This work was made possible in part by National Science Foundation grant (#DGE-9550152) to the first author. 

Address comments to:
A. Lillard,
Department of Psychology,
Gilmer Hall,
University of Virginia,
Charlottesville,
VA 22903-2477
USA

Email: lillard@virginia.edu
Telephone: 804-982-4750.
Fax: 804-982-4766

Many have thought that children have an early appreciation of the mind in the case of pretend play. Results from several experiments are against this (Joseph, in press, Exps. 2-3; Lillard, 1993a; Lillard, 1996; Lillard, in press; Rosen, Schwebel & Singer, 1997). However, an experiment by Lillard (1996, Exp. 4) suggested that when a pretense is about a fantasy character, instead of a real entity, children might have a better understanding of the mind's involvement. The present experiment tested this, and found that indeed, when pretending to be a fantasy character is at issue, 4-year-olds are significantly more apt to indicate the mind's involvement.

The Influence of Fantasy on Children's Understanding of Pretense

The field of early social cognition has been re-energized in recent years by the study of how children develop an understanding of minds. Particularly central has been the issue of when children understand that minds represent the world. Minds are not simply mirrors of reality; they are interpreters. So one might interpret another's gargling gesture as an insult, when in fact the other person was just clearing his throat. People's ability to interpret situations in myriad ways seems obvious to adults, but interestingly, children under 4 do not seem to appreciate it.

Evidence often credited with suggesting that this ability develops in early childhood is from Wimmer and Perner (1983). They showed children a doll named Maxi, who hid his chocolate in a cupboard and then went out to play. While he was out, his mother moved the chocolate to a new location. Children were asked, when Maxi returned, where he would look for his chocolate. Surprisingly, many children under 5 claimed he would look in the new location. Later experiments using more refined methods usually find that 4-year-olds respond correctly on such tasks, but most 3-year-olds do not (Astington, Harris & Olson, 1988; Flavell & Miller, 1998; Moses & Flavell, 1990; Wellman, 1990; but see Holmes, Black, & Miller, 1996, regarding income group differences). This result is very solid; other methods, like those in which the child hides an object herself, sometimes show better performance and sometimes do not (e.g., the discrepancy of Chandler, Fritz & Hala, 1989 and Sodian, Taylor, Harris & Perner, 1991). Anecdotally as well, younger children sometimes seem to show evidence of understanding false belief (Dunn, 1988; Reddy, 1991), but the understanding does not seem reliable or omnipresent. Young children do not consistently demonstrate understanding that minds represent, rather than mirror, reality.

Pretend play is of particular interest in this regard (Leslie, 1987). When one pretends, one mentally represents the pretend situation or object, and projects it onto the existing, real one (Lillard, 1993b). Children pretend as early as 18 months of age, and they appear to understand pretense in others by 28 months (Harris & Kavanaugh, 1993). For example, they understand that if someone pretends to spill pretend tea, the location is "wet". Many have assumed that engaging in pretense requires not only having mental representations, but also knowing that one has them (Flavell, 1988; Forguson & Gopnik, 1988; Leslie, 1988; Taylor & Carlson, 1997). If it were the case that pretenders know that they mentally represent the pretend situation, then this would constitute an early appreciation of mental representation in the pretense domain. Several theorists (those just cited, except Leslie) have suggested that this decalage does occur.

Against this speculation, some work suggests that young children do not understand pretense involves mental representation (Lillard, 1993a; Rosen et al., 1997), or even the mind at all (Lillard, 1996). For example, when asked if various events could occur using only one's body, or only one's mind, or whether they required both a body and a mind, 4-year-olds correctly placed think events (like think about your teacher) in the mind box, but placed pretend events (like pretend you are a rabbit), along with physical events (like getting wet in the rain) in the body box. This categorizing of pretend events with purely physical, mindless events suggests young children do not perceive mental involvement in pretense.

Interestingly, though, one experiment in the study by Lillard (1996, Exp. 4) suggested that for certain types of pretense, children might have somewhat more understanding that the mind is involved: namely, for pretending to be fantasy characters. Sixteen 4- and 5-year-olds (mean age: 4-11) were administered 19 trials, and asked to place various events in the mind, body, or both boxes just mentioned. Several items were controls, asking about instances that should surely go in a certain box. Others were nonsense items ("Proct a pram" and "Foss you are a feasehosh"), to check where children would place items about which they were merely uncertain. The remaining items focused on pretense: pretend you are a puppy, a hippopotamus, the Lion King, a mommy/daddy, in your bedroom, in an airplane, and in the jungle. The control items were put in their proper boxes, and roughly a third of the nonsense items were placed in each box. Of the pretense items, 13% were placed in the both box, 34% were placed in the body box, and 53% were placed in the mind box. Using the both box implies that the child knows that the event requires the mind. Therefore, this amounts to children claiming 66% of pretense events required a mind. This was the best performance seen in any experiment in this line, for this age group; in other experiments, performance has hovered around 40% (Lillard, 1996). Why might this be?

Children did particularly well on two items: pretend to be the Lion King and pretend you are in the jungle (averaging 81% mind and both box designations for each; for simplicity, hereafter "mind boxes" will be used to designate "mind and both boxes"). On closer inspection, the relative improvement in performance was particularly strong for children in one of the two item order groups. (See Table 1). Children in the first group put pretend items in the mind boxes on 90% of trials (range: 63-100%), as opposed to 46% of trials for Group 2 (range: 25-63%). What was different about these two groups? The mean ages of the two groups is the same, ruling out the possibility of a slightly older group affecting the means. Another possibility is that the placement of the Lion King fantasy character early in Group 1's order boosted their performance on subsequent items.

Table 1. Performance by Order in Lillard, 1996, Exp. 4
 

Group 1	plane	LionK	puppy	jungle	bedrm	hippo	mom
	.63	1.0	1.0	1.0	.75	1.0	.88
Group 2	puppy	mom	hippo	plane	bedrm	LionK	jungle
	.38	.25	.38	.50	.25	.63	.63

As can be seen in Table 1, performance on the Lion King item and all subsequent pretense items (only the jungle item came later for Group 2) was higher than on all prior pretense items for both groups. To be specific, Group 2's best performance was on the Lion King and jungle items; for each of these, 63% of children chose mind boxes. On the other pretense items (all of which came prior) Group 2's performance ranged from 25% to 50% mind box choices. Further, the groups were not so very different prior to the Lion King item: on the only pretend item Group 1 heard prior to the Lion King (pretend you are an airplane), it performed at 63%, slightly but not significantly better than Group 2's 50% on that item. Group 1's worst performance was on the airplane item. On the Lion King item and subsequent items it scored 75-100% (mean 94%). Granted, the groups were probably also somewhat different due to random variation: other experiments (cited earlier) have shown that about 40% of 4-year-olds do understand that pretending involves the mind, and probably that more competent subset was overrepresented in Group 1. But the pattern also suggests that children might be more insightful about the mind's being necessary to pretense when the pretense involves fantasy characters. The experiment described here aimed to test that possibility.

This experiment employs the method of Lillard (1996). Children were shown boxes designating mental and physical activities, and were essentially asked to categorize pretense and other activities by placing them in one of the boxes.

Participants. Twenty-four children from two urban area preschools were tested. Participants ranged in age from 4;1 to 5;7, with a mean age of 4;8, and there were 8 boys and 16 girls. Participants were from middle class families in a metropolitan area of the United States, spoke English as or as if it were their native language (as judged by the experimenter) and were mostly white, although a range of ethnic backgrounds was represented. Eight additional 4-year-olds were dropped from the experiment because they failed control items (explained later). Several of these were tested under noisy conditions and appeared to be distracted. Four younger children were also excluded as they appeared not to understand the control items.

Materials. Three large matchbox-sized boxes (7x12x4 cm), each with a slot on top into which index cards (4x6 cm) could be placed, were used. On the front of each box was a picture and a label. One box was labeled "Mind" and had a picture of a light bulb. One box was labeled "Body" and had a picture of a body. The third (the both box) was labeled "Mind and Body" and had both pictures. Twenty-nine index cards were used, each with a short phrase written on it. Twelve of these were training cards, 9 were controls, and 8 were test items. The test stimuli are listed in Table 2.

Table 2. Stimulus items
 

Ordinary Pretend to be a puppy

Pretend to be a bunny rabbit

Pretend to be a king/queen

Pretend to be a mommy/daddy

Exotic

Pretend to be Pocahantus

Pretend to be Ariel, The Little Mermaid

Pretend to be Batman

Pretend to be the Lion King Both

Write your name

Sing a song

Bake a cake Think Control

Think about a flower

Think about a cat

Think about your teacher Physical Event Control

Fall over if you were pushed

Slide down a slippery hill

Get wet in the rain

Procedure

Children were brought into a private game room or area of their school. They were given the same introduction to the boxes as children in Lillard (1996, Experiments 3-5). Children were first asked if they knew where their mind and body were. They were then told that their mind was used for certain things like dreaming and remembering and that their body was used for other things like being under a bed. Children were then told that the game was about choosing what went inside each of the boxes. The mind box was described as being for things that you can do with your mind, things that do not require a body at all, and the body box as being for things that you can do with just your body and that do not require a mind at all. The both box was described as being for things that absolutely needed both your mind and body.

Children then received up to 12 training phrases, one on each index card, and were asked in which box each card belonged. Two examples are "Imagine an ice cream cone" and "Get blown over by the wind." During the training phase, if children chose the wrong box, they were given feedback on which was the right box. Words like "think" and "pretend" that were crucial to the test phase were not used in the training phase. The training phase ended when children correctly responded on 5 training trials in a row or when all 12 cards had been used.

The test phase consisted of 17 phrases (4 ordinary pretend, 4 exotic pretend, 3 thinking, 3 body, and 3 both). Children received the 17 phrases in one of four quasi-random orders. Half of the subjects received the four ordinary pretend cards before any of the exotic pretend cards. The other half received the blocks in the opposite order. A further stipulation was that for each of these two groups, half had the Lion King early in the group of exotic items, and half had it late. This allowed for assurance that a category of items, not the Lion King alone, was responsible for any effect exotic items might have. A final stipulation was that there were never more than two consecutive items from a single category (no more than two physical event controls in a row, for example). Only children who placed at least 5 or the 6 think and physical event control cards in the correct box (mind or body, respectively) were included in the final set of 24 children.

Body box choices were scored 0, and mind and both box choices were both scored 1, since to place an item in the both box indicates understanding that the mind is required. Scores were summed for each category. Children chose the mind boxes for an average of 2.5 of 4 (57%) exotic items and for 1.9 of 4 (43%) ordinary items. To ensure that children were not responding haphazardly, Chi-Square Goodness of Fit tests were conducted. These indicated systematic responding both for the ordinary items, [chi]2(4) = 26.85, and the exotic items, [chi]2(4) = 40.67, both ps < .01.

A preliminary t-test showed that having the Lion King item early or late within the group of exotic items had no effect. A repeated measures analysis of variance was performed, with item type order (ordinary or exotic first) as the between subjects factor and item type (ordinary or exotic) as the within subjects factor. There was no significant effect for item order, but the effect of item type was significant: F(1,22) = 11.35, p <.005. Exotic items went into the mind boxes significantly more often than did ordinary items. Indeed, the ranges for the two types of item were nonoverlapping: 54-58% mind box choices for exotic items, and 37-46% for ordinary items.

Looked at another way, whereas 14 children scored the same on exotic and ordinary items, every one of the remaining 10 children performed better on exotic than ordinary items (p= .001, Binomial distribution). No specific exotic item appeared to carry this effect; instead it was simply that many children made one or two more mind or both box choices for the exotic items than they made for the ordinary items. In other words, there was an increased tendency to view exotic items as requiring a mind for pretense.

As these data show, young children tend to claim that the mind is involved in pretending to be certain types of fantasy characters more often than they make such claims for pretending to be more ordinary characters. These results seem consistent with those of Saltz, Dixon, and Johnson (1977). In that study, disadvantaged preschoolers who were trained over the course of a year to enact fairy tales performed better at year-end on tests of language, empathy, inhibitory control, and other skills than did a second group trained to enact ordinary routines like grocery shopping. This is also consistent with the work of Taylor and her colleagues, who find that children who engage in more fantasy play perform better on measures of understanding minds (Taylor & Carlson, 1997; Taylor, Cartwright & Carlson, 1993). It fits in as well with the work of Dias and Harris (1988), who showed that when logical syllogisms are set in a fantasy context, young children's performance at arriving at their conclusions improves (but see Leevers & Harris, 1997, for indications that this was not simply due to fantasy). This finding could also explain a result by Holmes et al. (1996). They found that children performed better on false belief tasks concerning locations than those concerning contents of boxes, but their locations tasks were enacted by fantasy characters like Big Bird.

Although consistent with earlier work, the present study breaks new ground in showing that children have early insight into pretense's mental qualities via certain types of pretense, but not others. The insight that pretending involves the mind does not occur for all fantasy characters for all children, but there is a consistent bias towards better insight. Perhaps it is via pretending to be fantasy characters that children come to appreciate pretense's mental qualities, and children who pretend about fantasy most often have the insight that pretending is mental earliest. Future research is needed to investigate that possibility.

One question that arises from this research is just how children conceptualize the exotic fantasy stimuli. The fact that children show better insight about them suggests that at some level children categorize those items differently from the ordinary stimuli. But perhaps they are categorizing them differently than adults might suppose. Whereas we included them as examples of exotic fantasy characters, perhaps another feature like emotional content caused children to categorize them differently. Children might not even be attending to the fact that the exotic stimuli are not real. Indeed, Samuels & Taylor (1992) showed that children are less likely to correctly categorize events as not real when they are emotionally charged (see discussion in Lillard, 1994, and Woolley, 1997), and one could argue that these big-screen characters are emotionally charged for young children. Another important task for further research, then, is to identify what features of the exotic stimuli led to children's better understanding. Following that, one would be in a better position to identify how frequently children pretend in such ways, and whether indeed this might be an avenue to mental insight.

There are many possible features of the exotic items that might account for their eliciting a higher level of understanding in young children. For example, a stimulus being cast as a cartoon rather than a realistic-looking figure (reality type) might lead children to consider the mind's involvement in pretense. Because one has to imagine cartoons, children are better able to appreciate that a mind would be needed to pretend to be one. One must project the pretense mentally since it does not exist in the world.

Another possibility has to do with the exotic characters in this experiment having a specific identity. The exotic characters are all referred to by name rather than simply by their object type ("a cat"), as the ordinary items are. Perhaps once something has a name, it becomes specific and a standard script will no longer serve for emulating it. One would need a mind to pretend a custom, nonscripted event, and having a name makes one an individual whose life is custom. This would seem to fit with the Saltz et al. (1977) data as well, since the fairy tale group was enacting stories about specific characters, whereas the other group was enacting daily routines not involving any particular characters.

Alternately, the effect might be carried by being presented in a specific emotional plot. Emotional stories are involving: one must follow the intricacies of a character's evaluations of events to engage in such a story. Perhaps children clue in early to the fact that these more involved stories would require a mind to act out, whereas simply being a cat, they suppose, could be done on "automatic pilot". Future work should examine these and alternative possibilities about what sorts of pretense enable early insight into pretense's mental aspects. What the present experiment establishes, building on Lillard (1996), is that children do seem to be more cognizant of pretense's mental qualities when the pretense concerns certain types of fantasy characters.

Astington, J. W., Harris, P. L., & Olson, D. R. (Eds.). (1988). Developing theories of mind. New York: Cambridge University Press.

Chandler, M., Fritz, S. A., & Hala, S. (1989). Small-scale deceit: Deception as a marker of two-, three-, and four-year-olds' early theories of mind. Child Development, 60, 1263-1277.

Dias, M. G., & Harris, P. L. (1988). The effect of make-believe play on deductive reasoning. British Journal of Developmental Psychology, 6, 207-221.

Dunn, J. (1988). The beginnings of social understanding. Cambridge, MA: Harvard University Press.

Flavell, J. H. (1988). The development of children's knowledge about the mind: From cognitive connections to mental representations. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories of mind (pp. 244-271). Cambridge: Cambridge University Press.

Flavell, J. H., & Miller, P. H. (1998). Social cognition. In D. Kuhn & R. S. Siegler (Eds.), Handbook of child psychology: Vol. 2. Cognition, perception, and language development. (5th Ed.) (pp. 851-898). New York: Wiley.

Forguson, L., & Gopnik, A. (1988). The ontogeny of common sense. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories of mind (pp. 226-243). New York: Cambridge University Press.

Harris, P. L., & Kavanaugh, R. D. (1993). Young children's understanding of pretense. Monographs of the Society for Research in Child Development, 58, (1, Serial No. 231).

Holmes, H., Black, C., & Miller, S. (1996). A cross-task comparison of false belief understanding in a Head Start population. Journal of Experimental Child Psychology, 63, 263-285.

Joseph, R. M. (in press). Intention and knowledge in preschoolers' conception of pretend. Child Development.

Leevers, H. J., & Harris, P. L. (1997). Persisting effects of instruction on young children's syllogistic reasoning. Unpublished manuscript, University of Oxford.

Leslie, A. M. (1987). Pretense and representation: The origins of "theory of mind". Psychological Review, 94, 412-426.

Leslie, A. M. (1988). Some implications of pretence for mechanisms underlying the child's theory of mind. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories of mind (pp. 19-46). New York: Cambridge University Press.

Lillard, A. (1996). Body or mind: Children's categorizing of pretense. Child Development, 67, 1717-1734.

Lillard, A. S. (1993a). Young children's conceptualization of pretense: Action or mental representational state? Child Development, 64, 372-386.

Lillard, A. S. (1993b). Pretend play skills and the child's theory of mind. Child Development, 64, 348-371.

Lillard, A. S. (1994). Making sense of pretence. In C. Lewis & P. Mitchell (Eds.), Children's early understanding of mind: Origins and development (pp. 211-234). Hillsdale, NJ: Lawrence Erlbaum.

Lillard, A. S. (in press). Wanting to be it: Children's understanding of intentions underlying pretense. Child Development.

Moses, L. J., & Flavell, J. H. (1990). Inferring false beliefs from actions and reactions. Child Development, 61, 929-945.

Reddy, V. (1991). Playing with others' expectations: Teasing and mucking about in the first year. In A. Whiten (Ed.), Natural theories of mind (pp. 143-158). Oxford: Basil Blackwell.

Rosen, C. S., Schwebel, D. C., & Singer, J. L. (1997). Preschooler's attribution of mental states in pretense. Child Development, 68, 1133-1142.

Saltz, E., Dixon, D., & Johnson, J. (1977). Training disadvantaged preschoolers on various fantasy activities: Effects on cognitive function and impulse control. Child Development, 48, 367-380.

Samuels, A., & Taylor, M. (1992). Children's ability to distinguish fantasy events from real-life events Unpublished manuscript, University of California at Santa Cruz.

Sodian, B., Taylor, C., Harris, P. L., & Perner, J. (1991). Early deception and the child's theory of mind: False trails and genuine markers. Child Development, 62, 468-483.

Taylor, M., & Carlson, S. M. (1997). The relation between individual differences in fantasy and theory of mind. Child Development, 68, 436-455.

Taylor, M., Cartwright, B. S., & Carlson, S. M. (1993). A developmental investigation of children's imaginary companions. Developmental Psychology, 29, 276-285.

Wellman, H. M. (1990). The child's theory of mind. Cambridge, MA: Bradford Books/MIT Press.

Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children's understanding of deception. Cognition, 13, 103-128.

Woolley, J. D. (1997). Thinking about fantasy: Are children fundamentally different thinkers and believers from adults? Child Development, 6, 991-1011.