Developmental Psychology Copyright 2007 by the American Psychological Association

2007, Vol. 43, No. 5, 1227–1238 0012-1649/07/$12.00 DOI: 10.1037/0012-1649.43.5.1227

Do 9-Month-Old Infants Expect Distinct Words to Refer to Kinds?

Kathryn Dewar and Fei Xu
University of British Columbia

In 3 experiments, 9-month-old infants’ expectations for what distinct count noun labels refer to were
investigated. In Experiment 1, a box was opened to reveal 2 objects inside during familiarization: either
2 identical objects or 2 different objects. Test trials followed the same procedure, except before the box
was opened, the contents were described using 2 distinct labels (“I see a wug! I see a dak!”) or the same
label twice (“I see a zav! I see a zav!”). Infants who heard a label repeated twice looked longer at 2
different objects versus 2 identical objects, whereas infants who heard 2 distinct labels showed a different
pattern of looking. Experiments 2 and 3 presented infants with object pairs that only differed in shape or
color, and it was found that infants expected the different-shaped (but not the different-colored) objects
to be labeled by distinct count nouns. Because the property of shape is a cue to kind membership and the
property of color is not, these results suggest that even at the beginning of word learning, infants may
expect distinct labels to refer to distinct kinds of objects.

Keywords: sortal concepts, labeling, object representations

In everyday life, adults and children keep track of the kinds of the same dog encountered earlier or whether we have seen a
objects they encounter and the individuals and individual objects different dog on each occasion.
they interact with. We categorize objects such as dogs, chairs, and To appreciate the conceptual role of sortals, consider two ques-
people, and we have certain cognitive capacities to keep track of tions. First, how many are there? And second, is that the same as
them over time and space: “Is this the same coffee mug that I left what was here before? It is impossible to answer either question
on my desk yesterday? I would not want to drink from someone without specifying a sortal— how many of what? One can count
else’s mug! Is that my best friend Sally? Looks like she changed cups, desks, people, pages, or fingers, but one cannot count the
her hair and she is wearing contacts instead of glasses.” Our blue, the sleeping, or the metal. Only sortals provide criteria of
perceptual and conceptual system cares a great deal about the individuation. Similarly, “same” in the sense of numerical identity,
number of objects in our mental model, the kinds of objects we indicates the same one, and a sortal is required to specify the
represent, and how to keep track of the objects over time and individual being traced through time. Max the puppy grows to
space. become an adult dog, changes size, coloring, shape, and location
A subset of our concepts—namely, sortal concepts—refers to but is still the same dog. Max’s identity is traced by the sortal dog.
kinds and supplies the criteria for individuation (where one object Various studies have sought to determine when young children
ends and another one begins) and identity (whether an object is the begin to represent sortal/kind concepts. Spelke, Kestenbaum, Si-
same one as was seen on a different occasion; Gupta, 1980; Hirsch, mons, and Wein (1995) and Xu and Carey (1996) found that at
1982; Macnamara, 1986; Macnamara & Reyes, 1994; Wiggins, both 4 and 10 months of age, infants are able to use spatiotemporal
1980; Xu, 1997; Xu & Carey, 1996). Sortal/kind information evidence for object individuation, thus demonstrating the ability to
specifies categorization under concepts such as dog, ball, and car, represent the sortal concept “physical object.” In contrast, it is not
categories of objects united by functional/causal features as well as until 10 –12 months that infants are able to use basic-level sortal/
by perceptual features. The sortal concept “dog,” for example, kind information for object individuation (Bonatti, Frot, Zangl, &
provides criteria for deciding whether we see one or two dogs; it Mehler, 2002; Van de Walle, Carey, & Prevor, 2000; Wilcox &
also provides criteria for deciding whether the dog we see now is Baillargeon, 1998a [Experiments 1 and 2]; Xu & Carey, 1996; Xu,
Carey, & Quint, 2004). More specifically, by 12 months of age,
infants are able to use the differences between a duck and a ball,
Kathryn M. Dewar and Fei Xu, Department of Psychology, University or a cup and a bottle, to establish a representation of two objects.
of British Columbia, Vancouver, British Columbia, Canada. In Xu and Carey (1996), infants were shown an event in which an
This work was supported by grants from the Natural Sciences and object, a toy duck, emerged from behind a screen then returned
Engineering Research Council of Canada and the Social Sciences and behind it, followed by another object, a ball, emerging from behind
Humanities Research Council to Fei Xu. We thank Susan Birch, Geoff the same screen then returning behind it. Adults draw on kind
Hall, and Janet Werker for helpful discussion, and members of the Uni-
differences (ducks and balls are two different kinds of objects) to
versity of British Columbia Infant Cognition Lab for their help in data
conclude that two distinct objects are involved in this event. The
collection.
Correspondence concerning this article should be addressed to Kathryn screen was then removed to reveal either both objects (the duck
Dewar, Department of Psychology, University of British Columbia, 2136 and the ball: the expected outcome) or just one of the two objects
West Mall, Vancouver, British Columbia V6T 1Z4, Canada. E-mail: (the duck or the ball: the unexpected outcome). At 10 months,
[email protected] infants did not look longer at the unexpected outcome of a single
1227
1228 DEWAR AND XU

object, suggesting that they did not use the kind differences be- nouns for object categories (Bloom, 2000; Hall, 1993; Nelson,
tween these objects to conclude that there were two distinct objects 1973). Recent studies provide evidence that language may play a
behind the screen. However, by 12 months of age, infants suc- causal role in this process (Balaban & Waxman, 1996; Waxman &
ceeded at this task, looking longer at the unexpected, single-object Braun, 2005; Xu, 2002; Xu, Cote, & Baker, 2005).
outcome. Xu and Carey (1996) hypothesized that 10-month-olds Balaban and Waxman (1996) found that words, but not tones,
do not represent basic-level sortal/kind concepts such as duck and facilitate categorization in 9-month-old infants. Infants were fa-
ball. Subsequently, using both looking time and manual search miliarized to pictures of a given category (e.g., rabbits). Some of
measures, Xu, Carey, and Welch (1999) and Van de Walle et al. the infants heard a word when shown a picture. For other infants,
(2000) provided convergent evidence for this shift between 10 and a tone accompanied the presentation of the picture. The findings
12 months (see also Bonatti et al., 2002; Krojgaard, 2000; Wilcox indicate that although both the presentation of the words and the
& Baillargeon, 1998a [Experiments 1 and 2]).1 tones effectively heightened the infants’ attention to the objects, it
A recent study has provided evidence that 12-months-olds suc- was only in the label condition that infants succeeded in catego-
ceed at this object individuation task through the use of sortal/kind rizing the objects. When provided a label, it seems that infants
information as opposed to property or featural differences. Xu et group exemplars into a single category more readily than they do
al. (2004), using the paradigm of Xu and Carey (1996), conducted in the absence of a label. Waxman and Braun (2005) showed that
a series of experiments with 12-month-old infants to determine only consistent labeling, not variable labeling, facilitated catego-
whether the ability to use the differences between a duck and a ball rization in 12-month-old infants. A post hoc analysis from the
for object individuation is based on property differences (i.e., study by Xu and Carey (1996) provides further suggestion that
yellow, irregularly shaped, and rubbery vs. red, round, and shiny) language may augment the acquisition of sortal/kind concepts.
or kind differences (i.e., a member of the kind duck vs. a member Although, as a whole, the 10-month-olds failed at the task of object
of the kind ball). They asked if infants would individuate objects individuation, those 10-month-olds with some linguistic knowl-
on the basis of property differences alone—for example, color edge of the objects were able to perform more like their 12-month-
differences (a red ball vs. a green ball), size differences (a small old counterparts. Perhaps having labels for objects is a means of
red ball vs. a big red ball), or combinations of these properties (a establishing that they belong to different kinds.
small red ball vs. a big green ball). The results indicated that To more directly examine the role of language in such a task, Xu
infants failed to use these property differences for object individ- (2002) presented 9-month-olds with the same object individuation
uation. These findings suggest that certain salient shape differ- task as was used in Xu and Carey (1996), with one crucial
ences enter into the computation of the numerical distinctness of manipulation: As each object emerged from behind the screen, the
objects before other property differences such as color and size. infants heard a label for the object in infant-directed speech (e.g.,
When within-kind shape differences (e.g., a regular cup vs. a sippy “Look, a duck!” or “Look, a ball!”). In the one-label condition, the
cup) were contrasted with cross-kind shape differences (e.g., a infants heard a single label applied to both objects (e.g., “Look, a
regular cup vs. a bottle), infants failed to use the within-kind shape toy”). In the two-label condition, but not in the one-label condition,
differences for object individuation, whereas they succeeded in infants looked longer at the unexpected outcome. Thus, upon
using cross-kind shape differences, even when the two types of hearing two contrastive labels when seeing the emerging objects,
shape differences were roughly equally salient to them. Since even 9-month-old infants were able to use the differences in object
shape differences are often correlated with object kind differences, kind to establish a representation of two distinct objects. In addi-
these results converge with previous findings suggesting that at 12 tion, 9-month-olds succeeded on this task when a pair of unfamil-
months, infants’ represent some basic-level sortals/kinds, and it is iar objects labeled with nonsense words (e.g., “a fendle” or “a
sortal/kind distinctions that support object individuation. Thus, at toma”) was used. The fact that familiarity with the objects and the
12 months of age, infants’ success in a complex object individu- labels was not necessary to succeed at this task suggests that it is
ation task may be based on object kind representations as opposed the presence of distinct labels, per se, that allows infants to
to property representations (see Xu, 2005, for a review). establish a representation of two distinct individual objects.
Additional evidence from other laboratories also suggests that Are these facilitation effects language specific? Would other
by the end of the 1st year of life, infants are able to distinguish types of auditory information be equally helpful for this individ-
property and kind information. Waxman and Markow (1995) and uation task? Subsequent experiments in Xu (2002) used two tones,
Waxman (1999) reported that by 13 months of age, the distinction two distinct sounds, or two emotional expressions instead of two
between property and kind plays a role in infants’ categorization.
In these studies, infants were sensitive to whether they heard a
1
count noun or an adjective while examining objects. If the children There is evidence to suggest that infants younger than 12 months of age
heard an adjective, they were more likely to generalize on the basis are able to use property information for establishing the representation of
of a property, such as color or texture; however, if they heard a two distinct objects when task demands are reduced. For example, Wilcox
count noun, the infants were more likely to generalize to objects of and Baillargeon (1998a [Experiments 7 and 8], 1998b, Wilcox &
the same kind on the basis of shape. These findings are consistent Schweinle, 2002), using a simplified object individuation procedure,
showed that 9-month-olds or even younger infants were able to use featural
with the findings of Xu et al. (2004) that the distinction between
information to individuate objects. Xu and Baker (2005), using a simplified
kinds and properties is present by about 12–13 months and maps manual search measure, demonstrated this ability in 10-month-olds. The
onto the linguistic distinction between count nouns and adjectives. present discussion focuses on when infants’ begin to use sortal/kind infor-
How do infants acquire sortal/kind concepts? Many have noted mation for object individuation, which is a related but distinct question
that infants begin to comprehend and produce their first words from when they begin to use property information for this purpose (see Xu,
toward the end of the 1st year, and many of these first words are 2005, for a review).
 WORDS AND KINDS 1229

words. Nine-month-olds did not look longer at the one-object, identical objects or two different objects. The test trials followed
unexpected outcome during the test trials when tones, sounds, or the same procedure, and used the same objects, as familiarization
emotional expressions were used. These findings suggest that trials except that before the box was opened, the experimenter
infants are able to use distinct labels to help them to succeed earlier looked into the top of the box and described its contents using two
(i.e., at 9 months as opposed to 12 months) in a task of object distinct object labels (e.g., “I see a wug!” and “I see a dak!”) or the
individuation, and these facilitation effects appear to be language same label twice (e.g., “I see a zav!” and “I see a zav!”). The
specific. experimenter looked intently into the box during labeling, provid-
How powerful are words in guiding object individuation? Xu et ing abundant intentional and referential cues, as in Xu et al. (2005).
al. (2005), using a manual search procedure, asked if 12-month-old The box was then opened to reveal the object-pair outcome (either
infants could use labeling to establish object representations even two identical objects or two different objects). The question of
when the objects were never shown to them. On a two-word trial, interest was whether the number of distinct labels would help
an experimenter looked into the opening of a box and referred to infants determine the nature of objects inside the box. If infants
what was inside (e.g., “Look, a fep!” and “Look, a wug!”). On a expect that distinct labels refer to distinct kinds of objects, when
one-word trial, the experimenter looked into the box and used the two distinct labels are used to refer to unseen objects inside a box,
same word twice (e.g., “Look, a zav!”). Infants spontaneously they should look longer when shown two identical objects inside.
reached into the box and always retrieved one object. The box was However, if infants expect only that distinct labels refer to distinct
then empty and sat within reach of the infant. If they had heard two individual objects, when two distinct labels are used, they should
distinct labels, an adult would reach in a second time to look for look equally long whether two identical objects or two different
another object but would not reach in again if he or she had heard objects are revealed because, in both outcomes, two distinct indi-
only a single label. Twelve-month-old infants behaved similarly. vidual objects are present.
Even without having seen the objects beforehand, the act of In a second experiment, using the same procedure, infants were
referring led the 12-month-olds to posit objects inside the box, and presented with either identical pairs of objects or pairs of objects
the number of distinct labels appeared to inform the infants of the that were the same in every property but shape. The property of
number of objects to be expected inside the box. In addition, this shape is a very salient cue to kind membership (Landau, Smith, &
effect seems to be specific to linguistic expressions since the Jones, 1988; Soja, Carey, & Spelke, 1991). Infants may expect that
number of emotional expressions did not help infants establish the objects that differ in shape should have different labels. However,
number of objects inside the box. Thus, labeling, when presented if different objects differ along a property dimension not linked to
with abundant intentional and referential cues, can guide the pro- kind membership, infants may show a different looking pattern.
cess of establishing the number of distinct objects involved in an This possibility was examined in a third experiment, in which
event for infants as young as 12 months of age. objects pairs were either identical or differed only in color. The
One empirical question remains unanswered from the results of current set of experiments addressed the question of whether
Xu (2002) and Xu et al. (2005): Did the infants in this study infants, at the beginning of word learning, expect distinct words to
interpret the distinct words as referring to distinct kinds of objects refer to distinct kinds of objects.
(types) or distinct individual objects (tokens)? Studies investigat-
ing the facilitating effect of language on object individuation (Xu,
2002; Xu et al., 2005) provide evidence that the use of two distinct Experiment 1
labels leads infants to posit two objects involved in the event, but
Method
what cannot be determined from the empirical evidence to date is
whether infants expect there to be merely two individual objects Participants. Participants were 24 full-term infants (12 male,
present or whether the objects involved must be of two different 12 female; mean age ⫽ 9 months, 0 days; range ⫽ 8 months, 8
kinds. For adults and older children, distinct words (in the form of days, to 9 months, 15 days). All infants were recruited from the
count nouns) tend to refer to distinct kinds of objects. Further greater Vancouver (British Columbia, Canada) area by mail and
empirical evidence is required to establish that the presence of two subsequent phone calls. Most infants came from a middle-class,
distinct labels leads infants to expect two different kinds of objects. non-Hispanic White background, with 21% of infants Asian, 4%
The current study sought to address this question directly by African Canadian, and 4% Hispanic. The infants received a token
examining the effect of labeling on 9-month-old infants’ expecta- gift (a T-shirt or bib with a university logo) after the study. English
tions regarding the nature of named objects. Since previous studies was the primary language spoken at home for all infants. An
have established that providing two labels leads 9-month-olds to additional 4 infants were tested but excluded due to fussiness (3)
expect two objects to be involved in an event, in the current study, or parental interference (1).
we focused on infants’ specific expectations about the two objects Materials. Objects were presented in a 28 ⫻ 19 ⫻ 23-cm box
themselves. Instead of asking infants to decide how many objects constructed out of foam core. The top of the box had an opening
there were in an event, the current study always presented two that measured 18 ⫻ 10 cm and was covered by spandex. The front
objects, and what was manipulated was the degree of difference of the box swung open in a doorlike movement. Four pairs of
between the two objects. Infants were asked to use labeling infor- objects were used in the study: a toy frog (approximately 8 ⫻ 14
mation as a cue to predict which of the two-object outcomes would cm in size) and a plush toy lion (approximately 14 ⫻ 8 cm in size),
be revealed. a plush toy dog (approximately 10 ⫻ 12 cm in size) and a plush
In the first experiment, infants watched events presented on a bird (approximately 9 ⫻ 12 cm in size), a toy train (approximately
puppet stage. During the familiarization trials, a box was opened to 9 ⫻ 6 cm in size) and a star-shaped puzzle (approximately 10 ⫻
reveal two objects inside. The revealed objects were either two 10 cm in size), and a double pinwheel (approximately 9 ⫻ 17 cm
1230 DEWAR AND XU

in size) and a ball figure (approximately 8 ⫻ 15 cm in size). Each looking for the observer. During the experiment, the experimenter
of the eight objects had a duplicate. Each object sat atop a square sat behind the stage in view of the infant at all times.
of white foamcore with a magnet on the bottom so that the objects To begin, the experimenter looked into the top of the empty box,
could be spaced a uniform distance apart (see Figure 1 for an pulling apart the spandex covering so as to look inside. The front
example of the identical and different object outcomes). door of the box was opened to show the infant that the box was
Apparatus. The events were presented on a stage with a dis- empty. In infant-directed speech, the experimenter said, “Look, it’s
play area that measured 94 cm in width and 55 cm in height. The empty! There’s nothing in there!” The door of the box was then
infant sat in a high chair about 30 cm from the stage, with eye level closed and the box turned around. The familiarizations trials were
slightly above the floor of the stage (about 8 cm). The parent sat then started.
next to the infant with his or her back toward the stage. A video Familiarization trials. Each infant received eight familiariza-
camera, set up under the stage, focused on the infant’s face and tion trials. Two objects were placed inside the box, out of view of
recorded the entire session. The video camera was connected to a the infant. The box was then turned to face the infant. The
19-in (48.26-cm) TV placed in one corner of the room. An ob- experimenter pulled apart the spandex on top of the box, so she
server watched the infant on the TV monitor and recorded the was looking at the objects inside the box. While looking inside, the
infant’s looking times. The observer was not able to see what was experimenter said, “I see something! There’s something there!
presented on the stage, nor was she aware of the order of the trials. [Infant’s name], look!” The front door of the box was then opened
A key on a laptop computer was pressed during infants’ on-target to reveal the objects inside. The experimenter lowered her head
looking. A computer program written specifically for looking time and eye gaze to ensure that she was not making eye contact with
studies (Hypercard [Version 2.4.1]; Pinto, 1995) was used to the infant while the box contents were visible. Infant looking times
record the looking times. were recorded. When the infant turned away for 2 consecutive
Design and procedure. The experimenter began by waving a seconds, the trial ended. The door of the box was closed, and the
set of keys at all corners of the stage to define the window of box was turned around so that the door of the box was facing the

Figure 1. Examples of identical and different (or different-shaped, or different-colored) object outcomes from
Experiments 1–3.
 WORDS AND KINDS 1231

experimenter and the infant was facing the back of the box. New
objects were placed inside the box to begin the next trial. Objects
from each of the four pairs (frog–lion, dog– bird, train–puzzle,
ball–pinwheel) were shown during the familiarization trials (either
both objects of the pair were shown [different objects], or one
object from the pair was shown with its duplicate [identical ob-
jects]). Familiarization Trials 5– 8 were a repetition of Trials 1– 4.
Which objects were shown, which side of the box an object was
positioned on, and the order of object pairs were counterbalanced
across infants.
Test trials. Test trials followed an identical procedure to that
of the familiarization trials, with one critical difference: Before
opening the front of the box, the experimenter looked into the top
of the box and labeled the objects inside with either the same label
two times (e.g., “I see a zav! I see a zav! There’s a zav! There’s a
zav! [Infant’s name], a zav! [Infant’s name], a zav!”) or with two
different labels (e.g., “I see a wug! I see a dak! There’s a wug!
There’s a dak! [Infant’s name], a wug! [Infant’s name], a dak!”).
Each sentence was spoken in infant-directed speech as the exper-
imenter looked into the box. Each infant was shown four test trials.
Infants were shown two label/object combinations for both the
expected and unexpected outcomes. For an expected outcome, an
infant either heard two different labels applied to the objects inside
the box and two different objects were revealed when the box was
opened, or, conversely, an infant heard one repeated label applied
to the objects inside the box and two identical objects were
revealed when the box is opened. For an unexpected outcome, an
infant either heard two different labels applied to the objects inside
the box and two identical objects were revealed when the box was
opened, or, conversely, an infant heard one repeated label applied
to the objects inside the box and two different objects were
revealed when the box was opened. The 8 objects were labeled
with nonsense words (“fep,” “zav,” “wug,” “dak,” “toma,”
“blicket,” “muba,” and “tupple”). Throughout the study, a partic-
ular object was always labeled with the same nonsense word; thus,
each object pair was always labeled with the same label pair. The
same objects from each of the four pairs (frog–lion, dog– bird,
train–puzzle, ball–pinwheel) that were shown during the familiar-
ization trials were shown on the test trials, but the order of object
presentation differed from that of the familiarization trials. The
four test trials included two instances of expected and unexpected
outcomes such that all four of the Label ⫻ Object Pair outcomes Figure 2. A schematic representation of the experimental procedure for
were shown (repeated label/identical objects [expected], different Experiment 1.
labels/different objects [expected], repeated label/different objects
[unexpected], different labels/identical objects [unexpected]). The
order of outcome (whether the infant saw an expected or unex-
Results
pected trial first, whether an identical or different object outcome
was presented first, or whether objects were first labeled with The results of Experiment 1 are shown in Figure 3. An alpha
repeated or different labels) was counterbalanced across infants level of .05 was used in all statistical analyses. Preliminary anal-
(see Figure 2 for a schematic representation of the experimental yses found no effects of gender, test trial order (whether the
procedure). expected or unexpected trial was presented first), or animacy
To evaluate whether any unintentional variations in the exper- (animate vs. inanimate) of the object pairs. Subsequent analyses
imenter’s actions or utterances during the test trials could have were collapsed over these variables. All infants were offline ob-
biased infants’ responses, a coder blind to the order of object-pair served by a second observer who was completely blind to the order
outcome reviewed the test trials for each infant and guessed the of object outcome. Interscorer reliability averaged 96%.
object outcome for each trial on the basis of the experimenter’s Familiarization trials. Averaging across all eight of the famil-
labeling and referential behavior. The coder was able to guess the iarizations trials, it was found that infants looked slightly longer
object-pair outcome (identical or different) on only 50% of the test when two different objects were revealed (Mdifferent ⫽ 14.04 s, SD ⫽
trials (48 out of 96 trials), which is not different from chance. 6.03) than when two identical objects were revealed (Midentical ⫽
1232 DEWAR AND XU

Discussion
When infants hear one repeated label, they expect the duplicated
label to refer to identical objects. Conversely, a trend in the present
looking data suggests that when infants hear two distinct labels,
they seem to expect those labels to refer to distinct objects. Note
that the latter effect seemed weaker (i.e., the t test was not, by
itself, statistically significant). A possible reason why infants
might not have seemed quite as surprised to see two identical
objects after hearing two distinct labels is that this looking pattern
required infants to overcome their slight baseline preference for
the different-object outcome. Thus, for infants to look longer at the
identical object pair (the unexpected outcome) after hearing two
distinct labels, they had to overcome this preference. It is not the
case, however, that infants simply preferred to look longer at the
Figure 3. Mean looking times as a function of the number of distinct different-object outcome in general, because there was no main
labels heard and the revealed object-pair outcome for Experiment 1.
effect of object outcome.
Up until now, studies investigating the facilitating effect of
language on object individuation (i.e., Xu, 2002; Xu et al., 2005)
12.61 s, SD ⫽ 4.91); however, this difference was not statistically have provided evidence that the use of two distinct labels leads
significant, t(23) ⫽ ⫺1.14, p ⫽ .27. infants to posit two objects involved in an event. The results of
Test trials. Infants’ looking times to the test outcomes were Experiment 1 indicate that infants may not only expect two distinct
compared by means of a 2 ⫻ 2 repeated measures analysis of labels to refer to two individual objects; a trend in the looking data
variance (ANOVA), with number of labels (one vs. two) and suggests that infants may expect those individual objects to be
object-pair outcome (identical vs. different) as within-subject different.
factors. The analysis revealed a significant interaction, F(1, What remains to be demonstrated is whether the infants expect
23) ⫽ 5.06, p ⫽ .03, effect size (␩p2) ⫽ .18. We performed these different objects to vary from each other in a specific fashion.
planned comparisons for each label number (one repeated label In other words, must the two objects simply be dissimilar to each
vs. two different labels) to determine whether infants looked other in some respect, or is the way in which the objects differ
longer to one of the two object outcomes (either identical important? We use distinct words to designate different kinds of
objects or different objects). When infants heard the box con- things (i.e., objects within a kind share the same label). And, in
tents described using one label repeated twice, they looked general, objects that vary along a property dimension that does not
significantly longer when two different objects were revealed affect kind membership are not marked by different basic-level
(the unexpected outcome; M ⫽ 10.02 s, SD ⫽ 7.86) than when labels. When we consider various perceptual dimensions of ob-
two identical objects were revealed (the expected outcome; jects, some are better correlated with kind membership than others.
M ⫽ 5.41 s, SD ⫽ 3.57), t(23) ⫽ ⫺2.93, p ⬍ .01. When infants The property shape is closely connected with kind membership
heard the box contents described using two distinct labels, they (Landau et al., 1988; Rosch, Mervis, Gray, Johnson, & Boyes-
looked longer when two identical objects were revealed (the Braem, 1976; Soja et al., 1991). Broadly speaking, objects that
unexpected outcome; M ⫽ 11.20 s, SD ⫽ 10.04) than when two differ in shape are usually different kinds of things and are marked
different objects were revealed (M ⫽ 8.38 s, SD ⫽ 9.53), by different basic-level count noun labels (Samuelson & Smith,
although this difference was not significant, t(23) ⫽ 1.13, p ⫽ 2005). In contrast, the property of color is not closely connected
.27. Examination of individual infants’ patterns of looking, by with kind membership and it is not usually marked by different
means of nonparametric analyses, yielded similar results be- basic-level count nouns.
tween these conditions. When one repeated label was applied to If infants expect two distinct labels to mark two different kinds
the box contents, 16 of 24 infants looked longer when two of objects, they will look longer at the unexpected outcomes if the
different objects were revealed (unexpected outcome) than property difference between the objects implies a difference in
when two identical objects were revealed (expected outcome; kind (e.g., shape), but infants will not look longer at the unex-
Wilcoxon signed ranks test: z ⫽ ⫺2.23, p ⫽ .01 [one-tailed]). pected outcomes if the property difference between the objects
Conversely, when two distinct labels were applied to the box does not imply a difference in kind membership (e.g., color). In
contents, 16 of the 24 infants looked longer when two identical Experiments 2 and 3, we used the same procedure as Experiment
objects were revealed (unexpected outcome) than when two 1, except that the different object pairs differed only in shape or
different objects were revealed (expected outcome; Wilcoxon color, respectively.
signed-ranks test: z ⫽ ⫺1.66, p ⬍ .05 [one-tailed]). The use of
one-tailed tests for the nonparametric analysis was justified Experiment 2
because we had clear predictions for the infant looking patterns.
Method
When one repeated label was heard, we expected infants would
be surprised to see two different objects revealed but not Participants. Participants were 16 full-term infants (8 male, 8
surprised to see two identical objects. The opposite was true female; (mean age ⫽ 9 months, 2 days; range ⫽ 8 months, 15
when two different labels were heard. days, to 9 months, 13 days). All infants were recruited from the
 WORDS AND KINDS 1233

same population as in Experiment 1, but none had participated in number of labels (one vs. two) and object-pair outcome (identical
the first experiment. Most infants came from a middle-class, vs. different) as with-subject factors. The analysis revealed a
non-Hispanic White background, with about 13% of infants Asian significant interaction, F(1, 15) ⫽ 8.16, p ⫽ .01, effect size
and 6% Hispanic. English was the primary language spoken at (␩p2) ⫽ .35. Planned comparisons were performed. When infants
home for all infants. An additional 5 infants were tested but heard the box contents described using a single label repeated
excluded due to fussiness (4) or parental interference (1). twice, they looked significantly longer when two different objects
Materials and apparatus. All materials and apparatus were the were revealed (the unexpected outcome; M ⫽ 10.34 s, SD ⫽ 6.06)
same as those used in Experiment 1 except for the objects pre- than when two identical objects were revealed (the expected out-
sented to the infants. Four pairs of objects were used in this come; M ⫽ 6.89 s, SD ⫽ 3.41), t(15) ⫽ ⫺2.14, p ⫽ .05. When
experiment. The objects in each pair were identical in material, infants heard the box contents described using two distinct labels,
texture, and color; thus, the objects in each pair differed from one they looked longer when two identical objects were revealed (the
another only in shape. Every object had a duplicate (see Figure 1). unexpected outcome; (M ⫽ 10.04 s, SD ⫽ 8.84) than when two
Design and procedure. These were identical to those of Ex- different objects were revealed (M ⫽ 5.72 s, SD ⫽ 3.72), although
periment 1. A coder blind to the order of object-pair outcome this difference was only marginally significant, t(15) ⫽ 2.03, p ⫽
reviewed the test trials for each infant and guessed the object .06. Examination of individual infants’ patterns of looking, by
outcome for each trial on the basis of the experimenter’s labeling means of nonparametric analyses, provided similar results. When
and referential behavior. The coder was able to guess the object- one repeated label was applied to the box contents, 12 of 16 infants
pair outcome (identical or different) on only 44% of the test trials looked longer when two different objects were revealed (unex-
(28 out of 64 trials), which is not different from chance. pected outcome) than when two identical objects were revealed
(expected outcome; Wilcoxon signed-ranks test: z ⫽ ⫺1.86, p ⫽
Results .03 [one-tailed]). Conversely, when two distinct labels were ap-
plied to the box contents, 9 of the 16 infants looked longer when
The main results of Experiment 2 are shown in Figure 4. two identical objects were revealed (unexpected outcome) than
Preliminary analyses found no effects of gender, test trial order when two different objects were revealed (expected outcome;
(whether the expected or unexpected trial was presented first), or Wilcoxon signed-ranks test: z ⫽ ⫺ 1.55, p ⫽ .06 [one-tailed]).
animacy (animate vs. inanimate) of the object pairs. Subsequent
analyses were collapsed over these variables. Half of the infants
were randomly selected and offline observed by a second observer Discussion
who was completely blind to the order of object outcome. Inter-
In this experiment, the only dissimilarity between the different
scorer reliability averaged 97%.
object pairs was the property of shape. Therefore, the different-
Familiarization trials. Averaging across all eight of the famil-
shaped object pairs shared many more features than the different
iarizations trials, it was found that infants looked equally long
object pairs of the first experiment. In Experiment 1, the different
whether two different-shaped objects were revealed (Mdifferent ⫽
object pairs differed maximally along all property dimensions. Yet
12.52 s, SD ⫽ 3.42) or two identical objects were revealed
the results of the current experiment mirror those of the previous
(Midentical ⫽ 12.93 s, SD ⫽ 3.75), t(15) ⫽ .44, p ⫽ .66.
experiment. Thus, it seems that for 9-month-old infants, objects
Test trials. Infants’ looking times to the test outcomes were
that differ in shape, like objects that differ completely, are ex-
compared by means of a 2 ⫻ 2 repeated measures ANOVA, with
pected to be marked by distinct labels.
It could be the case, however, that infants expect that objects
that differ along any perceptual dimension should be marked by
distinct labels. Perhaps the infants displayed looking time patterns
similar to those in the first experiment when shown object-pair
outcomes that alternated between identical objects and objects that
differed in shape not because shape is closely tied to kind mem-
bership but because any perceived difference between the different
object pair would have produced this effect. This possibility was
explored in Experiment 3, in which the different object pairs
differed only in color. For the kinds of objects we used in this
study, color is a property difference that does not correlate with
kind membership.

Experiment 3
Method
Participants. Participants were 16 full-term infants (8 male, 8
female; mean age ⫽ 9 months, 0 days; range ⫽ 8 months, 15 days,
to 9 months, 14 days). All infants were recruited from the same
Figure 4. Mean looking times as a function of the number of distinct population as in Experiment 1, but none had participated in Ex-
labels heard and the revealed object-pair outcome for Experiment 2. periments 1 or 2. Most infants came from a middle-class, non-
1234 DEWAR AND XU

Hispanic White background, with about 12% of infants being Test trials. Infants’ looking times to the four test outcomes
Asian. English was the primary language spoken at home for all were averaged and compared by means of a 2 ⫻ 2 repeated
infants. An additional 6 infants were tested but excluded due to measures ANOVA, with number of labels (one vs. two) and
fussiness (5) or parental interference (1). object-pair outcome (identical vs. different) as with-subject fac-
Materials and apparatus. All materials and apparatus were the tors. The analysis revealed a significant main effect of object-pair
same as those used in Experiment 2 except for the objects pre- outcome, F(1, 15) ⫽ 7.61, p ⫽ .02, effect size (␩p2) ⫽ .34. Thus,
sented to the infants. Four pairs of objects were used in this collapsing over number of labels heard, infants looked longer
experiment. The objects used in this experiment were almost when different-colored objects were revealed (Mdifferent ⫽ 9.20 s,
identical to the objects used in Experiment 2. Each object used in SD ⫽ 6.24) than when two identical objects were revealed
Experiment 2 was duplicated, but every color of the original object (Midentical ⫽ 6.46 s, SD ⫽ 4.22). More important, there was no
was changed, resulting in an identical pair that differed only in interaction between the number of labels and object-pair outcome,
color. Every object used in this experiment also had a duplicate F(1, 15) ⫽ 1.88, p ⫽ .19.
(see Figure 1). The data from Experiments 1, 2, and 3 were combined and
Design and procedure. These were identical to those of Ex- infants’ looking times to the test outcomes were compared by
periments 1 and 2. A coder blind to the order of object-pair means of a 3 ⫻ 2 ⫻ 2 repeated measures ANOVA, with experi-
outcome reviewed the test trials for each infant and guessed the ment (1, 2, or 3) as the between-subjects factor and number of
object outcome for each trial on the basis of the experimenter’s labels (one vs. two) and object-pair outcome (identical vs. differ-
labeling and referential behavior. The coder was able to guess the ent) as within-subject factors. The analysis revealed a significant
object-pair outcome (identical or different) on only 50% of the test three-way interaction, F(1, 53) ⫽ 3.77, p ⫽ .03, effect size (␩p2) ⫽
trials (32 out of 64 trials), which is not different from chance. .13. Thus, infants’ patterns of looking to each of the test outcomes
differed between Experiments 1, 2, and 3.
Results Comparing infants in Experiments 1 and 2, no Experiment ⫻
Number of Labels ⫻ Object-Pair Outcome interaction was found,
The main results of Experiment 3 are shown in Figure 5. F(1, 38) ⫽ .005, p ⫽ .94. Thus, infants’ pattern of looking to each
Preliminary analyses found no effects of gender, test trial order of the test outcomes did not differ between Experiments 1 and 2.
(whether the expected or unexpected trial was presented first), or However, a significant two-way interaction was found between
animacy (animate vs. inanimate) of the object pairs. Subsequent number of labels and object-pair outcome, F(1, 38) ⫽ 10.81, p ⬍
analyses were collapsed over these variables. Half of the infants .01, effect size (␩p2) ⫽ .22. Planned comparisons were performed.
were randomly selected and offline observed by a second observer When infants heard the box contents described using one label
who was completely blind to the order of object outcome. Inter- repeated twice, they looked significantly longer when two different
scorer reliability was 97%. (either completely different or different in shape) objects were
Familiarization trials. Averaging across all eight of the famil- revealed (the unexpected outcome; M ⫽ 10.15 s, SD ⫽ 7.11) than
iarization trials, it was found that infants looked equally long when two identical objects were revealed (the expected outcome;
whether two different-colored objects were revealed (Mdifferent ⫽ M ⫽ 6.00 s, SD ⫽ 3.54), t(39) ⫽ ⫺3.66, p ⬍ .01. When infants
9.68 s, SD ⫽ 4.11) or two identical objects were revealed heard the box contents described using two distinct labels, they
(Midentical ⫽ 11.24 s, SD ⫽ 7.11), t(15) ⫽ .92, p ⫽ .37. looked longer when two identical objects were revealed (the un-
expected outcome; M ⫽ 10.73 s, SD ⫽ 9.48) than when two
different (either completely different or different in shape) objects
were revealed (M ⫽ 7.31 s, SD ⫽ 7.79), t(39) ⫽ 2.00, p ⫽ .05.
Comparing infants in Experiments 1 and 3 only, a significant
Experiment ⫻ Number of Labels ⫻ Object-Pair Outcome interac-
tion was found, F(1, 38) ⫽ 5.52, p ⫽ .02. Therefore, infants’
patterns of looking to each of the test outcomes differed between
Experiment 1 and Experiment 3. Furthermore, comparison of only
those infants in Experiments 2 and 3 revealed a similar finding; the
Experiment ⫻ Number of Labels ⫻ Object-Pair Outcome interac-
tion was significant, F(1, 30) ⫽ 9.55, p ⬍ .01. Thus, infants’
patterns of looking to each of the test outcomes did differ signif-
icantly between Experiment 2 and Experiment 3. Because the
looking time patterns for infants in Experiments 1 and 2 are
statistically equivalent, it can be said that the looking patterns
obtained in Experiment 3 are different from those obtained in
Experiments 1 and 2.

Discussion
The results of Experiment 3 differed from those of both Exper-
Figure 5. Mean looking times as a function of the number of distinct iments 1 and 2. Infants did not expect two distinct labels to refer
labels heard and the revealed object-pair outcome for Experiment 3. to two different-colored objects. The number of labels used did not
 WORDS AND KINDS 1235

have an effect on which object-pair outcome the infants expected whether the infants in these studies interpreted the distinct words
to see inside the box. Therefore, it appears that not just any as referring to distinct kinds of objects (types) or distinct individ-
perceptual dissimilarity between different object pairs is sufficient ual objects (tokens). The current set of experiments provides
for infants to expect that they should be marked by distinct labels. evidence that even young infants expect distinct words to refer to
Moreover, it seems that infants expect that only objects that differ kinds. In these experiments, the object-pair outcomes alternated
in kind-relevant features (i.e., shape) should be marked by distinct between two identical objects or two objects that differed in some
labels. respect. If it were the case that infants merely expect distinct words
How do we interpret the looking time pattern in Experiment 3? to refer to individual objects, then when infants hear two distinct
In the previous two experiments, the object outcomes alternated labels, they should look equally at two identical objects as they do
between identical objects and different kinds of objects. Thus, at two different objects, because both outcomes represent two
labeling was a cue to object outcome: Hearing distinct labels led to individual objects. However, the results of the current study sug-
the expectation of seeing two kinds of objects (different[-shaped] gest that this is not the case. Furthermore, it is not enough to say
object outcome), whereas hearing one repeated label led to the that infants expect distinct words to refer to different objects: It is
expectation of seeing one kind of object (identical object out- evident from these results that the ways in which the different
come). However, in this experiment, when infants were shown, objects differ is important. The property of shape is a very salient
during familiarization, that the object outcomes alternated between cue to kind membership (Landau et al., 1988; Rosch et al., 1976;
identical and different-colored objects, both outcomes involved Soja et al., 1991), and it was hypothesized that infants may expect
only one kind of object. Thus, if infants expect labels to pick out that objects that differ in shape should have different labels. The
kinds, then the labeling in Experiment 3 was uninformative, be- results of the second experiment (identical objects vs. different-
cause both object outcomes involved only one kind of object. In shaped objects) lend support to the idea that, even for young
this case, the labeling was not predictive of object outcome, and infants, objects that differ in shape are seen as different kinds of
therefore, it may have been ignored. things. Here, infants expected two different-shaped objects to be
labeled with two distinct count nouns. In both of the first two
General Discussion experiments, the two object-pair outcomes alternated between one
kind (identical objects) and two kinds (different[-shaped] objects).
After seeing the contents of a box alternate between either two Our results show that labeling can be used to distinguish between
identical objects or two different objects (Experiment 1), 9-month- these two object outcomes. However, if the different objects varied
old infants who subsequently heard the hidden contents of the box along a property dimension not linked to kind membership (in this
referred to using one repeated label looked reliably longer when case, color), it was hypothesized that infants might show a differ-
the box was opened to reveal two different objects inside. Con- ent looking pattern. This prediction was confirmed in the third
versely, on hearing the hidden contents referred to using two experiment (identical objects vs. different-colored objects). Here,
distinct labels, infants did not show a reliable preference for one infants did not use labeling information to help them predict
outcome over the other and, if anything, showed a nonsignificant object-pair outcome. We suggest that the reason that labeling does
preference for the test event with two identical objects. A stronger not assist infants when the objects are either identical or different-
pattern of results was obtained when infants were shown that the colored is because both object-pair outcomes represent only one
box contents alternated between either two identical objects or two kind of object. Because distinct labels are used to mark distinct
objects that differed only in shape (Experiment 2). Here again, kinds, if the box contents do not alternate between same and
infants who heard the contents of the box referred to using two different kinds, the labeling information cannot be used to predict
identical labels looked longer when the box was opened to reveal object-pair outcome. It is clear from the results of this series of
two different-shaped objects. In contrast, when two distinct labels experiments that infants are sensitive to perceptual dimensions that
were used to label the hidden contents, infants in this experiment are related (or unrelated) to kind membership.
looked significantly longer when two identical objects were re- It should be made clear that there is no evidence that the
vealed. For infants watching these events, labeling acts as a cue to 9-month-olds in the current study actually learned the labels for the
object-pair outcome. Under these circumstances, when infants hear objects, nor were they expected to. The events presented to the
one repeated label used to describe the (unseen) box contents, they infants involved both novel objects and novel labels. In addition,
expect that the box will open to reveal two identical objects, and the labels were only presented when the objects were not visible
when they hear the contents described using two distinct labels, (i.e., when the box was closed). These certainly are not optimal
they expect to see two different-shaped objects inside the box conditions for word learning, especially for infants of this age.
(Experiment 2). However, these expectations are not upheld when However, even in the absence of specific word– object mappings,
infants are first shown that the box contents alternate between infants as young as 9 months seem to expect distinct words to refer
either two identical objects or two different-colored objects (Ex- to different sortal concepts, which in turn map onto different kinds
periment 3). In this case, infants simply look longer when the box of objects in the world.
is opened to reveal two different-colored objects, independent of Other research in the literature provides convergent evidence
the number of labels heard. Here, labeling does not appear to add that infants expect words to pick out sortal kinds (Xu, 2005, in
any predictive cues as to what infants can expect to see when the press). For example, Balaban and Waxman (1996) and Waxman
box is opened. and Braun (2005) found that consistent labels, but not tones or
An empirical question that remained unanswered from the re- variable labels, facilitate categorization in 9-month-old infants,
sults of previous studies examining the effect of labeling on suggesting that when provided a label, infants group exemplars
infants’ object representations (Xu, 2002; Xu et al., 2005) was into a single category more readily than they do in the absence of
1236 DEWAR AND XU

a label. In addition, if two kinds of objects are inferred in an event children have shown that children pay close attention to the inten-
because two distinct labels are heard, there must be two distinct tional act of referring and that factors such as joint attention and
tokens of objects present, as in the object-individuation task (e.g., speaker’s gaze play a critical role in how children decide which
Xu, 2002). Moreover, objects that share a label and, thus, are the object is the referent of a new word (e.g., Baldwin, 1991, 1993;
same kind of thing should share certain nonobvious properties, as Bloom, 2000; Tomasello, Strosberg, & Akhtar, 1996). It has been
in an inductive inference task. For example, at 13 and 18 months, previously shown that infants as young as 12 months of age are
when objects share a common label, infants are more likely to able to use labeling to generate object representations even when
produce a nonobvious property (e.g., squeeze it to make a sound; the objects are never shown to them (Xu et al., 2005). The results
Graham, Kilbreath, & Welder, 2004; Joshi & Xu, 2007; Welder & of the current study provide indirect evidence that even younger
Graham, 2001). The current study is consistent with these lines of infants are able to use intentional and referential cues (i.e., speak-
research as it indicates that when 9-month-old infants hear a er’s gaze) to guide their expectations about objects on the basis of
repeated label, they expect the referents of that label to be identical labeling information. The 9-month-olds in the current study did not
(i.e., the same kind of thing), but when they hear two distinct see the objects while they were being labeled. The experimenter
labels, they expect the referents of those labels to be different in labeled the object pairs while the pairs were inside a closed box.
kind. Thus, the infants had to infer that the labels being provided
The labels used in these experiments were all presented in a referred to the unseen contents of the box. Future studies may
count noun frame. It remains an open question whether the expec- manipulate more directly the direction of the experimenter’s gaze
tation that words refer to sortal kinds is specific to count nouns. It to see if infants are sensitive to that information in this context.
is not clear whether 9-month-old infants have made a syntactic Second, infants begin to comprehend words for object catego-
distinction between count nouns and adjectives, so it would not be ries at around 9 months of age. One controversial issue is whether
easy to address the question at this young age. By 13 months, these early words are count nouns that refer to kinds or proper
however, some evidence suggests that infants have learned that names that refer to individual objects. According to the literature,
count nouns correlate with objects that share a common shape, older children (2- to 4-year-olds) seem to assume that a novel word
whereas adjectives tend to correlate with properties such as color applied to an unfamiliar object refers to a basic-level kind
or texture (Waxman, 1999). Word-learning studies with older (Golinkoff, Hirsh-Pasek, Bailey, & Wenger, 1992; Hall, 1991;
infants have demonstrated that infants’ sensitivity to object cate- Imai, Gentner & Uchida, 1994; Markman & Hutchinson, 1984;
gory (kind) and object property (color) is influenced by the gram- Soja et al., 1991; Waxman & Gelman, 1986; Waxman & Hall,
matical form class of the novel label. Booth and Waxman (2003) 1993; Waxman & Kosowski, 1990) and will be extended to objects
found that 14-month-olds’ construal of objects is influenced by of similar shape (and kind; Baldwin, 1989; Jones, Smith, &
naming. They found that infants extended novel nouns (“This one Landau, 1991; Landau et al., 1988; Landau, Smith, & Jones, 1998;
is a blicket”) specifically to object categories (e.g., animal) and not Smith, Jones, & Landau, 1992, 1996). Thus, for expert word
to object properties (e.g., purple things). Conversely, infants ex- learners, novel words are thought of as count nouns that can be
tended novel adjectives (“This one is blickish”) to object proper- readily generalized to like objects. But what remains an open
ties when the property was color. The results of the current study question is whether this representation of novel words operates
demonstrate that when novel count nouns are presented, 9-month- similarly in children just beginning to learn words. In other words,
olds expect distinct count nouns to correspond to a category- whether infants show evidence that their earliest labels for objects
relevant commonality among the objects (shape) but not to a are count nouns that map onto kinds.
property-based commonality (color). Although the current study did not involve actual word learning,
Do young children understand that shape is a salient cue to kind the findings do inform the literature on early word extension.
membership, or have they simply formed an association between Research on word generalization has found that infants, as young
labeling and shape such that objects that are referred to by the as 13 to 15 months of age, will extend a newly learned label
same label share the same shape and objects that are marked by beyond the target object to other members of like kind (Wood-
distinct labels have different shapes? It has been suggested that ward, Markman, & Fitzsimmons, 1994). What has not yet been
young children demonstrate a “shape bias” in word learning such demonstrated is whether younger infants, who are just on the cusp
that objects that share the same shape should also share a label of word learning, also understand that words for objects are count
(i.e., Landau et al., 1988; Smith, Jones, Landau, Gershkoff-Stowe, nouns that refer to kinds. It could be that these very young infants
and Samuelson, 2002). However, other studies examining pre- expect that object labels simply designate individuals and, thus,
schoolers’ naming of pictorial representations strongly suggest that function as proper names (Hirsh-Pasek, Golinkoff, Hennon, &
children do not always apply labels on the basis of shape alone Maguire, 2004; Smith, 2000). However, the current findings sug-
(Bloom & Markson, 1998; Gelman & Ebeling, 1998). If shape gest that a “proper-name bias” in young word-learners may be
alone does not determine labeling, perhaps shape may simply be a untenable. If, for 9-month-olds at least, words serve as proper
proxy for kind membership. With 9-month-old infants, however, names and map to individuals, it should have been the case that
we do not yet have evidence of whether shape functions as a proxy there was no differentiation between looking times in each of the
for kind membership or whether words simply pick out groups of conditions (because in each trial, the box opened to reveal two
objects that have distinct shapes. individual objects). This scenario was not supported by the evi-
Although the current study did not require infants to learn the dence. When two labels were used, 9-month-olds expected to see
specific mappings between words and objects, it nevertheless objects that differed along a kind-relevant property dimension
offers insight into the nature of infants’ representation of early (shape), but they did not expect to see objects that differed along
words in two ways. First, many word-learning studies with young a property dimension that was independent of kind membership
 WORDS AND KINDS 1237

(color). Thus, even for young infants, the presence of different Macnamara, J. (1986). A border dispute: The place of logic in psychology.
labels leads to the assumption that those labels correspond to Cambridge, MA: MIT Press.
objects that are distinct in kind. Such a finding is clearly suggestive Macnamara, J., & Reyes, G. E. (1994). The logical foundations of cogni-
that, at the beginning of word learning, infants understand words tion. Oxford, England: Oxford University Press.
for objects as count nouns that designate kinds. Markman, E. M., & Hutchinson, J. E. (1984). Children’s sensitivity to
constraints on word meaning: Taxonomic versus thematic relations.
Cognitive Psychology, 16, 1–27.
References Nelson, K. (1973). Structure and strategy in learning to talk. Monographs
of the Society for Research in Child Development, 38(12, Serial No.
Balaban, M., & Waxman, S. (1996). Words may facilitate categorization in 149).
9-month-old infants. Journal of Experimental Child Psychology, 64, Rosch, E., Mervis, C. B., Gray, W., Johnson, D., & Boyes-Braem, P.
3–26. (1976). Basic objects in natural categories. Cognitive Psychology, 8,
Baldwin, D. A. (1989). Priorities in children’s expectations about object 382– 439.
label reference: Form over color. Child Development, 60, 1291–1306. Samuelson, L. K., & Smith, L. B. (2005). They call it like they see it:
Baldwin, D. A. (1991). Infants’ contribution to the achievement of joint Spontaneous naming and attention to shape. Developmental Science, 8,
reference. Child Development, 63, 875– 890. 182–198.
Baldwin, D. A. (1993). Infants’ ability to consult the speaker for clues to Smith, L. B. (2000). Learning how to learn words: An associative crane. In
word reference. Journal of Child Language, 20, 395– 418. R. M. Golinkoff, K. Hirsh-Pasek, L. Bloom, L. Smith, A. Woodward, N.
Bloom, P. (2000). How children learn the meanings of words. Cambridge, Aktar, et al. (Eds.), Becoming a word learner: A debate on lexical
MA: MIT Press. acquisition (pp. 51– 80). New York: Oxford University Press.
Bloom, P., & Markson, L. (1998). Intention and analogy in children’s Smith, L. B., Jones, S. S., & Landau, B. (1992). Count nouns, adjectives,
naming of pictorial representations. Psychological Science, 9, 200 –204. and perceptual properties in children’s novel word interpretations. De-
Bonatti, L., Frot, E., Zangl, R., & Mehler, J. (2002). The human first velopmental Psychology, 28, 273–286.
hypothesis: Identification of conspecifics and individuation of objects in Smith, L. B., Jones, S. S., & Landau, B. (1996). Naming in young children:
the young infant. Cognitive Psychology, 44, 388 – 426. A dumb attentional mechanism? Cognition, 60, 143–171.
Booth, A. E., & Waxman, S. R. (2003). Mapping words to the world in Smith, L. B., Jones, S. S., Landau, B., Gershkoff-Stowe, L., & Samuelson,
infancy: Infants’ expectations for count nouns and adjectives. Journal of L. (2002). Object name learning provides on-the-job training for atten-
Cognition and Development, 4, 357–381. tion. Psychological Science, 13, 13–19.
Gelman, S. A., & Ebeling, K. S. (1998). Shape and representational status Soja, N. N., Carey, S., & Spelke, E. S. (1991). Ontological categories guide
in children’s early naming. Cognition, 66, B35–B47. young children’s inductions of word meaning: Object terms and sub-
Golinkoff, R. M., Hirsh-Pasek, K., Bailey, L., & Wenger, N. (1992). stance terms. Cognition, 38, 179 –211.
Young children and adults use lexical principles to learn new words. Spelke, E. S., Kestenbaum, R., Simons, D. J., & Wein, D. (1995). Spatio-
Developmental Psychology, 28, 99 –108. temporal continuity, smoothness of motion and object identity in in-
Graham, S. A., Kilbreath, C. S., & Welder, A. N. (2004). Thirteen-month- fancy. British Journal of Developmental Psychology, 13, 113–142.
olds rely on shared labels and shape similarity for inductive inferences. Tomasello, M., Strosberg, R., & Akhtar, N. (1996). Eighteen-month-old
Child Development, 75, 409 – 427. children learn words in non-ostensive contexts. Journal of Child Lan-
Gupta, A. (1980). The logic of common nouns. New Haven, CT: Yale
guage, 23, 157–176.
University Press.
Van de Walle, G. A., Carey, S., &. Prevor, M. (2000). Bases for object
Hall, D. G. (1991). Acquiring proper names for familiar and unfamiliar
individuation in infancy: Evidence from manual search. Journal of
animate objects: Two-year-olds’ word learning biases. Child Develop-
Cognition and Development, 1, 249 –280.
ment, 62, 1442–1454.
Waxman, S. R. (1999). Specifying the scope of 13-month-olds’ expecta-
Hall, D. G. (1993). Basic-level individuals. Cognition, 40, 199 –221.
tions for novel words. Cognition, 70, B35–B50.
Hirsch, E. (1982). The concept of identity. New York: Oxford University
Waxman, S. R., & Braun, I. E. (2005). Consistent (but not variable) names
Press.
as invitations to form categories: New evidence from 12-month-old
Hirsh-Pasek, K., Golinkoff, R. M., Hennon, E. A., & Maguire, M. J.
infants. Cognition, 95, B59 –B68.
(2004). Hybrid theories at the frontier of developmental psychology: The
emergentist coalition model of word learning as a case in point. In D. G. Waxman, S. R., & Gelman, R. (1986). Preschooler’s use of superordinate
Hall & S. R. Waxman (Eds.), Weaving a lexicon (pp. 173–204). Cam- relations in classification and language. Cognitive Development, 1, 136 –
bridge, MA: MIT Press. 156.
Imai, M., Gentner, D., & Uchida, N. (1994). Children’s theories of word Waxman, S. R., & Hall, D. G. (1993). The development of a linkage
meaning: The role of shape similarity in early acquisition. Cognitive between count nouns and object categories: Evidence from fifteen- to
Development, 9, 45–75. twenty-one-month-old infants. Child Development, 64, 1224 –1241.
Jones, S. S., Smith, L. B., & Landau, B. (1991). Object properties and Waxman, S. R., & Kosowski, T. D. (1990). Nouns make category relations:
knowledge in early lexical learning. Child Development, 62, 499 –516. Toddlers’ and pre-schoolers’ word-learning biases. Child Development,
Joshi, A., & Xu, F. (2007). Inductive inference, artifact kind concepts, and 61, 1461–1473.
language. Manuscript submitted for publication. Waxman, S. R., & Markow, D. R. (1995). Words as invitations to form
Krojgaard, P. (2000). Object individuation in 10-month-old infants: Do categories: Evidence from 12- to 13-month-old infants. Cognitive Psy-
significant objects make a difference? Cognitive Development, 15, 169 – chology, 29, 257–302.
184. Welder, A. N., & Graham, S. A. (2001). The influence of shape similarity
Landau, B., Smith, L. B., & Jones, S. S. (1988). The importance of shape and shared labels on infants’ inductive inferences about non-obvious
in early lexical learning. Cognitive Development, 3, 299 –321. object properties. Child Development, 72, 1653–1673.
Landau, B., Smith, L. B., & Jones, S. S. (1998). Object perception and Wiggins, D. (1980). Sameness and substance. Oxford, England: Basil
object naming in early development. Trends in Cognitive Sciences, 2, Blackwell.
19 –24. Wilcox, T., & Baillargeon, R. (1998a). Object individuation in infancy:
1238 DEWAR AND XU

The use of featural information in reasoning about occlusion events. Xu, F. (in press) Count nouns, sortal concepts, and the nature of early
Cognitive Psychology, 37, 97–155. words. In J. Pelletier (Ed.), New directions in cognitive science (Vol.
Wilcox, T., & Baillargeon, R. (1998b). Object individuation in young 13). Oxford, England: Oxford University Press.
infants: Further evidence with an event-monitoring paradigm. Develop- Xu, F., & Baker, A. (2005). Object individuation in 10-month-old infants
mental Science, 1, 127–142. using a simplified manual search method. Journal of Cognition and
Wilcox, T., & Schweinle, A. (2002). Object individuation and event Development, 6, 307–323.
mapping: Developmental changes in infants’ use of featural information. Xu, F., & Carey, S. (1996). Infants’ metaphysics: The case of numerical
Developmental Science, 5, 132–150. identity. Cognitive Psychology, 30, 111–153.
Woodward, A., Markman, E., & Fitzsimmons, C. (1994). Rapid word Xu, F., Carey, S., & Quint, N. (2004). The emergence of kind-based object
learning in 13 -and 18-month-olds. Developmental Psychology, 30, individuation in infancy. Cognitive Psychology, 49, 155–190.
553–566. Xu, F., Carey, S., & Welch, J. (1999). Infants’ ability to use object kind
Xu, F. (1997). From Lot’s wife to a pillar of salt: Evidence that physical information for object individuation. Cognition, 70, 137–166.
object is a sortal concept. Mind & Language, 12, 365–392. Xu, F., Cote, M., & Baker, A. (2005). Labeling guides object individuation
Xu, F. (2002). The role of language in acquiring kind concepts in infancy. in 12-month-old infants. Psychological Science, 16, 372–377.
Cognition, 85, 223–250.
Xu, F. (2005). Categories, kinds, and object individuation in infancy. In L.
Gershkoff-Stowe & D. Rakison (Eds.), Building object categories in Received July 21, 2006
developmental time: Papers from the 32nd Carnegie Symposium on Revision received April 26, 2007
Cognition (pp. 63– 89). Mahwah, NJ: Erlbaum. Accepted May 16, 2007 䡲

Call for Nominations

The Publications and Communications (P&C) Board of the American Psychological Association
has opened nominations for the editorships of Psychological Assessment, Journal of Family
Psychology, Journal of Experimental Psychology: Animal Behavior Processes, and Journal of
Personality and Social Psychology: Personality Processes and Individual Differences (PPID),
for the years 2010-2015. Milton E. Strauss, PhD, Anne E. Kazak, PhD, Nicholas Mackintosh, PhD,
and Charles S. Carver, PhD, respectively, are the incumbent editors.
Candidates should be members of APA and should be available to start receiving manuscripts in
early 2009 to prepare for issues published in 2010. Please note that the P&C Board encourages
participation by members of underrepresented groups in the publication process and would partic-
ularly welcome such nominees. Self-nominations are also encouraged.

Search chairs have been appointed as follows:

• Psychological Assessment, William C. Howell, PhD, and J Gilbert Benedict, PhD
• Journal of Family Psychology, Lillian Comas-Diaz, PhD, and Robert G. Frank, PhD
• Journal of Experimental Psychology: Animal Behavior Processes, Peter A. Ornstein,
PhD, and Linda Porrino, PhD
• Journal of Personality and Social Psychology: PPID, David C. Funder, PhD, and Leah
L. Light, PhD

Candidates should be nominated by accessing APA’s EditorQuest site on the Web. Using your
Web browser, go to http://editorquest.apa.org. On the Home menu on the left, find “Guests.” Next,
click on the link “Submit a Nomination,” enter your nominee’s information, and click “Submit.”
Prepared statements of one page or less in support of a nominee can also be submitted by e-mail
to Emnet Tesfaye, P&C Board Search Liaison, at [email protected].
Deadline for accepting nominations is January 10, 2008, when reviews will begin.