Symposium
Chair: Judy S. DeLoache
Discussant: Judy S. DeLoache
The extraordinary power of human cognition arises in large part from our creative andflexible use of symbols. Adults employ a vast array of symbol systems to facilitatecommunication, from language to a variety of special-purpose objects and general-purposemedia. A crucial aspect of the creation and use of symbols is intentionality. When I use asymbolic representation of some sort to communicate something to you, it is my intention thatyou interpret the relation between the symbol and its referent in the same way that I do. If youdo not, my communicative effort will have failed; the symbol will not convey what I wanted itto. If, for example, I show you a map and tell you it shows the location of buried treasure, itwill be of no use to you if you don't know how maps relate to reality in general or if you don'tunderstand how you are supposed to use this particular map. The central role accorded to intentionality in this view of symbolizationsuggests that both the onset and subsequent development of symbolic skills may be related tothe development of the understanding of intentionality. To interpret a symbol correctly,children have to have some idea of why the other person is using it. This view thus suggests that some understanding of human intention would precede theuse of symbols in general. The first presentation in the symposium speaks directly to thisissue. It argues that chimpanzees' lack of understanding of communicative intent precludes theuse of symbols for communicative purposes. In contrast, human infants draw on theirunderstanding of human intention to begin to master the use of symbols for communication. The second paper underscores the importance of intentionality in early symbol use byshowing that in a rich communicative situation, the precise nature of a symbol is not asimportant to young infants as the context in which it is used. In a context in which an adultmakes it clear that a given symbol--whether a word, gesture, picture--stands for a particularentity, the infant will learn the symbol-referent relation. The third presentation focuses on young children's explicit understanding of the role ofintentionality in the production of a symbolic artifact. A series of studies reveals that theirprimary criterion for judging the content of a picture is what the artist intended to draw,regardless of what the picture actually looks like. In the last presentation, the role of intentionality is examined in very young children'suse of a symbolic object in two different problem-solving contexts. One set of studies revealsthat the intended relation between a symbolic object (a scale model) and its referent (a largerspace) must be explicitly explained and demonstrated for young children to detect the relation. A second set of studies using a different task reveals a clear relation between the extent towhich young children understand an adult's intention to give them a clue using a symbolicobject and their ability to solve the problem.
Details of individual items:
paper
Chimpanzee communication is complex in many ways. However, recent studies haveshown that it does not involve one individual understanding the communicative intentions ofanother. Communicative intentions have the structure: I intend [for you [to share attentiontoX]], and so the comprehender must understand that the sender has intentions towards hisintentional states. Chimpanzees apparently do not have this kind of understanding, and so theycommunicate with signals, not symbols. Human children, on the other hand, begin to understand the communicative intentionsof other persons from soon after their first birthdays, including in situations in which thosecommunicative intentions are expressed in language. In addition, children engage in rolereversal imitation of communicative acts in which communicative intentions are expressedtowards them; that is, children learn to use towards others the same communicative devicesthat others have used towards them. The central theoretical argument of this paper is that the understanding ofcommunicative intentions (and learning to use them via role reversal imitation) transformscommunicative signals into communicative symbols by making them intersubjectively shared. The central empirical argument of the paper--resting on two empirical studies of chimpanzees'and human children's behavior in communicative situations--is that chimpanzees do notunderstand communicative intentions and so only use communicative signals, whereas humanchildren from 12-18 months of life have such an understanding and so do use communicativesymbols. This understanding also displays itself in the interpretation of the symbolic use ofobjects.
paper
Recent research has begun to converge on a striking new notion about the relationbetween symbolic knowledge and early word learning. This work suggests that children beginthe word-learning endeavor with a general understanding of how symbols map to objects,actions and events in the world and only later accord a priority to words over othercommunicative symbols. A primary goal of this research program has been to explore factorsinfluencing children's symbolic abilities early in development and how children's expectationsabout which signals in their environment are intended as symbols changes with development.Three lines of relevant research will be discussed. First, what is the range of signals that infants interpret as object names, and how doesthis range change over development? We compared children's interpretations of verbal versusnon-verbal symbols (including gestures, non-verbal sounds, and pictograms). These studiesindicate that infants accept a wide range of symbolic forms as object names at 18 months, butby 26 months, they accept verbal symbols more readily than non-verbal symbols. We arguethat initially, children interpret any signal that is presented in a socially referential context as asymbol and that developmental change is tied to children's increasing reliance on language astheir primary communicative tool. Second, does iconicity facilitate early symbol acquisition? Although many would expectinfants' first symbols to be iconic, we have found that 18-month-olds map novel gesturalsymbols to object categories, regardless of whether the gestures are iconic or arbitrary. Incontrast, preliminary evidence suggests that 26-month-olds are more successful at mappingiconic than arbitrary symbols. This finding confirms that younger children accept a broaderrange of symbolic forms than older children, and also challenges the traditional perspectivethat infants' earliest symbols derive from routines with objects. The similarity between asymbol and its referent appears to exert a greater influence on non-verbal symbol acquisitionlater in development than at its onset. Third, what aspects of parental input facilitate or inhibit the acquisition of non-verbalsymbols? In free-play sessions with their infants, parents frequently produced gesturalsymbols, and they tended to produce these gestural symbols within the same joint-attentioncontexts in which they provided verbal labels. Furthermore, the frequency of gestures in theparental input reliably predicted the frequency of symbolic gestures in the infant lexicon. These findings suggest that infants interpret both adults' verbal and non-verbal input asintending to stand for the objects to which they refer. Current investigations focus on howchildren's interpretations of adults' gestural communication changes as children become moreproficient language users. Taken together, these findings suggest that infants initially focus more on the social-referential context in which a signal occurs than the form of the signal itself in determiningwhether the signal is intended as a symbol. These findings also indicate marked developmentalchanges in children's symbol interpretation over time. An initial general ability to mapsymbols to objects develops into a more complex, differentiated system of verbal versus non-verbal symbols with age and experience.
paper
Philosophers and psychologists have long been interested in the naming andcategorization of artwork. What makes us call one drawing 'a dog', and another 'a pencil'? How do we name pictures that do not resemble what they depict? And how do these abilitiesemerge in the course of development? I will discuss three series of studies that bear on theseissues. (1) At least for adults, the ability to infer the intentions of the artist underlies thenaming of some pictures. Consider a simple line drawing of a face. The precise name that anadult will give this drawing depends on a host of factors, including what the adult believes theartist intended for the drawing to depict. For instance, if the artist is intently looking at herselfin a mirror as she creates the drawing, it would be natural to think that the drawing is a pictureof the artist. A series of studies finds this same reliance on intentional cues in young children. Even 2-year-olds name their own pictures not on the basis of what these pictures resemble, buton the basis of what they are intended to represent. This same pattern exists when children areasked to name pictures drawn by other people. When doing so, children attend to cuesregarding the artist's representational intent, often ignoring salient perceptual properties of thepictures. (2) One prominent theory of autism is that such individuals have an impairment intheory of mind, hence have difficulties reasoning about people as intentional beings. If picturenaming involves reasoning about the intentions of other people, one might expect autisticindividuals to have special problems in this domain. To explore this, we investigated picturenaming in a population of high-functioning autistic individuals. We found that although suchindividuals were very good at 'creating' pictures (much better than their mental-age matchedcontrols), they were quite poor at naming pictures and did not spontaneously exploit cues aboutthe artist's intention. (This was true both for their naming of pictures that other people drawand for pictures that they themselves draw). This deficit supports the view that theory of mindis relevant to picture naming. (3) How special is art? That is, do the same intentional factors that underlie thenaming of pictures also affect the naming of non-representational entities, such as cups andshoes? To address this question, children and adults were asked to name a series of picturesand tools. In one condition, the objects were described as purposefully created; in another, theobjects were described as resulting from an accident. This manipulation had significant effecton the subjects' responses; they were more likely to name the entity ('dog', 'knife') when itwas described as being the result of a goal-directed process. This suggests that the importanceof the creator's intention for naming is not limited to the domain of artwork, even for youngchildren.
paper
A primary function of symbols is to serve as a source of information; through symbols,we can learn about reality without directly experiencing it. From a naval chart, for example,we can find out where the rocks are that must be avoided when entering a harbor. We then usethis knowledge to guide our behavior. To obtain such information, we must understandsomething about the intended relation between the symbol and its referent. The use of mapsand charts depends crucially on assuming that the creator of the map intended and tried toprovide an accurate representation of an existing reality. This presentation focuses on young children's developing understanding of the intendeduse of symbolic objects in two different tasks. In both of them, very young children are giveninformation via a symbol about where to find a target object. In both cases, the children haveto understand something about why the adult is presenting them with the symbolic object. Ifthey do not understand the experimenter's intentions, they are unlikely to make appropriate useof the information the symbol provides for solving the problem. In one task, a scale model is used to try to communicate to children where a toy ishidden in a room. As the child watches, a miniature toy is hidden in a scale model of a room,and the child is then asked to find a larger toy that is hidden in the corresponding location inthe room itself. This paper will report dramatic differences in performance that occurred in aserries of studies as a function of the completeness of the instructions children received aboutthe nature of the model-room relation. Even with a high degree of iconicity between modeland room, very young children have to be explicitly told in detail how the two spaces arerelated. Slightly older children can get by with less information about the intended relation,but only with high iconicity. With less similarity between the two spaces, they require fullerinstructions. When no explicit instructions about the model-room relation are given, not untilthe age of five do children figure out the intended relation on their own. In research currentlyunderway, we are attempting to increase the performance of 2-1/2-year-old children in themodel task by highlighting the intentional origin and function of the model. Preliminaryresults are very promising. In the second task, an adult uses a replica object or a picture to give very youngchildren a clue as to which of three objects contains a sticker hidden by a different adult. Analysis of video tapes reveals few signs of understanding of the intention of the adult on thepart of some of the children, but clear evidence of awareness by others. For example, somechildren carefully study the adult's actions with the replica or picture, looking back and forthbetween it and the larger object (the hiding place) it represents. Some even smileconspiratorially. The children whose behavior suggests that they understand why the adult isshowing them the object or picture clue perform more successfully (i.e., they more often findthe hidden object) than do children who give no signs of 'intention awareness.'