necessarily lies down this path of more and more deeply13nested recursive operations. The recent study by Sullivan, Zaitchik, and Tager-Flusberg (1994),for example, showing that, when task demands are held at a minimum, even 4- and 5-year-oldsquickly grasp second order recursions involving beliefs about beliefs, offers support to theposition that mature views of mental life rest upon a groundwork that is somehow more genuinelyinterpretive than that required to deal with second order false beliefs.Even if, as the previous paragraphs are meant to imply, much of the original social role-taking literature now seems flawed because of what, in retrospect, appears to have been tooundisciplined a reliance on tests of false belief understanding plus other things, such criticismshardly apply to a small handful of more contemporary studies that have worked quite self-consciously to avoid confusing matters of interpretation with the simpler consequences ofdifferential ignorance. In particular, the work of both Taylor (1988a, 1988b) and Pillow (1991,in press) seem to fit this new and improved mold. While both of these investigators are matchedby their common interest in bringing out the details of young children’s growing awareness of therole of bias and prior knowledge in shaping the process of meaning making, there is sufficientoverlap in their methods to perhaps excuse focusing only on the details of the recent work byPillow (see Appendix A, pages 82-9 1, for a further review of the literature).Pillow, like Taylor before him, has taken some pains to ensure that both his researchsubjects, and one or more other stimulus persons whose views were to be inquired into, all endedup being exposed to one and the same stimulus event (e.g., a toy being trod upon or a cartoonfigure caught in the frozen moment of either taking a rabbit out of, or returning it to, its cage).What was actively manipulated in these studies was, however, nothing to do with present events,which were conscientiously kept self-same for all corners, but turned instead on past sentimentsbuilt up about these stimulus persons, or background details from the past that might reasonablybe expected to influence one’s reading of present events. Using such procedures, Pillow hasfound interesting evidence to indicate that preschoolers, all of whom might be expected to passstandard false belief tests, regularly fail to see any relevance to such prior sentiments orbackground information, and naively assume instead that anyone whose immediate experiences14are the same will end up holding to identical beliefs. Not until the age of 6 or 8 did Pillow’ssubjects show any clear appreciation that belief and sentiment from the past could influencejudgments in the present. Pillow’s findings are consistent with those of Taylor, who has likewisereported that only 8-, but not 6-year-olds showed evidence of beginning to shift the locus of theorigin of knowledge from the object to the subject, and are closely in line with the expectationsthat guided the studies reported here.Despite obvious similarities, there remain, however, important differences that dividePillow and Taylor’s account of epistemic development from the account presented here. Mostpointedly, what both of these investigators have done is to expand the horizon of relevantinformation to include, not just those particular facts available at some frozen epistemic moment,but also various bits of relevant background information that are differentially available to differentstimulus persons. While such procedures require a way of thinking about mental life that appearsclearly unavailable to most preschoolers, it is also true, as Ruffman, Olson and Astington (1991)point out, that tasks of these sorts are really “false belief tasks at heart.” That is, interpretation, inthe concrete sense explored by Pillow, is really equivalent to “misrepresentation,” in that whatsuch accounts seem to make room for is the prospect that bits of ignorance or prejudice from thepast can come forward to shape, or perhaps miss-shape, our readings of the present. This is animprovement over other still more restricted views that seem to make false belief understandingthe only possible epistemic achievement, and simple ignorance the only available intellectualcrime. What is still needed, in my view, however, is a way of thinking about and empiricallyinvestigating matters of interpretation that allows for the possibility that interpretation is more thanthe effects of cumulative ignorance or bias, but rather the ordinary constructive way in whichmeanings are always acquired. The studies described below strive to bring out evidence about thedevelopment of interpretation in that broader sense.15Assessing an Interpretive Theory ofMindAlthough there is a lack of research on the development of children’s understanding ofinterpretation, there is converging evidence from research in several areas, such as humor,5irony,6 linguistic ambiguity,7referential communication,8and pictorial ambiguity (Keil, 198O),For example, in the study of humor, McGhee (1979) has described a series of stages in the development ofchildren’s sense of humor. It is at McGhee’s fourth stage that children’s humor begins to resemble the humor ofadults, much of which is based on ambiguity in meanings. According to McGhee, children reach this fourth stageat about 7 or 8 years of age, because it is at this time that they first begin to appreciate multiple meanings. Punsare a classic example of humor which is based on a key word having more than one meaning. (Consider, forexample, the following classic joke: “Hey, did you take a bath?” “No. Why, is one missing?” [McGhee, 1979, p.76].) Because humor typically involves contrasting a normal situation with an alternative, incongruous situation itdoes not tend to offer equally compelling evidence for two alternative interpretations, and thus, may not be the mostappropriate situation to begin assessing young children’s early understanding of interpretation.6 Irony is an aspect of communication which also depends on the understanding of multiple meanings. In order tounderstand irony one must be able to distinguish the speaker’s intention from the literal meaning of the utterance.Although irony is common is adults’ conversation, young children under the age of approximately 6 to 7 years aretypical said to fail to understand ironic utterances (Winner & Leekam, 1991).‘ Shultz’s research on the development of changes in the appreciation of riddles (Shultz, 1974) and the ability todetect linguistic ambiguity (Shultz & Pilon, 1973) is generally supportive of McGhee’s claims about humor. Theresearch on children’s understanding of riddles has been employed to reveal their metalinguistic awareness ofambiguity because understanding certain types of riddles turns on an appreciation of the possibility of multiplemeanings (Fowles & Glamz, 1977; Hirsh-Pasek, Gleitman, & Gleitman, 1978). As will be brought out in theMethod section examples can be drawn from this literature on linguistic ambiguity in which a word in a messagecould be interpreted as having two possible meanings that are equally likely.8 Research on children’s understanding of referential communication offers a further example of an area thatinvolves children’s understanding of ambiguity. It has been shown in a number of studies that young children of 5to 6 years of age often fail to recognize that ambiguous messages may not sufficiently specify the referent to allowfor correct identification. Much of this research has employed variations of referential communication tasks inwhich children are given directions for making a block building, selecting an item from an array, or finding a hiddenobject. In a typical referential communication task a speaker must describe a referent object and the task of thelistener is to correctly select the referent from among a group of candidate objects. If the message is ambiguous,two or more objects may fit the speaker’s description equally well. Of course, older children notice this ambiguityand either request further description or blame the speaker’s inadequate message for the failure to select the correctobject. However, young children (typically children younger than 5- or 6-years-old) seem relatively unaware of theinadequacies of ambiguous messages and they tend to blame the listener rather than the speaker if a mistake is madebecause of an ambiguous message. However, these children that judge an ambiguous message to be adequate maystill display some verbal or nonverbal signs of uncertainty (Flavell, Speer, Green, & August, 1981).When young children fail to recognize that ambiguous messages are inadequate they seem to think thatreceiving the message is equivalent to understanding the speaker’s intention. In other words, they do not seem tounderstand the distinction between the literal meaning of the “very words” of the message and the speaker’s intendedmeaning (Beal, 1988). This difficulty with the distinction between literal and speaker’s meaning results in aninability to deal with misinterpretations. “Young children appear to identify one utterance with one interpretation,and therefore they are incapable of dealing with ambiguous sentences in which one utterance may result in two ormore interpretations” (Bomtatibus, 1988, p. 326). Until about the age of 7 to 8 children do not understand that amessage’s literal meaning can be somewhat independent of what the speaker intended to mean by the message. Thisshort coming is revealed in first graders’ difficulty in detecting ambiguous messages when they were already awareof the speaker’s intended meaning. Second graders, on the other hand, were more competent at detecting ambiguousmessages even when they had prior knowledge of the speaker’s intentions. It seems that they are able to distinguish16that it is not until approximately 7 or 8 years of age that children begin to understand that oneobject or message may have multiple meanings. The research in these various areas appears tobetween what the speaker intended and the literal meaning of the words used in the message, and thus, they werebetter able to evaluate the adequacy of the message (Beal, 1988).Robinson and Robinson (1983) found that even when children were made aware of the two possibleinterpretations of ambiguous referential communication, because they had to make these interpretations for puppets,the subjects were still confident that they themselves had made the correct interpretation. These results fromresearch on children’s understanding of communication are generally consistent with the assumption that youngchildren have a copy theory of knowledge and it is not until 7 or 8 years that children acquire the understanding thatthe same message can have two meanings. Young children overlook problems with messages and assume that ifthe message is received the listener should possess the speaker’s knowledge. Not until they are about 7- or 8-years-old do children seem to consciously recognize ambiguity in messages--i.e., messages can have more than oneinterpretation. This insight involves a recognition that knowledge is not simply the result of a direct transmissionof information from one’s environment. Rather, since messages can have more than one meaning the person musthave some role in deciding on that meaning. This insight involves a shift in one’s understanding of the source ofknowledge from the environment to a recognition of the complementary role of the mind in the acquisition ofknowledge.As studied in the psychological literature, ambiguous figures are typically line drawings that can be seen as eitherof two familiar objects. One ambiguous figure with a long history is a drawing known as the “duck-rabbit”--a linedrawing that can be seen as either a duck or a rabbit that was used by the psychologist Joseph Jastrow in 1900(Attneave, 1974). Another example of an ambiguous figure from the psychological literature is Bugelski’s “rat-man”, which can be seen as either a rat or a man with glasses (Bugelski, 1960; Bugeiski & Alampay, 1961). Thisambiguous figure was used in earlier research on “perceptual set”. When adults are shown this “rat-man” afterviewing a series of pictures of animals they tend to see the rat, whereas if they had previously viewed pictures ofhuman figures they were more likely to see the man (Bugelski & Alampay, 1961). Initially, it appeared that thisresult did not hold for preschool children (Reese, 1963; Reese & Ford, 1962), but later research did reveal evidenceof perceptual sets in children of 4 years and under when the children were more actively involved in classifying themembers of the classes (West & Abravanel, 1972).Ambiguous figures have also been used with preschoolers by Rock, Gopmk and Hall (Gopnik, 1995) totest the idea that subjects must be informed about the figure’s ambiguity in order to experience reversals whenlooking at the figure. None of their 3- and 4—year-old subjects experienced reversals when uninformed of thepossibility, and even when informed few of the 3-year-olds experienced reversals, but many of the 4-year-olds did.Gopnik attributed the difference between 3- and 4-year-olds, to 4-year-old’s grasp of the representational nature ofmind.These well known ambiguous figures survive because they are particularly good examples of drawings thatcan appear to be two different objects. But ambiguous figures can be thought of as just particularly strikingexamples of a phenomenon that could conceivable occur between many other images as well. These figures occurat a point of ambiguity at which the lines on the paper could equally well depict two different images. It ispossible to transform any image into another image with a computer technique called “morphing”, in which oneimage is transformed into another image in a series of small steps. The resulting image at the mid-point in thisprocess would be ambiguous as to which of the two objects the image depicts, but classic ambiguous figures areespecially good examples because of the ingenious choice of shared contours.Because images in ambiguous figures share contours they are somewhat akin to images involving figure-ground reversal, such as the well known example of the picture that could be a goblet or a pair of faces. However,ambiguous figures do not depend on the reversal of figure and ground (Attneave, 1974). Instead, the various linesmust be interpreted in different ways or be given different meanings to construct another image from the samemarks on the paper.The nature of these ambiguous figures can also be illuminated by considering their similarity to some ofthe cards from the Rorschach test. One of the dimensions on which the Rorschach test is scored involves thefrequency of particular responses to a card. A response is considered “popular” if over twenty percent of the peopletaking the test make this response to a particular card, and some cards have two responses that are classified aspopular (Rapaport, 1946). Ambiguous figures can be thought of as stronger examples of this bimodality of“popular” responses because they have been intentionally designed to elicit two common responses.17reveal separate manifestations of the underlying development of children’s insight into theinterpretive nature of knowledge.’°The present program of research is predicated on the assumption that a useful way oftrying to get at young children’s earliest understanding of the interpretive character of the knowingprocess is to begin by confronting them with instances of that special class of interpretationproblems characterized by stimulus situations that seem to especially call out for two and only twodifferent readings. That is, I judged it reasonable to suppose that some of the best chances forwitnessing early instances of a view of mental life as interpretive in character were to be found intesting situations that require subjects to consider circumstances in which available evidencesupport each of two interpretations of one and the same thing to approximately the same degree.Under such circumstances, at least, it would be possible to avoid the chance that young childrenwho are already well aware of the interpretive nature of knowing might still choose to discount,and thus fail to mention, what they choose to regard as remote and unlikely interpretations.With such prospects in mind, I choose a set of six measurement problems: two eachdrawn from the areas of lexical ambiguity, ambiguous referential communication, and ambiguousfigures. In all of those problem cases I worked to introduce stimulus materials that could be fairly10 Much of this research on the development of an understanding of humor, irony, and communication, has beenconcerned with the development of the ability to detect ambiguity or, in some cases, to intentionally constructambiguous messages (e.g., Sodian, 1990). Although this is all part of an understanding of interpretation, thepurpose of the present research is to address the question of how children make sense of such ambiguities once theyrecognize them, and whether or not they understand the implications of ambiguity for predictions (i.e., that given anambiguous stimulus it is not possible to know in confidence how someone else will interpret it) and thelimitations of interpretation (i.e., interpretations are not just made up, they must be grounded in evidence and goodreasons). Understanding the nature of interpretation involves understanding the epistemic implications ofambiguity, such as the fact that an ambiguous stimuli can often be legitimately interpreted in more than one way.There are several problems with using only children’s ability to detect ambiguity as a measure of their understandingof ambiguity and its implications for multiple interpretations. First, tests of the ability to detect ambiguity may betests of creativity or flexibility in thinking rather than tests of a particular understanding of the nature of knowledge.Even adults may have some difficulty seeing the second entity in an ambiguous figure, recognizing the alternativemeaning in an ambiguous statement, or getting a joke based on multiple meanings, yet we would not argue thatthese adults lack an understanding of the problems of interpretation that ambiguity can lead to. Thus, in some casesemploying the ability to detect ambiguity may underestimate children’s understanding of interpretation.Alternatively, in other cases this measure may overestimate children’s understanding because the mere capacity tonotice the possibility for multiple interpretations does not indicate whether or not these children believe that theother interpretation is in any sense legitimate, or if one is right and the other is completely wrong.18said to provide equal support for both of two distinctive interpretations.” These same stimulusmaterials were used in each of the three studies reported below.11 A reason for choosing three types of problems and not more or less is that, on one hand, this research is notmeant to be an exploration of a particular ability to understand interpretation in one specific domain; rather, theseabilities in various domains are thought to be a manifestation of an underlying conception of knowledge. For thisreason it is important to include a range of different types of tasks in this study. But, on the other hand, thisresearch cannot consist of an exhaustive cataloguing of all the ways in which children’s understanding of mind andknowledge is reflected across all domains of action and interaction. Such an extensive investigation would, at aminimum, include, in addition to ambiguous referential communication, pictorial ambiguities, and various forms oflinguistic ambiguity either lexical or at the level of the sentence, such other areas as, humor, irony, indirect speechacts, politeness (Brown & Levinson, 1988), and ambiguities in narratives.Since these problems are drawn from what have been considered separate areas of research, it is importantto consider the similarities between these problems as well as the differences. All three types of tasks involveambiguous stimuli that afford two approximately equally likely interpretations. In the lexical task this ambiguityis at the level of words that have two meanings. The ambiguity is at the level of the message in the referentialcommunication task. And the third type of task involves pictorial ambiguities that could depict two images withapproximate equal likelihood.In the case of the ambiguous referential communication task there is a hidden object, although its locationis unclear because of the ambiguous message. Thus, there is a right and a wrong interpretation of the ambiguousmessage and there is a fact of the matter that could be appealed to in order to arbitrate between the twointerpretations (even though it is not possible to decide which one is right on the basis of the message). In the caseof the lexical ambiguity task there may be one interpretation that the speaker intended, but it depends on thespeaker’s intentions and there is no simple fact of the matter that can be checked. With the ambiguous figures thereis no fact of the matter about which interpretation is correct (unless features of the drawings are argued about) andthe artist’s intentions were to create an image that make both interpretations equally likely.Both the lexical ambiguities and the ambiguous referential communication task concern ambiguitiesinvolving communication. However, in the case of the lexical ambiguities the words have two meanings and it isthe context and the speaker’s use that normally disambiguates the terms. On the other hand, in the case of thereferential communication task the words in the message are normally unambiguous and it is the context thatrenders the message ambiguous (i.e., the fact that there are, for example, two “red blocks”). It is not clear whetheror not these differences between the tasks will translate into different levels of difficult for the children.19STUDY 1In this first study, young children’s understanding of the interpretive nature of knowledgewas assessed by presenting them with situations in which each of two characters (represented bypuppets’ 2) were shown to have interpreted one and the same object or message in sharplydifferent ways. Making proper sense of these competing knowledge claims requires, I reasoned,some fledgling appreciation of the concept of interpretation in order to formulate what an adultwould take to be a reasonable explanation for these multiple readings. More particularly, the earlyschool age subjects of this study were asked not only to explain how such different interpretationsmight arise, but also whether or not they thought it possible to correctly predict how some thirdperson might go on to interpret these same objects or messages. As a control question, subjectswere also asked to evaluate a “deviant” interpretation of the same stimulus materials in order toassess their ability to appreciate that the very idea of interpretation necessitates that there always belimits to what can count as a warrantable reading of ambiguous matters. To assess children’sunderstanding that people’s tastes differ, subjects were also presented with situations in which thetwo characters again disagreed about something, but in this case the disagreement involved likingor disliking the stimulus in question. Finally, because a central purpose for this research was toevaluate the distinction between false belief understanding and the concept of interpretation, astandard the false-belief task involving both explanation and prediction questions was alsoadministered.’312 Puppets have generally been employed in research on children’s understanding ofmind and no differences havebeen found when these tasks have been compared to tasks involving real people (e.g., Hala, 1994; Russell, et al.,1991).13 The series of tasks introduced above are primarily explanation tasks, this is in contrast to the standard falsebelief tasks which are generally in the form of a prediction task. Both prediction and explanation are presumablybased on the child’s understanding of the mind. Prediction involves reasoning forward based on beliefs and desires,whereas explanation involves reasoning backwards (i.e., accounting for action in terms of the actor’s beliefs anddesires) (Wellman, 1990). It might be argued that these two abilities are separate competencies that may appear atdifferent points in the child’s development. And if explanation requires more in the way of verbal abilities it mayappear later in the child’s development for this reason alone. However, Bartsch andWellman (1988; Wellman,1990) found that when presented with an explanation task, even 3-year-old children, who would not normally be20In brief, then, the primary purposes of Study 1 were: a) to help identify the developmentalpoint at which young children first begin to show clear evidence of understanding the interpretivenature of the knowing process; b) to evaluate the competing claim that children who appreciate thepossibility of false beliefs also automatically understand the process of interpretation; and c) toassess the relationship between children’s understanding of taste and their concept ofinterpretation.MethodThe 8 boys and 12 girls (ten 5- to 6-year-olds [M=5.6 years, range 60 to 72 months] andten 7- to 8-year-olds [M= 7.5 years, range 82 to 104 months]) who served as subjects in this firststudy were given: a) a standard false-belief task; b) a task featuring an example of personal taste(i.e., a drawing depicting either a painting or a bowl of soup said to be liked by one puppetcharacter and disliked by another), and c) at least three tasks involving matters of interpretation.In brief, the problems of interpretation consist of an object or message that is viewed in differentways by two people (represented by puppets). The three types of problems involve: (1) problemsprompted by ambiguous figures (the classic “duck-rabbit” and “rat-man” drawings, see Figures 1and 2); (2) matters of lexical ambiguity (i.e., the homophones “pear/pair” and “ring”); and (3)ambiguous referential communication tasks (i.e., an object said to be hidden under one or theother of two equally “large” or “red” blocks). The three response measures collected with regardto these tasks included: a) questions used to elicit subjects’ explanations of the multipleinterpretations presented, (b) questions intended to get at the implications of these discrepantinterpretations for future efforts to predict how some third person might view the same object orexpected to pass standard false belief tasks, did talk about false beliefs in their explanations of behavior. However,Moses and Flavell (1990) did not replicate this result.Weliman (1990) argued that explanations could be very revealing because they can potentially provideinformation about the child’s spontaneous use of various ideas. Similarly, children’s explanations of events thatinvolve interpretation may also reveal their spontaneous use of the idea of interpretation. In Bartsch and Wellman’s(1988) study they followed up children’s spontaneous explanations with further probes.21message, and (c) a final question aimed at bringing out possible limitations on the range oflegitimate, as opposed to “deviant” interpretations of these same stimulus materials. Moreconcretely, subjects were first asked to explain the difference in interpretation offered by twopuppets figures. Next, subjects were queried regarding their understanding of the implications ofsuch divergent views for any future predictions they might make by asking them to predict howsome third person would interpret the self-same message or picture. Finally, in all but theproblems involving matters of taste, each subject’s understanding of the limitations ofinterpretation (i.e., the fact that legitimate interpretations are not just made up, but must begrounded in evidence and good reasons) was assessed by presenting them with a “deviant”interpretation offered by a third puppet figure. This control question focused on how subjectsundertook to explain their views about whether some extravagant interpretation did or did not“make sense.”False-belieftestThe false-belief test was modeled on Wimmer and Perner’s (1983) classic “Maxi task.”Subjects were introduced to two puppets (Mary and Maxi). In this version of the task, Maxi wasshown to be playing with a toy which he then placed in a blue container before going outside toplay. While Maxi was out (actually under the table) Mary took the toy out and played with it.Before leaving she put the toy away, not, however, in the original resting place, but in a second,yellow container. Maxi was then shown to return and the subjects were told that he wanted histoy. They were then asked the following prediction questions: “Where will Maxi look for his toywhen hefirst comes in?” (“look” question), and “Where will Maxi think his toy is when he firstcomes in?” (“think” question). Next, they were asked the following explanation question: “Whydo you think Maxi looked in the blue/yellow box?” These “explanation” questions, while not aroutine part of this so-called “unexpected change” task, were asked in an effort to ensure a closeparallel between the false belief test and the tests of interpretation detailed below.22Problems ofInterpretation1) Lexical Ambiguity. In one instance of the task involving lexical ambiguity theexperimenter told the subject about another game in which the puppets have to “wait for a ring.”Then the experimenter asked Maxi and Mary what they were waiting for. Each was made toreport one of the common meanings of “ring,” illustrated with pictures of a ringing telephone anda “diamond” ring. In the second instance of this task the puppets were told to “wait for apear/pair.” Again, Maxi and Mary were made to assume different meanings of “pear/pair,” andeach meaning was again illustrated with a picture on a card (i.e., “a pear to eat,” and “a pair ofshoes”). This procedure was loosely based on related test problems earlier employed by Shultzand Pilon (1973).2) Ambiguous referential communication. The referential communication task used in thisstudy was based on a hiding game previously introduced by Sodian (1990). A penny was hiddenunder one of three cards, the backs of which were distinctively marked by either a large red block,a large blue block, or a small red block. Two test trials using these ambiguous referentialmaterials were administered to subjects. These consisted of showing these materials to the twopuppets and explaining: “The penny is under the card with a big block,” or “The penny is underthe card with a red block.” In each of the two trials one puppet was made to endorse one of thetwo equally reasonable interpretations of the ambiguous message and the other puppet endorsedthe other interpretation.3) Ambiguous Figures. Subjects were shown one of two ambiguous line drawings: a)Jastrow’s (1900) “duck-rabbit”; and b) Bugeiski’s (1960) “rat-man” (See Figures 1 and 2). Maxiand Mary were again each made to endorsed one of the two readily available interpretations ofthese classic ambiguous figures.Matters of tasteThe subjects were presented with two vignettes involving: a) a disagreement aboutwhether a particular soup tastes “good” or “bad,” and, b) a disagreement about whether a given23picture was “nice” or “ugly” (each problem was illustrated with a single picture). Again, the Maryand Maxi puppets each took up opposite sides in the disputes.Order of the TasksThe false belief task was always presented first because, given the age of the subjects inthis study, it was anticipated that they would have no difficulty in successfully passing this test.The children’s enjoyment of their easy success on the false belief test was meant to serve tointerest them the subsequent series of tests of interpretation and taste. The remaining four tasksinvolving interpretation and matters of taste were randomized in their order of presentation.ProcedureSubjects were presented with one of the problems of interpretation or taste, and then askeda series of explanation and prediction questions, and a control question involving improbable or“deviant” interpretations (see Appendix B for a complete protocol for Study 1).Explanation Questions. First, the experimenter ensured that the child understood bothinterpretations. For example, in the case of the ambiguous figures the experimenter confirmedthat the child could see both the duck and the rabbit, or the rat and the man. Then theexperimenter restated the problem in the form of the following questions:“Is it okay for Mary to say ... and for Maxi to say ...“Why is it okay . . . /Why isn’t it okay ...Prediction Questions. After the explanation questions the children were asked a series ofprediction questions. This was done in order to assess their understanding that one implication ofsuch ambiguous messages and objects is that it is not possible to confidently predict which of thetwo equally reasonable interpretations any given person might choose to give. Specifically, thesubjects were asked if they could predict how a child from another school would interpret theproblems. If the subject took a decisive position by saying, for example, that the other person24would think that the “duck-rabbit” is a rabbit, the experimenter then asked: “How can you tellwhat they will think?”; and “How sure are you that they will think that?” If the subject said “Idon’t know,” then the experimenter asked: “Why is it hard to tell what they will think?”Deviant interpretations. After the prediction questions the subjects were told that a thirdpuppet, Josef, held an improbable or deviant interpretation as to what the picture or message was(e.g., when told to “wait for a pair/pear,” Josef says he is waiting for an apple). The childrenwere asked if it “makes sense” for Josef to give the deviant answer, and why or why not. Therewas no opportunity to ask questions of this sort on the tasks involving matters of taste.Scoring the False Belief testFollowing scoring conventions common in the literature, subjects were scored as passingthe “look” and “think” questions on the false belief test if they accurately stated that Maxi wouldmistakenly “believe” that his toy was in the container in which he had originally placed it. Failingthese questions was indexed by wrongly concluding that Maxi would somehow know that the toywas in its new location, a fact which he had no legitimate way of appreciating. Subjects werejudged to have passed the explanation question on the false belief test if they correctly explainedthe source of Maxi’s false belief (e.g., “because that’s where he left it and he didn’t know thatMary moved it”).Scoring the Problems ofInterpretationFor all problems involving two trials subjects were assigned the highest score theyreceived on either trial. This strategy avoids discounting any fledgling understanding ofinterpretation and is regarded as conservative in that it works against the present hypothesis thatthe concept of interpretation is achieved later than false belief understanding.Explanation questions. Children were scored as failing the explanation question if theywrongly stated that it was not possible or reasonable for the two puppets to offer different25interpretations because one of them was wrong, or if they considered both interpretations to be“okay,” but could not justify this judgment, or justified it purely in terms of internal individualdifferences without noting the ambiguous nature of the stimulus (e.g., “because they havedifferent thoughts. She has purple eyes and he has brown eyes so they see different things”).Children’s responses were scored as passing if they judged both interpretations to be legitimateand explained the reported difference in interpretation as due to the ambiguous nature of themessage or object (e.g., “because you said it’s under the red block and there are two red blocks, abigger one and a smaller one,” or “because you didn’t say what kind of ring you mean,” or“because it looks like both of those things [a duck and a rabbit]”).Prediction questions. Children were scored as failing the prediction questions if theymade a clear and specific prediction on behalf of a third person, or if they failed to make aprediction but could not explain why it would be impossible to do so with any degree of certainty.Children were scored as passing if they refused to make a prediction and explained why it wouldbe difficult to do so (e.g., “I don’t know. There are lots of different kinds of rings”), or indicatedthat some people might endorse one interpretation while other people would endorse the otherinterpretation (e.g., “a little bit [of the people] will say it looks like a duck and a little bit more willsay it looks like a rabbit”).Deviant interpretation questions. Children were scored as failing these test questions ifthey alleged that the deviant interpretation makes sense, or if they indicated that the deviantinterpretation does not make sense, but could not explain why. Children were scored as passingif they claimed that the deviant interpretation does not make sense, and justified this judgment byreferring to the ambiguous nature of the stimulus and the lack of good reasons for such aninterpretation (e.g., “No. [it doesn’t make sense] because the card that Josef showed isn’t a bigblock,” or “because it [a necklace] has nothing to do with a ring,” or “because it doesn’t look like26one [a dog]”, or “because, for one, a dog’s head goes like that ... [the subject points at thepicture]”).Scoring the Matters ofTasteExplanation questions. Children were scored as failing the explanation question on theproblems involving matters of taste if they believed that one of the puppets was correct and theother one was wrong (e.g., “I think she’s [Mary] right”), or if they showed some recognition thatpeople have different tastes by correctly judging that it was okay for Mary and Maxi to disagreeabout the soup or the painting, but were unable to justify or explain this judgment in any way, ortook the question in a moral sense (e.g., it is not okay for Maxi to express his dislike for the soupbecause, “everyone should try new things,” or “it would hurt people’s feelings,” or because he“won’t grow into a healthy boy”). Subjects were scored as passing if they were offered somejustification for the disagreements involving taste (e.g., “Yeah it’s okay because she likes differentthings than he likes,” or “that’s fine, different people like different things”).Prediction question: Subjects were scored as failing the prediction question on the tests ofmatters of taste if they made an unqualified prediction with a high degree of certainty regardingwhether another person would like or dislike the soup or painting, or if they correctly stated thatthey could not predict whether or not another person would like or dislike the soup or painting,but they could not explain why. Children were scored as passing if they stated that they could notpredict another person’s tastes (e.g., “I wouldn’t know because I wouldn’t know if he likesvegetables,” or “because he never told us before”, or “I can’t read his mind”).ReliabilityThe responses from six subjects were rescored by another rater. There was 91%agreement between the raters on the tests of interpretation, Cohen’s kappa = .82. On theproblems involving matters of taste there was 91% agreement, Cohen’s kappa = .82.27ResultsThe primary hypotheses to be evaluated in this study are: a) passing recognized measuresof false belief understanding is not the same thing as succeeding on tasks tapping anunderstanding of the interpretive character of the knowing process; b) that children develop aninitial understanding of interpretation some years after they acquire an understanding of falsebelief; and c) that a recognition of the diversity of tastes is earlier developing and distinct from theconcept of interpretation. The alternative hypothesis, supported by Perner (1991), Ravell (1988),and others (e.g., Ruffman et al., 1991), is that false belief understanding already presupposes anunderstanding of the mind as “interpretive,” and so children will succeed at problems requiring anunderstand of interpretation as soon as they can pass standard measures of false belief.The clear result of this study regarding the first two hypotheses concerning interpretationis that all the subjects passed the false belief test with no difficulty, but virtually none of the 5- to6-year-old subjects passed the interpretation tasks. This summary conclusion can be seen to besupported by the results even without the benefits of inferential statistics, but to further specifythis observation it is possible to compare the 5- to 6-year-old subjects’ performance on theexplanation question on the false belief test with their performance on the comparable explanationquestion on the interpretation tasks. Whereas all the subjects passed the explanation question onthe false belief test (100%), none of the younger subjects passed the explanation question on thelexical ambiguity test (0%), one of the 5- to 6-year-old subjects correctly answered this questionon the referential communication task (10%), and only two of these same subjects passed theambiguous figures task (20%), binomial tests, p < .01. Thus, even though children in the 5- to 6-years-old age group had no difficulty with the false belief test, and presumably had acquired thisunderstanding some years earlier, they had still not developed equivalent competence with tasksrequiring an understanding of interpretation.28On the control question in which subjects were presented with a “deviant” interpretation47% of the 5- to 6-year-olds and 77% of the 7- to 8-year-olds were judged to have passed. Manyof the subjects could reject the deviant interpretations because they were not based on the stimulusin question, but this does not tell us whether or not these subjects can understand that twodifferent interpretations of one and the same stimulus can be legitimate. To address this questionthe explanation and prediction questions on the tests of interpretation must be considered. Therewas no significant difference between the explanation and prediction questions, t < 1. Therefore,these questions were combined and a mean score across both questions was used in furtheranalyses.Next, the subjects’ performance on the tests involving interpretation and taste wascompared across the two age-groups. The mean score across both the explanation and predictionquestions on each task was used as the dependent variable in a 2 x 4 (Age Group x Task: lexicalambiguity, referential communication, ambiguous figures, and matters of taste) repeated measuresanalysis of variance with repeated measures on the last factor. This analysis revealed a significantmain effect for age-group, F(1, 18) = 21.06, p < .001. As expected the proportion of testquestions passed by the 7- to 8-year-old subjects (.61) was significantly higher than theproportion passed by the 5- to 6-year-old subjects (.19). There was a significant main effect fortask, F(3, 54) = 8.16, p < .00 1, with no significant task by age-group interaction, F < 1.Follow-up t-tests revealed that only the test of children’s understanding of taste differedsignificantly from the other tests, involving lexical ambiguity, referential communication, andambiguous figures, ts(19) = 3.68, 4.27, and 2.98, respectively, p < .05, corrected for familywiseerror rate. In addition, there was no significant difference between the proportion of questionsacross all the tests passed by the boys and the proportion passed by the girls, t < 1.Insert Figure 3 about here29DiscussionThe results of this study are consistent with the hypothesis that the understanding ofinterpretation is not equivalent to the understanding of false beliefs. Rather, interpretation wouldappear to involve a more complex understanding of knowledge and the mind than is implied in agrasp of the simpler possibility of mere ignorance, as evidenced by the fact that success on tests ofinterpretation typically occurs a number of years after children can pass a standard false belief test.The subjects were also significantly more competent at explaining the source ofdisagreement on the matters of taste than they were on the tests of interpretation. And there wasalso a significant improvement in performance with age for the matters of taste. The fact thatsubjects’ performance on these tests concerning differences in taste was less than perfect does notcontradict Flavell et al.’s research (1990), and the findings of other investigators (e.g., Hart &Goldin-Meadow, 1984; Zahn-Waxler, Radke-Yarrow, & Brady-Smith, 1977), showing that wellbefore the age of 5, children have an understanding that people can and do differ in their tastes.Flavell et al. (1990) showed that even 3-year-olds enjoy some success in tasks requiring only thatthey keep track of another person’s “non normative value beliefs” (e.g., the fact that anexperimenter disliked a cookie that tasted good to the child subject), but it should be noted that inthe clearest of those several experiments the children were reminded of the other person’s distinctpreferences just before they were asked to report on these same preferences. Ravell et al. (1990,p. 927), comment that whether their subjects also sensed that “there need be no definite right orwrong in such differences of opinion is not known, but would be worth finding out.” On thetests involving matters of taste employed in the present study subjects were required to firstacknowledge that it is “okay” that the two puppet characters disagree over whether they like ordislike a type of soup or a particular painting (essentially the question posed by Ravell et al.)before going on to attempt to offer an explanation for this state of affairs. Thus, the majority ofthe subjects in Study 1 (90%) revealed an understanding that people legitimately differ in theirtastes by acknowledging that it was “okay” for Mary and Maxi to disagreed about whether the30soup tasted good or bad. Many of these subjects, however, were unable to go on and offerreasons for these differences in taste. (The two 5-years-old subjects who did not agree that it was“okay” that the puppet characters expressed different tastes did not offer an explanation, and as aconsequence it is difficult to know how to interpret their failures.) This recognition that the twopuppets have acceptable different tastes is clearly more demanding than the criteria required forpassing Flavell et al.’s tasks.Subjects in the present study were required to provide counterpart explanations in order topass the tests featuring differences in taste, just as they also were required to provide explanationson the false belief task, and the tests of interpretation. In contrast to those testing situationsinvolving matters of belief, it is difficult to know what ought to count as an adequate explanationfor differences in taste. In this study, the solution strategy adopted by many of the 5- to 6-year-olds (45%), and significantly more of the 7- to 8-year-olds (80%), consisted of simplyannouncing as a matter of fact that people differ among themselves in this way. That is, peoplesimply prove to have different tastes. These results are comparable with a study conducted byMansfield and Clinchy (1985) in which they found that 35% of 4-year-olds, 60% of 7-year-olds,and 97% of 10-year-olds explained differences in taste in terms of individual personalpreferences. Similarly, on a somewhat simplified perspective-taking task, Zahn-Waxler, RadkeYarrow, and Brady-Smith (1977) found that 61% of 3-year-olds and 100% of 6-year-oldscorrectly selected the appropriate food choice for a confederate with different tastes from theirown. Explanations of these sorts are, of course, really just a way of restating the problem.Nevertheless, it is a form of explanation that both children and adults seem to count as acceptable.As noted above, subjects’ responses to tests of interpretation, on the one hand, and testsbased on matters of taste, on the other, are different in quantitative ways, but they also differ inqualitative ways. In their responses to questions concerning matters of taste the subjects did not,for example, attempt to locate the source of the difference in opinions in the nature of the stimulus.Rather, they undertook to locate responsibility for such differences in esthetic judgements entirelywithin the person making the judgement. Conversely, the subjects who passed the tests of31interpretation did so almost exclusively by referring to the ambiguous nature of the message orpicture itself. Thus, there was some clear evidence on the part of at least the older subjects ofsome recognition that, whereas tastes require no extrinsic justification, differences ininterpretations must be warranted by an appeal to grounded evidence in the material world. Thatis, interpretations, unlike esthetic preferences, must be based on evidence and reasons groundedin the stimulus in question. This qualitative difference in the way children responded to theproblems of interpretation as compared to the tasks involving matters of taste is consistent with theargument outlined above that the domain of taste may rest primarily on affective evaluations thatare distinct from the epistemic matters underpinning interpretation. This attempt to drive a wedgebetween matters of taste and interpretation is consistent with Flavell et al. s argument that beliefsdirect one’s attention outward, toward the world, whereas preferences direct one’s attentioninward, and goes some distance toward walling off the early arising ability to process individualdifferences in taste, without otherwise threatening the conclusion that children are slow to movetoward a realization of the interpretive nature of knowing.Further research regarding the development of various forms of children’s understandingof taste is obviously required. Nevertheless, the data of this initial study seem to offer compellingempirical support for the theory driven conclusion that matters of taste have a developmentalcourse that is different from the distinct nature of maturing thoughts about matters ofinterpretation.In search for further support for a developmental difference between matters of taste andinterpretation, plans were made to undertake a second study of interpretation, this time involvingonly adults. That is, even though it seemed appropriate to assume that the kinds of justificationsadults typically employ for differences in taste would be comparable to those of intermediateschool-aged children in that they would differ in a qualitative way from their justifications fordifferences in interpretation, it nonetheless seemed worthwhile to confirm this intuition beforegoing on. This issue was addressed in Study 2.32STUDY 2In Study 2, a group of 15 adults was presented with a subset of the questions concerningmatters of taste and problems of interpretation already employed in Study 1. The purpose of thisstudy was to evaluate the strong expectation that adults will respond, as did especially the olderchildren in Study 1, in a qualitatively different manner when justifying differences in taste, on theone hand, and differences in interpretation, on the other hand.MethodSubjectsThe 15 adults who participated in this quick check upon what adults take to be thedistinction between epistemic matters and matters of personal taste represent diverse educationalbackgrounds, and were selected on the basis of convenience from groups of students, graduatestudents and their spouses (8 women and 7 men). Their ages ranged from 18 to 43, and theireducational backgrounds ranged from not having completed high school to undergraduate andgraduate degrees in such diverse topics as computer science, medicine, history, psychology,geography, and social work.ProcedureThe subjects were presented with the two problems turning on matters of taste and two ofthe problems of interpretation previously employed in Study 1 (i.e., a problem based onambiguous referential communication and an ambiguous figure) (see Appendix C for a completeprotocol for Study 2). The order of these two types of tasks was counterbalanced, and these briefinterviews were audio-taped and later transcribed. The general adequacy of these subjects’responses were classified according to the scoring system used in Study 1. More to the presentpoint, however, these same responses were also divided into those explanations that located33responsibility for the different conclusions reached by the different stimulus characters either inthe stimulus environment or these story characters themselves. That is, responses were eitherclassified as accounting for the character& disagreements predominately in terms of somethingabout the person (internal source) or something about the stimulus event in question (externalsource). The subject& responses were rescored by a second rater with 100% agreement betweenraters on both the pass/fail scoring system and on the types of explanation.ResultsThe results of this study are simple and straightforward. All the adult subjects withoutexception (100%) passed both the tests of interpretation and taste. Similarly, each and every oneof these adult subjects (100%) explained the differences in taste as justifiable differences inpersonal preference, and uniformly located the source of these differences within the personsexpressing these conflicting opinions. These responses were generally indistinguishable from theresponses offered by the children in Study 1 in that they simply acknowledged the differences intastes that were reported and treated such differences as endemic to the different individuals whoexpressed such distinctive preferences. For example, subjects justified the puppets’ likes anddislikes of a certain food item and a particular picture with statements such as: “everybody hasdifferent tastes...”; “personal preferences”; and “different people like different things.” In directcontrast to their way of understanding differences in personal tastes, all the adult subjects justifieddifferences in interpretation in terms of the ambiguity of the message or the picture presented. Forexample, when subjects were attempting to explain the fact that the two puppets had differentlyinterpreted the ambiguous message that the experimenter was thinking of “the card with the redblock,” they said, “there are two red blocks.” When subjects were justifying the disagreementbetween the puppets over whether the ambiguous figure was a duck or a rabbit, they stated, forexample: “there are two things, or images in this picture”;” because it can be two things”; “Maryis seeing this as being the ears and Maxi is seeing it as being the bill. So they are seeing the same34thing differently.” The point of interest here is that in their responses these adult subjects, liketheir child counterparts in Study 1, consistently located the source of the difference ininterpretation in the ambiguity of the message or the picture, not in the personal attributes of thepuppet characters.DiscussionIn brief, this short study provides further evidence of a difference between understandingmatters of taste and the concept of interpretation, and, consequently, works to help justify adecision, taken up in Study 3, to focus attention more exclusively on the concept of interpretation.Thus, in Study 3 children’s understanding of taste was not further explored and, instead, attentionwas focused exclusively on epistemic matters and the development of children’s understanding ofinterpretation and its implications.A number of issues, nevertheless, were raised in Studies 1 and 2 that are still in need ofdiscussion. It seems clear, for example, that the stark distinction between subjects’ explanationsof matters of disputed taste in terms of internal personal preferences, and their explanations ofconflicting beliefs on the tests of interpretation by making reference to the ambiguity of the stimuliis in some part an artifact of the assessment method employed. As defined here, “passing” a testof interpretation turned upon the subjects having made some reference to the ambiguous nature ofthe stimulus. This follows for the straightforward reason that the test questions posed concernedhow multiple interpretations are in fact possible. Having uniformly met this simple standard,some subjects (27% of the adults), however, went on to speculate about why the puppetcharacters endorsed those particular interpretations, and in doing so proceeded to discuss thepuppets’ possible internal attributes. This mention of internal factors when discussing differencesin interpretation is not taken here to imply that the internal versus external distinction should bediscarded. Rather, it seems evident from their responses that when subjects mentioned internalfactors they were considering the related question of why the puppets voiced those particular35interpretations, which goes beyond the question of how it is that two interpretations were possibleto begin with. In other words, the possibility of multiple interpretations depends on the ambiguityof the stimulus in question. The further, although related question of why particularinterpretations are endorsed by different characters, given the already mentioned stimulusambiguity, may require speculation about the internal attributes of these characters.Although further research into the question of how children and adults view the relationbetween matters of taste and matters of belief is clearly needed, the results of this study add to thefindings of Study 1 and suggest that an understanding of the diversity of taste differs qualitativelyfrom a grasp of the concept of interpretation. Since young children’s recognition that people’stastes differ does not reflect an early understanding of interpretation, attention in Study 3 will befocused on the development of children’s understanding of interpretation. At the top of any list ofcritical next steps beyond studies 1 and 2 is the need to replicate these findings and to test theconclusions drawn against other more reductive interpretations. The results of Study 1 needed tobe replicated for two reasons. First, improvements in the procedure and questions needed to beintroduced to rule out the possibility that the 5-year-old subjects’ poor performance might havebeen due to a simple lack of proper understanding of the test questions posed. Second, a largersample size was judged to be required for assessing the possibility of gender differences andinteractions between age and type of question, further assessing possible differences between thetypes of questions concerning interpretation, and counterbalancing the order of the tests.36STUDY 3A possible objection that could be raised to Study 1 is that the emphasis placed on mattersof justification gave undue importance to issues of linguistic sophistication, rather than to realdifferences in the complexity of subjects! understanding of the knowing process. This commondilemma in developmental psychology has its roots in the distinction between competence andperformance, and there are a number of ways to respond to this potential criticism. The relevantresults from Study 1 that engage this problem need to be summarized briefly, however, beforeconsidering how Study 3 was designed to address these concerns.All the subjects in the first study proved themselves to be competent at explaining thesource of false beliefs in the false belief task, but failed similar questions that inquired into mattersof interpretation. Thus, the subjects’ problems with the tests of interpretation employed do notseem to be simply due to problems in constructing explanations in general. It is only when suchexplanations require an understanding of interpretation that younger subjects encounter difficulty.Still, it is possible to suppose that there was something obfuscating or otherwise “tricky” aboutthe particular wording of these interpretation questions that worked as a stumbling block toartificially trip up the younger subjects in Study 1. This alternative reading of the results of Study1 was evaluated in Study 3 by modifying the explanation question on the false belief test in orderto make it almost identical in structure to the explanation question on one of the tests ofinterpretation. This procedural modification would allow a more direct comparison of children’sexplanations of the two types of task. In the false belief task already employed in Study 1, Maxiand Mary disagree about the location of an object, and this difference of opinion (i.e., Maxi’sfalse belief) is due to the differential degree to which they are informed about the whereabouts ofthe toy. There is a similar disagreement between Maxi and Mary regarding the location of anobject in the referential communication task originally employed in Study 1, but this difference inopinion is due this time to a difference in interpretation of an ambiguous message. In otherwords, on the referential communication task both Mary and Maxi possess the same information,37but they interpret it differently. Even without additional modifications similar levels of linguisticskill would already seem to be required to answer both questions, while at the same time adifferent conception of knowledge would be needed to appropriately answer the questionconcerning interpretation. This discussion underscores and helps bring out the fact that whereasRuffman et al. (1991) claim that the type of ambiguity tasks employed in their own study arereally “false belief tasks at heart,” the problems of interpretation employed in the present researchare fundamentally different from false belief tasks because the differences in beliefs being featuredhere arise due to different readings of precisely the same infonnation, rather than the possessionof different amounts of available information.For the purposes of Study 3, a further modification was made in the false belief test tomake it more similar to one of the tests of interpretation employed. In the explanation question onthe false belief test used in Study 1 the subjects were asked only about the source of Maxi’s falsebelief (i.e., “why does Maxi think his toy is in ... ?“). In Study 3 the subjects were asked what toan adult would seem a fuller version of the same thing, “Why does Maxi think his toy is in X andMary think it’s in Y?” This modification of the explanation question makes it closer (in factalmost identical) in structure to the explanation question on the ambiguous referentialcommunication task already employed in Study 1. Thus, the subjects are required to giveexplanations to similar questions concerning a difference in beliefs. In the case of the false belieftask this difference in beliefs arises because of Maxi’s partial ignorance. In the referentialcommunication task, by contrast, the difference in belief concerning the location of an object isdue to different interpretations of the same ambiguous message.The foregoing methodological nicety is in response to the potential criticism that the poorperformance of the young subjects in Study 1 on questions about interpretation might be due totask complexity (performance factors), rather than lack of competence, and illustrates one strategyfor dealing with this ubiquitous line of criticism (i.e., make the false belief test similar to theinterpretation tests). Although criticisms of this sort can never be completely countered, one38additional defensive strategy is to simplify and clarify the questions on the interpretation tests.This solution strategy was taken up, as follows, in Study 3.In the primary explanation question in Study 1 the children were asked “Is it okay forMary to say ... and Maxi to say ... ?“ One problem with using this form of question to elicitexplanations is that some subjects took these questions in a moral or normative sense, by hearingthem as equivalent to, “is it right or good for Mary and Maxi to think different things?” What wewant to know in the place of such a reading is if children believe that it is legitimate (i.e.,supportable with good reasons) for Mary and Maxi to disagree in this case. A more naturalquestion might appear to be, for example, “Why is Mary waiting for a ring for her finger andMaxi waiting for a telephone to ring?” Pilot testing indicated that such questions caused somechildren to construct reasons for those specific interpretations. Therefore, the critical test questionin Study 3 was phrased in a more general way to avoid the focus on those specific interpretationsand to elicit instead the reasons that two different interpretations are possible. The form adoptedfor this question was, “Why does Mary say she’s waiting for one thing and at the same time Maxisay he’s waiting for another thing?” A probe question read as follows, “Does is make sense forMary to say one thing and Maxi to say something else?”, and “Why does (doesn’t) it makesense?”The subjects in Study 1 were also questioned concerning the possibility of making aprediction about how another child might interpret the ambiguous stimuli. In order to make theprediction question as parallel as possible to the explanation question, subjects were asked tomake predictions about a third puppet character rather than a real person. In Study 1, all thosesubjects who did make predictions about how someone else would interpret the ambiguousmessage or picture were also asked how sure they were about their predictions. It is important torecord subjects’ level of certainty regarding such predictions because some children may feelinclined to make a prediction even though they may recognize that the circumstances permit orily asimple guess. To improve the assessment of children’s degree of certainty about their predictions,in Study 3 subjects were asked to use a pointer (see Robinson & Robinson, 1983; Robinson &39Whittaker, 1985) to select one of three categories of certainty: 1) “really sure”; 2) “not quite sure”;and 3) “don’t know at all.”4Finally, none of the subjects in Study 1 had any difficulty with either the “look” or the“think” questions on the false belief test. In general, previous research has not found consistentdifferences between these questions, although there is some possibility that the “think” questionmay be slightly more difficult for 3-year-old subjects (Hala, 1994). Since the youngest subjectsin Study 3 were to be 5 years old, only the more difficult “think” question was retained in order toshorten the task.The procedure used in Study 3 was then quite similar to that employed in Study 1 exceptthat problems involving matters of taste were not presented to the subjects.MethodSubjectsA total of 48 boys and girls between the ages of 5 and 8 participated in this study. Therewere equal numbers of boys and girls at each of the four ages (5, 6, 7, and 8 years). There were12 5-year-old children (mean age, 5 years 6 months, with a range of 58 to 71 months), 12 6-year-olds (mean age, 6 years 5 months, with a range of 72 to 83 months), 12 7-year-olds (meanage, 7 years 5 months with a range of 84 to 94 months), and 12 8-year-olds (mean age 8 years 4months, with a range of 96 to 107 months). This was a predominantly middle-class, althoughracially and ethnically diverse, group of children drawn from day-cares and after-school-cares.14 As discussed above, one difference between the referential communication tasks and the lexical ambiguity tasksis that in the hiding game used as a referential communication task in this study (from Sodian, 1990) there is a factof the matter about where the object is hidden that could be checked to decide which of the two alternativeinterpretations of the ambiguous message is correct. On the other hand, in the lexical ambiguity tasks (talk ofrings and pears/pairs), if there is a correct interpretation it depends on the speaker’s intentions. In Study 3 one ofthe two referential communication tasks was a hiding game, as in Study 1. The other task, however, was aselection task in which the experimenter gave the puppets an ambiguous clue about which one of the cards he is“thinking of.” This type of selection task is often employed in research on referential communication, and in thistype of task, like the lexical ambiguity tasks, there is no “objective” fact of the matter. This modification of one ofthe referential communication tasks allows for a comparison between tasks in which there either is or is not an“objective” fact of the matter.40ProcedureThe subjects were first presented with a standard false belief test, and then with sixproblems involving interpretation slightly modified from those used in Study 1. These six tasksconsisted of two examples of each of the following types of problems: lexical ambiguity,ambiguous referential communication, and ambiguous figures (see Appendix D for a completeprotocol for Study 3). The three types of problems of interpretation were completelycounterbalanced in six different orders. The order of the two examples of each type of task wasalternated within these six orders. The interviews were taped and later transcribed in order toscore the subjects’ responses.ReliabilityAll the subjects’ responses were rescored by a second rater who was blind to the subjects’ages and sex. Disagreements were resolved by a third rater. Overall (on a total of 864 responses)the agreement between raters was 92%; Cohen’s kappa = .81. The agreement between raters onthe individual questions: explanation, prediction, and deviant interpretation, was 91%, 88%, and95%, respectively; Cohen’s kappas = .81, .75, and .81, respectively.Results and DiscussionAs in Study 1, the central hypothesis to be evaluated was that there is an importantdistinction between the understanding of false beliefs and an insight into the interpretative natureof mind and the knowing process. In contrast to this view, other theorists (e.g., Perner, 1991)claim that false belief understanding entails understanding interpretation. Were this true it wouldfollow that all subjects passing the false belief test should also pass the tests of interpretation. Inthis study, although all the subjects easily passed the false belief test, on average the 5-year-oldsubjects passed only 36% of the explanation of interpretation questions, and 13% of the prediction41of interpretation questions. The 8-year-olds, in contrast, passed 92% of the explanation ofinterpretation questions, and 67% of the prediction questions. This evidence offers good supportfor the claim that, although children at the age of 5 easily deal with problems concerning falsebeliefs, acquiring an insight into the interpretative nature of mind is a development that does notoccur until some years later.Twenty-seven of the 48 subjects passed both the false belief test and the tests ofinterpretation (i.e., passed more than 50% of the explanation questions on the tests ofinterpretation). The remaining 21 subjects who responded differently on the two tests all passedthe false belief test while failing the interpretation tests, McNemar’s X2 (Siegel & Castellan, 1988)(1, N = 48) = 19.05, p < .001. This difference in level of difficulty between the two tests is alsosignificant when only the 5-year-old subjects are considered, McNemars X2 (1, N = 12) = 7.11,p < .01.The mean performance on the control question concerning deviant interpretations wasclose to the ceiling. On average the percentage of these questions passed by the 5-, 6-, 7-, and 8-year-olds, was 76%, 78%, 82%, and 96%. As expected, children of all ages in this study werefairly competent at recognizing and rejecting deviant interpretations, and justifying this rejectionbased on a lack of support for the deviant interpretation in the stimulus materials. The fact that thesubjects passed these control questions demonstrates that they knew the deviant interpretationswere wrong because they could not be warranted on the basis of the stimulus material in question.However, this result leaves unanswered the question of whether these subjects did or did notconceive of the possibility that two interpretations of the same stimulus could both be right. Toaddress this issue we must turn to a more detailed examination of the subjects’ responses to theexplanation and prediction questions.Preliminary analyses were first done to check for possible effects of order and sex. A 6 x3 (Order x Task) repeated measures analysis of variance with repeated measures on the last factorrevealed no significant effects, F < 1. A similar analysis with sex as the between subject factorrevealed no significant main effect for sex, or task by sex interaction, Fs < 1. To check for the42effect of the order of presentation of the three types of tasks (i.e., first, second, third) onperformance, a 4 x 3 (Age: 5, 6, 7, and 8 years x Order of Task: First, Second, and Third)repeated measures analysis of variance with repeated measures on the last factor and number ofcorrect explanations as the dependent variable was conducted. This analysis revealed a maineffect for age, F(3, 44) = 5.60, p < .002, and a main effect for order, F(2, 88) = 6.39, p < .003,which was qualified by an age x order interaction, F(6, 88) = 2.82, p <.02. An analysis ofsimple effects revealed that the improvement in performance with order was significant only forthe 6-year-old subjects, F(2, 88) = 10.73, p < .001 (see Figure 4). This effect of the order ofpresentation on the explanation question did not extend to the prediction question. A similaranalysis for the prediction question revealed no main effect for order, or interaction with age, Fs <1.Insert Figure 4 about hereBinomial tests revealed that subjects’ performance on the two instances of each type ofambiguity (lexical, communicative, and pictorial) did not differ significantly. Therefore, subjects’scores on the two instances of each type of task (e.g., the “duck-rabbit” and the “rat-man”) werecombined and mean scores across both instances were used for further analysis.The next step in the analysis of these results was to compare the subjects’ performance onthe tests of interpretations across age. To address this question a 4 x 3 x 2 (Age x Task xQuestion: Explanation vs. Prediction) repeated measures analysis of variance was conducted withrepeated measures on the last two factors and mean correct responses as the dependent variable.This analysis revealed significant main effects for age, F(3, 44) = 6.80, p < .00 1, and type ofquestion, F(1, 44) = 18.35, p < .001, with no other significant main or interaction effects, Fs <1. Performance on the prediction question was significantly lower than performance on theexplanation question. Follow-up trend analysis showed that there was a significant linear increasein performance from 5 years of age to 8, for both the explanation questions, F(1, 44) = 15.9, p. <43.001, and for the prediction questions, F(1, 44) = 13.9, p. <.001, with no significant departuresfrom linearity. These results are further strengthened by significant positive correlations betweensubjects’ total correct scores on the explanation and prediction questions (scores ranged from 0 to6, combined across all six tasks) and the subjects’ ages in months, r = .59, p < .01, and .58, p <.01, respectively.Insert Figure 5 about hereIt is not surprising that the prediction questions were harder for the subjects, even theolder subjects, than the explanation questions. Understanding the difficulty of predicting howsomeone else would interpret an ambiguous message or picture seems to require first an ability toexplain the basis for such multiple interpretations, plus the ability to make use of an implication ofthis interpretive diversity. (There also was no practice effect for the prediction questions.) Itshould be noted, in passing, that subjects who failed this question did not do so merely byattempting to make a prediction. Some of the subjects did make a prediction as to how anotherpuppet might interpret the message or picture, but went on to justify what would seem, at firstblush, to be an in appropriate response by answering in ways that were meant to go more deeplyinto the factors responsible for such differences in interpretation. For example, one subjectargued that it is more likely that someone told to wait for a pear/pair would be waiting for a pear toeat “because you didn’t say a pair of something, so it makes sense for it to be a pear.” Thus,although some subjects did make a prediction, their argument in support of such predictionsrevealed an understanding of the interpretive nature of the problem in that they reflected a searchfor differences in the stimuli that could warrant the diverse interpretations assigned to the puppetfigures.’515 When they were presented with the ambiguous figures, the last 36 subjects were asked what they thought itwas before the two images were pointed out. In response to the “rat-man” figure, 25% first saw the rat, 44% sawthe man, 28% saw something else (e.g., “frog,” “fish,” “baby chick?”), and one subject saw both the rat and theman. In response to the “duck-rabbit” figure, 61% saw the duck, 36% saw the rabbit, and one subect saw both.44The results reported above clearly show that whereas the 5-year-old subjects pass the falsebelief test, children generally do not develop competence with tests of interpretation until severalyears later. As mentioned above, a potential criticism to which conclusions of this sort are open isthat the assessment of children’s understanding of interpretation is too dependent on their ability toprovide the verbal explanations required. Thus, it could be argued that the failure of the youngsubjects in this study to provide explanations merely reflects imperfect performance due to taskcomplexity, and does not adequately assess the subjects’ underlying competence. In other words,it could be argued that children’s conceptual competence may not be revealed because ofperformance factors such as difficulties in understanding the questions or in an inability toverbally construct appropriate explanations.With respect to the ability of these young subjects’ to verbally construct appropriateexplanations regarding the problems of interpretation, it should be noted that all the subjects in thepresent study (and in Study 1) had no difficulty in providing explanations of the source ofdiffering beliefs in the false belief task. It was only when the difference in the puppets’ beliefswas the result of different interpretations of the same ambiguous stimuli that the young subjects inthis study failed to provide coherent explanations. Study 3 also incorporated minor modificationsof the questions used in Study 1 intended to make the false belief and one of the interpretationquestions very similar in structure. In Study 3 the false belief question was: “Why does Maxithink his toy is in the yellow (blue) container and Mary think it’s in the blue (yellow) container?”And the question on one of the referential communication tasks was: “Why does Maxi think thesticker is under the card with the big red block and Mary think it’s under the card with the big blueblock?” A typical 5-year-old subject’s response to the false belief question was: “Because Maxiwent outside and he put it away, and then Mary wanted to play with it and she put it away in theyellow container.” A response typical of the 8-year-olds who were scored as passing the similarquestion on the interpretation problem is: “Cause there’s two big blocks.” The point to be madehere is that a typical and perfectly adequate explanation for interpretive diversity can be shorter andsimpler than an equally adequate explanation for a similar difference in beliefs that arise due to the45puppets’ access to differential information. The mean number of words actually used by the 26subjects who passed the interpretation test (M = 19, SD = 18.3), while smaller, did not differsignificantly from the mean number of words used by these subjects in their explanations on thefalse belief test (M = 24, SD = 14.6). In framing their answers to questions about differences ininterpretation some subjects went on to speculate about why the puppets had endorsed thoseparticular interpretations (e.g., “maybe Mary likes the colour red”). If attention is directed only tothe number of words actually required to pass the two tests, then the mean number of words onthe test of interpretation (M = 9, SD = 5.8) is actually significantly lower than the mean number ofwords required to pass the false belief test (M = 24, SD = 14.6), t(25) = 4.90, p < .00 1.With respect to the issue of the younger subjects’ ability to even understand the questionson the problems of interpretation, note should be taken of the fact that the questions in Study 3were changed from the questions employed in Study 1 in order to make them easier for thechildren to understand. As discussed above, the general form of the primary explanation questionin Study 1 was, “Is it okay for Mary to say ... and Maxi to say ... ?“ The subjects of Study 1generally seemed to have understood this question, but because some of them tended to take thisquery in a moral sense, it was changed in Study 3 to “Why does Mary say she is waiting for onething and at the same time Maxi say he’s waiting for another thing?” The question was phrased inthis general way because it was found in pilot testing that when children were asked a morespecific question such as, “Why is Mary waiting for a ring for her finger and Maxi waiting for atelephone to ring?” some children attempted to offer reasons for those specific interpretations.Although these modifications served to better clarify the question in Study 3, some subjects stillgave specific answers for those particular interpretations (e.g., “Cause she wants a ring and hewants a telephone to ring,” or “probably their favourite colours”). Younger subjects seemed toassume that they were really being asked two separate questions (i.e., “Why is Mary... ?“ and“Why is Maxi ... ?“), to which they responded by offering explanations such as “she’s probablyhungry and he needs a new pair of shoes.” Such responses would be perfectly adequate formatters of taste because they concern individual differences in preference, but they do not46constitute an adequate explanation for problems involving multiple interpretation because nomention is made of the ambiguous nature of the stimulus being responded to. Thus, the youngersubjects did not tend to recognize that the real problem to be explained was the conjunction ofdifferent answers to one and the same interpretive problem.In Study 3, however, two primary explanation questions were employed in order to givesubjects a second chance and offer them every opportunity to reveal an understanding ofinterpretation. If subjects failed to give an adequate answer to the first question they were askedthe follow-up question (i.e., “Does it make sense for Mary to say one thing and Maxi to saysomething else?” and “Why does it (doesn’t it) make sense?”). In response to this secondexplanation question some subjects went on to mention the ambiguity of the message or thepicture and were scored as having passed the test item, but many of the younger subjects were notable to go beyond their initial explanation of the difference in interpretation in terms of thepuppets’ personal preferences.47CONCLUSIONSThe results of this program of research are consistent with the hypothesis that theunderstanding that knowledge is interpretive in nature is not equivalent to the understanding of thepossibility of false beliefs. Rather, as anticipated, the concept of interpretation would appear toinvolve a more complex and significantly later arriving understanding of the constructive nature ofthe knowing process than is implied by the altogether simpler insight that persons who aredifferently informed may legitimately hold different beliefs. The understanding of interpretation isalso distinct from the superficially similar issue of differences in taste. When subjects of any ageso far tested discussed matters of taste, they assumed that the locus of responsibility for suchdifferences was to be found within the persons making such differentjudgments. By contrast,when older subjects of 7 or 8 discussed contrasting matters of interpretation they laidresponsibility for such differences on the ambiguous nature of the message or picture in question.These same qualitative differences between matters of taste and interpretation were confirmed in asecond study with 15 adults. Like the children, these adults tended to treat matters of taste asdifferent in kind from problems of interpretation, locating responsibility for differences in tastewithin persons, while seeing differences in interpretation as supported by the inherentlyambiguous nature of the stimulus environment. Altogether, then, this program of research showsthat an initial grasp of the notion of personal taste is both different from and significantly earlierarriving than is an appreciation of the possibility that one and the same thing can be assigneddifferent meanings by different persons.One possible criticism of the results of this series of studies is that the 5-year-old subjectsmay have failed the problems of interpretation just because they are not as articulate as the olderchildren. That is, they may possess, but not be able to reveal that they possess, the concept ofinterpretation because they lack the words to properly explain themselves. Several lines ofevidence speak against this dismissive reading of the results presented here. First, even theyoungest subjects tested here were perfectly competent at constructing elaborate explanations to48account for counterpart differences in beliefs on the false belief task. Rather, the 5-year-olds onlyhad difficulty offering up explanations when the stimulus characters different beliefs were basedon the same information. Second, it is important to note that the primary purpose of this researchis to contrast what were hypothesized to be two distinct forms of understanding knowledge: falsebelief understanding versus understanding the interpretive nature of mind and knowledge. As theevidence presented here shows, the 5-year-old subjects’ understanding of false beliefs was not inthe least vague or inarticulate. Rather, their grip on the notion of false belief seemed crystal clear,implying, as the literature on children’s developing theories of mind would suggest, that thesechildren had most likely developed this insight some years before. That is, it was clear from theirresponse to the false belief story that even the youngest of these subjects was easily able toanticipate the outcome of the story (e.g., when Mary moved the toy while Maxi was outside, thechildren often laughed or smiled in anticipation). Nothing like this was true, however, of theyounger subjects’ general puzzlement over the fact that, against all reason, two different puppetcharacters were shown giving different responses to one and the same stimulus event.The fact that the 8-year-old subjects in the studies reported above had largely mastered thesimple problems of interpretation employed in this research does not, however, necessarily implythat these same children have a mature or adult-like understanding of the complex issue ofinterpretation. Even though the 8-year-old subjects in these studies were competent at explainingdifferences in interpretation on these very simple tasks, they were still not equally competent atdrawing the implication that it is difficult to predict how another person will interpret anambiguous stimulus. The competence demonstrated in this research is seen, then, to indicateonly a fledgling insight into the interpretative nature of the knowing process--a developmentalprocess that very likely continues through adolescence, and probably early adulthood as well(Chandler, 1987; Kuhn, Pennington, & Leadbeater, 1983; Kuhn, Amsel, & O’Loughlin, 1988;Perry, 1970, 1981).1616 A key issue to explore in future research in coming to a fuller appreciation of children’s developingunderstanding of the concept of interpretation is the matter of their knowledge regarding the limitations of thisconcept. That is, as children are beginning to develop an understanding of when they can appropriately invoke the49Limitations and Suggestionsfor Future ResearchIn demonstrating a distinction between false belief understanding and a beginning grasp ofthe more complex notion of interpretation, and in so doing providing evidence of a moredifferentiated view of the development of children’s understanding of mind than is commonlyassumed in the theories of mind literature, this thesis also brings to light a number of interestingpotential directions for future research. In what follows I will discuss several of these possibleavenues of future study as well as return to some of the issues raised in the Introduction.In an attempt to demonstrate the early appearance of a concept, one available methodologyis to determine whether or not training can improve young children’s performance on these testsof interpretation. Although such a training study would be an important advance in future effortsto better map the transition between false belief understanding, on one hand, and an achievementof an interpretive theory of mind, on the other, the present study sequence already provides somenotion of interpretation as a explanation for disagreement, do they also recognize that there are situations in whichthis possibility can not be made to fit? If a knowledge claim is to be accorded the status of a legitimate“interpretation” it must be understandably based on the object or event in question, and thus, interpretations can beevaluated in terms of whether or not they are justifiable. This question of the evaluation of interpretations isunavoidable once the issue of interpretation has been raised (Bernstein, 1988), and adolescents’ attempts to deal withthe implications of this problem and the potential for relativism has already been explored by a number of authors(e.g., Chandler, 1987; Kuhn & Leadbeater, 1983). Young school-age children would not be expected to grasp, as doadolescents, the potentially relativistic implications of the interpretative character of the knowing process, but theymay still recognize that some interpretations may be harder to judge or defend than others and that there arelimitations as to what can count as a legitimate or warrantable interpretation.One aspect of this matter of “legitimacy” that has been explored in a handful of studies is that of children’sdeveloping understanding of deviance, disorder and defense mechanisms (Chandler, Paget, & Koch, 1978; Coie &Pennington, 1976; Dollinger & McGuire, 1981; Roberts, Beidleman, & Wurteie, 1981). Tn these studies children’spotential understanding of deviance typically has been assessed by presenting them with vignettes involvinginstances of abnormal behavior, some of which evidently turn on the deviant interpretations of these storyprotagonists. The broad conclusions arising from these studies is that there are clear age-graded changes in thereadiness or ability of children to identify certain behaviors as falling outside the bounds of normalcy. None ofthese studies, however, has involved preschool children. In future research subjects could be asked more questionsabout “deviant” interpretations of the problems presented to them. Interpretations can be deviant in at least twoways. First, they may not be sufficiently based on the stimulus in question, and, therefore, could be thought of asjust wrong. Second, interpretations could be partially based on the stimulus in question but clearly drawconclusions that go beyond what can be justified in terms of the stimulus, that is, they involve confabulation.Subjects could also be questioned about their perception of the individual offering such an abnormal interpretation.The subjects’ responses to these questions will serve as an indication of the degree to which they have begun torecognize the limitations of interpretation and the possibility that interpretations that fall outside of some acceptablerange may suggest something is seriously amiss about the person offering up such a view.50evidence relevant to this matter of plasticity. As indicated, the significant interaction between ageand order of presentation of the tests of interpretation in Study 3, suggest that while training mightbe effective for 6-year-olds, and perhaps, to a lesser extent, 7-year-olds, who are alreadybeginning to acquire some insight into the interpretive nature of knowledge, the 5-year-oldsubjects, by contrast, showed no improvement across the trials. Clearly, there is some supporthere for the prospect that a carefully designed training study could advance our understanding ofthis transitional period.A second line along which future research might be carried out with profit is suggested bycertain of the methodological choices adopted in the present study sequence. A defensible, butnecessarily limiting choice made here involved presenting children with situations in which oneobject or message affords two interpretations that are both equally well supported by the availableevidence. As argued in the Introduction, the basis for this methodological choice was that, incontrast to situations involving two unequally supported alternatives, where young children mightwell discount, and thus fail to mention, interpretations that are only weakly supported by theavailable evidence, the present strategy would be less likely to produce false negative results. Analternative strategy, however, would be to employ situations that hold the potential for manydifferent interpretations, such as the Rorschach test, or situations in which the prospects for somealternative interpretations are slim, but potentially devastating in their consequences. These typesof situations may be fairly commonly encountered in children’s everyday life. Another suchpromising prospect is afforded by the study of humor. It is common, for example, for jokes toturn on the fact that a single word can have two or more meanings. The listener is set up tointerpret the word in one way, and then, when the punch line is delivered, must search for asecond meaning of the ambiguous word in order to make sense of the punch line. As mentionedin footnotes to the Introduction, earlier research on the development of children’s humor andunderstanding of riddles tends to support the conclusions of this thesis. That is, children begin toappreciate the potential for such multiple meanings at about 7 or 8 years of age. A further type ofcommonly experienced situation in which children might realize the potential for multiple readings51of one and the same events are those social occasions on which it becomes apparent that all theevidence points to an interpretation of a situation that presents oneself in a particularly bad light.In order to save face, one is then obliged to take steps to ward off the possibility of being wronglysuspected or accused (see, for example, Austin, 1961; Backman, 1985; Scott & Lyman, 1966).A third major opportunity for future research is suggested by comments made earlier inreporting on the social role-taking literature. The distinction between false belief understandingand the more complex concept of interpretation that is demonstrated in this thesis should prove tobe a useful tool in the task of re-examining the older role-taking literature. As suggested in theIntroduction, some of the role-taking tasks that dominated the research literature in the 1970s mayprove, on analysis, to be essentially false belief tasks plus something else, whereas othercounterpart measures may come closer to being true measures of an interpretive theory of mind.The work presented here has the potential, then, to serve as a bridge between the earlier literatureon social role-taking and the more recent research on children’s theories of mind. There are, Ibelieve, important insights to be salvaged from this earlier and massive research effort directedtoward coming to a better understanding of children’s developing social intelligence. In supportof the possible benefits of such efforts, some investigators working in the area of children’sdeveloping theories of mind are now turning new attention to development in children’sconception of the mind after the age of four (Flavell, Green, & Flavell, 1995; Wellman &Hickling, 1994).An additional area of concern raised by, but not followed up in this thesis is the prospectthat young children’s understanding that different people are entitled to their differences in opinionregarding matters of taste, might be an early indication of their beginning insight into theinterpretive nature of knowledge. In the first two studies reported above some evidence waspresented to show that both children and adults treat matters of taste as being qualitatively differentfrom issues of interpretation. These findings are relevant to, but go no distance toward helping toresolve, the long-standing philosophical debate concerning whether or not there is anythingfoundational or “epistemic” that would allow for an objective arbitration of differences in matters52of taste (e.g., Gadamer, 1982; Herrnstein-Smith, 1988). This research has demonstrated adistinction between beliefs about matters of fact and matters of value by showing that they developat different points in the ontogenetic course. The possibility, however, of a genetic relationshipbetween an insight into the individualized nature of differences in taste and the subsequentappreciation of the possibility of interpretation remains to be explored. Although, after Hume andMoore, many philosophers have tended to assume that the distinction between values and facts isunbridgeable (Doeser, 1986), others, such as Putnam (1987), reject a simple dichotomy and argueinstead for a continuum between values and facts. There is some suggestive evidence fromchildren’s responses to the present tests of interpretation that before such concepts are achieved,children may attempt to account for interpretational differences by resorting to explanations thatare really only appropriate for differences in taste. For example, some subjects in these studiesattempted to account for different interpretations of the ambiguous messages in terms of thepuppets’ personal preferences (e.g., “that’s his favourite colour”). With further probe questionsthe older subjects would go on to point out the ambiguous nature of the message as the likelysource of these different interpretations, but the younger subjects were unable to offer these or anyother additional explanation.One final aspect of the present results that tends to be overlooked when attention isfocused, as it was in the present study sequence, only on the mean performance of age-groups, isthe great deal of variability evident in the 6- to 7-year-old subjects’ performance on the tests ofinterpretation. This variability, which is consistent with earlier studies on children’sunderstanding of riddles that also rely on an appreciation of multiple meanings (Fowles & Glanz,1977), suggests two additional directions for future research. First, this more complex insightinto the interpretive nature of knowledge should make a difference for children in their attempts tonavigate their social world. Presumably a more mature understanding of other people would leadto success in relations with both peers and parents, teachers and other adults. This hypothesis,while given indirect support by the old literature connecting social role-taking ability to variousmeasures of interpersonal competence, still remains to be evaluated. Second, it would be53important to determine what this variability is related to in children’s background (e.g.,intelligence, parental style, number of siblings). This approach is beginning to be applied in thecase of children’s understanding of false beliefs. For example, Perner, Ruffman and Leekam(1994) have reported finding that the more siblings their young subjects had the better theirperformance on false belief tests. This suggests that social interaction is important in thedevelopment of social intelligence, but, of course, children may experience social interaction fromsources other than their siblings, such as parents or day-cares and schools. All of this researchwould be interesting and important in its own right, and could, perhaps have applied implications,but it could also be used in the service of answering more fundamental questions concerning thebest way to evaluate competing explanations of children’s knowledge of the mind. This thesis isprimarily a description of the development of children’s understanding of knowledge, but it alsoraises questions about what it is that changes with development that enables children to understandthe possibility and implications of multiple interpretations. One possibility is that rather thanmerely sequentially recognizing the possibility of alternative readings of the same stimulus,children must develop the ability to simultaneously hold two perspectives in mind in order tocoordinate these perspectives.Although many potential avenues of further exploration are raised by this thesis, theresearch presented here does clearly suggest a more differentiated view of children’s epistemicdevelopment than the view commonly endorsed in the theories of mind literature. This thesis, it ishoped, will help serve to shift the attention of the field away from a continued preoccupation withthe counterintuitive results of standard false belief tests, and promote instead a fresh line ofresearch into what preschoolers have yet to learn about the mind (Ravell, Green, & Ravell,1995), and what new insights develop during the early school years and beyond.54REFERENCESAstington, J. W., Hams, P. L., & Olson, D. R. (Eds.). (1988). Developing theories ofmind.New York: Cambridge University Press.Attneave, F. (1974). Multistability in perception. In R. Held (Ed.), Image, object, and illusion:readingsfrom Scientfic American (pp. 90-99). San Franciso: W. H. Freeman andCompany.Austin, J. L. (1961). Philosophicalpapers. Oxford: Clarendon Press.Backman, C. W. (1985). Identity, self presentation, and the resolution of moral dilemmas. In B.R. Schienker (Ed.), The selfand social bfe (pp. 261-289). New York: McGraw-I-Till.Barthes, R. (1992). The death of the author. In H. Adams (Ed.), Critical theory since Plato (pp.1130-1133). New York: HBJ. (Original work published 1960)Bartsch, K., & Weliman, H. (1988). Young children’s attribution of action to beliefs and desires.ChildDevelopment, 60, 946-964.Beal, C. R. (1988). Children’s knowledge about representations of intended meaning. In J. W.Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories ofmin4 (pp. 315-325). New York: Cambridge University Press.Bernstein, R. J. (1983). Beyond objectivism and relativism: Science, herm.eneutics, andpraxis.Philadelphia: University of Pennsylvania Press.Bernstein, R. J. (1988). Interpretation and its discontents: The choreography of critique. In S. B.Messer, L. A. Sass, & F. L. Woolfolk (Eds.), Hermeneutics andpsychological theory:Interpretive perspectives on personality, psychotherapy, andpsychopathology (pp. 87-108). New Brunswick: Rutgers University Press.Bernstein, R. J. (1991). The new constellation: The ethical-political horizons ofmodernity/postmodernity. Cambridge, MA: The MIT Press.Bernstein, R. J. (Ed). (1985). Habermas andmodernity. Cambridge, MA: The MIT Press.Bohman, J. F., Hiley, D. R., & Shusterman, R. (1991). Introduction: The interpretive turn. InD. R. Hiley, J. F. Bohman, & R. Shusterman (Eds.), The interpretive turn: Philosophy,science, culture (pp. 1-14). Ithaca, New York: Cornell University Press.Boring, E. G. (1930). A new ambiguous picture. American Journal ofPsychology, 42, 444-445.Bonitatibus, G. (1988). What is said and what is meant in referential communication. In J. W.Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories ofmind (pp. 326-338). New York: Cambridge University Press.Bugeiski, B. R. (1960). An introduction to the principles ofpsychology. New York: Holt,Rinehart, and Winston.Bugeiski, B. R., & Alampay, D. A. (1961). The role of frequency in developing perceptual sets.Canadian Journal ofPsychology, 15, 205-211.55Butterworth, G., Harris, P. L., Leslie, A. M., & Weliman, H. M. (Eds.). (1991). Perspectiveson the child’s theory ofmind. Oxford: Oxford University Press.Byrne, R. W., & Whiten, A. (Eds.). (1988). Machiavellian intelligence: Social expertise and theevolution of intellect in monkeys, apes, and humans. New York: Oxford UniversityPress.Carpendale, J. I. (1995, February). The development ofchildren’s conception of interpretation.Paper presented at “Current Directions in Theories of Mind Research,” Eugene, Oregon.Chandler, M. J. (1977). Social cognition. In W. F. Overton & J. M. Gallagher (Eds.),Knowledge and development: Vol. 1, Advances in research and theory (pp. 93-147). NewYork: Plenum Press.Chandler, M. J. (1987). The Othello effect: Essay on the emergence and eclipse of skepticaldoubt. Human Development, 30, 137-159.Chandler, M. J. (1988). Doubt and developing theories of mind. In J. W. Astington, P. L.Harris, & D. R. Olson (Eds.), Developing theories ofmind (pp. 387-413). New York:Cambridge University Press.Chandler, M. J. (1992). Early steps toward an interpretive theory ofmind. Invited address to theBritish Psychological Society, Edinburgh.Chandler, M. J., & Boyes, M. (1982). Social cognitive development. In B. B. Wolman (Ed.),Handbook ofdevelopmental psychology (pp. 387-402). Englewood Cliffs, NJ: PrenticeHall.Chandler, M. J., Carpendale, J. I., & Lalonde, C. (1995, February). interpretive turns en routeto a mature theory ofmind. Paper presented at “Current Directions in Theories of MindResearch,” Eugene, Oregon.Chandler, M. J., & Helm, D. (1984). Developmental changes in the contribution of sharedexperience to social role-taking competence. internationalJournalofBehavioralDevelopment, 7, 145-156.Chandler, M. J., & Lalonde, C. E. (1991). Surprising, miraculous, and magical turns of events.Paper presented as part of the symposium on Rationality and Early Reactions to Magic, atthe XIth biennial meetings of the International Society for the study of BehaviouralDevelopment, Minneapolis.Chandler, M. J., Paget, K. F., & Koch, D. A. (1978). The child’s demystification ofpsychological defense mechanisms: A structural and developmental analysis.Developmental Psychology, 14, 197-205.Chandler, M., & Chapman, M. (Eds.). (1991). Criteriafor competence: Controversies in theconceptualization and assessment ofchildren’s abilities. Hillsdale, NJ: Lawrence ErlbaumAssociates.Chapman, M. (1987). Inner processes and outward criteria: Wittgenstein’s importance forpsychology. In M. Chapman & R. A. Dixon (Eds.), Meaning and the growth ofunderstanding: Wittgenstein ‘s signficancefor Developmental Psychology (pp. 103-127).Berlin: Springer-Verlag.56Chapman, M. (1988). Constructive evolution: Origins and development ofPiaget’s thought.New York: Cambridge University Press.Coie, J. D., & Pennington, B. F. (1976). Children’s perceptions of deviance and disorder. ChildDevelopment, 47, 407-413.Doeser, M. C. (1986). Can the dichotomy of fact and value be maintained? In M. C. Doeser & J.N. Kraay (Eds.), Facts and values: Philosophical reflectionsfrom Western and non-Western perspectives (pp. 1-19). Dordrecht, Netherlands: Martinus Nijhoff Publishers.Dollinger, S. J., & McGuire, B. (1981). The development of psychological-mindedness:Children’s understanding of defense mechanisms. Journal ofClinical Child Psychology,10, 117-121.Enright, R. D., & Lapsley, D. K. (1980). Social role-taking: A review of the constructs,measures, and measurement properties. Review ofEducational Research, 50, 647-674.Enright, R. D., Lapsley, D. K., Franklin, C. C., & Steuck, K. (1984). Longitudinal and cross-cultural validation of the belief-discrepancy reasoning construct. DevelopmentalPsychology, 20, 143-149.Fabricius, W. V., & Schwanenflugel, P. J. (in press). The older child’s theory of mind. In A.Demetriou & A. Efidides (Eds.), The structure and development ofmind: Models,methods, and applications. Amsterdam: Elsevier.Feyerabend, P. (1988). Against method (revised edition). London: Verso.Ravel!, J. H. (1988). The development of children’s knowledge about the mind: From cognitiveconnections to mental representations. In J. W. Astington, P. L. Harris, & D. R. Olson(Eds.), Developing theories ofmind (pp. 244-267). New York: Cambridge UniversityPress.Flavell, J. H. (1992). Perspectives on perspective taking. In H. Beilin & P. B. Pufall (Eds.)Piaget’s theory: Prospects andpossibilities (pp. 107-139). Hilisdale, NJ: LawrenceEribaum Associates, Publishers.Ravel!, J. H. (1995, February). The development of children’s knowledge about thinking andconsciousness. Paper presented at “Current Directions in Theories of Mind Research,”Eugene, Oregon.Ravel!, J. H., Everett, B. A., Croft, K., & Ravel!, E. R. (1981). Young children’s knowledgeabout visual perception: Further evidence for the Level 1 -- Level 2 distinction.Developmental Psychology, 17, 99-103.Ravel!, J. H., Ravel!, E. R., Green, F. L., & Moses, L. J. (1990). Young children’sunderstanding of fact beliefs versus value beliefs. ChildDevelopment, 61, 9 15-928.Ravel!, J. H., Green, F. L., & Ravel!, E. R. (1986). Development of knowledge about theappearance-reality distinction. Monographs of the Society for Research in ChildDevelopment, 51(1, Serial No. 212).Ravel!, J. H., Green, F. L., & Flavell, E. R. (1993). Young children’s understanding of thestream of consciousness. Child Development, 64, 387-398.57Flavell, J. H., Green, F. L., & Ravel!, E. R. (1995). Young children’s knowledge aboutthinking. Monographs of the Society for Research in Child Development., 60 (1, SerialNo. 243).Ravell, J. H., Miller, P. H., & Miller, S. A. (1993). Cognitive development. Englewood Cliffs,NJ: Prentice Hall.Ravell, J. H., Speer, J. R., Green, F. L., & August, D. L. (1981). The development ofcomprehension monitoring and knowledge about communication. Monographs of theSocietyfor Research in Child Development, 46 (5, Serial No. 192).Fodor, J. A. (1992). A theory of the child’s theory of mind. Cognition, 44, 283-296.Forguson, L., & Gopnik, A. (1988). The ontogeny of common sense. In J. W. Astington, P. L.Harris, & D. R. Olson (Eds.), Developing theories ofmind (pp. 226-243). New York:Cambridge University Press.Fowles, B., & Glanz, M. E. (1977). Competence and talent in verbal riddle comprehension.Journal ofChild Language, 4, 433-452.Frye, D., & Moore, C. (Ed.). (1991). Children’s theories ofmind: Mental states and socialunderstanding. Hillsdale, NJ: Lawrence Eribaum Associates.Furth, H. G. (1969). Piaget andknowledge: Theoreticalfoundations. Englewood Cliffs: Prentice-Hall.Furth, H. G. (1987). Knowledge as desire: An essay on Freud and Piaget. New York: ColumbiaUniversity Press.Gadamer, H. G. (1982). Truth and method. New York: The Crossroad Publishing Company.Gopmk, A. (1990). Developing the idea of intentionality: Children’s theories of mind. CanadianJournal ofPhilosophy, 20, 89-114.Gopnik, A. (1993). How we know our minds: The illusion of first-person knowledge ofintentionality. Behavioral andBrain Sciences, 16, 1-14.Gopnik, A. (1995, February). Psychopsychology. Paper presented at “Current Directions inTheories of Mind Research,” Eugene, Oregon.Habermas, J. (1984). The theory ofcommunicative action: Vol. 1, reason and the rationalizationof society. Boston: Beacon Press (original work published in 1981).Hala, S. M. P. (1994). The role ofpersonal involvement in accessingfalse-beliefunderstanding.Unpublished doctoral dissertation, University of British Columbia, Vancouver.Hanson, N. R. (1958). Patterns ofdiscovery. London and New York: Cambridge UniversityPress.Hart, L. M., & Goldin-Meadow, S. (1984). The child as nonegocentric art critic. ChildDevelopment, 55, 2122-2129.Herrnstein-Smith, B. (1988). Contingencies ofvalue: Alternativeperspectivesfor critical theory.Cambridge, MA: Harvard University Press.58Hirsh-Pasek, K., Gleitman, L. R., & Gleitman, H. (1978). What did the brain say to the mind?A study of the detection and report of ambiguity by young children. In A. Sinclair, R. J.Jarvella, & W. J. M. Levelt (Eds.), The child’s conception of language (pp. 97-132).Berlin: Springer-Verlag.Jastrow, J. (1900). Fact andfable in psychology. Boston: Houghton-Mifflin.Keil, F. (1980). Development of the ability to perceive ambiguities: Evidence for the taskspecificity of a linguistic skill. Journal ofPsycholinguistic Research, 9, 219-229.Kuhn, D., Amsel, E., & O’Loughlin, M. (1988). The development of scientific thinking skills.San Diego: Academic Press.Kuhn, D., Pennington, N., & Leadbeater, B. (1983). Adult thinking in developmentalperspective. In P. Baltes & 0. Brim Jr. (Eds.), Life-span development and behavior(Vol. 5, pp. 157-195). New York: Academic Press.Kuhn, T. S. (1962). The structure ofscientific revolutions. Chicago: University of ChicagoPress.Kuhn, T. S. (1991). The natural and the human sciences. In D. R. Riley, J. F. Bohman, & R.Shusterman (Eds.), The interpretive turn: Philosophy, science, culture (pp. 17-24).Ithaca, New York: Cornell University Press.Lakoff, (1987). Women, fire, and dangerous things: What categories reveal about the mind.Chicago: University of Chicago Press.Lalonde, C., Chandler, M., & Moses, L. (1992, June). Early steps toward an interpretive theoryofmind: On recognizing the possibility of being wrong in more than one way. Paperpresented the 22nd Annual Symposium of the Jean Piaget Society, Montreal, Quebec.Mansfield, A., & Clinchy, B. (1985, June). The early growth ofmultiplism in the child. Paperpresented at the 15th Annual Symposium of the Jean Piaget Society, Philadelphia.Masangkay, Z. S., McCluskey, K. A., McIntyre, C. W., Sims-Knight, J., Vaughn, B. E., &Flavell, J. (1974). The early development of inferences about the visual percepts ofothers. ChildDevelopment, 45, 357-366.McGhee, P. E. (1979). Humor: Its origin and development. San Franciso: W. H. Freeman andCompany.Moses, L. J., & Ravell, J. H. (1990). Inferring false beliefs from actions and reactions. ChildDevelopment, 61, 929-945.Nietzsche, F. (1968). The will to power. New York: Vintage.Overton, (1994). Contexts of meaning: The computational and the embodied mind. In W. F.Overton & D. S. Palermo (Eds.), The nature and ontogenesis ofmeaning (pp. 1-18).Hilisdale, NJ: Lawrence Eribaum Associates, Publishers.Overton, W. F. (1994). The arrow of time and cycles of time: Concepts of change, cognition, andembodiment. Psychological Inquiry, 5, 215-237.59Pemer, J. (1988). Developing semantics for theories of mind: From propositional attitudes tomental representation. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.),Developing theories ofmind (pp. 141-172). New York: Cambridge University Press.Perner, J. (1991). Understanding the representationalmind. Cambridge, MA: MIT Press.Perner, J., & Astington, J. W. (1992). The child’s understanding of mental representation. In H.Beilin & P. B. Pufall (Eds.) Piaget’s theory: Prospects andpossibilities (pp. 141-160).Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.Perner, J., & Davies, G. (1991). Understanding the mind as an active information processor: Doyoung children have a “copy theory of mind”? Cognition, 39, 5 1-69.Perner, J., Ruffman, T., & Leekam, S. R. (1994). Theory of mind is contagious: You catch itfrom your sibs. Child Development, 65, 1228-1238.Perner, J., & Wimmer, H. (1985). “John thinks that Mary thinks that ...“: Attribution of second-order beliefs by 5- to 10-year-old children. Journal ofExperimental Child Psychology,39, 437-47 1.Perry, W. G. (1970). Forms ofintellectual and ethical development in the college years. NewYork: Holt, Rinehart, & Winston.Perry, W. G. (1981). Cognitive and ethical growth: The making of meaning. In A. W.Chickering (Ed.), The modern American College (pp. 76-116). San Francisco, CA:Jossey-Bass.Piaget, J. (1967) Six psychological studies. New York: Vintage. (Original work published 1964)Piaget, J. (1970). Piaget’s theory. In P. Mussen (Ed.), Carmichael’sManual ofChildPsychology (3rd ed.) (pp. 703-732). New York: Plenum Press.Pillow, B. H. (1991). Children’s understanding of biased social cognition. DevelopmentalPsychology, 27, 539-551.Pillow, B. H. (in press). Two trends in the development of conceptual perspective-taking: Anelaboration of the passive-active hypothesis. International Journal ofBehavioralDevelopment.Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? TheBehavioral andBrain Sciences, 4, 5 15-526.Price, R. (1953). Droodles. New York: Simon & Schuster.Putnam, H. (1981). Reason, truth and history. New York: Cambridge University Press.Putnam, H. (1987). The manyfaces of realism. LaSalle, IL: Open Court Publishing Company.Putnam, H. (1988). Representation and reality. Cambridge, MA: MIT Press.Putnam, H. (1990). Realism with a humanface. Cambridge, MA: Harvard University Press.60Rapaport, D. (1946). Diagnosticpsychological testing: The theory, statistical evaluation, anddiagnostic application ofa battery of tests, Volume II. Chicago: The Year BookPublishers, Inc.Reese, H. W. (1963). “Perceptual set” in young children. ChildDevelopment, 34, 151-159.Reese, H. W., & Ford, L. H. (1962). Expectancy and perception of an ambiguous figure inpreschool children. Journal ofVerbalLearning and VerbalBehavior, 1, 188-191.Roberts, M. C., Beidleman, W. B., & Wurteie, S. K. (1981). Children’s perceptions of medicaland psychological disorders in their peers. Journal ofClinical Child Psychology, 10, 76-78.Robinson, E. J., & Robinson, W. P. (1983). Children’s uncertainty about the interpretation ofambiguous messages. Journal ofExperimental Child Psychology, 36, 81-96.Robinson, E. J., & Whittaker, S. J. (1985). Children’s responses to ambiguous messages andtheir understanding of ambiguity. Developmental Psychology, 21, 446-454.Rock, I., Gopnik, A., & Hall, S. (1994). Do young children reverse ambiguous figures?Unpublished manuscript.Rouse, J. (1991). Interpretation in natural and human science. In D. R. Hiley, J. F. Bohman, &R. Shusterman (Eds.), The interpretive turn: Philosophy, science, culture (pp. 42-56).Ithaca, New York: Cornell University Press.Ruffman, T., Olson, D. R., & Astington, J. W. (1991). Children’s understanding of visualambiguity. British Journal ofDevelopmental Psychology, 9, 89-102.Russell, J., Mauthner, N., Sharpe, S., & Tidswell, T. (1991). The ‘windows task’ as a measureof strategic deception in preschoolers and autistic subjects. British Journal ofDevelopmental Psychology, 9, 331-349.Schwanenflugel, P. J., Fabricius, W. V., & Alexander, J. (1994). Developing theories of mind:Understanding concepts and relations between mental activities. Child Development, 65,1546- 1563.Schwanenllugel, P. J., Fabricius, W. V., Noyes, C. R., & Bigler, K. D. (in press). Theorganization of mental verbs and folk theories of knowing. Journal ofMemory andLanguage.Scott, M. B., & Lyman, S. M. (1966). Accounts. American Sociological Review, 33, 46-62.Searle, J. R. (1969). Speech acts: An essay in the philosophy of language. New York: CambridgeUniversity Press.Searle, J. R. (1983). Intentionality: An essay in the philosophy ofmind. Cambridge: CambridgeUniversity Press.Shantz, C. V. (1983). Social cognition. In J. H. Flavell & E. M. Markman (Eds.), Handbook ofchild psychology: Cognitive development (Vol. 3, p. H. Mussen, General Ed., pp. 495-555). New York: Wiley.61Shultz, T. R. (1974). Development of the appreciation of riddles. Child Development, 45, 100-105.Shultz, T. R., & Pilon, R. (1973). Development of the ability to detect linguistic ambiguity. ChildDevelopment, 44, 728-733.Shusterman, R. (1991). Beneath interpretation. In D. R. Riley, J. F. Bohman, & R. Shusterman(Eds.), The interpretive turn: Philosophy, science, culture (pp. 102-128). Ithaca, NewYork: Cornell University Press.Siegel, S., & Castellan, N. J. (1988). Nonparametric statisticsfor the behavioral sciences. NewYork: McGraw-Hill Book Company.Sodian, B. (1988). Children’s attributions of knowledge to the listener in a referentialcommunication task. Child Development, 59, 378-385.Sodian, B. (1990). Understanding verbal communication: Children’s ability to deliberatelymanipulate ambiguity in referential messages. Cognitive Development, 5, 209-222.Sullivan, K., Zaitchik, D., & Tager-Flusberg, H. (1994). Preschoolers can attribute second-orderbeliefs. Developmental Psychology, 30, 395-402.Taylor, M. (1988a). Conceptual perspective taking: Children’s ability to distinguish what theyknow from what they see. ChildDevelopment, 59, 703-718.Taylor, M. (1988b). The development of children’s understanding of the see-knowing distinction.In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories ofmind (pp.207-225). New York: Cambridge University Press.The Oxford English Dictionary, second edition (1989). Oxford: Clarendon Press.Varela, F. J., Thompson, E., & Rosch, E. (1991). The embodiedmind: Cognitive science andhuman experience. Cambridge, MA: The MIT Press.von Glasersfeld, E. (1982). An interpretation of Piaget’s constructivism. Revue Internationale dePhilosophie, 142-143, 612-635.von Glasersfeld, E. (1988). The reluctance to change a way of thinking. The Irish Journal ofPsychology, 9, 83-90.von Glasersfeld, E. (1984). An introduction to radical constructivism. In P. Watzlawick (Ed.),The invented reality (pp. 17-40). New York: Norton.Wellman, H. M. (1988). First steps in the child’s theorizing about the mind. In J. W. Astington,P. L. Harris, & D. R. Olson (Eds.), Developing theories ofmind (pp. 64-92). New York:Cambridge University Press.Wellman, H. M. (1990). The child’s theory ofmind. Cambridge, MA: The MIT PressWellman, H. M., & Hickling, A. K. (1994). The mind’s “I”: Children’s conceptions of the mindas an active agent. Child Development, 65, 1564-1580.Wellman, H. M., & Woolly, J. D. (in press). From simple desires to ordinary beliefs: The earlydevelopment of everyday psychology. Cognition.62West, H., & Abravanel, B. (1972). Evidence for class-concept mediation of perceptual sets inpreschool children. ChildDevelopment, 43, 1242-1248.Whiten, A. (Ed.). (1991). Natural theories ofmind: Evolution, development and simulation ofeveryday mind-reading. Oxford: Basil Blackwell.Wimmer, H., & Hartl, M. (1991). Against the Cartesian view on mind: Young children’sdifficulty with own false beliefs. British Journal ofDevelopmental Psychology, 9, 125-138.Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining functionof wrong beliefs in young children’s understanding of deception. Cognition, 13, 103-128.Winner, E., & Leekam, S. (1991). Distinguishing irony from deception: Understanding thespeaker’s second-order intention. British Journal ofDevelopmental Psychology, 9, 257-270.Zahn-Waxler, C., Radke-Yarrow, M., & Brady-Smith, J. (1977). Perspective-taking andprosocial behavior. Developmental P.sychology, 13, 87-88.FIGURES63C H Ct I 0IFIGURE 2. The “Rat-Man”6566• TASTEA INTERPRETATIONFIGURE 3. Study 1. Proportion Correct Responses to the tests of Taste and Interpretation byAge-Group.1 .00.8S 0.60.400.20.05-6 YEARS 7-8 YEARSAGE-GROUP67FIGURE 4. Study 3. Proportion Correct Responses by Age and Order of Presentation.1 .oILU 0.8-00.6-00.4-00.2-o.o• FIRSTSECONDTHIRD5 6 7AGE IN YEARS8IQuJ0Qz0I0068FIGURE 5. Study 3. Proportion Correct Responses by Age and Question Type.1 .00.80.60.40.20.00 EXPLANATION• PREDICTIONI • I • I •5 6 7 8AGE IN YEARSAPPENDICES6970APPENDIX A: Conceptual and Theoretical Background, and Empirical Evidencefor the Study ofChildren’s Understanding ofInterpretationInterpretation andRepresentationPerner (1991) concludes that passing a false-belief test is a demonstration of aninterpretive understanding of mind. This conclusion is reached, it would seem, because Pernerbegins with the assumption that the mind is necessarily representational in nature. That is,knowledge is seen by him to consist of mentally representing objects in a pregiven external world.From this assumption about the nature of the mind and the process of knowledge acquisition, theproblem simply becomes a matter of charting young children’s developing insight into the natureof the representational process. The end-point of this development is the possession of a“representational theory of mind”. Since the false belief test is read as a clear demonstration of anunderstanding of the process of representation, children who pass this test are assumed to havedemonstrated an understanding of metarepresentation. That is, they demonstrate that theyunderstand the nature of the representing process, and thus, they must also understandinterpretation (Perner, 1991).According to Perner (1991), “a representation represents something as being a certainway” (p. 19). Further characteristics of the representing relation between medium and contentfollow from this definition, and one of these characteristics is misrepresentation. Perner statesthat “for any representation it is possible to misrepresent. For instance, using a flash oftenproduces photos in which people have red eyes. Such a photo would misrepresent your beautifulblue eyes as being red” (p. 20). To illustrate the idea that a false belief is a misrepresentation,Perner uses an example of a general who falsely believes that a soldier is on duty. This belief“misrepresents the real situation in the field, and ... is characterized by the divergence between thereal situation it represents (referent) and how it represents that situation as being (sense)” (p. 30).For Perner, an essential characteristic of representation is interpretation, and when young children71demonstrate their metarepresentational competence by showing their understanding of false belief,it necessarily follows that they must also understand interpretation. Once a child has achieved anunderstanding of metarepresentation the “child can understand cases of misrepresentation byseparating sense (interpretation) from referent” (Perner, 1991, p. 284). Thus, for Pernerinterpretation is just misrepresentation.Although misrepresentation may be an aspect of some uses of the concept ofinterpretation, it is really just about different ways of getting it wrong. Implicit in the assumptionthat the mind is representational is the assumption of a pregiven external world, and the mind’stask is to accurately represent it internally. Of course, there may be mistakes in the process ofrepresenting the world, and these mistakes constitute misrepresentation or interpretation. Sincefalse beliefs are misrepresentation it would appear to follow logically that if a child demonstratesan insight into the representing process, by passing a false belief test, this child must alsounderstand interpretation.There are two interrelated aspects that I wish to question about this line of argument.First, Pemer equates interpretation with misinterpretation. That is, by his reckoning there aremany ways to get things wrong but there is only one way to get it right. This is because, in hisaccount, there is a pregiven reality which one is trying to accurately represent. Second, thisunderstanding of interpretation is tied to an understanding of the nature of the mind and theknowing process as representational. There are two senses of representation at work here. Firstthere is the notion of representation as construal. This is the relatively uncontroversial claim that“cognition always consists in construing or representing the world a certain way” (Varela,Thompson, & Rosch, 1991, p. 134). Although Perner sometimes uses representation in this wayhe also relies upon a second stronger sense of representation in which we assume that “the worldis pregiven, that its features can be specified prior to any cognitive activity. Then to explain therelation between this cognitive activity and a pregiven world, we hypothesize the existence ofmental representations inside the cognitive system” (Varela, Thompson, & Rosch, 1991, p. 135).72Furth (1969) showed that this is the “crux of the empiricist position: knowledge has its adequatesource in external reality or external actions and resides in internal re-presentations” (p. 81).Although the assumption that the mind is representational in the second sense outlinedabove, is clearly the dominant view in cognitive science it is not above question, and Perner(1991, chapter 5) acknowledges that whether or not the mind is best thought of as representationalis controversial. This representational or computational view of the mind is being increasinglyquestioned (e.g., Overton, 1994a, b; Varela et al., 1991; Furth, 1987; Glasersfeld, 1979, 1982,1984, 1988; Lakoff, 1987). For example, Overton (in press) claims that “interpretation is anecessary feature of all human activity from the reflex of the infant to the reflective abstractcategories of the epistemologist” (in press, p. 8). This position on interpretation fromembodiment theory or constructivistic epistemology converges with the approach to interpretationand the rejection of the correspondence theory of truth and the representational theory ofknowledge in recent philosophy (e.g., Putnam, 1987, 1988; Rorty, 1979, 1991).While the difference of opinion outlined above will not be decided by empirical researchreported in this thesis, I have introduced the idea of interpretation and the argument that the mindis constructivistic rather than representational in nature in order to show its relationship to theclaim that a commonsense understanding of the mind includes some notion that knowledge isinterpretive. Although I have presented arguments for a constructivistic view of knowledge, itshould not be necessary to resolve this ongoing debate before empirically exploring the possibilityof locating a commonsense understanding of the interpretive nature of knowledge in children’sand adults’ understanding of the mind. I do not mean to argue that most adults hold to an explicitconstructivistic theory of mind in which insight into the constructivistic nature of knowledge isconsistently applied. But I do propose that most adults achieve some commonsenseunderstanding that the same object or event can be legitimately interpreted in different ways, andthat this insight is different from and later developing than the understanding that beliefs can bemistaken.73In the next section I will turn to a brief review of some of the literature on the theory ofinterpretation in order to provide some intellectual context for this thesis.The Concept ofInterpretationBernstein (1988), in a discussion of Freud, remarked that the 20th century could belabeled “The Age of Interpretation” (p. 88). The importance of the concept of interpretation in thecurrent intellectual climate is reflected in the fact that a simple computer search for books with“interpretation” or “interpretive” as a title word revealed over 3,000 such books in the Universityof British Columbia’s library collection (not including older titles still in the card catalogue). Yeteven this large number of volumes concerned with the concept of interpretation significantlyunderestimates the importance of this concept in the current intellectual climate because it does notinclude book chapters, articles or titles referring to hermeneutics. And, of course, many of theimportant works on interpretation do not include the word in the book’s title. Perhaps the mostsignificant example of this is Gadamer’s (1982) magnum opus, “Truth and method”, which is oneof the most important books on interpretation. However, this crude measure of the interest in theconcept of interpretation does reveal how, in one form or another, the notion of interpretationpermeates almost all fields of inquiry. A sample of the topics collected in this library searchinclude justice and legal theory, music, literary studies, architecture, culture, religion,environmental studies, politics, history, film, geology, medicine, sociology, anthropology,psychology, psychoanalysis, astronomy, translation, poetry, and, of course, philosophy.A discipline that has been greatly influenced by the “epistemology of interpretation” isanthropology. Crapanzano (1992), in a series of essays, frequently characterizes theanthropologist as Hermes, in Greek mythology the messenger of the gods. Hermes, when heagreed to be Zeus’ messenger, promised not to lie, but he did not promise to tell the whole truth(Crapanzano, 1992, p. 45). The anthropologist, as a messenger sharing in “Hermes’ dilemma”,must make the message “relevant” to the listener, yet this will change the message. Carrying the74message necessarily involves interpretation because the anthropologist cannot simply repeat whatshe has heard, she must understand it, and thus, must translate and interpret it.Interpretation is, of course, a central concern in literary criticism, and an important debatein the philosophy of literary theory concerns the relevance of the author’s intentions for theinterpretation of literary texts (Iseminger, 1992). From the traditional intentionalist position it isassumed that the text means what its author intended. In opposition to intentionalism, Barthes(1960/1992) has proclaimed “the death of the author”, and Beardsley has argued for “the authorityof the text”, by which he means that an interpretation may be independent of what the authormeant when composing the text. In what Beardsley (1970/1992) termed an “intentionalistbacklash”, Hirsch (1967/1992) has argued, “in defense of the author”, that the meaning of a textis determined by the author’s intentions, and Knapp and Michaels (1985/1992) maintain that it isimpossible to have intentionless meaning. Shusterman (1992) acknowledges that intention isnecessary for meaning, but he points out that this intention is not necessarily the author’s, it couldalso be on the part of the reader.In an inquiry concerned with interpretation I must consider, if only briefly, the extensivehistory of this concept in many diverse fields, and especially in modem philosophy. In the nextsection I will briefly review the history of the concept of interpretation in philosophy.“The Interpretive Turn”: Interpretation inModern PhilosophyIn an introduction to a collection of papers on the “interpretive turn” in philosophy,Bohman, Hiley and Shusterman (1991) acknowledge that the more “philosophy and theinterpretive disciplines proclaim the importance of interpretation in all inquiry, the less there isagreement about what it is, what interpretive practices presuppose, and how to judge interpretivesuccesses and failures” (p. 1). To understand these disagreements it will be helpful to approachthe concept of interpretation in philosophy from an historical perspective.Originally, hermeneutics referred to the interpretive discipline that emerged in the 17thcentury concerned with the proper interpretation of biblical scripture. The goal of this discipline75was to interpret these biblical texts in order to uncover the message from God that was assumed tobe contained in the Bible (Meichenbaum, 1988; Solomon, 1985). The term “hermeneutics” comesfrom Hermes, the messenger of the Greek gods. To accomplish his task of carrying messagesfrom the gods to mortals, Hermes had to be able to understand and interpret the gods’ messagesso that he could translate them and convey the messages to the mortals (Mueller-Vollmer, 1985).The term was brought into philosophy by Dilthey and it now has a much broader meaning.Dilthey, who brought the hermeneutic tradition to the interpretation of texts in general, maintainedthat historical events must be interpreted in their context (Meichenbaum, 1988). Hermeneuticsnow has a much broader meaning than the practice of interpreting religious texts, and it has beendescribed as the art of interpretation. This approach is largely based on the work of Heideggerand his student Gadamer. Hermeneutics was first applied in the social sciences because it wasassumed that understanding human action required interpretation, whereas the natural scienceswere thought to be in the business of explanation. The hermeneutical approach is in clearopposition to the positivist philosophy of science that has dominated the natural sciences for muchof this century. Dilthey, and others, resisted the imperialistic spread of positivism to the humansciences and advocated hermeneutics instead (Woolfolk, Sass & Messer, 1988).It was thought that the distinction in philosophy between explanation and interpretationformed a clear demarcation between the natural sciences and the human sciences. The goal of thenatural sciences was assumed to be explanation, that is, achieving an explanation of the natural,pre-given world. The human sciences, on the other hand, were thought to be based oninterpretation because meaning must be attributed to human action. The rejection of thisdemarcation between the natural and human sciences came from within the philosophy of thenatural sciences. Hanson (1958), Kuhn (1962), Feyerabend (1988) and others rejected thepositivistic claim of the “neutrality of observation” and the “giveness” of experience--i.e., theindependence of data from theory. Instead they argued that “data is theory laden”. With the claimthat data is already theory laden, even the natural sciences are now seen to involve interpretation76and the distinction between explanation and interpretation as a way to separate the natural andhuman sciences no longer holds.If there is no longer a clear demarcation between the natural and human sciences, and ifinterpretation is involved in all inquiry, then “is interpretation the only game in town?” (Bohman,Hiley, & Shusterman, 1991). This “hermeneutic universalism” (Shusterman, 1991) is assumedby some philosophers, such as Gadamer who claimed that “all understanding is interpretation”(1982, p. 350), and Nietzsche (1968, para 481) who wrote that “facts are precisely what there isnot, only interpretations”. In response to this claim of “hermeneutic universalism”, Taylor (1980)remarked that he and other “old-guard Diltheyeans, their shoulders hunched from years-longresistance against the encroaching pressures of positivist natural science, [will now] suddenlypitch forward on their faces as all opposition ceases to the reign of universal hermeneutics” (p.26).However, others, such as Shusterman (1991), believe that interpretation is a deliberate orreflective activity that is only required when we are faced with a problem situation. Similarly,Kuhn believes that interpretation is called for only when we are “confronted with texts or practicesthat are unfamiliar or puzzling” (Rouse, 1991, p. 45). Shusterman (1991), in his attempt to arguethat there is something “beneath interpretation”, used the concept of interpretation in a way thatmay be closer to a commonsense understanding than the current claim in philosophy that it is“interpretation all the way down”. Shusterman argues that interpretation implies some “deliberateor at least conscious thinking, whereas understanding does not” (p. 114). Interpretation,according to Shusterman, also characteristically involves a problem-situation (p. 126).Historically, a major division has been assumed between the natural and the humansciences (Naturwissenschaften and Geisteswissenschaften). Interpretation was thought to applyonly to the human sciences, whereas the natural sciences were assumed to be in the business ofexplanation because natural science deals with empirically verifiable fact. Historically,hermeneutics was applied to the human or interpretive disciplines, and this demarcation betweenthe scientific enterprise and the interpretive disciplines has been reinforced by positivistic77philosophy of science for much of this century (Bohman, Hiley & Shusterman, 1991). It hasonly been in the last 20 or 30 years that philosophers of the natural sciences, such as Hanson(1958), Kuhn (1962) and others, have begun to argue that even the domain of what was oncethought to be the cold, hard, objective facts of the natural sciences are also interpretive.Even though it is now generally acknowledged that knowledge in the natural sciences alsoinvolves interpretation the issue of whether or not Geisteswissenschaften andNaturwissenschaften are different kinds is still debated. Rorty (1980) believes that “the demise oflogical empiricism means that there is no interesting split between the Natur- andGeisteswissenschaften” (p. 39). On the other hand, Dreyfus (1980) and Taylor (1980) argue thatimportant differences remain between the human and natural sciences. Kuhn (1991) agrees withTaylor that the natural and the human sciences are different and that a line can be drawn betweenthem, but he differs in how that line should be drawn. Kuhn (1991) argues that interpretationapplies in the natural sciences just as in the human sciences, because “no more in the natural thanin the human sciences is there some neutral, culture-independent, set of categories within whichthe population--whether of objects or of actions--can be described” (p. 21). Thus, the “normalscience” of a particular period is based on a hermeneutic framework, or a “paradigm”, a term thatKuhn introduced but now seldom uses, “having totally lost control of it” (Kuhn, 1991, p. 22).However, Kuhn argues that what practitioners of natural science do seldom involves hermeneuticsbecause most of their work is “normal science” within a “paradigm” that has been passed on tothem from their teachers. Thus, according to Kuhn (1991), the natural sciences require ahermeneutic base, but they “are not themselves hermeneutic enterprises” (p. 23). In contrast tothe natural sciences, Kuhn argues that in the social sciences “new and deeper interpretations arethe recognized object of the game” (p. 23).Thus, although the interpretive nature of the subject matter of the human sciences (i.e.,human action) is recognized, the question of whether or not this applies to the facts of naturalscience is still debated by experts. It is possible that the development of children’s growing78awareness of what domains of knowledge are open to interpretation parallels, in some roughfashion, the course of this awareness in the history of philosophy.One implication of an understanding of interpretation is that interpretations must bewarranted. That is, to count as legitimate, an interpretation must be based on evidence and goodreasons. It is to this issue of the evaluation of interpretations that I turn to in the next section.Deviant interpretationsBernstein (1988) points out that the present concern with interpretation raises a questionthat “resonates throughout most of modern philosophy” (p. 89), that is, how are interpretations tobe evaluated? Bernstein rejects the possibility of any rigorous set of standards or rules with whichto judge interpretations. But he does not believe that this rejection of any permanent “legitimizingmatrix of cognitive evaluations “(p. 90) leads inevitably to the conclusion that all interpretation,and thus all knowledge, is trapped in relativism. Rather, Bernstein (1988) argues for an elaborate“choreography of critique”; a “logic of argumentation and counter argumentation” (p. 90) withwhich it is possible to critically evaluate interpretations, because these interpretations must besupported with evidence and good reasons. Thus, he concludes that the refutation of objectivismdoes not necessarily lead to relativism (Bernstein, 1983).It is beyond the scope of this thesis to further explore the moral and political implicationsof this fundamentally important question of if or how alternative interpretations can be evaluatedthat is at the centre of many debates in modern philosophy, such as the modern-postmoderndebate (Bernstein, 1985, 1991). But I raise this issue because, as Bernstein (1988) points out, itis a question that cannot be avoided. And, also, because even in a common sense understandingof interpretation there must be some understanding of the limitations of interpretation--i.e., someinsight that not all interpretations are equally valid.The necessity for some insight into the limits of interpretation in order to be judged asunderstanding interpretation can be illustrated with reference to the understanding of belief. Anadequate understanding of the concept of belief includes the understanding that beliefs can be79mistaken. If it is evident that someone assumes that beliefs always and necessarily match the stateof the world we would have to say that he or she fails to understands the nature of belief. Just asa complete understanding of belief entails the idea that beliefs are sometimes false, an adequateunderstanding of interpretation includes the idea that interpretations can be more or less adequate.That is, interpretations are not just made up, they are not equivalent to guesses, and they must begrounded in evidence and good reasons. Thus, in order to be judged as having achieved somebeginning understanding of interpretation children should be able to evaluate interpretations andstate which interpretations are good and which are “silly”. That is, they should have the ability todetermine if interpretations could potentially be supported with evidence and reasons if the claimto validity is challengedHow do people evaluate alternative interpretations? How do people judge whether twoalternative interpretations are both legitimate or whether one of them is deviant? To address thisquestion I will appeal to Bernstein’s (1988) discussion of this issue in philosophy. As notedabove, Bernstein argued that although there may be no set of rules for judging interpretations, it isnonetheless possible to evaluate alternative interpretations through a process of logicalargumentation and counter argumentation, or a “choreography of critique”. That is, it should bepossible to support interpretations with good reasons or evidence, Of course, Bernstein sharesthis insight with Habermas. In Habermas’ (1981/184) theory of communicative rationality hedistinguishes communicative rationality from cognitive-instrumental rationality. In instrumentalrationality the relationship between the subject and the object is of primary importance and thisform of rationality refers to the instrumental mastery of the environment. This form of rationalitytends to dominate modern life and Habermas has referred to this domination as the pathology ofmodern life. But he argues that this is not all there is to rationality. In a communicative contextwe regard someone as rational if they are able to support an assertion, if it is challenged, withevidence and good reasons. When people make assertions they implicitly make a claim tovalidity, and if that claim is challenged they would be expected to justify their statement with achain of good reasons.80It is this process through which the validity of interpretations can be evaluated that allowsus to distinguish legitimate interpretations from deviant interpretations. A deviant interpretationcannot be supported with good reasons and evidence. The distance between the interpretationoffered and the object being interpreted is simply too far, that is, the interpretation lacks sufficientgrounds. And, if challenged, its validity cannot be supported through a “choreography ofcritique”. This shows that the difference between valid interpretations and deviant interpretationsis qualitative rather than quantitative because the difference concerns the process of reasoningrather than the content of reasoning.A similar conclusion can be reached with Searle’s language of “direction of fit” or with theolder Piagetian concepts of assimilation and accommodation. A deviant interpretation involves toomuch assimilation, there are just not enough grounds for the interpretation. There is no basis for a“choreography of critique”, no way to support a validity claim if it is challenged. Just as a balancebetween assimilation and accommodation is necessary, so too is a balance required in terms of“direction of fit”. If a statement is based completely on a “world to mind direction of fit” (i.e., theworld is made to fit the mind) then it cannot form part of a rational discourse with other peoplebecause there are not sufficient grounds for others to recognize that interpretation as legitimate.Theories ofMind and Social Perspective-TakingThe recent interest in children’s “theories of mind” appears to be addressing similarquestions to that addressed in the earlier literature on social perspective taking. This earlierliterature was also concerned with the development of children’s understanding of persons andrelationships (e.g., Chandler, 1977; Selman, 1976, 1980). Social perspective-taking generated agreat deal of interest and research in the early 1970s, yet it is now almost totally eclipsed by therecent work on children’s “theories of mind”. What happened to this earlier research tradition,and what is its relationship with the theories of mind research? For some developmentalpsychologists the theories of mind bandwagon is an industry that is founded on repackagingearlier research on perspective-taking, and it is sometimes suggested that this “old wine in new81bottles” has little to offer in the way of fresh insights regarding the development of children’ssocial intelligence. However, for some commentators there is a continuity between the literatureson perspective-taking and children’s theories of mind. Since Flavell’s career spans theintervening decades between the 1960s and the early 1990s, he tells a fairly continuous story fromperspective taking research in the 1960s, and 1970s to the theories of mind work in the 1980s and1990s (Flavell, 1992). And Flavell asks the question that, given the fact that the recent researchon theories ofmind and the work on metacognitive and social-cognitive development all concernthe child’s folk psychology, how are we to integrate the three literatures?On the other hand, Perner (1991; Perner & Astington, 1992) is far more dismissive of theearlier Piagetian inspired research. He makes a sharp distinction between an explanation ofchildren’s social intelligence in terms of the possession of a theory of mind versus the notions ofegocentrism and role-taking. Perner and Astington (1992) characterize the Piagetian position asCartesian in orientation, that is, it is assumed that the mind is transparent to itself and weunderstand our mental states through direct introspection. Thus, “children should have littledifficulty understanding their own minds [and] the principal developmental difficulty should lie intrying to figure out other people’s minds” (p. 151). Empirical evidence it then offered that favorsthe theory view rather than the Cartesian view (Gopnik, 1990, 1993; Perner, 1991; Perner &Astington, 1992; Wimmer & Hartl, 1991).Some background is needed to provide a context for this discussion. The research onrole-taking was largely inspired by Piaget’s theory, and in particular his concept of “egocentrism”.However, Chandler and Boyes (1982) suggest that “PiageVs theory appears to have functionedmore as a springboard to, than a solid theoretical foundation for” (p. 389) most of the role-takingresearch. The concept of egocentrism was often taken as a “fact” of childhood that could bestudied in a theory-neutral way, rather than as a theoretical characterization of children’s thinking(Chandler & Boyes, 1982). The construct of “egocentrism” refers to a lack of differentiationbetween the self and other people (Shantz, 1983). And, as Shantz (1983) points out, egocentrismtended to be equated with a lack of role-taking ability. However, the ability to differentiate82between self and other may be only one of the abilities required to take other people’s roles(Shantz, 1983).The social perspective-taking approach encountered difficulties when it became apparentthat different measures of perspective taking did not correlate well and it seemed that differentmeasures were assessing different competencies (e.g., Borke, 1971, 1972; Chandler &Greenspan, 1972; Urberg & Docherty, 1976). Although one proposed solution to this empiricalproblem of a lack of correlation between different measures of role-taking was the suggestion thatthis problem could be solved by more rigorous measurement procedures (Enright & Lapsley,1980), others believed that the problem could be traced to a lack of theoretical clarity concerningthe concept of egocentrism (Shantz, 1983). The assumption that egocentrism is a unitaryconstruct is usually attributed to Piaget. And Piaget did not distinguish different types ofegocentrism, but he did use the term in varying ways in different contexts, and a reading of themultiple uses of egocentrism can convey some of the complexity packed into this single concept.Piaget (1923/1955) first used the term egocentrism in his book, The language and thought ofthechild. In the studies reported in this early book, Piaget categorized children’s speech as eitheregocentric or socialized. According to Piaget, “this talk is ego-centric, partly because the childspeaks only about himself, but chiefly because he does not attempt to place himself at the point ofview of his hearer” (Piaget, 1923/1955, p. 32). Piaget found that about 45% of the language ofthe 6-year-old children in his sample was egocentric, and this percentage declined after about theage of 7 or 8. However, it is not clear that these children completely lacked the ability to adapttheir speech to the needs of others because 55% of their speech was socialized in this way. Asecond clue that what Piaget was referring to was not just a lack of ability, but also a lack of efforton the child’s part, is Piaget’s perhaps humorous or cynical suggestion that this form of speechalso occurs in adults. The child “feels no desire to influence his hearer nor to tell him anything:not unlike a certain type of drawing-room conversation where every one talks about himself andno one listens” (Piaget, 1923/1955, p. 32).83Piaget (193211965) also discussed egocentrism in his study of moral thought. Hereegocentrism occurs in the context of a morality of constraint between children and adults becausethe inequality between them leads to a unilateral respect which prevents the children from takingthe adults’ perspective, and the children effectively “remain imprisoned in their own perspective”(Chapman, 1988, p. 62). On the other hand, with peers, the child is able to take another child’sperspective, and thus, is able to achieve a morality of cooperation. Here the use of egocentricthought appears to be influenced by the social context, and, perhaps, the child’s rights inparticular social situations. The age norms for the decline of egocentric thought in the context ofmorality are later than for the egocentric speech reported earlier (Piaget, 1923/1955).Yet another example of Piaget’s use of the concept of egocentrism can be found in hisdiscussion of adolescent thought. Piaget used the term “Messianism” to characterize adolescentthought, because, according to Piaget, “the adolescent in all modesty attributes to himself anessential role in the salvation of humanity and organizes his life plan accordingly” (Piaget,1964/1967, pp. 66-67). These three examples of Piaget’s use of the concept of egocentrism,while not even considering the far more common use of this concept in the context of concrete-operational reasoning such as solving the “three-mountain” task, give some indication of thecomplexities that are packed into one concept. However, this discussion is not meant to suggestthat we can look to Piaget to unravel these complexities. I merely suggest that even in Piaget’sintuitive use of the concept of egocentrism it may not have been the unitary construct that is wastaken for in the role-taking literature. It remains to be a task inherited by the theories of mindliterature to more clearly specify the nature of the knowledge or abilities that underlie these variousmanifestations of role-taking.It is hoped that this thesis may serve as a bridge between the earlier research on socialperspective-taking and the more recent work on children’s theories of mind, because both of theseliteratures have much to offer. The theories of mind work offers a more differentiated approach tothe abilities required to pass various measures of perspective taking. On the other hand, the socialperspective-taking literature offers a far richer and more complex description of the development84that awaits children after they have achieved the insight that beliefs can be false than can be foundin most work on theories of mind (Chandler, 1988).The focus of the theory of mind approach seems to have been captured by the verycompelling empirical demonstration that young preschoolers seem to lack a concept of false belief.As mentioned above, this is read as a clear demonstration of an understanding that the mind isrepresentational in nature, which is taken to entail an understanding of interpretation, even at theage of 4 years. Yet theorists working within the social perspective-taking approach did notattribute an understanding that two people may interpret the same thing in different ways tochildren at this young age. And theorists such as Chandler and Selman described many otherinsights into the nature of persons and relationships that are achieved by children after the age offour. Selman (1976, 1980) described five levels of social perspective taking competence, andChandler (1987, 1988) discussed the development of children’s understanding of the nature ofknowledge in terms of a movement from “retail doubt” to “wholesale doubt”, with several stationsbeyond this in efforts to deal with the insight that doubt permeates knowledge.However, it would be fair to say that the great interest generated by both the role-takingand the theories of mind literatures reflects the intuition that the question of children’sunderstanding of other people is of fundamental importance. One source of this interest inchildren’s developing social intelligence can be found in Piaget’s (1923/1955, 1924/1928,1932/1965) early work. From another perspective it could be argued that since humans livetogether in social groups, the ability to successfully navigate one’s social environment may havebeen a more powerful selection factor in the evolution of intelligence than the physicalenvironment (Byrne & Whiten, 1988). This is not to imply that social knowledge of persons is inany way innate, as Fodor (1992) would have it, rather that humans must have the ability todevelop this knowledge if they are to be successful, because our environment is so fundamentallysocial in nature. This just reinforces the intuition that may be responsible for the great popularityin developmental psychology of both the social perspective taking and the theories of mind85literature, that we cannot underestimate the importance of children’s ability to understand the talkand action of other people.In the research reported here, I am attempting to assess only a first insight intointerpretation, and not the more developed insights that are achieved by older children andadolescents. Thus, I will be exploring the point at which children first begin to appreciate thepossibility of “retail doubt” (Chandler, 1987, 1988), or in Selman’s (1976, 1980) terms thiswould be the difference between level 0 and level 1.Evidence Regarding Children’s Understanding ofInterpretationIt is now clear that there are conflicting claims regarding the age at which children firstbegin to show an understanding of the notion of interpretation. Chandler (1988, 1992) arguesthat it is not until 6 or 7 years of age that children develop the insight that two people may interpretthe same thing in different ways, and Taylor (1988a, 1988b) has interpreted her research assupport for Chandler’s position. Pillow (1991) has also conducted research showing that it is notuntil approximately 6 or 7 years of age that children begin to understand that a person’s priorexpectations will influence their interpretation of events. Fabricius and his colleagues, beginningwith a very different approach and methodology, similarly reached the conclusion that it is notuntil 8 or 10 years of age that children begin to develop a “constructivistic theory of mind” (e.g.,Fabricius & Schwanenllugel, in press). In contrast, Perner and Davies (1991) and Ruffman,Olson and Astington (1991) disagreed with these conclusions, and instead claim that already bythe age of 4 children understand the need for interpretation. Perner (1991) also cites Flavell’sresearch on visual perspective taking as evidence that 4-year-old children understand that the samepicture can be interpreted in different ways. I will now turn to the evidence that has been offeredin support of these various claims regarding the age of onset of an interpretive “theory of mind.”To sort out these conflicting claims and the evidence offered in support for these positions it isnecessary to examine these studies in some detail.86While examining the empirical evidence supporting these conflicting claims regarding theonset of an interpretive understanding of mind it is also important to be aware of the differentways in which the concept of interpretation is used. What we see in this debate over youngchildren’s understanding of interpretation is a replay of a common theme in developmentalpsychology, that is, the issue of what criteria should be used to assess a particular form ofcompetence (Chandler & Chapman, 1991). As is often the case in these debates, differentinvestigators use different criteria in deciding when to assign competence. This is an example ofthe “conceptual confusion” that Wittgenstein described in psychology, because although the sameword is used by different investigators it takes on different meanings in different contexts(Chapman, 1987). Thus, we need to examine the ways in which these various theorists use thenotion of interpretation.Interpretation and Visual Perspective-TakingA line of research that appears relevant to the issue of children’s understanding ofinterpretation is Flavell and his colleagues’ (Favell, 1978; Flavell, Everett, Croft, & Ravell, 1981;Masangkay, McCluskey, McIntyre, Sims-Knight, Vaughn, & Ravell, 1974) studies of Level 1and Level 2 visual perspective-taking. In these studies, when 3- and 4-year-old subjects werepresented with a picture of a turtle in front of them on the table, they were able to correctly statewhether it was standing on its feet or lying on its back. The subjects were then asked how theturtle would seem to someone sitting opposite to them across the table. Four-year-old childrenresponded correctly that the other person would see the turtle lying on its back, but 3-year-oldchildren could not correctly answer this question. Perner (1991) cites this research as evidencethat at 4 years of age children already understand that the same picture can be interpreted indifferent ways.Children’s competence at stating how other people would see a picture of a turtle is ademonstration that they understand that people will have different visual experiences of the sameobject from different perspectives, but this simpler meaning of interpretation is conflated with87interpretation as meaning an appreciation of the fact that different people may attach differentmeanings to the same object or event. Thus, level 2 visual perspective taking is not equivalent toan understanding of interpretation (Chandler, 1992).Interpretation and Visual AmbiguityMost of the studies addressing the issue of approximately when children begin tounderstand that knowledge is interpretive have been based on a methodology that was originallyemployed in the study of role-taking (Chandler & Boyes, 1982; Chandler & Helm, 1984).Chandler and Helm (1984) used a series drawings, called “droodles”, based on a cartooning stylepopularized by Roger Price (1953). This procedure involves a series of line drawings over whichcan be placed a cardboard overlay in which a small rectangular viewing window has been cut out,allowing a restricted view of the larger drawing. This view is intentionally ambiguous regardingthe nature of the drawing of which it is a fractional part. The droodle’s caption allows one tomake sense of the ambiguous keyhole view in a way that would not be possible without suchcontextualization. For example, one droodle consists of two triangles extending into the window,one from the side and the other from the bottom. The caption, “a ship arriving too late to save adrowning witch”, allows the viewer to infer that the triangles are the bow of a ship and the top ofa witch’s hat.Chandler and Helm (1984) showed 4-, 7-, and 1 1-year-old children a set of expandeddrawings and then the subjects were asked how another child might interpret only the portion ofthe drawing that was visible through the viewing window. The subjects, because they had seenthe whole drawing, had more knowledge that was available to someone who could merely see therestricted view, and this knowledge would have to be set aside to accurately predict the experienceof a naive observer. Chandler and Helm found that 4-year-old children consistently attributedtheir own knowledge to the naive observer. Thus, they would assume that another child viewingthe droodle would know that the triangles were the bow of a ship and the top of a witch’s hat, andstate that the other child would say it was a ship and a witch. The 7-year-old subjects were more88competent at realizing the another child, lacking their own behind-the-scene knowledge, wouldnot be able to guess the nature of the larger drawing. However, it was only the 1 1-year-oldsubjects who were fairly consistent in their ability to predict that someone viewing just the droodleportion would have no chance of correctly guessing what the larger drawing depicted, unless theyhad prior knowledge of the larger drawing.These results are consistent with the description of young preschool children as “copytheorists”. That is, at this age children associate knowledge with things themselves, without theinsight that the subject also has an interpretive role in knowledge acquisition. Not until earlyschool age and school age do children shift from this object-centered understanding to a subject-centered position regarding the location of the origins of knowledge.Taylor (1988a, 1998b) described two levels of conceptual perspective-taking based onChandler and Boyes’ (1982) ideas about a copy theory of knowledge. At Taylor’s first level ofconceptual perspective-taking children equate seeing with knowing, and children have difficultyseparating their knowledge of a shared visual event from the interpretation of someone who hasless background knowledge than they possess. Children at this level tend to assume that if twopeople view an object or an event, their knowledge or interpretation will be identical. That is, theywill not consider the effect of their own background information on their understanding of anobject or an event, and they will attribute their own knowledge to the other person. With a copytheory of knowledge children think of knowledge as originating in things themselves, and thus, iftwo people see the same things they should acquire the same knowledge. At Taylor’s proposedlevel 2 of conceptual perspective-taking children make a shift in the locus of the origin ofknowledge from the object to the subject. Children now become able to take backgroundinformation that is not shared into account when deciding how someone else would interpret aperceptual display.Taylor used a modified version of Chandler and Helm’s (1984) droodle procedure to testthis “seeing = knowing” hypothesis. In this version of the droodle procedure several partialviews of a series of pictures were used that varied in informativeness. In addition to the identity89of the animals in the picture (e.g., a giraffe), children were also told what the animal was doing(e.g., the giraffe was sitting down) and personal information about the giraffe (e.g., the giraffe’sname was George). Children were then asked if a naive observer who just saw small,nondescript parts of the animals through a small viewing window would know what the animalswere, what they were doing and the personal information.When nothing was visible through the window 4-year-old children reported that anobserver would not be able to identify the object, but they “tended to claim that seeing a smallnondescript part, or sometimes even a tiny edge, was sufficient to allow the observer to identifythe object” (Taylor, 1988a, p. 71 1). These level 1 errors were still often made by the 6-year-oldsubjects, but the 8-year-old children had developed a level 2 understanding that seeing anondescript part would not enable an observer with no previous knowledge to correctly identifythe object. Although children did tend to attribute knowledge of the identity of the animal to thenaive observer, they were less likely to assume that the observer would also know what theanimal was doing and personal information (e.g., the giraffe was sitting down and his name wasGeorge).Taylor’s (1988a) second experiment was a training study in which children watched apuppet having difficulty guessing the identity of ambiguous portions of drawings. Apparently,being made aware that multiple interpretations of the same picture are possible enabled 4-year-oldchildren to perform almost at the level of the 6-year-old subjects who lacked the training. Traininghad less effect on the 6-year-old children, probably because they were already fairly competent atthis task.Pillow (1991) sought to extend the research on children’s understanding of visualambiguity (e.g., Taylor, 1988) and verbal ambiguity (e.g., Sodian, 1988), to consider children’sunderstanding of ambiguous actions. Rather than ask children to make judgments about a naiveobserver’s knowledge, Pillow asked children about an observer with prior beliefs that would biasthe observer’s interpretation of the action in question. In the first experiment children were told aseries of stories in which one character disliked another character, who committed an act of90ambiguous intention (e.g., a character bumped into a desk and knocked a toy airplane onto thefloor). This study was designed to investigate whether or not children could use an observer’sprior bias (positive or negative attitude) toward the actor to judge whether the observer woulddecide whether the ambiguous action was intentional or accidental. Pillow found that secondgraders (approximately 8 years of age) were competent at using an observer’s bias and previousbeliefs when predicting the observer’s judgment. This is a demonstration of their understandingthat prior expectations can influence the interpretation of action. On the other hand, kindergartners(approximately 6 years of age) were inconsistent in their correct use of an observer’s bias andprior expectations to predict ajudgment. They tended to be correct in some situations but not inothers.In a second experiment, Pillow (1991) attempted to simplify the task. Instead of requiringchildren to consider the intentional vs. accidental distinction, the ambiguous event was the act anactor was performing (i.e., putting something into a container or taking it out). In addition, thestories in this experiment all involved two observers with contrasting perspectives. For example,two children see a child holding a rabbit in front of its cage. The subjects must decide, based onthe two children’s prior expectations whether the children will believe the child is taking the rabbitout of the cage, as they have been told not to, or if the child found that the rabbit had escaped andis returning it to the cage. In this experiment, kindergartners (approximately 6 years of age)proved to be able to correctly predict how an observer with prior expectations would interpret thescene, but preschoolers performed near chance levels on these questions. Only the secondgraders were able to appropriately justify their answers.Schwanenflugel, Fabncius, and Alexander, (1994; Fabricius & Schwanenflugel, in press)studied the development of children’s understanding of mind by studying children’s knowledge ofthe relations among various mental activities. They argue that this reveals children’sunderstanding of the nature of the mind. From their studies with 8 and 10 year old children, andadults, they suggest that at 8 years of age children think of the mind as a sort of informationprocessor, and at 10 years they are beginning to develop the constructivistic theory of mind that91adults possess. They use a very different approach and methodology than has been used in mostresearch on children’s theories of mind and it is interesting that they reach a similar conclusion.In contrast to the position that an understanding of interpretation is not achieved bychildren until they reach the age of 6 or 7, several other investigators claim that this insight isalready in place by most children’s fourth birthday. Perner and Davies (1991) claim that 4-year-old children understand that the mind has an active role in evaluating the truth of verbalinformation. Perner (1991) claims that 4-year-old children “start to understand the need forinterpretation. What is lacking is adult accuracy in judging how much and what needs to bevisible before the depicted object can be identified” (p. 274). Pemer and Davies’ (1991) goal wasto investigate the age at which children acquire a “notion of mind as an active informationprocessor”. They take this to be equivalent to an understanding of interpretation, but what Pernerand Davies mean by an understanding of mind as involving the active processing of information isthe insight that the truth or falsity of statements must be evaluated in terms of previously heldbeliefs, or background knowledge. The question they addressed in their first experiment was“whether children understand that belief formation is not a passive copying of the most recentinformation received but involves active evaluation of that information in view of existing beliefs”(Perner & Davies, 1991, p. 56).In this first experiment, Perner and Davies (1991) questioned 3- and 4-year-old childrenabout the likely response of a character, portrayed by a puppet, when presented with informationregarding the location of a football or the identity of an object (i.e., whether a brick was real orfake). The puppet either had no opinion about these matters or already held firm beliefs. Pernerand Davies found that most 4-year-old children were competent at predicting that a naive personwithout a previously held firm belief would accept as true a message from someone claiming tohave relevant knowledge. But these children also expected that a person with already firmpreconceptions regarding the location of a football or the identity of a brick would reject, asmisguided, a statement that conflicted with his or her pre-existing beliefs. This, according toPerner and Davies, is clear evidence that even 4-year-old children understand the active role of92existing beliefs in evaluating new information. Furthermore, Perner and Davies (1991) argue that“even children at this early age cannot be described as entertaining a copy theory of mind in whichthe mind is seen as a passive recorder of information” (p. 58). Although Perner and Daviesqualify their claims somewhat by acknowledging that this “nascent interpretive theory of mind is,of course, not very accurate” (p. 65), they advocate a quantitative difference between a 4-year-oldchild’s understanding of mind and an adult’s understanding rather than a qualitative difference.There are two aspects to unpack in these conclusions. First, Pemer and Davies’ (1991)use of the idea of interpretation is somewhat liberal, because there is no construal of experience,there is only an evaluation of the truth or falsity of information. Second, Perner and Daviesmisconstrue the idea of a copy theory. Perner and Davies (1991) state that “Chandler and Boyesput the emphasis on direct informational contact with reality” (p. 53, italics in original). Althoughthere are certainly grounds for this interpretation from Chandler and Boyes’ (1982) vividmetaphors of “projectile firings from things themselves” (p. 391), and facts as being the“epistemological equivalent of a bullet in the brain”, the important part of the idea of a copy theoryis that knowledge comes from things, rather than being an interpretive product. The importantachievement is a shift in the locus of knowledge from objects to subjects, and “knowing begins tobe recognized as a constructive, meaning-generating, human activity” (Chandler & Boyes, 1982,p. 395). Thus, while Perner and Davies’ research is of interest, it is not a demonstration of what Imean by an interpretive understanding of mind, nor is it evidence against Chandler’s notion of acopy theory.Ruffman, Olson and Astington (1991) reach a similar conclusion to Perner. They believethat their results “show that an understanding of visual ambiguity begins around age 4 and isfirmly in place by age 5” (p. 100). Ruffman et al. take this as evidence that instead ofinterpretation being a stage that is reached after false-belief understanding, there is only one stageand this is reached when children acquire an understanding of false-belief. Further, Ruffman etal. argued that the reason that “ambiguity tasks are solved around age 4 is that they are false belief93tasks at heart--in recognizing that the other would not know, children were required to recognizethat the other could hold a false belief” (p. 101).However, Pillow (in press) points out that Ruffman et al.’s 4-year-old subjects performedabove chance only when both they and the doll, acting as the observer, were uninformed about theidentity of the objects, and they could correctly respond in this condition merely by attributingtheir own ignorance to the doll. Ruffman et al. claim that their procedure assesses youngchildren’s understanding of the need for interpretation, and there is a sense in which it isappropriate to use the term “interpretation” in regard to Ruffman et al.’s task, because they wereasking the children if the stimulus was “interpretable”. That is, the children had to evaluatewhether or not there was enough information presented for a naive observer to correctly identifythe object. However, in this use of interpretation there is something that the object really is.Whereas when Chandler uses the idea of interpretation he is referring to the understanding thatthere may be no single fact of the matter, but rather there are situations in which there may be twointerpretations of an object that may both be legitimate. The achievement that we are interested inis a shift in the locus of knowledge from objects to subjects.The studies by Perner and Davies (1991) and Ruffman et al. (1991) are about children’sunderstanding of whether or not another person will have enough information to correctly identifyan ambiguous object. Most of the studies of ambiguity have been based on a methodologyemployed by Chandler and Helm (1984) in which subjects are given restricted informationalaccess to a drawing. With this methodology there really is some fact of the matter, although it ishidden behind a screen, and subjects must decide if the portion visible in the restricted view issufficient to correctly judge the identity of the drawing or object. Although this procedure mayassess a form of understanding of interpretation, what is really need is a procedure in whicheveryone is presented with the same information.Most of the studies described above on children’s understanding of interpretation orassimilation have involved asking children to infer the knowledge of either a naive observer, or aninformed or biased observer, after the observer has been exposed to incomplete or ambiguous94information (Pillow, in press). The logic behind this research is that if children understand therole of prior information in the assimilation of new information, then they should be able toaccurately predict the conclusions that an observer will draw from ambiguous information whenthe observer either lacks or possesses certain key bits of information. Children’s understandingof the role of prior expectations or knowledge on the conclusions drawn from ambiguousinformation has been studied in the area of visual ambiguity and ambiguity in human action andcommunication (Pillow, 1991).An important exception to this line of research is a study that was conducted by Lalonde,Chandler and Moses (1992). Lalonde eta!. conducted a series of experiments using the droodleprocedure and found that 7-year-old and some 6-year-old children, but not 5-year-old children,are able to come up with different interpretations of the same ambiguous restricted portion of a linedrawing. This could be read as an understanding that the object is ambiguous, since it allowsdifferent interpretations, but it is different from evaluating whether or not a character has enoughinformation to correctly identify an ambiguous object. This is a demonstration of a form ofunderstanding of interpretation because the children must understand that the picture can beinterpreted in multiple ways since they have just done so, but it is not yet a demonstration of thechildren’s ability to use this idea of interpretation in predicting or explaining people’s action.Another relevant study was conducted by Mansfield and Clinchy (1985). They presented3-, 4-, 7- and 10-year-old children with vignettes in which two people disagreed about something.For the 3- and 4-year-old children, the source of the disagreement involved either matters ofimmediately verifiable fact or subjective opinion. The 7- and 10-year-old children were alsopresented with versions of these vignettes, as well as stories involving more points on the“continuum from the most objective to the most subjective” (p. 6). The stories involved fourdomains: verifiable fact, debatable fact, interpretation, and taste or value. However, the examplethat they give of a vignette involving interpretation consists of a mother and her son disagreeingabout whether another child is “yucky” or “nice.” This seems similar to their story involvingmatters of taste in which a father and his son disagree about whether or not a new soup is “gross.”95These short stories were presented to the children with one picture and then the subjects wereasked a series of questions: “Why do you think they would disagree about this?”; “Who do youthink is right?”; “Why?”; “Why do you think that he said [whatever he said]?”; “Can they find outfor sure who is right?”; “How?”; “Can they come to an agreement?”; “Does one have to be rightand one wrong?”One of the ways Mansfield and Clinchy (1985) coded the children’s responses to thesequestions was whether the children believed that there was a single answer (“Absolutism”), orwhether they believed that both positions were legitimate (“Multiplism”). The 3- and 4-year-oldchildren were almost completely absolutist in their responses, that is, they believed that only oneof the protagonists could be right (although 24% of the time they said they did not know). Thisheld even in matters of taste where the preschoolers maintained that experience would prove oneof the protagonists right.Mansfield and Clinchy (1985) created a category they called “objective multiplism” inwhich the children suggested that the “disputants were referring to different aspects of the samephenomenon” (p. 12). The multiplism is “objective” because it is located in different aspects ofthe same phenomenon. The examples from this category involve the ingenious stretching of theconcept of “absolutism” to account for occasions involving disagreement, while still maintainingtheir faith in absolute empirically-derived truth. The category of “subjective multiplism” originatesin the subject rather than the object, and they found that this type of response increased with age.These results from the Mansfield and Clinchy (1985) study are similar to what I wouldpredict, except that I would expect 4-year-olds to be more successful on matters of taste.However, it seems that in Mansfield and Clinchy’s vignettes involving taste the opinions wereexpressed before the food was actually tasted, so the preschoolers claimed that the issue could bedecided by actually tasting the food. If this was corrected I would expect 4- and 5-year-oldchildren to be successful in this domain. This prediction is consistent with a study by Flavell,Flavell, Green, and Moses (1990) in which they also examined the “fact to value” continuum.Flavell et al. were interested in whether young children would have less difficulty inferring that96someone else holds an odd, nonnormative belief about a matter of taste or value than they do ininferring a false belief about a matter of fact. As predicted, Ravell and his colleagues found thatyoung children do have more success when the beliefs refer to values rather than matters of fact.EpistemologicalDevelopment in AdolescenceAdults’ understanding of interpretation has been studied by Kuhn, Pennington, andLeadbeater (1983) under the rubric of “cognitive relativism.” Following Perry’s (1970, 1981)pioneering work on the developmental progression from absolutist to relativist thought in collegestudents, Kuhn et al. developed a measure with which they could observe subjects applying theirimplicit assumptions about knowledge rather than asking them to describe these epistemologies.This measure consists of presenting subjects with two discrepant accounts of a fictitious historicalevent. The subjects were asked to evaluate the conflicting accounts of the “Fifth Livian War”from a “South Livian” historian and a “North Livian” historian. Kuhn et al. described a sequenceof developmental levels in the understanding of knowledge, and they found these levels to bestrongly age-linked in a sample of school children in grades 6 to 12 (Kuhn, Amsel, &O’Loughlin, 1988, Chapter 12).The work of Enright and his colleagues (Enright, Lapsley, Franklin, & Steuck, 1984) on“belief-discrepancy” reasoning has some relevance for the present research. They presentedchildren with dilemmas, asked them to resolve the dilemmas, and then presented the subjects witha statement from another child who had resolved the dilemma in the opposite way. They wereinterested in how children of different ages dealt with differences of opinion, in particular theywere interested in how the children evaluated or judged another person who had expressed anopinion that differed from their own. They were studying the development of tolerant viewstowards disagreeing others, and their dilemmas involved moral considerations. In contrast, theresearch reported here is concerned with how children account for the fact that people can reachdiffering interpretations of the same event or object. That is, it is concerned with the development97of children’s conceptions of knowledge, but I would argue that one source of tolerance may be anappreciation of the interpretive nature of knowledge.98APPENDIX B: Protocol for Study]False-belief test:The false-belief test is modeled on the Maxi task. The experimenter will introduce thesubjects to two puppets (Mary and Maxi). Maxi is playing with a toy and he places it in a red boxwhen he goes out to play. While Maxi is out (under the table) Mary takes the toy out and playswith it. Then she puts it away in the blue box and goes out. Maxi comes back in:“When Maxi comes in he will want his toy.”The subjects will be asked the following prediction questions:“Where will Maxi think his toy is when he first comes in?”“Where will Maxi lookfor his toy when he first comes in?”[These questions will be counterbalanced.]“And where is the toy really?”“You’re right, Maxi did look in the red box.”If the subjects do not answer the questions correctly, they will be told that, in fact, Maxi did lookfor his toy in the red box:“Well, Maxi actually looked in the red box.”All the subjects will then be asked the following explanation question:“Why do you thinkMaxi looked in the red box?”If the subjects are overly brief they will be asked the follow-up, probe question:“Tell me more about that.”991) Matters of taste: (first example)The experimenter will show the subjects a picture of a bowl of soup:“(Subject’s name) this is apicture of a bowl ofsoup that Mary andMaxi’s mothers makesometimes. Mary do you like this soup?”[with Maxi under the table]“Yes, it tastes good. I like it.”[Mary goes under the table and Maxi comes out.]“Maxi, do you like this soup?”“No. it tastes bad. I don’t like it.”Explanation Questions“Is it okayforMary to say the soup tastes good andMaxi to say it tastes bad?”“Why is it okay .../Why isn’t it okay ... ?“Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Now, Mary likes the soup, but Maxi doesn’t like it. Ifwe tell children in another schoolabout the soup do you think that they will like the soup or not like it, or wouldn’t you know whatthey would think?”If the subject takes a position the experimenter will ask:“How can you tell what someone else will think about the soup?”“How sure are you that they will say that?”If the subject says “I don’t know”, then ask“Why is it hard to tell what someone else will think?”100Matters of taste: (2nd example)The experimenter will show the subjects and the puppets a picture of a painting, and ask:“(Subject’s name) here is apicture ofa painting. Maxi do you like this picture? [withMary under the table]“Yes, I like it. I think it’s nice. I would like to have it in my room.” [Maxi goes under thetable and Mary comes out]“Mary do you like this picture?”“No. I don’t like it I think it’s ugly. I wouldn’t like to have it in my room.”Explanation Questions“Is it okayfor Maxi to say he would like to have the picture in his room and Mary to sayshe wouldn’t like to have it in her room?”“Why is it okay .../Why isn’t it okay ... ?“Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Now, Maxi likes the picture, but Mary doesn’t like the picture. Ifwe show this picture tochildren in another school do you think they will like the picture or not like, or don’t you knowwhat they would say?”If the subject takes a position the experimenter will ask:“How can you tell what they will say?”“How sure are you that they will say that?”If the subject says, “I don’t know”, then ask“Why is it hard to tell what they will say?”1012) Ambiguous referential communication: (first example)“Now we are going to play another game with Maxi andMary. While they are under thetable and can’t see I will hide a penny under one of these cards. Then they can come out and Iwill give them a clue about where to lookfor the penny.”The experimenter will hide the penny under the card with a large red block.“Okay, Maxi andMary, the clue is; the penny is under the card with the big block.”“Mary show us where you think the penny is.” [with Maxi under the table]“I think it’s under the card with the big red block.” [Mary goes under the table and Maxicomes out]“Now Maxi, show us where you think the penny is.”“I think it’s under the card with the big blue block.”Explanation QuestionsThe experimenter must ensure that the subject understands both of the possible locationsallowed by the ambiguous message.“I told them that the penny is under the card with the big block. Is it okayfor Mary to saythe penny is under the card with the big red block andfor Maxi to say it’s under the card with thebig blue block?”“Why is it okay .../Why isn’t it okay ... ?“Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Now, Mary thinks it’s under the card with the big red block, andMaxi thinks it’s underthe card with the big blue block. Ifwe told children in another school that the penny is under thecard with the big block would they think it’s under the card with the big blue block or the cardwith the big red block, or wouldn’t they know what to think? [counterbalance]If the subject takes a position then ask“How can you tell what they will think?”“How sure are you that they will think that?”If the subject says “I don’t know”, then ask:“Why is it hard to tell what they will think?”Deviant interpretation:“Well, Josefsays it’s under the card with the small red block. Does it make sense forJosef to say that or does it not make sense?”“Why . . .102Referential Ambiguity: (2nd example)The procedure described above will be repeated for a second ambiguous message:“Now we are going to play the game again. We will put Maxi andMary under the tableand I will hide the penny. Then they can come out I will give them a clue about where to lookforthe penny.”The experimenter will hide the penny under the card with a small red block.“Okay, Maxi andMary, the clue is; the penny is under the card with the picture ofa redblock.”“Maxi show us where you think the penny is. [with Maiy under the tablel 1 think it’sunder the card with the big red block.[Maxi goes under the table and Mary comes out]“Now Mary show us where you think the penny is. I think it’s under the card with the little redblock.”Explanation QuestionsThe experimenter must ensure that the subject understands both of the possible locationsallowed by the ambiguous message. Then the experimenter will restate the problem in thequestion:“1 told them the penny is under the card with the red block. Is it okay for Maxi to say thepenny is under the card with the big red block, andfor Mary to say it s under the card with thelittle red block?”“Why is it okay .../Why isn’t it okay ... ?“Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Now, Mary thinks it’s under the card with the big red block, and Maxi thinks it’s underthe card with the little red block. If we told children in another school that the penny is under thecard with the red block would they think it’s under the card with the big red block or the card withthe little red block, or wouldn’t they know what to think?If the subject takes a position ask:“How can you tell what they will think?”“How sure are you that they will think that?”If the subject says “I don’t know”, then ask“Why is it hard to tell what they will think?”Deviant interpretation:“Well, Josefsays the penny is under the card with the big blue block. Does it make sensefor Josef to say that or does it not make sense?”“Why . . . ?“1033) Lexical Ambiguity: (first example, homonym)“Now we are going to play another kind ofgame with the puppets. Mary and Maxi, inthis game you both iwed to wait here for a ring. Now Mary, tell us what you are waiting for.”“I’m waiting for a ring for my finger.” [Show illustration; with Maxi under the table.Mary goes under and Maxi comes out]“Okay, Maxi, tell us what you are waiting for.”“I’m waiting for a telephone to ring.” [Show illustration]The experimenter will show the subject illustrations of the two possible meanings of“ring” to ensure that the subject understands both meanings.Explanation Questions“I told them to waitfor a ring. Is it okay for Mary to say she is waitingfor a ring for herringer, andMaxi to say he is waitingfor the telephone to ring?”“Why is it okay .../Why isn’t it okay ... 7”Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Now, Mary says she is waiting for a ring for herfinger, but Maxi says he is waiting forthe telephone to ring. Ifwe told children in another school to waitfor a ring would they bewaiting for a ringfor theirfinger or the telephone to ring, or wouldn’t they know what to think?”[counterbalance]If the subject takes a position, ask:“How can you tell what they would think?”“How sure are you that they would think that?”If the subject says “I don’t know”, ask:“Why is it hard to tell what they would think?”Deviant interpretation:“Well, Josefsays he is waiting for a necklace Does it make sense for Josef to say that ordoes it not make sense?“Why . . . ?“104Lexical Ambiguity: (Second example, homophone).“Now we are going to play another game with the puppets. Mary andMaxi, in this gameyou both need to wait here for a pear/pair. Now Maxi, tell us what you are waitingfor.”“I’m waitingfor a pear to eat” [Show illustration, with Mary under the table. Then Maxigoes under and Mary comes out]“Okay, Mary tell us what you are waitingfor.”“I’m waiting for a pair ofshoes.” [Show illustration]The experimenter will show the subject illustrations of the two possible meanings ofpear/pair to ensure that the subject understands both meanings.Explanation Questions“I told them to waitfor apear/pair. Is it okay forMaxi to say he is waitingfor apear toeat, and Mary to say she’s waitingfor a pair ofshoes?”“Why is it okay .../Why isn’t it okay ... ?Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Now, Maxi says he’s waiting for a pear to eat, but Mary says she’s waiting for a pair ofshoes. Ifwe told children in another school about waitingfor a pear/pair would they waitfor apear to eat orfor a pair ofshoes, or wouldn’t they know what to think?” [counterbalance]If the subject takes a position, ask:“How can you tell what they will think?”“How sure are you that they will think that?”If the subject says, “I don’t know”, ask:“Why is it hard to tell what they will think?”Deviant interpretation:“Well, Josefsays he’s waiting for an apple. Does it not make sense for Josef to say thator does it make sense?”“Why . . .1054) Ambiguous Figures: (first example)a) The subjects will be shown the “duck-rabbit”.“Now we will show Mary and Maxi a picture. Mary what do you think this is?” [withMaxi under the table]“I think that it’s a duck” [Mary goes under and Maxi comes out]“Okay, Maxi, what do you think it is?”“I think that it’s a rabbit.”Explanation QuestionsThe experimenter must ensure that the subject sees both entities.“Is it okayfor Mary to say it’s a duck and Maci to say it’s a rabbit?”“Why is it okay .../Why isn’t it okay ... ?“Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Now, Mary says it’s a duck and Maxi says it’s a rabbit. Ifwe showed this picture tochildren in another school would they think it’s a duck or a rabbit, or wouldn’t they know what tothink?” [counterbalancelIf the subject takes a position, ask:“How can you tell what they will think?”“How sure are you that they will think that?”If the subject says, “I don’t know”, then ask“Why is it hard to what they will think?”Deviant interpretations:“Well, Josefsays it’s an elephant. Does ifmake sense for Josef to say that, or does it notmake sense?”“Why . . . ?“106Ambiguous Figure: 2nd example.The subjects will be shown the “rat-man”:“Now we will show Mary andMaxi another picture. Maxi what do you think this is?”[with Mary under the table]“I think that it’s a rat” [Maxi goes under and Mary comes out]“Okay, Mary, what do you think it is?”“I think that it’s a man with glasses.”Explanation QuestionsThe experimenter must ensure that the child sees both entities. Then the experimenter willrestate the problem in the question:“Is it okayfor Maxi to say it’s a rat and Mary to say it’s a man with glasses?”“Why is it okay . ../Why isn’t it okay ...Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Now, Maxi says it’s a rat and Mary says it’s a man with glasses. If we showed thispicture to children in another school would they think that it’s a rat or a man with glasses, orwouldn’t they know what to think?”If the subject takes a position, ask:“How can you tell what they will think?”“How sure are you that they will think that?”If the subject says, “I don’t know”, then ask“Why is it hard to tell what they will think?”Deviant interpretations:“Well, Josefsays it’s really a dog Does it not make sense for Josef to say that, or does itmake sense?”“Why . . . ?“107APPENDIX C: Protocolfor Study 2Matters of taste: (Example 1)The experimenter will show the subjects a picture of a bowl of soup:“(Subject’s name) this is a picture of a bowl ofsoup that Mary andMaxi ‘s mothers makesometimes. Mary do you like this soup?”[with Maxi under the table]“Yes, it tastes good. I like it.”[Mary goes under the table and Maxi comes out.]“Maxi, do you like this soup?”“No. it tastes bad. I don’t like it.”Explanation Questions“Does it make sense that one of them likes the soup and the other one doesn’t like it?”“Why does it (doesn’t it) make sense?”Probe questions:“Is it okayforMary to say the soup tastes good andMaxi to say it tastes bad?”“Why is it okay .../Why isn’t it okay ...Standard probe if the subject is vague: “You said .... Can you tell me more about that?”Matters of taste: (Example 2)The experimenter will show the subjects and the puppets a picture of a painting, and ask:“(Subject’s name) here is a picture ofa painting. Maxi do you like this picture? [withMary under the table]“Yes, I like it. I think it’s nice. I would like to have it in my room.” [Maxi goes under thetable and Mary comes out]“Mary do you like this picture?”“No. I don’t like it I think it’s ugly. I wouldn’t like to have it in my room.Explanation Questions“Does it make sense that one of them likes the picture and the other one doesn’t like it?”“Why does it (doesn’t it) make sense?”Probe questions:“Is it okayfor Maxi to say he would like to have the picture in his room andMary to sayshe wouldn’t like to have it in her room?”“Why is it okay ... !Why isn’t it okay ...Standard probe if the subject is vague: “You said .... Can you tell me more about that?”108Referential Communication: (Example 2: Selection task)The procedure described above will be repeated for a second ambiguous message for aselection task:“Now we are going to play a different game. I will think ofa card, and then I will give thepuppets a clue about which card I’m thinking about.” [The experimenter selects the card with asmall red block.]“Okay, Maxi andMary, the clue is; I’m thinking of the card with the picture ofa redblock.”“Maxi show us which card you think I chose [with Mary under the table] I think it’s thecard with the big red block.[Maxi goes under the table and Mary comes out]“Now Mary show us which card you think I chose. I think it’s the card with the little red block.”Explanation QuestionsThe experimenter must ensure that the subject understands both of the possible meaningsof the ambiguous message. Then the experimenter will restate the problem in the question:“I told them that I was thinking of the card with the red block. “Does is make sense forMary to say one thing andMaci to say something else?”“Why does it (doesn’t it) make sense?”Probe questions:“Is it okayforMary to say one thing andMaxi to say something else?”“Why does Mary think it’s one card and Maxi think it’s another card?”Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Ambiguous Figure: (Example 1: “Duck-Rabbit”)a) The subjects will be shown the “duck-rabbit”.“Now we will show Mary and Maxi a picture. Mary what do you think this is?” [withMaxi under the table]“I think that it’s a duck” [Mary goes under and Maxi comes out]“Okay, Maxi, what do you think it is?”“I think that it’s a rabbit.”Explanation QuestionsThe experimenter must ensure that the subject sees both entities.“Does is make sense forMary to say one thing andMaxi to say something else?”“Why does it (doesn’t it) make sense?”Probe questions:“Is it okayfor Mary to say one thing andMaxi to say something else?”“Why does Maxi think it’s one thing andMary think it’s another thing?”Standard probe if the subject is vague.“You said .... Can you tell me more about that?”109APPENDIX D: Protocol Study 3False-belief test:The false-belief test is modeled on the Maxi task. The experimenter will introduce thesubjects to two puppets (Mary and Maxi). Maxi is playing with a toy and he places it in a yellowcontainer when he goes out to play. While Maxi is out (under the table) Mary takes the toy outand plays with it. Then she puts it away in the blue container and goes out. Maxi comes back in:“When Maxi comes in he will want his toy.”The subjects will be asked the following prediction question:“Where will Maxi think his toy is when he first comes in?”“And where is the toy really?”“You’re right, Maxi did look in the yellow container.”If the subjects do not answer the questions correctly, they will be told that, in fact, Maxi did lookfor his toy in the yellow container:“Well, Maxi actually looked in the yellow container.”All the subjects will then be asked the following explanation question:“Why does Maxi think his toy is in the yellow container and Mary think it’s in the bluecontainer?”If the subjects are overly brief they will be asked the follow-up, probe question:“Tell me more about that.”110Referential Communication: (Example 1: Hiding task)“Now we are going to play another game with Maxi andMary. While they are under thetable and can’t see and you look away I will hide a sticker under one of these cards. Then theycan come out and I will give them a clue about where to lookfor the sticker.”The experimenter will hide the sticker under the card with a large red block.“Okay, Maxi andMary, the clue is; the sticker is under the card with the big block.”“Mary show us where you think the sticker is.” [with Maxi under the table]“I think it’s under the card with the big red block.” [Mary goes under the table and Maxicomes out]“Now Maxi, show us where you think the sticker is.”“I think it’s under the card with the big blue block.”Explanation QuestionsThe experimenter must ensure that the subject understands both of the possible locationsallowed by the ambiguous message.“I told them that the sticker is under the card with the big block. Why does Mary think thesticker is under the card with the big red block and at the same time Maxi think it’s under the cardwith the big blue block?”“Does is make sensefor Mary to say one thing andMaxi to say something else?”“Why does it (doesn’t it) make sense?”Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Mary thinks it’s under the card with the big red block, andMaxi thinks it’s under the cardwith the big blue block. Now we will tell Josefand Ann that the sticker is under the card with thebig block.Ask about the two puppets’ predictions consecutively.“Do you think Josef(Ann) will say it’s under the card with the big blue block or the cardwith the big red block, or wouldn’t you know what he would say?” [counterbalance]If the subject takes a position then ask“How can you tell what they will think?”“How sure are you that they would think that?” [use pointerl“Would they both think the same thing?”If the subject says “I don’t know”, then ask:“Why is it hard to tell what they will think?”Repeat for the second puppet, alternate order of puppets.Deviant interpretation:“Well, Josefsays the sticker is under the card with the small red block. Does it makesense for Josef to say that or does it not make sense?”“Why does it (doesn’t it) make sense?”If this question needs to be repeated rephrase it with “... is that silly or not silly ...“[counterbalance].111Referential Communication: (Example 2: Selection task)The procedure described above will be repeated for a second ambiguous message for aselection task:“Now we are going to play a dfferent game. I will think ofa card, and then I will give thepuppets a clue about which card I’m thinking about.” [The experimenter selects the card with asmall red block.]“Okay, Maxi and Mary, the clue is; I’m thinking ofthe card with the picture ofa redblock.”“Maxi show us which card you think I chose [with Mary under the table] I think it’s thecard with the big red block.[Maxi goes under the table and Mary comes out]“Now Mary show us which card you think I chose. I think it’s the card with the little red block.”Explanation QuestionsThe experimenter must ensure that the subject understands both of the possible meaningsof the ambiguous message. Then the experimenter will restate the problem in the question:“I told them that I was thinking of the card with the red block. Why does Mary say it’sone card and at the same timeMaxi say it’s another card?”“Does is make senseforMary to say one thing andMaxi to say something else?”“Why does it (doesn’t it) make sense?”Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Mary thinks it’s the card with the big red block, andMaxi thinks it’s the card with thelittle red block. Now we will tell Josef and Ann that I was thinking of the card with the redblock.”Ask about the two puppets’ predictions consecutively.“Do you think Josef (Ann) will say it’s the card with the big red block or the card with the littlered block, or wouldn’t you know what he would say?” [counterbalance]If the subject takes a position ask:“How can you tell what they will think?”“How sure are you that they would think that?” [use pointer]“Would they both think the same thing?”If the subject says “I don’t know”, then ask“Why is it hard to tell what they will think?”Repeat for second puppet.Deviant interpretation:“Well, Josefsays it’s the card with the big blue block. Does it make sense for Josef to saythat or does it not make sense?”“Why does it (doesn’t it) make sense?”If this question needs to be repeated rephrase it with “... is that silly or not silly[counterbalance].112Lexical Ambiguity: (Example 1: “Ring”, homonym)“Now we are going to play another kind ofgame with the puppets. Mary and Majci, inthis game you both need to wait here for a ring.”“Now Mary, tell us what you are waiting for.”“I’m waiting for a ring for myfinger.” [Show illustration; with Maxi under the table.Maiy goes under and Maxi comes out]“Okay, Maxi, tell us what you are waiting for.”“I’m waiting for a telephone to ring.” [Show illustration]The experimenter will show the subject illustrations of the two possible meanings of“ring” to ensure that the subject understands both meanings.Explanation Questions“I told them to waitfor a ring. Why does Mary say she’s waiting for one thing and at thesame time Maxi say he’s waitingfor another thing?”“Does is make senseforMary to say one thing andMaxi to say something else?”“Why does it (doesn’t it) make sense?”Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Mary says she is waitingfor a ringfor herfinger andMaxi says he is waitingfor thetelephone to ring. Now we will tell Ann and Josef to waitfor a ring.”Ask about the two puppets’ predictions consecutively.“Do you think Josef (Ann) will be waiting for a ring for hisfinger or the telephone to ring, orwouldn’t you know what he would say?” [counterbalance]If the subject takes a position, ask:“How can you tell what they would think?”“How sure are you that they would think that?” [use pointer]“Would they both think the same thing?”If the subject says “I don’t know”, ask:“Why is it hard to tell what they would think?”Repeat for second puppet.Deviant interpretation:“Well, Josefsays he is waiting for a necklace Does it make sense for Josef to say that ordoes it not make sense?“Why does it (doesn’t it) make sense?”If this question needs to be repeated rephrase it with “... is that silly or not silly[counterbalance].113Lexical Ambiguity: (Example 2: “Pear/Pair”, homophone).“Now we are going to play another game with the puppets. Mary andMaxi, in this gameyou both need to wait here for a pear/pair. Now Maxi, tell us what you are waitingfor.”“I’m waitingfor a pear to eat” [Show illustration, with Mary under the table. Then Maxigoes under and Mary comes out]“Okay, Mary tell us what you are waitingfor.”“I’m waiting for a pair ofshoes.” [Show illustrationiThe experimenter will show the subject illustrations of the two possible meanings ofpear/pair to ensure that the subject understands both meanings.Explanation Questions“I told them to waitfor a pear/pair. Why does Mary say she’s waiting for one thing and atthe same time Maxi say he’s waitingfor another thing?”“Does is make sensefor Mary to say one thing andMaxi to say something else?”“Why does it (doesn’t it) make sense?”Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Maxi says he’s waitingfor a pear to eat andMary says she’s waiting for a pair ofshoes.Now we will tell Ann and Josef to waitfor a pear/pair.”Askabout the two puppets’ predictions consecutively.“Do you think Josef (Ann) will waitfor apear to eat, orfor apair ofshoes, or wouldn’t youknow what they would say?” [counterbalance]If the subject takes a position, ask:“How can you tell what they will think?”“How sure are you that they would think that?” [use pointer]“Would they both think the same thing?”If the subject says, “I don’t know”, ask:“Why is it hard to tell what they will think?”Repeat for second puppet.Deviant interpretation:“Well, Josefsays he’s waiting for an apple. Does it not make sense for Josef to say thator does it make sense?”“Why does it (doesn’t it) make sense?”If this question needs to be repeated rephrase it with “... is that silly or not silly[counterbalance].114Ambiguous Figure: (Example 1: “Duck-Rabbit”)a) The subjects will be shown the “duck-rabbit”.“Now we will show Mary andMaxi a picture. Mary what do you think this is?” [withMaxi under the table]“I think that it’s a duck” [Mary goes under and Maxi comes out]“Okay, Maxi, what do you think it is?”“I think that it’s a rabbit.”Explanation QuestionsThe experimenter must ensure that the subject sees both entities.“Why does Maxi say it’s one thing and at the same time Mary say it’s another thing?”“Does is make senseforMary to say one thing andMaxi to say something else?”“Why does it (doesn’t it) make sense?”Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Mary says it’s a duck andMaxi says it’s a rabbit. Now we will showed this picture toAnn and Josef”Ask about the two puppets’ predictions consecutively.“Do you think Josef (Ann) will think it’s a duck or a rabbit, or wouldn’t you know what theywould say?” [counterbalance]If the subject takes a position, ask:“How can you tell what they will think?”“How sure are you that they would think that?” [use pointer]“Would they both think the same thing?”If the subject says, “I don’t know”, then ask“Why is it hard to tell what they will think?”Repeat for second puppet.Deviant interpretations:“Well, Josefsays it’s an elephant. Does fmake sense for Josef to say that, or does it notmake sense?”“Why does it (doesn’t it) make sense?”If this question needs to be repeated rephrase it with “... is that silly or not silly[counterbalance].115Ambiguous Figure: (Example 2: “Rat-Man”)The subjects will be shown the “rat-man”:“Now we will show Mary and Maxi another picture. Maxi what do you think this is?”[with Mary under the table]“I think that it’s a rat” [Maxi goes under and Mary comes out]“Okay, Mary, what do you think it is?”“I think that it’s a man with glasses.”Explanation QuestionsThe experimenter must ensure that the child sees both entities. Then the experimenter willrestate the problem in the question:“Why does Maxi say it’s one thing and at the same time Mary say it’s another thing?”“Does is make senseforMary to say one thing andMaxi to say something else?”“Why does it (doesn’t it) make sense?”Standard probe if the subject is vague.“You said .... Can you tell me more about that?”Prediction Questions“Maxi says it’s a rat andMary says it’s a man with glasses. Now we will showed thispicture to Ann and Josef.”Ask about the two puppets’ predictions consecutively.“Do you think that Josef (Ann) will say it’s a rat, or a man with glasses, or wouldn’t you knowwhat he would say?”If the subject takes a position, ask:“How can you tell what they will think?”“How sure are you that they would think that?” [use pointer]“Would they both think the same thing?”If the subject says, “I don’t know”, then ask“Why is it hard to tell what they will think?”Repeat for second puppet.Deviant interpretations:“Well, Josefsays it’s really a dog Does it not make sense for Josef to say that, or does itmake sense?”“Why does it (doesn’t it) make sense?”If this question needs to be repeated rephrase it with “... is that silly or not silly ...“[counterbalance].

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088295/manifest

Comment

Related Items