1
In press: Developmental Science
DO YOU SEE WHAT I SEE?
INFANTS’ REASONING ABOUT OTHERS’ INCOMPLETE PERCEPTIONS
Yuyan Luo
and
Whitney Beck
University of Missouri
Address correspondence to Yuyan Luo, Department of Psychological Sciences, University of Missouri,
Columbia, MO 65211, USA; phone: (573) 882-0429; fax: (573) 882-7710; email: [email protected].
2
Abstract
Twelve-month-olds realize that when an agent cannot see an object, her incomplete perceptions
still guide her goal-directed actions. What if the agent had incomplete perceptions because she could only
see one part of the object, e.g., one side of a screen? In the present research, 16-month-olds were first
shown an agent who always pointed to red objects, as opposed to a black or a yellow object, suggesting
that she preferred red over the other colors. Next, two screens were introduced while the agent was absent.
The screens were (1) red or green on both sides; (2) red on the front (infants’ side) but green on the back
(the agent’s side) or vice versa; or (3) only colored red or green on the front. During test, the agent, who
could only see the back of the screens, pointed to one of the two screens. The results revealed that while
infants expected the agent to continue acting on her color preference and point to the red rather than the
green screen during test, they did so in accord with the agent’s perceptions of the screens, rather than their
own perceptions: they expected the agent to point to the red screen in (1), but to the green-front screen in
(2), and they had no prediction of which screen the agent should point to in (3). The implications of the
present findings for early psychological reasoning research are discussed.
3
We know that infants possess some understanding about the psychology of agents (i.e., objects
that can act and react to their environment) (e.g., Gergely & Csibra, 2003; Leslie, 1995; Premack &
Premack, 1995). From a young age, infants interpret the actions of agents in terms of goals and dispositions
(e.g., Bíró & Leslie, 2007; Gergely, Nádasdy, Csibra, & Bíró, 1995; Kuhlmeier, Wynn, & Bloom, 2003; Luo &
Baillargeon, 2005; Phillips & Wellman, 2005; Sommerville, Woodward, & Needham, 2005; Woodward, 1998;
Woodward & Guajardo, 2002). Recent research suggests that at least by the second year, infants also
appear to consider an agent’s perceptions and beliefs to make sense of the agent’s goals and dispositions,
even when those perceptions and beliefs are incomplete or false (e.g., Allard & Onishi, 2008; Luo &
Baillargeon, 2007; Luo & Johnson, in press; Onishi & Baillargeon, 2005; Sodian, Thoermer, & Metz, 2007;
Song & Baillargeon, in press; Southgate, Senju, & Csibra, 2007; Surian, Caldi, & Sperber, 2007). In the
present research, we focused on how infants considered an agent’s incomplete perceptions to interpret her
disposition-guided behavior.
In her ground-breaking study, Woodward (1998) demonstrated that infants at 5 to 6 months of age
encode a human agent’s repeated actions of grasping an object as goal-directed. In one experiment, infants
were first habituated to an event in which they saw the human agent’s arm and hand reach for and grasp
object-A, as opposed to object-B. Following habituation, the objects’ positions were reversed. During test,
infants saw the agent’s arm and hand reach for and grasp object-A in its new location (old-goal event) or
object-B in the position previously occupied by object-A (new-goal event). Infants looked reliably longer at
the new- than at the old-goal event. These and control results suggested that infants interpreted the agent’s
actions during habituation as directed toward the goal of approaching object-A, and that they expected the
agent to continue acting on this goal during test and hence responded with increased attention when the
agent’s arm and hand reached for and grasped object-B instead in the new-goal event.
A few recent studies (e.g., Luo, 2008; Luo & Baillargeon, 2005; Song, Baillargeon, & Fisher, 2005b)
showed that in the Woodward-type task, infants also interpret agents’ goal-directed actions as stemming
from particular dispositions. A disposition is defined as a tendency or state that helps us interpret and
predict an agent’s goal-directed behavior. In these studies, infants were assigned to a two-object or a one-
4
object condition. The two-object condition was essentially similar to the Woodward (1998) study described
above. The one-object condition was identical to the two-object condition except that when the agent
approached object-A, object-B was absent. During test, infants again looked reliably longer at the new- than
at the old-goal event in the two-object condition. However, they looked about equally at the two events in the
one-object condition. These results suggested that in the two-object condition, infants attributed to the agent
a goal of approaching object-A as well as a particular disposition, a preference for object-A over object-B.
These attributions led infants to expect the agent to maintain the preference and form the same goal of
approaching object-A as before. They thus responded with increased attention when the agent approached
object-B instead during test. The results of the one-object condition suggested that although infants
attributed to the agent the goal of approaching object-A, since the agent did so repeatedly, this goal
attribution was insufficient for infants to predict what the agent should do when a new object, object-B, was
introduced during test. They thus accepted that the agent could approach either object. Therefore, in some
cases at least, a disposition may be a more powerful predictor of agents’ actions than a goal – it can help
predict what an agent will do to pursue her goal in a relatively different context. Knowing that an agent
prefers one object over another can lead to predictions that the agent will still approach its preferred object
when the positions of the two objects change. However, only knowing the agent’s goal directed towards one
object provides no information as to the agent’s actions when a new object is added.
Luo and Baillargeon (2007) found that 12.5-month-olds also consider what an agent perceives and
remembers when interpreting the agent’s preference-guided actions. In their study, infants watched an agent
repeatedly reach for and grasp object-A, which was visible to the agent through a transparent screen, as
opposed to object-B. Object-B was either visible to the agent through another transparent screen; hidden
from the agent (but not the infants) by an opaque screen; or placed by the agent herself behind the opaque
screen, so that even though she could no longer see object-B, she knew of its presence there. The results
showed that infants responded based on the agent’s perceptions: they interpreted the agent’s repeated
actions toward object-A as revealing her preference for object-A over object-B only when she could see
object-B or when she was aware of its presence, but not when object-B was hidden from her. Recent
5
investigations suggest that when the agent holds false perceptions or beliefs, rather than incomplete
perceptions as in Luo and Baillargeon (2007), 13- to 15-month-old infants also use the agent’s perceptions
and beliefs to interpret her actions (Onishi & Baillargeon, 2005; Song & Baillargeon, in press; Surian et al.,
2007).
The present research sought new evidence indicating that infants consider an agent’s incomplete
perceptions to make sense of her disposition-guided actions. In Luo and Baillargeon (2007), the agent held
incomplete perceptions because she could not see an entire object. We designed situations in which the
agent could only see one part of the object. We reasoned that these situations might be more sophisticated
and hence tested older, 16-month-old infants. In addition, we asked whether infants could make inferences
about an agent’s color preference, a new type of dispositions. As discussed above, infants can interpret
agents’ goal-directed actions as stemming from a particular disposition. Infants have been found to attribute
to agents several simple dispositions: a positive disposition towards an agent (e.g., Hamlin, Wynn, & Bloom,
2007; Kuhlmeier et al., 2003; Premack & Premack, 1997) or an object (e.g., Luo & Baillargeon, 2005, 2007;
Luo & Johnson, in press; Repacholi & Gopnik, 1997; Song & Baillargeon, in press), or an inclination to
perform certain actions (e.g., Song & Baillargeon, 2007; Song, Baillargeon, & Fisher, 2005a). In these cases,
after seeing an agent’s behavior in one context, infants use disposition attributions to make predictions of
the agent’s actions in a relatively different context. Specifically, after seeing how a helper or a hinderer
interact with an agent going up a hill, infants form expectations as to whether the agent should later
approach the helper or the hinder (e.g., Hamlin et al., 2007; Kuhlmeier & Wynn, 2003; Kuhlmeier et al.,
2003); after seeing an agent repeatedly choose one object over another, infants expect the agent to
approach the same object when the positions of the two objects change (e.g., Luo & Baillargeon, 2005,
2007; Luo & Johnson, in press; Song et al., 2005b); after seeing an agent slide different objects, infants
expect the agent to later pick a new object that is also slideable (e.g., Song & Baillargeon, 2007; Song et al.,
2005a). We examined whether infants could ascribe to agents yet another type of simple disposition
involving features of objects: a preference for a certain color. In particular, after seeing an agent point to red
6
objects as opposed to black and yellow objects, would infants then generalize this to a relatively new
context and expect the agent to also point to a red but not a green screen?
Experiment
Sixteen-month-olds were assigned to one of three conditions, a congruent, an incongruent, or an
ignorance condition (see Fig. 1). In the congruent condition, infants first watched two familiarization displays.
In each, a female human agent sat at a window behind and between two distinct objects, a red object on the
right (from the infant’s perspective) and an object of a different color on the left; she pointed to the red object.
A red toy pepper and a black cup, a red cardboard pyramid and a yellow toy house, were used in the first
and second familiarization displays, respectively. Next, the window was closed and infants watched an
orientation event in which two screens, the left one red and the right one green, were introduced. The
screens were mounted on a handle, which extended outside the apparatus, so that a hidden experimenter
could use the handle to rotate the screens. The screens first stood upright. The experimenter lowered the
screens to lay them flat on the floor and infants could see that the back of each screen was of the same
color as the front; she then rotated the screens 900 upward. This sequence was repeated once, after which
the screens were kept upright. The window was then opened and the agent sat at the window, as in
familiarization, behind and between the two upright screens. Infants saw a test event in which the agent
pointed to the red-front (red-front event) or the green-front screen (green-front event).
The incongruent condition was identical to the congruent condition except that the screens were of
different colors on the two sides: the left screen was green on the front and red on the back (green-front
screen), while the right screen was red on the front and green on the back (red-front screen).
If infants in the congruent condition realized that the agent’s actions during familiarization revealed
her preference for the color red over other colors, and expected her to maintain this color preference during
test when she saw the screens, then infants who saw the green-front event should look reliably longer than
those who saw the red-front event. In contrast, if infants in the incongruent condition again attributed to the
agent a preference for red over other colors, and recognized that the agent could only see the back but not
the front of the screens (since she was absent when the screens were introduced and hence should be
7
unaware of the fact that the screens were also red or green on the front), then they should expect the agent
to point to the green-front screen to act on her color preference because it was red on the back. Infants who
saw the red-front event should therefore look reliably longer than those who saw the green-front event.
Note that according to our predictions, infants should find the green-front event in the congruent
condition and the red-front event in the incongruent condition unexpected. Both screens were positioned on
the right. Therefore, infants might simply respond to which direction, left or right, the agent pointed during
test, ignoring the colors of the screens: they might find it more interesting when the agent pointed at the right
than at the left screen. The ignorance condition was thus designed to rule out this alternative interpretation.
This condition was identical to the congruent condition except that the screens were only colored on the
front: the left screen was red (red-front screen) and the right green (green-front screen); the back of the
screens were covered with the same light-brown contact paper that covered the apparatus floor. If infants
simply responded to which direction the agent pointed at during test, those in the ignorance condition should
respond like those in the congruent condition and look reliably longer at the green-front than at the red-front
screen. However, if infants responded by taking the agent’s perspectives, infants in the ignorance condition
should behave differently. They should realize that although the agent had a preference for the color red
over other colors, as revealed during familiarization, she should be ignorant of the colors on the front of the
screens since she was absent when the screens were introduced. Therefore, infants should have no
expectation as to which of the two screens the agent would point to during test, for both screens appeared
identical to her and neither was red. Infants who saw the red-front event and those who saw the green-front
event should thus look about equally.
Method
Participants
Participants were 42 healthy term infants, 21 male and 21 female (range: 15 months, 5 days to 17
months, 29 days, M = 16 months, 6 days). Seven infants were assigned to each of the six experimental
groups formed by crossing the three conditions (congruent, incongruent, or ignorance) and the two test
events (red- or green-front). Another 8 infants were tested but eliminated, because of fussiness (2),
8
distraction (2), observer difficulties (1), procedural problems (1), or test looking times that were more than 3
SDs from the mean (2).
Apparatus
The apparatus consisted of a wooden display box (114 cm high x 104 cm wide x 47.5 cm deep)
mounted 76 cm above the room floor. The infant sat on a parent's lap and faced an opening (53 cm x 102
cm) in the front of the apparatus. Between trials, a curtain consisting of a muslin-covered frame (61 cm x
104 cm) was lowered in front of the opening. The side walls of the apparatus were painted white, and the
floor was covered with light-brown contact paper. The back wall of the apparatus was made of a white foam
board. A rectangular window (35 cm x 45 cm) was created in its midsection. During familiarization and test,
the window was opened and a human agent, wearing a white shirt, sat centered on the apparatus and
behind the window. During orientation, the window was closed.
During familiarization, two pairs of colored toys were used. One pair consisted of a red toy pepper
(7.6 cm x 12.7 cm x 7.6 cm) and a black cup (10 cm high and 8 cm in diameter), the other consisted of a red
cardboard pyramid (13 cm high and 9.5 cm x 9.5 cm at the bottom) and a yellow toy house (19 cm x 12.5
cm x 12.5 cm).
During orientation and test, two foam board screens (20.5 cm x 20.5 cm, 0.5 cm thick and 24 cm
apart) mounted on a foam board handle (5 cm high, 102 cm long, and 0.7 cm thick) were used. The handle
was fastened to the apparatus floor by Velcro; its right end extended outside the apparatus through a slit in
a muslin-covered window (31 cm x 30.5 cm) in the right side wall. An experimenter rotated this end to raise
and lower the screens. In the congruent condition, the left screen was covered with red tape and the right
green on both sides. In the incongruent condition, the left screen was covered with green tape on the front
and red tape on the back, and vice versa for the right screen. In the ignorance condition, the left and right
screens were covered with red and green tape, respectively, only on the front; the back of the screens were
covered with the same light-brown contact paper as the floor.
Procedure
9
Two naïve observers monitored each infant's looking behavior through peepholes in large cloth-
covered frames on either side of the apparatus. The primary observer's looking times were used to
determine the endings of the trials (see below). Interobserver agreement was measured for 38 of the 42
infants and averaged 92% per trial per infant.
All infants first received four familiarization trials; they saw the familiarization display 1 on the first
and third trials in which the agent pointed at the red toy pepper and the familiarization display 2 on the
second and fourth trials in which the agent pointed at the red cardboard pyramid. The agent remained in this
position until the trial ended when infants (1) looked away for 2 consecutive seconds after having looked for
at least 2 cumulative seconds, or (2) looked for 30 cumulative seconds.
Next, infants received two orientation trials in which the two screens appropriate for their condition
were introduced. The screens remained upright until the trial ended by the same criteria as in familiarization.
Finally, all infants received one test trial in which the agent pointed to the red- (red-front event) or
the green-front screen (green-front event), remaining in this position until the trial ended. Half of infants in
each condition saw the red-front event; the other saw the green-front event. Each test trial ended when
infant (1) looked away for 2 consecutive seconds after having looked for at least 5 cumulative seconds, or (2)
looked for 60 cumulative seconds.
At the start of each familiarization and test trial, the agent sat at the window in the back wall of the
apparatus, with her eyes focused on a neutral mark between the two objects. While pointing to an object,
the agent kept her gaze on the object. Thus, the agent did not make eye contact with the infant during the
experiment.
Infants attended well to the 2-s pre-trials in both familiarization and test (means range: 1.9-s to 2.0-
s) and to the 12-s pre-trials in orientation (means range: 11.1-s to 11.9-s). Preliminary analyses of the test
main-trial data revealed no significant interaction among condition, event, and sex, F(2, 30) = 0.41; the data
were therefore collapsed across sex in subsequent analyses.
10
Results
The analysis of infants’ familiarization and orientation main-trial looking times (see Fig. 2) revealed
that infants in the six experimental groups did not differ in their looking behavior during these two phases.
The only significant effect was that of familiarization display, suggesting that across conditions, infants
looked reliably longer at display 1 on the first and third trials than at display 2 on the second and fourth trials.
Infants' test main-trial looking times (see Fig. 2) were analyzed by a 3 x 2 ANOVA with condition
(congruent, incongruent, or ignorance) and event (red- or green-front) as between-subjects factors. The
analysis yielded a significant condition x event interaction, F(2, 36) = 6.74, p < .005. Planned comparisons
revealed that (1) in the congruent condition, infants who saw the green-front event (M = 24.0, SD = 7.2)
looked reliably longer than those who saw the red-front event (M = 13.3, SD = 7.0), F(1, 36) = 8.71, p < .01,
Cohen’s d = 1.5, (2) in the incongruent condition, infants who saw the red-front event (M = 18.3, SD = 8.3)
looked reliably longer than those who saw the green-front event (M = 10.6, SD = 2.5), F(1, 36) = 4.52, p
< .05, d = 1.3, and (3) in the ignorance condition, infants who saw the green-front (M = 14.2, SD = 7.6) and
red-front (M = 16.1, SD = 6.7) events looked about equally, F (1, 36) = 0.26, d = -0.3. The main effects of
condition, F(2, 36) = 1.53, p > .20, and event, F(1, 36) = 0.03, were not significant. Examinations of
individual infants’ looking times confirmed these results (congruent condition: Wilcoxon rank-sum W = 34, p
< .025; incongruent condition: W = 36, p = .05; ignorance condition: W = 48, n.s.).1
Infants’ pointing behavior
In reviewing videotaped records of the testing sessions, it became evident that infants sometimes
pointed during the experiment. We therefore examined their pointing behavior. A pointing action was defined
as when “infants extended their arm (either fully or slightly bent) and index finger or hand” (p. F3) in the
direction of the apparatus (Liszkowski, Carpenter, & Tomasello, 2007). Two coders blind to infant’s condition
coded 41 of the 42 infants’ videotapes. Inter-coder agreement on whether pointing happened on each trial
was 93%. The judgments by the more experienced coder were used.
During familiarization, pointing happened on 11 of the 56 trials in the congruent and in the
incongruent condition (56 was the product of 14 times 4 as the 14 infants each received 4 familiarization
11
trials), and on 12 of the 52 trials in the ignorance condition (52 was the product of 13 times 4 because only
13 infants could be coded from the videotape). During orientation, pointing happened on 2 of the 28 trials in
the congruent condition (28 was the product of 14 times 2 as the 14 infants each received 2 orientation
trials), on 7 of the 28 trials in the incongruent condition, and on 3 of the 26 trials in the ignorance condition.
During test, pointing happened on 3 of the 14 trials in the congruent condition (the 14 infants each received
one test trial), on 4 of the 14 trials in the incongruent condition, and on 6 of the 13 trials in the ignorance
condition. A loglinear model (Agresti, 1996) with condition (congruent, incongruent, or ignorance), trial type
(familiarization, orientation, or test), and outcome (pointing or no-pointing) as factors and count (number of
trials) as the dependent variable best fit these data (χ212 = 10.75, p > .50). The model revealed that infants’
pointing pattern among the three conditions during orientation (χ21 = 26.26, p < .0001) and test (χ21 = 63.04,
p <.0001) differed significantly from that of familiarization. In other words, infants in the three conditions
pointed about the same amount during familiarization, but their pointing differed among conditions during
orientation and test.
Discussion
Infants in the congruent condition who saw the green-front event looked reliably longer than those
who saw the red-front event during test, suggesting that they (1) attributed to the agent a preference for the
color red over other colors after watching her point to a red toy as opposed to a black or a yellow one during
familiarization, (2) expected her to continue acting on this color preference and point at the red but not the
green screen during test, and (3) responded with increased attention when she failed to do so in the green-
front event. Infants in the incongruent condition who saw the red-front event looked reliably longer than
those who saw the green-front event, indicating that although infants attributed to the agent a color
preference based on her actions during familiarization, they (1) also realized that the agent was absent
when the screens with different colored front and back were introduced during orientation, (2) expected the
agent, who could only see the back of the screens during test, to point to the green-front screen which was
red to her, and therefore (3) responded with prolonged looking when this expectation was violated in the
12
red-front event. Infants in the ignorance condition who saw the green- and red-front events looked about
equally, suggesting that they recognized that the agent could not act on her color preference during test,
since she could only see the uncolored back of the screens, and hence accepted that the agent could point
at either screen. Together, these results revealed that infants learned and remembered the colors on both
sides of the screens during orientation but used the agent’s perceptions, which only included the back of the
screens, to predict and interpret her actions during test, even when the agents’ perceptions differed from
their own.
Infants also pointed during the experiment. Prior research on infant pointing suggests that in the
second year of life, infants point to show their attention to objects or to “communicate” with others, e.g., to
share mutual interests or to inform others of things they do not know about (e.g., Ganea, 2005; Legerstee &
Barillas, 2003; Liszkowski, Carpenter, Henning, Striano, & Tomasello, 2004; Liszkowski et al., 2007; Moore
& D'Entremont, 2001; Saylor, 2004; Tomasello, Carpenter, & Liszkowski, 2007). The pointing behavior
observed in the present research seemed to be consistent with these findings. During familiarization, infants
in the three conditions pointed about the same amount: they might simply do so to imitate the agent, or to
express or share their interests in the situation. During orientation, the condition-appropriate screens were
introduced with the agent absent. Infants’ pointing in this phase might indicate their recognition of the
novelty of the screens: they pointed the most in the incongruent condition, less so in the ignorance condition,
and least in the congruent condition. During test, infants’ pointing might suggest that they attempted to “tell”
the agent something: they pointed the most in the ignorance condition when the agent could not see the
colors on the screens, least in the congruent condition when the agent’s perceptions of the screens were
identical to theirs; whereas in the incongruent condition, infants might be pointing to inform the agent of the
fact that the screens were of different colors on the two sides, while accepting that the agent acted
according to what colors she saw. Therefore, these post-hoc patterns of infant pointing might provide
converging evidence that infants considered the agent’s incomplete perceptions, but not their own more
complete perceptions, to make sense of the agent’s actions. However, the present research was not
specifically designed to elicit infant pointing. Pointing only happened on a small portion of the trials. A
13
proportion of infants did not point at all, and it was difficult to determine whether infants were pointing at the
screens or the agent when they did point. Future research could use situations more conducive to infants’
pointing and/or other behavior in looking-time tasks to provide various measures of infants’ knowledge.2
The present results have implications for two lines of research. First, these results extend prior
findings that infants consider agents’ incomplete or false perceptions and beliefs to make sense of agents’
behavior (e.g., Allard & Onishi, 2008; Luo & Baillargeon, 2007; Onishi & Baillargeon, 2005; Sodian et al.,
2007; Song & Baillargeon, in press) to new situations in which the agent holds incomplete perceptions not
because the whole object is hidden from her, but because only one part of an object is visible to her.
Together with reports that infants also non-egocentrically take into account agents’ knowledge-states to
guide their own behavior or to interact with agents (e.g., Moll & Tomasello, 2004, 2006; O'Neill, 1996;
Tomasello & Haberl, 2003), these findings lend support to the claim that at least by the second year, infants
possess rudimentary perspective-taking abilities.
Flavell and his colleagues introduced an important distinction between two levels in the
development of children’s understanding of others’ perspectives (e.g., Flavell, 1978, 2004; Lempers, Flavell,
& Flavell, 1977; Masangkay et al., 1974). Level-1 involves understanding what objects agents can see;
children at this level realize, for example, that they may be able to see an object that another agent cannot
see or vice versa. Level-2 involves understanding how objects appear to agents; children at this level
recognize, for example, that the same object or scene may appear differently when viewed from different
perspectives or by agents with different perceptual experiences. When cast in these terms, the present
results revealed infants’ level-1 understanding that the agent could only see the back of the screens while
they themselves saw the front of the screens during test. The present findings could also be relevant to
infants’ level-2 understanding: infants, who saw both sides of the screens during orientation, realized that
the agent’s absence during orientation and her perspectives during test determined her knowledge about
the screens, which differed from their own.
Second, the present research reveals another type of simple dispositions infants ascribe to agents:
a color preference. In the present experiment, seeing four instances of two color pairings, red-black and red-
14
yellow, enabled 16-month-olds to attribute to the agent a preference for the color red, which they
generalized to the test event in a relatively new context, predicting that the agent should prefer red over a
new color, green. In Song et al. (2005a), 13.5-month-olds watched an agent slide three different objects and
realized that the agent had an inclination to slide things, which they also generalized to a relatively new
context, expecting the agent to slide a new object. Although it remains unclear whether 13.5-month-olds
could succeed with only two exemplars, as in the present study, Song and Baillargeon (2007) later found
that 9.5-month-olds in a similar study needed six exemplars. It thus seems reasonable that younger infants
need more exemplars to learn about an agent’s dispositions from her actions. These findings, together with
those showing that infants attribute to agents a positive disposition towards certain objects or agents (e.g.,
Hamlin et al., 2007; Kuhlmeier et al., 2003; Luo & Baillargeon, 2007; Repacholi & Gopnik, 1997), pose
exciting research questions. For instance, do dispositions differ in terms of how well a disposition predicts
agents’ goal-directed actions? In the present experiment, infants expect the agent to always point to a red
object. In Song et al. (2005a), infants expect the agent to always slide an object. However, in a Woodward
(1998) type of task, given the one-object condition results described in the Introduction, it seems unlikely
that infants expect an agent who prefers object-A over object-B to always approach object-A as opposed to
other objects. Therefore, it remains to be seen what evidence enables infants to attribute to an agent an
object-specific disposition, which predicts that the agent will always choose the particular object. In addition,
can infants distinguish dispositions specific to a certain person and those shared by a group (e.g., Buresh &
Woodward, 2007)? Future studies addressing these questions will shed light on the nature and development
of early psychological understanding.
15
Acknowledgments
This research was supported by research funds from the University of Missouri to the first author.
We thank Renée Baillargeon, Shawn Christ, Frieder Kohn, Lexin Li, Kris Onishi, Doug Steinley, and Kristy
vanMarle for helpful suggestions. We are also grateful to the University of Missouri Infant Cognition
Laboratory for their help with data collection, and the parents and infants who participated in the research.
16
References
Agresti, A. (1996). An Introduction to Categorical Data Analysis. New York: Wiley.
Allard, A., & Onishi, K. H. (2008). What you say is (not quite) what you see: Perspective taking in
14-month-old infants. Paper presented at the International Conference on Infant Studies.
Bíró, S., & Leslie, A. M. (2007). Infants’ perception of goal-directed actions: Development through
cue-based bootstrapping. Developmental Science, 10, 379-398.
Buresh, J. S., & Woodward, A. L. (2007). Infants track action goals within and across agents.
Cognition, 104, 287-314.
Flavell, J. H. (1978). The development of knowledge about visual perception. In C. B. Keasey (Ed.),
Nebraska Symposium on Motivation (Vol. 25, pp. 43-76). Lincoln: University of Nebraska Press.
Flavell, J. H. (2004). Development of knowledge about vision. In D. T. Levin (Ed.), Thinking and
Seeing: Visual Metacognition in Adults and Children (pp. 13-36). Cambridge, MA: MIT Press.
Ganea, P. (2005). Contextual factors affect absent reference comprehension in 14-month-olds.
Child Development, 76, 989-998.
Gergely, G., & Csibra, G. (2003). Teleological reasoning in infancy: The naïve theory of rational
action. Trends in Cognitive Sciences, 7, 287-292.
Gergely, G., Nádasdy, Z., Csibra, G., & Bíró, S. (1995). Taking the intentional stance at 12 months
of age. Cognition, 56, 165-193.
Hamlin, J. K., Wynn, K., & Bloom, P. (2007). Social evaluation by preverbal infants. Nature, 450,
557-559.
Kuhlmeier, V. A., & Wynn, K. (2003). 9-month-olds’ understanding of intentionality. Paper
presented at the Biennial meeting of the Society for Research in Child Development.
Kuhlmeier, V. A., Wynn, K., & Bloom, P. (2003). Attribution of dispositional states by 12-month-olds.
Psychological Science, 14, 402-408.
Legerstee, M., & Barillas, Y. (2003). Sharing attention and pointing to objects at 12 months: Is the
intentional stance implied? Cognitive Development, 18, 91-110.
17
Lempers, J. D., Flavell, E. R., & Flavell, J. H. (1977). The development in very young children of
tacit knowledge concerning visual perception. Genetic Psychology Monographs, 95, 3-53.
Leslie, A. M. (1995). A theory of agency. In D. Sperber, D. Premack & A. J. Premack (Eds.),
Causal cognition: A multidisciplinary debate (pp. 121-141). Oxford, UK: Clarendon Press.
Liszkowski, U., Carpenter, M., Henning, A., Striano, T., & Tomasello, M. (2004). Twelve-month-
olds point to share attention and interest. Developmental Science, 7, 297-307.
Liszkowski, U., Carpenter, M., & Tomasello, M. (2007). Pointing out new news, old news, and
absent referents at 12 months of age. Developmental Science, 10, F1-F7.
Luo, Y. (2008). Three-month-old infants attribute goals to a self-propelled box. Manuscript under
review.
Luo, Y., & Baillargeon, R. (2005). Can a self-propelled box have a goal? Psychological reasoning
in 5-month-old infants. Psychological Science, 16, 601-608.
Luo, Y., & Baillargeon, R. (2007). Do 12.5-month-old infants consider what objects others can see
when interpreting their actions? Cognition, 105, 489-512.
Luo, Y., & Johnson, S. C. (in press). Recognizing the role of perception in action at 6 months.
Developmental Science.
Masangkay, Z. S., McCluskey, K. A., McIntyre, C. W., Sims-Knight, J., Vaughn, B. E., & Flavell, J.
H. (1974). The early development of inferences about the visual percepts of others. Child Development, 45,
357-366.
Moll, H., & Tomasello, M. (2004). 12- and 18-month-old infants follow gaze to spaces behind
barriers. Developmental Science, 7, F1-F9.
Moll, H., & Tomasello, M. (2006). Level I perspective-taking at 24 months of age. British Journal of
Developmental Psychology, 24, 603-613.
Moore, C., & D'Entremont, B. (2001). Developmental changes in pointing as a function of
attentional focus. Journal of Cognition and Development, 2, 109-129.
18
O'Neill, D. K. (1996). Two-year-old children's sensitivity to a parent's knowledge state when making
requests. Child Development, 67, 659-677.
Onishi, K. H., & Baillargeon, R. (2005). Do 15-month-old infants understand false beliefs? Science,
308, 255-258.
Phillips, A., & Wellman, H. M. (2005). Infants’ understanding of object-directed action. Cognition,
98, 137-155.
Premack, D., & Premack, A. J. (1995). Origins of human social competence. In M. S. Gazzaniga
(Ed.), The cognitive neurosciences (pp. 205-218). Cambridge, MA: MIT Press.
Premack, D., & Premack, A. J. (1997). Infants attribute value +/- to the goal-directed actions of self-
propelled objects. Journal of Cognitive Neuroscience, 9, 848-856.
Repacholi, B., & Gopnik, A. (1997). Early understanding of desires: Evidence from 14- and 18-
month-olds. Developmental Psychology, 33, 12-21.
Saylor, M. M. (2004). Twelve- and 16-month-old infants recognize properties of mentioned absent
things. Developmental Science, 7, 599-611.
Sodian, B., Thoermer, C., & Metz, U. (2007). Now I see it but you don't: 14-month-olds can
represent another person's visual perspective. Developmental Science, 10, 199-204.
Sommerville, J. A., Woodward, A. L., & Needham, A. (2005). Action experience alters 3-month-old
infants’ perception of others’ actions. Cognition, 96, B1-B11.
Song, H., & Baillargeon, R. (2007). Can 9.5-month-old infants attribute to an agent a disposition to
perform an action on objects? Acta Psychologica, 124, 79-105.
Song, H., & Baillargeon, R. (in press). Infants' reasoning about others' false perceptions.
Developmental Psychology.
Song, H., Baillargeon, R., & Fisher, C. (2005a). Can infants attribute to an agent a disposition to
perform a particular action? Cognition, 98, B45-B55.
19
Song, H., Baillargeon, R., & Fisher, C. (2005b). The development of infants' use of verbal
information when reasoning about others' actions. Paper presented at the Biennial meeting of the Society
for Research in Child Development.
Southgate, V., Senju, A., & Csibra, G. (2007). Action anticipation through attribution of false belief
by two-year-olds. Psychological Science, 18, 587-592.
Southgate, V., van Maanen, C., & Csibra, G. (2007). Infant pointing: Communication to cooperate
or communication to learn? Child Development, 78, 735-740.
Surian, L., Caldi, S., & Sperber, D. (2007). Attribution of beliefs by 13-month-old infants.
Psychological Science, 18, 580-586.
Tomasello, M., Carpenter, M., & Liszkowski, U. (2007). A new look at infant pointing. Child
Development, 78, 705-722.
Tomasello, M., & Haberl, K. (2003). Understanding attention: 12- and 18-month-olds know what is
new for other persons. Developmental Psychology, 39, 906-912.
Woodward, A. L. (1998). Infants selectively encode the goal object of an actor's reach. Cognition,
69, 1-34.
Woodward, A. L., & Guajardo, J. J. (2002). Infants’ understanding of the point gesture as an object-
directed action. Cognitive Development, 17, 1060-1084.
20
Footnotes
1. The test data were subjected to an analysis of covariance (ANCOVA); the factors were as in the ANOVA,
and the covariates were the infants’ mean looking times at the display 1 and display 2 during the
familiarization trials and their mean looking times during the orientation trials. The results replicated those of
the ANOVA: the condition x event interaction was significant, F(2, 33) = 4.84, p < .025; and planned
comparisons yielded similar results to those of the ANOVA (congruent condition: F(1, 33) = 6.97, p < .025;
incongruent condition: F(1, 33) = 4.55, p < .05; ignorance condition: F(1, 33) = 0.20).
2. The pointing data suggest that infants might construe the present situation as communicative (e.g.,
Southgate, van Maane, & Csibra, 2007; Tomasello et al., 2007). Although the agent did not establish eye
contact or have verbal communication with the infant, the fact that she pointed might indicate her intention to
communicate. Infants might thus point to “respond” to her. However, infants even pointed during orientation
in which the agent was absent from the apparatus. It is thus also possible that infants pointed to
communicate with their parents, who held them sitting in front of the apparatus. Our data were simply
insufficient to determine which of these possibilities was at play.
21
Figure Captions
Figure 1. Schematic drawing of the events shown in the congruent, incongruent, and ignorance conditions.
Congruent condition. (a) Familiarization display 1. Each familiarization trial consisted of a 2-s pre-trial and a
main-trial. At the start of the pre-trial, the agent sat behind the back window between the toy pepper and the
cup, looking at a neutral mark between them; the two toys were 34 cm apart and each about 13 cm from the
back wall. When the computer signaled that the infant had looked for 2 cumulative seconds, the 2-s pre-trial
began in which the agent used her right index finger to point to the red toy pepper while turning to look at it;
she then paused, with her eyes focused on the pepper. In the main-trial, the infants watched this paused
scene until the trial ended. When this occurred, an experimenter hidden behind the apparatus lowered the
curtain in front of the apparatus. (b) Familiarization display 2. This display was similar to the familiarization
display 1 except that the toy pepper and the cup were replaced with the cardboard pyramid and the toy
house. The two toys were 27 cm apart; the backmost tip of the pyramid was 5.5 cm from the back wall and
the house was 13 cm from it. (c) Orientation event. Each orientation trial consisted of a 12-s pre-trial and a
main-trial. At the start of the pre-trial, the screens stood upright, 13 cm from the back wall. When the
computer signaled that the infant had looked for 2 cumulative seconds, the 12-s pre-trial began. After a 1-s
pause, the experimenter lowered the screens so that they lay flat on the floor (2 s). After another 1-s pause,
the experimenter rotated the screens 90 degrees upward (2 s). This sequence was repeated once;
afterwards, the screens were kept upright. In the main-trial, the infants watched this paused scene until the
trial ended. (d) Test events. Green-front event. This event was similar to the familiarization display 1 except
that the toys were replaced with the two upright screens and the agent pointed at the green-front screen.
Red-front event. This event was identical to the green-front event except that the agent pointed to the red-
front screen. Incongruent and Ignorance conditions. (a-d). The events shown in the two conditions were
identical to those of the congruent condition except that the screens appropriate for the condition were used
during orientation and test.
22
Figure 2. Mean looking times during the familiarization, orientation, and test trials of the infants in the six
experimental groups formed by crossing the congruent, incongruent, and ignorance conditions and the
green-front and red-front test events. Error bars represent standard errors. An asterisk (*) denotes p < .05.