Journal of Experimental Psychology: General
2015, Vol. 144, No. 3, 528 –533
© 2015 American Psychological Association
0096-3445/15/$12.00 http://dx.doi.org/10.1037/xge0000072
BRIEF REPORT
A Self-Serving Bias in Children’s Memories?
Arber Tasimi and Marcia K. Johnson
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Yale University
Although children’s initial perceptions and judgments about sociomoral situations are being actively
explored, little is known about what children remember about them. In four experiments testing over 400
children, we investigated children’s memories for small acts of giving and taking. When asked to recall
their own giving and taking, children were relatively accurate following a number of delays. In contrast,
when asked to recall a child’s giving or taking, children exaggerated the child’s taking after a 1-day or
1-week delay. Notably, this pattern of misremembering occurred only when children recalled the actions
of a child but not an adult. We consider the idea that children spontaneously engage in social comparison,
which colors their memories of the social world.
Keywords: memory, social cognition, cognitive development, morality
Although initial perceptions and judgments are important, our
views of the social world are also determined by what we remember. Do children’s memories also reflect a parallel sensitivity to
social comparison? When asked to report on past conflicts, children seem to shade the truth in self-serving ways (McGuire,
Manke, Eftekhari, & Dunn, 2000; Ross, Smith, Spielmacher, &
Recchia, 2004). For example, children asked to describe past
family disputes claim more innocence for themselves and more
harm done by their siblings (Wilson, Smith, Ross, & Ross, 2004).
These findings demonstrate that children “remember” the past in
ways that make them appear more favorably to others. However,
studies investigating children’s reports of past conflicts have not
clarified whether this pattern of remembering reflects misremembering or conscious shading of the truth in the interests of selfpresentation when one’s actions may be under scrutiny.
When asked to recall others, children, like adults, seem to
show a memory advantage for negative information (Baumeister, Bratslavsky, Finkenaurer, & Vohs, 2001). After learning
about individuals who engaged in nice (e.g., sharing) and mean
(e.g., stealing) behaviors, children show better memory for
mean people (Kinzler & Shutts, 2008). Children not only show
enhanced face recognition of wrongdoers, but they also better
remember the details of their mean behaviors compared to the
details of their nice behaviors (Baltazar, Shutts, & Kinzler, 2012).
One explanation for these findings is that children find it surprising
when people do not behave well; if children think that other people
are good, then individuals who behave in unusual (negative) ways
may be noticed and remembered. Another possibility relates to the
proposal that cognitive systems have evolved to detect and remember wrongdoers (Cosmides, Tooby, Fiddick, & Bryant, 2005).
According to this idea, children’s memory advantage for negative
information serves useful purposes because remembering individuals who have wronged in the past could reduce the possibility of
being wronged in the future. However, children’s superior memory
for bad over good may serve other purposes beyond knowing who
to avoid. In particular, this pattern of remembering may reflect a
When asked various kinds of questions, children appear to
interpret the world through a positive lens (Boseovski, 2010). As
early as kindergarten, children evaluate themselves very positively
(Stipek & Mac Iver, 1989) and continue to do so until the late
elementary school years (Benenson & Dweck, 1986). Children’s
evaluations of others are also positive. For example, children
consider an individual to be good even after hearing of a single
positive behavior followed by numerous negative behaviors performed by this individual (Rholes & Ruble, 1986). By age 5,
children think a character’s negative traits will become positive
later in life (Lockhart, Chang, & Story, 2002), a belief that has
been found cross-culturally (Lockhart, Nakashima, Inagaki, &
Keil, 2009). As late as age 8, children consider another’s goodness
to be more stable than their badness (Heyman & Dweck, 1998).
When placed in situations that evoke social comparisons, however, children become less positive toward their peers. For example, between the ages of 7 and 13, children are more satisfied with
their performance on a task after learning that another child failed
compared to when they succeeded (Steinbeis & Singer, 2013).
Indeed, 5- and 6-year-olds willingly incur a personal cost to ensure
that another child receives less than themselves, thus putting them
at a relative advantage (Sheskin, Bloom, & Wynn, 2014). Such
findings show that children want to fare better relative to their
peers and are consistent with the well-known feature of adult
social cognition of comparing oneself to others (Festinger, 1954;
Fiske, 2011).
Arber Tasimi and Marcia K. Johnson, Department of Psychology, Yale
University.
We thank the children, families, and staffs of the following elementary
schools: Frisbie, Kelley, Middlebury, Plantsville, Pomperaug, Strong, and
Thalberg. We also thank Andrei Cimpian, Chaz Firestone, Susan Gelman,
Joshua Knobe, Mara Mather, and Liane Young for feedback on this work.
Correspondence concerning this article should be addressed to Arber
Tasimi, Department of Psychology, Yale University, P.O. Box 208205,
New Haven, CT 06520. E-mail: arber.tasimi@yale.edu
528
SELF-SERVING MEMORIES
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
self-serving bias that extends beyond previous conflicts that children themselves were engaged in.
The present investigation tested this idea. Four experiments
examined children’s memory for small acts of giving and taking
that they, another child, or an adult engaged in. We focused on
children between the ages of 5 and 8 given previous findings
showing that children of this age are (a) remarkably positive when
reasoning about themselves and others and (b) likely to engage in
social comparison. Our primary question was whether children’s
memories reflect a self-serving bias even for unobtrusive events
that do not place children in competitive or explicitly comparative
contexts.
Experiment 1
Participants
The Human Subjects Committee at Yale University approved all
study procedures used in Experiments 1– 4. Children were recruited from the greater New Haven, Connecticut area. Parents of
participating children gave written informed consent; children also
provided oral assent. Sample size for all experiments was determined by how many parents provided consent and how many
children participating on the first day returned on the second day
(Experiments 1, 3, and 4). Children were tested individually in a
quiet room at their elementary school, and all sessions were
audio-recorded. Experiment 1 included 139 children (68 girls;
mean age ⫽ 6.80 years; range ⫽ 5.14 – 8.38 years). An additional
13 children were tested but excluded because they were absent on
the second day of testing.
Procedure
Children were told that they would play a short game. Children
were then randomly assigned to one of the following four transaction conditions, each involving photos of smiling White male
children (from LoBue & Thrasher, 2014): (a) to give stickers to a
child, in which the participant was handed five stickers and provided with the opportunity to place any number in front of a child’s
photo (e.g., “For coming in today, you get 5 stickers. This is Jack.
Jack has no stickers. Would you like to give Jack any of your
stickers?”); (b) to take away stickers from a child, in which five
stickers were placed in front of a child’s photo and the participant
was provided with the opportunity to take any number (e.g., “This
is Jack. Jack has 5 stickers. Would you like to take away any of
Jack’s stickers?”); (c) to observe a child give another child stickers, in which the experimenter manipulated a transfer from one
child’s photo to another child’s photo (e.g., “This is John. John has
5 stickers. This is Jack. Jack has no stickers. John gave Jack 3
stickers.”); or (d) to observe a child take away stickers from
another child, in which the experimenter manipulated a transfer
from one child’s photo to another child’s photo (e.g., “This is John.
John has no stickers. This is Jack. Jack has 5 stickers. John took
away 3 stickers from Jack.”). The design was wholly betweensubjects, and we matched the number of stickers given or taken in
the “self” and “other” conditions. For example, if a child gave
three stickers, then the next child observed a character give three
stickers. Similarly, if a child took no stickers, then the next child
observed a character take no stickers. Because we framed our
529
study as a game, all materials remained in the testing room, so no
child left with stickers.
The following day, the experimenter returned to the school and
asked children how many stickers were given or taken (e.g., “Do
you remember Jack? How many stickers did you take away from
Jack?”).
Results
Our primary measure was the difference between the number of
stickers given or taken compared to what children reported as
being given or taken (e.g., number recalled as given minus number
actually given; see Table 1). An analysis of variance (ANOVA) on
these difference scores, including Condition (self, other) ⫻ Behavior (giving, taking), revealed an interaction between condition
and behavior, F(1, 135) ⫽ 5.66, p ⫽ .019, p2 ⫽ .04. Children’s
reports did not differ between another child’s giving (M ⫽ 0) and
their own giving (M ⫽ ⫺0.03), t(67) ⫽ 0.09, p ⫽ .93, d ⫽ .02. In
contrast, children overestimated another child’s taking (M ⫽ 1.00)
in comparison to their own (M ⫽ ⫺0.09), t(68) ⫽ 3.56, p ⬍ .001,
d ⫽ .85. There was no correlation between age and difference
scores in the Other-Take condition, Pearson r ⫽ ⫺.26, p ⫽ .13.
Thus, children remembered the actions of another child relative to
their own in a way that exaggerated how much another child took.
Experiment 2
Experiment 2 addressed the possibility that children in Experiment 1 misremembered another child’s behavior simply because
they did not attend to their actions. The procedure was identical to
Experiment 1 with one exception: here, the delay was 1 min,
during which children were asked to list as many fruits and
vegetables (30 s) and animals (30 s) as they knew.
Participants
Experiment 2 included 54 children (21 girls; mean age ⫽ 7.04
years; range ⫽ 5.50 – 8.80 years).
Table 1
Children’s Reports of What Was Given or Taken by Themselves
or a Child in Experiments 1–3, M (SD)
1-min delay
(Experiment 2)
Self-give
Reported
Actual
Difference
Self-take
Reported
Actual
Difference
Child-give
Reported
Actual
Difference
Child-take
Reported
Actual
Difference
1-day delay
(Experiment 1)
1-week delay
(Experiment 3)
1.93 (1.00)
1.93 (1.00)
0.00 (0.00)
2.11 (1.05)
2.14 (1.06)
⫺0.03 (0.62)
2.92 (1.21)
2.75 (1.15)
0.17 (0.48)
1.54 (1.66)
1.54 (1.66)
0.00 (0.00)
1.71 (1.67)
1.80 (1.68)
⫺0.09 (0.28)
1.27 (1.46)
1.35 (1.65)
⫺0.08 (0.48)
1.86 (1.03)
1.93 (1.00)
⫺0.07 (0.27)
2.18 (1.57)
2.18 (1.19)
0.00 (1.81)
2.53 (1.20)
2.71 (1.08)
⫺0.18 (0.82)
1.54 (1.66)
1.54 (1.66)
0.00 (0.00)
2.77 (1.96)
1.77 (1.66)
1.00 (1.78)
3.04 (1.87)
1.17 (1.53)
1.87 (2.07)
TASIMI AND JOHNSON
530
Results
Remembered-Actual difference scores are shown in Table 1. Of
the 54 children, 53 reported the correct number of stickers that was
given or taken (one child reduced another child’s giving by one).
These findings demonstrate that children in Experiment 1 likely
paid sufficient attention to the stimuli to encode the relevant
information about the giving and taking by themselves and another
child.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Experiment 3
The difference in outcomes between Experiment 1 and Experiment 2 suggests that the tendency to inflate how much another
child took increases with time. To confirm that children’s memories exaggerate another child’s taking with time and assess whether
the effect grows over time, Experiment 3 replicated Experiments 1
and 2 with a longer delay of 1 week.
Participants
Experiment 3 included 101 children (51 girls; mean age ⫽ 6.97
years; range ⫽ 5.40 – 8.55 years). An additional 15 children were
tested but excluded because they were absent on the second day of
testing.
Results
Remembered-Actual difference scores are shown in Table 1. An
ANOVA including Condition (self, other) ⫻ Behavior (giving,
taking) revealed an interaction between condition and behavior,
F(1, 97) ⫽ 25.81, p ⬍ .001, p2 ⫽ .21. Children slightly minimized
another child’s giving (M ⫽ ⫺0.18) relative to their own (M ⫽
0.17), t(50) ⫽ 1.81, p ⫽ .076, d ⫽ .50, and, as in Experiment 1,
overestimated another child’s taking (M ⫽ 1.87) in comparison to
their own (M ⫽ ⫺0.08), t(47) ⫽ 4.65, p ⬍ .001, d ⫽ 1.33. There
was no correlation between age and difference scores in the
Other-Take condition, Pearson r ⫽ ⫺.01, p ⫽ .95. We also
conducted a linear trend on children’s reports of another’s taking
over the three delays (1 min, 1 day, 1 week), which confirmed the
impression that the tendency to inflate how much another child
took increased with time, F(1, 68) ⫽ 9.74, p ⫽ .003.
Experiment 4
Does children’s pattern of misremembering in the previous
experiments reflect an implicit comparative self-serving bias? Because people seek social comparisons with individuals who are
similar to themselves (Goethals & Darley, 1977; Suls, Martin, &
Wheeler, 2002; Wood, 1989)—and age is an important dimension
by which children engage in social comparison (Dijkstra, Kuyper,
van der Werf, Buunk, & van der Zee, 2008; Suls, Gastorf, &
Lawhon, 1978)—Experiment 4 tested whether children differentially remember the giving and taking behaviors of a child versus
an adult.
Participants
Experiment 4 included 104 children (50 girls; mean age ⫽ 6.33
years; range ⫽ 5.01– 8.32 years). An additional three children were
tested but excluded because they were absent on the second day of
testing.
Procedure
The procedure was identical to Experiment 3, with one exception: here, children were randomly assigned to observe the giving
or taking of a child or an adult (photos of smiling White male
young adults were taken from Tottenham et al., 2009), and their
memory was tested 1 week later. We matched “child” and “adult”
conditions by using the distributions of another child’s giving or
taking in Experiment 3.
Results
Remembered-Actual difference scores are shown in Table 2. An
ANOVA including Character (adult, child) ⫻ Behavior (giving,
taking) revealed an interaction between character and behavior,
F(1, 100) ⫽ 7.46, p ⫽ .007, p2 ⫽ .07. Children’s reports did not
differ between an adult’s giving (M ⫽ ⫺0.23) and a child’s giving
(M ⫽ ⫺0.20), t(50) ⫽ 0.10, p ⫽ .92, d ⫽ .03. In contrast, children
overestimated a child’s taking (M ⫽ 1.41) in comparison to an
adult’s taking (M ⫽ 0.16), t(50) ⫽ 3.21, p ⫽ .002, d ⫽ .89. There
was no correlation between age and difference scores in the
Other-Take condition, Pearson r ⫽ ⫺.24, p ⫽ .23. Thus, children’s
pattern of misremembering in the present investigation occurred
when children recalled the actions of another child but not an adult.
Discussion
These experiments provide strong evidence that children remember other children as having taken more than they actually
took. In contrast, children were quite accurate in remembering
their own giving and taking, as well as the giving and taking of an
adult. Taking is a canonical moral violation in childhood (e.g.,
Smetana, Killen, & Turiel, 1991), and the present findings suggest
that even this very mild negative action is subject to systematic
memory distortion.
Table 2
Children’s Reports of What Was Given or Taken by an Adult or
a Child in Experiment 4, M (SD)
M (SD)
Adult-give
Reported
Actual
Difference
Adult-take
Reported
Actual
Difference
Child-give
Reported
Actual
Difference
Child-take
Reported
Actual
Difference
2.44 (1.12)
2.67 (1.07)
⫺0.23 (0.80)
1.36 (1.44)
1.20 (1.47)
0.16 (0.55)
2.48 (1.08)
2.68 (1.11)
⫺0.20 (0.82)
2.56 (1.89)
1.15 (1.43)
1.41 (1.86)
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
SELF-SERVING MEMORIES
What might account for our findings? One possibility is that
people remember exceptional information (Hunt, 2006); if children
believe that others are good, then negative information may be
noticed and remembered. It is possible that findings showing a
negativity bias in children’s memories (Baltazar et al., 2012;
Kinzler & Shutts, 2008) reflect a memory advantage for unusual
information about others. If so, children in our experiments should
have been more, not less, accurate in remembering the surprising—taking— behavior.
Another possibility is that children’s memory reflects their
initial interpretation of a situation and/or the schemas invoked
during remembering (Bransford & Johnson, 1973). Introducing
negative information about an individual can distort memory for
specific behaviors, resulting in an exaggeration of their wrongdoing after a delay (Pizarro, Laney, Morris, & Loftus, 2006). Perhaps
children in the present study created their own explanations for
why the character gave or took and “remembered” consistent with
the explanation they generated (an interpretive narrative account).
For example, children observing a character that took stickers
might have judged this character as “bad” and thus, in remembering, exaggerated the number of stickers taken. Indeed, children
make behavior-to-trait inferences about the niceness and meanness
of others from an early age (Liu, Gelman, & Wellman, 2007).
However, several children reported a child taking one or more
stickers when no stickers were taken, suggesting the influence of
a factor that needs little evidence to get started. One possibility is
that children interpreted the situation as “sharing,” assuming it was
“fair” to take stickers and thus misremembered that stickers were
taken even when they were not. Again, children did not misremember that they or an adult had taken stickers when they had not.
What else, then, could explain our findings? Perhaps observing
another child giving or taking invokes implicit social comparison;
indeed, similarity in age influences whether children compare
themselves with others (Dijkstra et al., 2008; Suls et al., 1978).
Thus, children’s memory in this situation may reflect a self-serving
bias, suggesting that other children take more than they themselves
would. Although a self-serving bias is well documented among
adults (Pronin, 2007; Pronin, Gilovich, & Ross, 2004), little is
known about its developmental origins and its role in remembering. For example, previous work indicates that self-serving judgments are rooted in adults’ overly charitable view of themselves
(Epley & Dunning, 2000); might their self-serving memories be
based on an overly cynical view of others? One possibility is that
self-serving memories occur more often for negative than positive
behaviors in mild situations such as ours. For example, taking little
may imply generosity in a sharing situation; hence, taking may
need to be exaggerated in order to achieve a self-serving effect. If
someone gave little to begin with, it may not seem necessary to
minimize their giving further (i.e., their selfishness is self-evident).
If so, we would expect that another’s giving would be minimized
in situations where their generosity challenges one’s self-image.
Thus, the current findings highlight potential motivational factors that may influence memory. According to the sourcemonitoring framework (Johnson, Hashtroudi, & Lindsay, 1993),
memories are judgments about our subjective experience during
remembering that may reflect source misattributions contaminated
by information from other events or from prior schemas and
motives. For example, adults misremember past events based on
desired outcomes (Barber, Gordon, & Franklin, 2009; Gordon,
531
Franklin, & Beck, 2005), engage in choice-supportive memory
distortion about past decisions (Mather, Shafir, & Johnson, 2000),
and selectively recall positive information to regulate their mood
(Mather & Carstensen, 2005) or maintain a desirable self-view
(Sanitioso, Kunda, & Fong, 1990). However, relatively little work
has been done on such motivational factors in children’s memories.
Although children are sometimes motivated to derogate individuals—for example, out-group members (Buttelmann & Böhm,
2014)— children in the current study received no additional information beyond the giving or taking of another child. Unlike
previous studies that have put children in comparative contexts
(e.g., Steinbeis & Singer, 2013), each child in our study either
engaged in a behavior or observed another child engage in a
behavior. Thus, whatever motives were activated were relatively
implicit. Of course, it is possible that children purposefully exaggerated the taking of another child; children may have misreported
(i.e., lied) rather than misremembered, which is still consistent
with the idea that children spontaneously engage in social comparison. If so, future studies may find that fabricated responses are
later misremembered as what actually happened (e.g., Ackil &
Zaragoza, 1998).
In contrast to the few studies exploring the influence of motives
in children’s memory, many have shown that children’s memories
can be distorted by extra-event information, including stereotypes
and suggestions (Bruck & Ceci, 1999; Leichtman & Ceci, 1995;
Lindsay, 2002). Critically, our questioning was not suggestive;
children were not induced to generate inaccurate reports by exposing them to misleading or confusable additional information.
However, children were tested in schools, raising the question of
whether they exchanged information following their individual
sessions. For example, if a child heard from the experimenter that
John took one sticker and heard during the retention interval from
another participant that John took three stickers, then this child
might misremember “three stickers.” However, if there is no
self-serving bias operating, then a child who heard from the
experimenter that John took three stickers and heard during the
retention interval from another participant that John took one
sticker should sometimes misremember “one sticker,” which
should not produce an overall bias to exaggerate another child’s
taking. On the other hand, if negative behaviors are more salient
than positive ones (Baumeister et al., 2001; Rozin & Royzman,
2001), another child’s taking may be more likely to be passed
along to others than their giving. This gossip account could help
explain why children distorted only the taking, but not the giving,
of another child. Note, however, that children did not also distort
an adult’s taking; social group membership may not only induce
spontaneous comparison of self to another group member but also
prompt conversation that becomes a potential source of memory
distortion. Thus, understanding the ways in which self-serving
biases arise spontaneously from viewing others’ behaviors or may
prompt asymmetric communications presents many questions, especially when considering the ways in which these two factors
may or may not interact.
Taken together, these studies highlight the importance of memory in elucidating early social cognition. Studying what children
remember should contribute to a more complete understanding of
what, and how, they think about the everyday social world and the
mechanisms that contribute to their evolving views.
TASIMI AND JOHNSON
532
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
References
Ackil, J. K., & Zaragoza, M. S. (1998). Memorial consequences of forced
confabulation: Age differences in susceptibility to false memories. Developmental Psychology, 34, 1358 –1372. http://dx.doi.org/10.1037/
0012-1649.34.6.1358
Baltazar, N. C., Shutts, K., & Kinzler, K. D. (2012). Children show
heightened memory for threatening social actions. Journal of Experimental Child Psychology, 112, 102–110. http://dx.doi.org/10.1016/j.jecp
.2011.11.003
Barber, S. J., Gordon, R., & Franklin, N. (2009). Self-relevance and
wishful thinking: Facilitation and distortion in source monitoring. Memory & Cognition, 37, 434 – 446. http://dx.doi.org/10.3758/MC.37.4.434
Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001).
Bad is stronger than good. Review of General Psychology, 5, 323–370.
http://dx.doi.org/10.1037/1089-2680.5.4.323
Benenson, J. F., & Dweck, C. S. (1986). The development of trait explanations and self-evaluations in the academic and social domains. Child
Development, 57, 1179 –1187. http://dx.doi.org/10.2307/1130441
Boseovski, J. J. (2010). Evidence for “rose-colored glasses”: An examination of the positivity bias in young children’s personality judgments.
Child Development Perspectives, 4, 212–218. http://dx.doi.org/10.1111/
j.1750-8606.2010.00149.x
Bransford, J. D., & Johnson, M. K. (1973). Considerations of some
problems of comprehension. In W. G. Chase (Ed.), Visual information
processing (pp. 383– 438). New York, NY: Academic Press. http://dx
.doi.org/10.1016/B978-0-12-170150-5.50014-7
Bruck, M., & Ceci, S. J. (1999). The suggestibility of children’s memory.
Annual Review of Psychology, 50, 419 – 439. http://dx.doi.org/10.1146/
annurev.psych.50.1.419
Buttelmann, D., & Böhm, R. (2014). The ontogeny of the motivation that
underlies in-group bias. Psychological Science, 25, 921–927. http://dx
.doi.org/10.1177/0956797613516802
Cosmides, L., Tooby, J., Fiddick, L., & Bryant, G. A. (2005). Detecting
cheaters. Trends in Cognitive Sciences, 9, 505–506. http://dx.doi.org/
10.1016/j.tics.2005.09.005
Dijkstra, P., Kuyper, H., van der Werf, G., Buunk, A. P., & van der Zee,
Y. G. (2008). Social comparison in the classroom: A review. Review of
Educational Research, 78, 828 – 879. http://dx.doi.org/10.3102/
0034654308321210
Epley, N., & Dunning, D. (2000). Feeling “holier than thou”: Are selfserving assessments produced by errors in self- or social prediction?
Journal of Personality and Social Psychology, 79, 861– 875. http://dx
.doi.org/10.1037/0022-3514.79.6.861
Festinger, L. (1954). A theory of social comparison processes. Human
Relations, 7, 117–140. http://dx.doi.org/10.1177/001872675400700202
Fiske, S. T. (2011). Envy up, scorn down: How status divides us. New
York, NY: Russell Sage Foundation.
Goethals, G. R., & Darley, J. M. (1977). Social comparison theory: An
attributional approach. In J. M. Suls & R. L. Miller (Eds.), Social
comparison processes: Theoretical and empirical perspectives (pp.
259 –278). Washington, DC: Hemisphere.
Gordon, R., Franklin, N., & Beck, J. (2005). Wishful thinking and source
monitoring. Memory & Cognition, 33, 418 – 429. http://dx.doi.org/
10.3758/BF03193060
Heyman, G. D., & Dweck, C. S. (1998). Children’s thinking about traits:
Implications for judgments of the self and others. Child Development,
69, 391– 403. http://dx.doi.org/10.1111/j.1467-8624.1998.tb06197.x
Hunt, R. R. (2006). The concept of distinctiveness in memory research. In
R. R. Hunt & J. B. Worthen (Eds.), Distinctiveness and memory (pp.
3–26). New York, NY: Oxford University Press.
Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psychological Bulletin, 114, 3–28. http://dx.doi.org/10.1037/00332909.114.1.3
Kinzler, K. D., & Shutts, K. (2008). Memory for “mean” over “nice”: The
influence of threat on children’s face memory. Cognition, 107, 775–783.
http://dx.doi.org/10.1016/j.cognition.2007.09.005
Leichtman, M. D., & Ceci, S. J. (1995). The effects of stereotypes and
suggestions on preschoolers’ reports. Developmental Psychology, 31,
568 –578. http://dx.doi.org/10.1037/0012-1649.31.4.568
Lindsay, D. S. (2002). Children’s source monitoring. In H. L. Westcott,
G. M. Davies, & R. H. C. Bull (Eds.), Children’s testimony: A handbook
of psychological research and forensic practice (pp. 83–98). Chichester,
UK: Wiley. http://dx.doi.org/10.1002/9780470713679.ch6
Liu, D., Gelman, S. A., & Wellman, H. M. (2007). Components of young
children’s trait understanding: Behavior-to-trait inferences and trait-tobehavior predictions. Child Development, 78, 1543–1558. http://dx.doi
.org/10.1111/j.1467-8624.2007.01082.x
LoBue, V., & Thrasher, C. (2014). The Child Affective Facial Expression
(CAFE) set: Validity and reliability from untrained adults. Frontiers in
Psychology, 5, 1532.
Lockhart, K. L., Chang, B., & Story, T. (2002). Young children’s beliefs
about the stability of traits: Protective optimism? Child Development,
73, 1408 –1430. http://dx.doi.org/10.1111/1467-8624.00480
Lockhart, K. L., Nakashima, N., Inagaki, K., & Keil, F. C. (2009). From
ugly duckling to swan? Japanese and American beliefs about the stability and origins of traits. Cognitive Development, 23, 155–179. http://dx
.doi.org/10.1016/j.cogdev.2007.08.001
Mather, M., & Carstensen, L. L. (2005). Aging and motivated cognition:
The positivity effect in attention and memory. Trends in Cognitive
Sciences, 9, 496 –502. http://dx.doi.org/10.1016/j.tics.2005.08.005
Mather, M., Shafir, E., & Johnson, M. K. (2000). Misremembrance of
options past: Source monitoring and choice. Psychological Science, 11,
132–138. http://dx.doi.org/10.1111/1467-9280.00228
McGuire, S., Manke, B., Eftekhari, A., & Dunn, J. (2000). Children’s
perceptions of sibling conflict during middle childhood: Issues and
sibling (dis)similarity. Social Development, 9, 173–190. http://dx.doi
.org/10.1111/1467-9507.00118
Pizarro, D. A., Laney, C., Morris, E. K., & Loftus, E. F. (2006). Ripple
effects in memory: Judgments of moral blame can distort memory for
events. Memory & Cognition, 34, 550 –555. http://dx.doi.org/10.3758/
BF03193578
Pronin, E. (2007). Perception and misperception of bias in human judgment. Trends in Cognitive Sciences, 11, 37– 43. http://dx.doi.org/
10.1016/j.tics.2006.11.001
Pronin, E., Gilovich, T., & Ross, L. (2004). Objectivity in the eye of the
beholder: Divergent perceptions of bias in self versus others. Psychological Review, 111, 781–799. http://dx.doi.org/10.1037/0033-295X.111
.3.781
Rholes, W. S., & Ruble, D. N. (1986). Children’s impressions of other
persons: The effects of temporal separation of behavioral information.
Child Development, 57, 872– 878. http://dx.doi.org/10.2307/1130364
Ross, H., Smith, J., Spielmacher, C., & Recchia, H. (2004). Shading the
truth: Self-serving biases in children’s reports of sibling conflicts.
Merrill-Palmer Quarterly, 50, 61– 85. http://dx.doi.org/10.1353/mpq
.2004.0005
Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity dominance, and contagion. Personality and Social Psychology Review, 5,
296 –320. http://dx.doi.org/10.1207/S15327957PSPR0504_2
Sanitioso, R., Kunda, Z., & Fong, G. T. (1990). Motivated recruitment of
autobiographical memories. Journal of Personality and Social Psychology, 59, 229 –241. http://dx.doi.org/10.1037/0022-3514.59.2.229
Sheskin, M., Bloom, P., & Wynn, K. (2014). Anti-equality: Social comparison in young children. Cognition, 130, 152–156. http://dx.doi.org/
10.1016/j.cognition.2013.10.008
Smetana, J. G., Killen, M., & Turiel, E. (1991). Children’s reasoning about
interpersonal and moral conflicts. Child Development, 62, 629 – 644.
http://dx.doi.org/10.2307/1131136
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
SELF-SERVING MEMORIES
Steinbeis, N., & Singer, T. (2013). The effects of social comparison on
social emotions and behavior during childhood: The ontogeny of envy
and Schadenfreude predicts developmental changes in equity-related
decisions. Journal of Experimental Child Psychology, 115, 198 –209.
http://dx.doi.org/10.1016/j.jecp.2012.11.009
Stipek, D. J., & Mac Iver, D. (1989). Developmental change in children’s
assessment of intellectual competence. Child Development, 60, 521–
538. http://dx.doi.org/10.2307/1130719
Suls, J., Gastorf, J., & Lawhon, J. (1978). Social comparison choices for
evaluating a sex- and age-related ability. Personality and Social
Psychology Bulletin, 4, 102–105. http://dx.doi.org/10.1177/
014616727800400121
Suls, J., Martin, R., & Wheeler, L. (2002). Social comparison: Why, with
whom, and with what effect? Current Directions in Psychological Science, 11, 159 –163. http://dx.doi.org/10.1111/1467-8721.00191
533
Tottenham, N., Tanaka, J. W., Leon, A. C., McCarry, T., Nurse, M., Hare,
T. A., . . . Nelson, C. (2009). The NimStim set of facial expressions:
Judgments from untrained research participants. Psychiatry Research,
168, 242–249. http://dx.doi.org/10.1016/j.psychres.2008.05.006
Wilson, A. E., Smith, M. D., Ross, H., & Ross, M. (2004). Young
children’s personal accounts of their sibling disputes. Merrill-Palmer
Quarterly, 50, 39 – 60. http://dx.doi.org/10.1353/mpq.2004.0008
Wood, J. V. (1989). Theory and research concerning social comparisons of
personal attributes. Psychological Bulletin, 106, 231–248. http://dx.doi
.org/10.1037/0033-2909.106.2.231
Received January 8, 2015
Revision received February 24, 2015
Accepted March 1, 2015 䡲
Call for Papers: Behavioral Neuroscience
Special Issue on Behavioral Neuroscience of Sleep
Sleep is increasingly recognized as a major factor in awake psychological functioning. Nighttime
brain activation reflects information processing of stimuli encountered the previous day(s), and also
prepares the brain for optimal information processing the next day. Fundamental human and animal
studies have shown mechanisms of sleep-related interactions with cognitive processes; in neurological and psychiatric disorders, sleep abnormalities appear to contribute to cognitive and emotional disturbances. Novel imaging approaches and sensitive behavioral tasks have begun to show
the intricate relations between sleep and the functioning of the awake brain.
The main focus of this special issue will be to highlight progress made in the past 10 years in
understanding the behavioral neuroscience of sleep. This issue will provide an international forum
for researchers to report their most recent findings in the area, draw attention to exciting new
directions in the field, and emphasize new areas for future research. Topics to be covered may
include, but are not limited to:
●
●
●
●
●
●
Memory consolidation with sleep
Cognitive effects of sleep deprivation (experimentally-induced; insomnia)
Relation of psychological function and sleep in clinical conditions
Effect of normal aging on sleep and psychological functioning
Sleep quality and REM behavior disorder
Sleep-related psychological function and its relation to neurotransmitters or sex hormones
Manuscripts for this special issue should be submitted to Behavioral Neuroscience as usual through the
APA Online Submission Portal, accessible through http://www.apa.org/pubs/journals/bne/. The cover
letter should indicate that the authors wish the manuscript to be considered for publication in the Special
Issue on Sleep. Manuscripts must be received by July 01, 2015.
Inquiries about the special section can be directed to the Guest Editor, Alice Cronin-Golomb, at
alicecg@bu.edu.
Journal of Experimental Psychology:
General
How Do Children Learn Novel Emotion Words? A Study of
Emotion Concept Acquisition in Preschoolers
Holly Shablack, Misha Becker, and Kristen A. Lindquist
Online First Publication, December 30, 2019. http://dx.doi.org/10.1037/xge0000727
CITATION
Shablack, H., Becker, M., & Lindquist, K. A. (2019, December 30). How Do Children Learn Novel
Emotion Words? A Study of Emotion Concept Acquisition in Preschoolers. Journal of
Experimental Psychology: General. Advance online publication.
http://dx.doi.org/10.1037/xge0000727
Journal of Experimental Psychology: General
© 2019 American Psychological Association
ISSN: 0096-3445
2019, Vol. 1, No. 999, 000
http://dx.doi.org/10.1037/xge0000727
How Do Children Learn Novel Emotion Words? A Study of Emotion
Concept Acquisition in Preschoolers
Holly Shablack, Misha Becker, and Kristen A. Lindquist
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
University of North Carolina at Chapel Hill
Understanding emotion words is vital to understanding, regulating, and communicating one’s emotions.
Yet, little work examines how emotion words are acquired by children. Previous research in linguistics
suggests that children use the sentence frame in which a novel word is presented to home in on the
meaning of that word, in conjunction with situational cues from the environment. No research has
examined how children integrate these cues to learn the meaning of emotion adjectives (e.g., “happy,”
“sad,” “mad”). We conducted 2 studies examining the role of sentence frame and situational context in
children’s (ages 3–5) understanding of the meanings of novel words denoting emotions. In Study 1 (N ⫽
135) children viewed a conversation wherein a novel “alien” word was presented in 1 of 3 sentence
frames that varied in how likely the word was to denote an emotion (i.e., is daxy, feels daxy, or feels daxy
about). Children selected the image that represented the meaning of the word in a picture-pointing task.
Images depicted aliens experiencing an emotion, a physical state, or performing an action. In Study 2
(N ⫽ 113) situational context was added via cartoons depicting an emotional scenario. Findings suggest
that children are more likely to associate emotion images with a novel word with increasing age, more
informative sentence frames, and when the situational context implies that an emotion is present. This
provides important insight on how educational and clinical settings can use language and situational
context to aide in emotion understanding.
Keywords: emotion development, language, syntax, emotion concepts, concept acquisition
Supplemental materials: http://dx.doi.org/10.1037/xge0000727.supp
Understanding one’s own and others’ emotions is crucial to
social communication, interpersonal relationships, emotion regulation, and well-being (Eisenberg, Sadovsky, & Spinrad, 2005;
Hagelskamp, Brackett, Rivers, & Salovey, 2013; Kashdan, Barrett,
& McKnight, 2015; Lindquist & Barrett, 2008; Lindquist, Satpute,
& Gendron, 2015; Twenge, Catanese, & Baumeister, 2003).
Greater understanding of emotions in childhood predicts later
academic and interpersonal success (for reviews see Lindquist,
Gendron, & Satpute, 2016; Shablack & Lindquist, in press). Moreover, multiple forms of psychopathology are characterized by
difficulties in understanding emotions, such as autism (BaronCohen & Wheelwright, 2004; Baron-Cohen, 1991; Dapretto et al.,
2006; Wang, Dapretto, Hariri, Sigman, & Bookheimer, 2004),
depression (Berenbaum & Oltmanns, 1992; Joormann, 2010; Murphy et al., 1999; Phillips, Drevets, Rauch, & Lane, 2003), and
anxiety (Etkin & Wager, 2007; Mennin, McLaughlin, & Flanagan,
2009; Salters-Pedneault, Roemer, Tull, Rucker, & Mennin, 2006).
It is thus important to understand the mechanisms by which some
people gain a complex understanding of emotions, whereas others
fail to.
One hypothesis is that a complex understanding of emotions is
learned, primarily during early development via discourse with
caregivers (Castro, Halberstadt, Lozada, & Craig, 2015; Dunsmore
& Halberstadt, 1997; Garrett-Peters, Castro, & Halberstadt, 2017;
Halberstadt & Lozada, 2011; Shablack & Lindquist, in press.;
Weinberg, Tronick, Cohn, & Olson, 1999). In this view, interactions with caregivers, but particularly emotion word use by caregivers, helps children to acquire a rich cache of knowledge about
the emotion concepts most relevant to their culture (Campos,
Frankel, & Camras, 2004; Denham, Zoller, & Couchoud, 1994;
Editor’s Note. Cathie Tamis-LeMonda served as the action editor for this
article.—NC
Holly Shablack, Department of Psychology and Neuroscience, University of North Carolina at Chapel Hill; Misha Becker, Department of
Linguistics, University of North Carolina at Chapel Hill; X Kristen A.
Lindquist, Department of Psychology and Neuroscience, University of
North Carolina at Chapel Hill.
This research was supported by a grant from the Fostering Interdisciplinary Research Explorations program at UNC Chapel Hill, awarded to
Kristen A. Lindquist and Misha Becker. We also thank all of the research
assistants who helped finalize study materials and run participants. A
subset of these findings was presented at the Society for Affective Science
Conference in 2018 and Cognitive Development Society in 2019. This
research was supported by a grant from the Fostering Interdisciplinary
Research Explorations program at UNC Chapel Hill. Findings were presented at the Society for Affective Science Conference in 2018 and
Cognitive Development Society in 2019.
Correspondence concerning this article should be addressed to Holly
Shablack, Department of Psychology and Neuroscience, University of
North Carolina at Chapel Hill, 235 East Cameron Avenue, Chapel Hill, NC
27599. E-mail: hshab@live.unc.edu
1
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2
SHABLACK, BECKER, AND LINDQUIST
Ellis, Alisic, Reiss, Dishion, & Fisher, 2014; Fivush, Brotman,
Buckner, & Goodman, 2000; Fivush, Haden, & Reese, 2006; Fogel
et al., 1992; Halberstadt, Denham, & Dunsmore, 2001; Pons,
Harris, & de Rosnay, 2004). Indeed, research in cognitive science
suggests that words help infants, children, and adults acquire and
use concepts of all kinds (Lupyan, 2012a, 2012b; Xu & Kushnir,
2013). In particular, words facilitate the acquisition of socially
shared concepts such as emotions (Doyle & Lindquist, 2018),
color (Steels & Belpaeme, 2005), and spatial relations (Casasanto,
2008; Casasanto & Bottini, 2014) in adults. Yet the mechanisms
by which children learn emotion concepts and their labels remains
relatively unknown. Unlike object concepts (e.g., tiger, table),
which are concrete, stable, and typically labeled by nouns, emotion
concepts are abstract, transient, and typically labeled by adjectives,
which makes them a difficult class of concepts to learn labels for
(Gentner, 1982; Pinker, 1984). By combining research in the
development of emotion and psycholinguistics, the present studies
assess for the first time how children might learn to associate novel
words with emotion concepts. We hypothesize that like certain
other lexical categories, children use the syntactic structure of
language to infer the meanings of novel adjectives labeling emotion concepts. In addition, based on prior research on emotion
concept understanding (Kayyal, Widen, & Russell, 2015; Widen &
Russell, 2010, 2011) we hypothesize that situational cues from the
environment will be needed to understand that novel adjectives
denote emotion concepts.
The Role of Language in Emotion Development
Many psychological models of emotional development argue
that the ability to experience, perceive, and ultimately understand
specific emotion concepts follows a developmental trajectory and
is learned through social relationships and verbal communication
between children and caregivers (Castro et al., 2015; Dunsmore &
Halberstadt, 1997; Garrett-Peters et al., 2017; Halberstadt & Lozada, 2011; Shablack & Lindquist, in press; Weinberg et al., 1999).
In this view, infants start life with the ability to experience in their
own bodies and perceive in others very basic feelings such as
agitation and excitement (Bridges, 1932) or pleasantness and unpleasantness (Camras, 1992; La Barbera, Izard, Vietze, & Parisi,
1976; Lewis & Brooks, 1978; Russell & Bullock, 1985, 1986;
Widen, 2013). It is hypothesized that over time, infants and children learn to make more fine-grained discriminations among these
basic feelings. According to psychological constructionist approaches, concept knowledge about emotions is what ultimately
helps children to learn to differentiate between different types of
unpleasantness (e.g., fear vs. anger) or different types of pleasantness (e.g., joy vs. pride; Barrett, 2006, 2013; Barrett & Russell,
2015; Lindquist, 2013; Russell, 2003).
One perspective is that language helps children to acquire emotion concept knowledge because words serve as “essence placeholders” that cohere together instances (e.g., feeling aggressive,
feeling like one’s goals are blocked, feeling one’s heart beating
more quickly) as members of the same emotion category (e.g.,
anger; for a review, see Lindquist, MacCormack, & Shablack,
2015). This hypothesis shares much in common with hypotheses
about the role of language in the acquisition of other concept types
(Lindquist, MacCormack et al., 2015; Lupyan, 2012a, 2012b; Xu
& Kushnir, 2013). It suggests that as children develop a larger
emotion vocabulary, they develop more nuanced emotion concept
knowledge and are thus able to perceive, express and experience a
wider range of emotions.1
There is ample evidence to suggest that discourse with parents
about emotions predicts children’s greater emotional perception
and understanding. For instance, correlational evidence suggests
that as children age, emotion word knowledge increases (Bretherton & Beeghly, 1982; Ridgeway, Waters, & Kuczaj, 1985; Wellman, Harris, Banerjee, & Sinclair, 1995) as does performance in
emotion perception tasks (Astington & Jenkins, 1999; Cutting &
Dunn, 1999; Harris, De Rosnay, & Pons, 2005; Wellman et al.,
1995). Furthermore, in 3- to 6-year-olds, general verbal ability
(when controlling for age, attachment security, and gender) is an
important predictor of children’s ability to understand the emotions of others (de Rosnay & Harris, 2002; de Rosnay, Pons,
Harris, & Morrell, 2004; Pons, Lawson, Harris, & De Rosnay,
2003). When using emotion words more specifically, toddlers
initially describe their own feelings and the feelings of others in
broad valenced terms by using general words such as happy and
sad or mad. Yet by age 5 they additionally use words such as
afraid, surprised, and disgust to describe a more nuanced range of
emotional states (Widen & Russell, 2003). This effect appears to
go beyond mere labeling, as children’s ability to perceptually
identify emotional facial expressions in a nuanced manner increases as emotion labeling ability increases. For instance, early in
toddlerhood, children tend to use the words happy and sad2 in
daily discourse, and correspondingly can only reliably differentiate
pleasant and unpleasant facial expressions from one another. Yet,
around the time that children begin to use the words anger and fear
in daily discourse to differentiate between different negative states,
they also become able to perceive these negative emotions on faces
(i.e., distinguish anger from fear) in perceptual sorting tasks
(Widen & Russell, 2008). Importantly, asking children to match
emotional facial expressions to words, as opposed to other facial
expression exemplars, facilitates children’s performance, even
1
Psychological constructionist models of emotion stand in contrast to
“basic emotion” approaches that assume that certain emotions are innate,
such that children are able to experience in their own bodies and perceive
in others discrete emotion concepts such as anger, disgust, fear, happiness,
sadness, and surprise from birth (Ekman, 1992; Ekman & Cordaro, 2011;
Izard, 2007; Panksepp, 1998). Although a prominent view, there is relatively little empirical support for the idea that infants and young children
reliably perceive these emotions in others or themselves; the evidence is
more consistent with the hypothesis that infants and children can differentiate between dimensions such as valence and develop an understanding
of different emotion categories over early childhood (Ruba & Repacholi, in
press; Shablack & Lindquist, in press; Widen, 2013). Although some basic
emotion views acknowledge a role for learning in emotion understanding
and experience, they stipulate that language itself is unrelated to emotion
save for communication (Ekman & Cordaro, 2011). Nonetheless, these
approaches would require that children need to learn how to communicate
about emotions with words, a task which requires mapping experienced
concepts to word forms during early development.
2
Whereas some children use the word sad to label negative states, others
use the word mad instead (Widen, 2013). Nonetheless, most children at this
age only reliably use two words for emotion, meaning that if they use
happy and sad, they do not also use mad (and vice versa). It is unknown
why this idiolectal difference in use of sad versus mad occurs, but one
possibility is that this difference could be a product of socialized gender
stereotypes (Plant, Hyde, Keltner, & Devine, 2000), whereby parents use
the term sad more with girls and mad more with boys when explaining
emotions (Fivush, 1991; Fivush et al., 2000).
CHILDREN’S EMOTION ACQUISITION
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
among children as young as 2-years-old (Russell & Widen, 2002a,
2002b).
These findings suggest that caregivers help children to label
their own emotional states and the emotional states of others in an
effort to scaffold their acquisition of knowledge about the emotion
concepts relevant to the present context (Campos et al., 2004;
Denham et al., 1994; Ellis et al., 2014; Fivush et al., 2000, 2006;
Fogel et al., 1992; Halberstadt et al., 2001; Pons et al., 2004).
However, very little research has assessed how children learn that
a word denotes an emotion concept in the first place. This is an
important next step in understanding the relationship between
language and emotion and is the purpose of the present studies.
The Role of Linguistic Cues in Word Learning
Much of the experimental work on children’s vocabulary development focuses on how children acquire words for object concepts, which are primarily labeled by nouns (Bloom, 2000; Gentner, 1982; Huttenlocher & Smiley, 1987; Markman, 1990). This
emphasis is logical, as children’s earliest vocabulary items are
largely nouns that label people and basic objects (Bates et al.,
1994). However, words of different lexical categories (verbs, adjectives, etc.) tend to have very different kinds of meanings and are
learned in very different ways. For instance, verbs often label
actions and events, and adjectives, which modify nouns, typically
label properties or attributes. Emotions are internal states that are
most frequently labeled by adjectives in everyday speech
(Shablack, 2017), and verbs and adjectives are conceptually more
complex than nouns (Gentner, 1982). Moreover, while caregivers
may label salient objects for children ostensively (e.g., “Look!
That’s a dog!”), caregivers do this only rarely (if at all) with
properties and states of being (Gleitman, 1990).
As a product of their abstract and complex nature and lack of
ostensive instruction, verb and adjective meanings are less
straightforward than noun meanings for children to learn, and they
are learned later (Gentner, 1982). An influential approach called
the syntactic bootstrapping hypothesis (Fisher, Gleitman, & Gleitman, 1991; Gleitman, 1990; Gleitman, Cassidy, Nappa, Papafragou, & Trueswell, 2005; Naigles, 1996) points to the role of
syntactic structure in cueing the meanings of these types of words;
work within this approach has focused mainly on verbs. Both
within a language and across different languages, the number of
noun phrases (i.e., arguments) that occur with a verb in a sentence
is related in highly regular ways to the lexical meaning of the verb.
Children can thus use these arguments to infer the meaning of
verbs (see Table 1 for examples). For instance, the novel verb gorp
in the sentence Bill gorped Jane a rose could not mean something
like “sleep” or “kiss” (note the oddness of ⴱBill slept Jane a rose).3
The syntactic bootstrapping hypothesis has been tested (Fisher
et al., 1991; Gleitman, 1990; Gleitman et al., 2005; Naigles, 1990,
1996) by presenting young children with novel verbs used in
particular sentence “frames” (e.g., with one vs. two arguments).
Children are then asked to point to, or their eye gaze is tracked
toward a visual image depicting the relevant action (Fisher, 2002;
Hirsh-Pasek & Golinkoff, 1996; Naigles, 1990; Yuan & Fisher,
2009). The methodology has been adapted to test learners’ interpretations of not only verbs denoting observable actions but also
relatively more abstract kinds of verbs, such as those that could
have a meaning like think or seem, or to abstract adjectives such as
3
easy (Becker, 2006, 2014, 2015; Becker & Estigarribia, 2013;
Papafragou, Cassidy, & Gleitman, 2007). In some cases, such as
for abstract verbs such as think, the sentence frame is even more
informative than extragrammatical situational context (what is
going on in the world when a particular verb is uttered) for cueing
children to the predicate’s meaning (Gleitman et al., 2005; Papafragou et al., 2007). However, children can draw such inferences
about more abstract verbs only after age 4 (Becker, 2006, 2014).
Although syntactic bootstrapping is a likely mechanism by
which children learn predicate meanings, there is little work examining the relative role of sentence frame in the acquisition of
adjectives as a lexical category (Booth & Waxman, 2003, 2009;
Syrett & Lidz, 2010). No work to our knowledge looks at the
application of this approach to novel emotion concepts. Insofar as
emotions are most often labeled by adjectives (happy, sad, mad,
etc.; Shablack, 2017), children should be able to use syntactic
bootstrapping to understand the meaning of novel emotion words
and concepts. On the other hand, emotions are highly situated
concepts in which people are responding to real or imagined events
in the world around them (Wilson-Mendenhall, Barrett, Simmons,
& Barsalou, 2011). As such, access to details about the context in
which emotions are occurring (Harris, Johnson, Hutton, Andrews,
& Cooke, 1989; Kayyal et al., 2015; Widen & Russell, 2010, 2011,
2013) may influence a child’s inferences about the meaning of
emotion concept labels. In particular, situational details from the
immediate context, such as information about the causes of a
certain emotion (e.g., loss) and what a person did as a consequence
(e.g., cry) help children to accurately identify the emotion concept
being described in a story, often above and beyond other perceptual information such as facial expressions (Kayyal et al., 2015;
Widen & Russell, 2010, 2011). When the evidence is taken together, it is likely that children use both linguistic information
(sentence frames) and situational cues from the environment
(causes and consequences of emotion) in the acquisition of emotion words. Yet no work to our knowledge has assessed how
children use syntactic structure in combination with situational
context to learn the meaning of emotion concept labels.
Present Studies
We conducted two studies examining the extent to which sentence frames and situational context impact children’s (aged 3–5)
understanding that novel words denote emotion concepts. In Study
1, we manipulated the sentence frame a novel word was presented
in to examine whether sentence frame alone helps children perceive that a word denotes an emotion concept as opposed to other
predicate meanings such as a physical state or an action, and at
what age this occurs. Sentence frames were manipulated by presenting a novel word in a verb structure that limits the possible
meaning of the word. For example, when presenting a novel word
(binty) as the sole complement of the verb is (i.e., Susan is binty),
the meaning of binty is fairly unlimited. As an adjective binty can
denote an emotion, such as “happy,” a physical state, such as
“cold,” or even a physical characteristic, such as “tall” (note the
grammaticality of Susan is happy/cold/tall). However, if binty is
the complement of the verb feels (i.e., Susan feels binty), the
3
Linguists use the symbol
ability.
ⴱ
to indicate ungrammaticality or unaccept-
SHABLACK, BECKER, AND LINDQUIST
4
Table 1
Examples of Syntactic Structure Cueing Possible Meaning
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Number of
arguments (NP)
Examples
Verb denotation
Single NP
Two NPs
John slept; Mary is running
Sue kissed the baby; Kevin hugged his friend
Three NPs
Katrina gave Emily a pen; Roger told Marvin a story
meaning of binty can no longer denote a physical characteristic
like “tall” (excepting figurative uses of this adjective), though it
could still denote either an emotion (“happy”) or a temporary
physical state (“cold”; thus: Susan feels happy/cold/ⴱtall). Finally,
if binty is used within a sentence like Susan feels binty about
something, the meaning of binty is further restricted to denoting a
mental or emotional state; it can no longer denote a physical state
like “cold” (Susan feels happy/ⴱcold/ⴱtall about something). Therefore, we predicted that if children rely exclusively on linguistic
cues, the likelihood for a child to determine that a novel word is an
emotion concept label would increase as the sentence frame becomes more restrictive.
In Study 2, we examined the extent to which the sentence frame
helped children acquire the meaning of novel emotion concepts
when that acquisition was also scaffolded by situational context.
We manipulated the sentence frame as in Study 1 and also told
children a “background” story that highlighted the causes and
consequences of an experience that the main character underwent.
For example, prior to a sentence such as “Palooza feels binty,”
children heard a story about a character receiving a gold star for
her drawing, providing situational context for what she was feeling
and why. Study 2 thus examined whether the sentence frame a
word appeared in is especially useful for understanding the meaning of a novel emotion concept when that word appears in a
relevant situation.
Across both studies we hypothesized that older children would
be more likely to perceive a novel word as an emotion concept
label given appropriate linguistic and situational cues; however, as
we hypothesize that sentence frames and situational context provide specific cues in concept acquisition and word disambiguation,
we predicted age interactions with sentence frame (Study 1),
especially in the presence of situational context (Study 2).
Study 1
In Study 1, we predicted that children would use sentence frame
as a cue for their understanding of novel emotion words. Specifically, if children use sentence frame alone to drive emotion
concept word acquisition, then children would especially perceive
a novel word as labeling an emotion when it is presented in a more
restrictive sentence frame (e.g., feels or feels about). Mirroring
past work on syntactic bootstrapping (e.g., Papafragou et al., 2007)
we predict that children would become more adept at using the
sentence frame as age increases.
Method
Participants. One hundred sixty-two children participated in
the study at the Museum of Life and Science in Durham, North
An action/event that an individual engages in by themselves
An action/event that one individual does to another (i.e., the
object is affected by the action)
Generally means something about transfer or communication
Carolina. Children who did not meet our a priori inclusion criteria
were excluded from further analysis. Nine children were not within
our desired age range (i.e., were siblings of other children who
participated or were found to not be in the proper age range
following parent report). Three additional children were removed
from the analysis because they lived in households in which
languages other than English were spoken greater than 50% of the
time; we reasoned that in the case of bilingual children our sentence frame manipulation would be less effective since those
children might regularly experience more varied sentence frames.
Two children were removed from analysis because their parent/
guardian reported that they had a learning disability on the parent
questionnaires. An additional 12 children started, but did not
complete the video task because they either stopped participation
part way through (n ⫽ 9) or failed the screening task (n ⫽ 3). One
child completed an incorrect survey due to experimenter error and
is not included. The final sample consisted of 135 (Mage ⫽ 3.97,
SDage ⫽ .79; 72 female) children: 44 3-year-olds (25 female), 51
4-year-olds (27 female), and 40 5-year-olds (20 female).4
Children completed the study individually with a single experimenter. A second experimenter observed the interaction and answered any parent/guardian questions. Parents and/or legal guardians completed a packet of questionnaires about their child’s home
life and development. Sixteen individuals did not report family
income; of the 119 who did, 19.3% reported a household income ⬍$70,000; 21% reported between $70,000 and $100,000;
29.4% reported between $100,000 and $150,000; and 30.3% reported $150,000 or higher. Thirty-one individuals did not report
the race/ethnicity of one or both parents. Of the 104 who did,
83.7% of the children were Caucasian/White, 13.5% multiracial,
1.9% African American, and .9% Asian.
Materials.
Novel word videos. Children viewed videos of two animal
hand puppets conversing about aliens (see Table 2 for general
dialogue) followed by a picture pointing task with three image
options. These brief dialogues were modeled after those used by
Yuan and Fisher (2009) and Arunachalam and Waxman (2010) in
a syntactic bootstrapping study with 2-year-olds. Videos were on
average 16-s long and contained voices from two out of four
different individuals. Voice identity was quasirandomly distrib4
Because the museum’s staff wished to create an atmosphere of inclusivity, we permitted all interested children to participate in the task.
Children were excluded post hoc only for the reasons listed above (out of
age range, not monolingual, learning impairment, etc.; i.e. reasons that
would have led to them not participating if we had prescreened them for
eligibility). We excluded children who had already started the task if they
failed to continue the task at that point without making children feel
excluded.
CHILDREN’S EMOTION ACQUISITION
Table 2
Study 1 Sentence Frames and General Puppet Video Dialogue
Sentence frame
Is
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Feels
Feels about
Fillers
General script
Puppet 1: I know an alien who is [novel word].
Puppet 2: Really? You know an alien who is
[novel word]?
Puppet 1: Yes! This alien is [novel word]
Puppet 2: Wow! You know an alien who is
[novel word]
Puppet 1: I know an alien who feels [novel
word].
Puppet 2: Really? You know an alien who feels
[novel word]?
Puppet 1: Yes! This alien feels [novel word]
Puppet 2: Wow! You know an alien who feels
[novel word]
Puppet 1: I know an alien who feels [novel
word] about brushing his teeth.
Puppet 2: Really? You know an alien who feels
[novel word] about brushing his teeth.?
Puppet 1: Yes! This alien feels [novel word]
about brushing his teeth.
Puppet 2: Wow! You know an alien who feels
[novel word] about brushing his teeth.
Puppet 1: I saw an alien who was [novel word].
Puppet 2: Really? You saw an alien who was
[novel word]?
Puppet 1: Yeah! I saw an alien who was [novel
word]
Puppet 2: Oh! You saw an alien who was
[novel word]
uted across all videos. Conversations about aliens were used both
to engage children’s interest and to allow the framing of the novel
words as new “alien words.” This limits any undue influence of
already known emotion concept words via mutual exclusivity (i.e.,
the belief that two words cannot have the same meaning; Clark,
1987; Hutchinson, 1986; Markman, 1990).
Prior to the experiment, each child completed three screening
trials. Each screening trial had a single video that discussed an
alien engaging in an action (i.e., “I know an alien who likes to eat
pizza”) and an accompanying descriptive sentence using a verb in
the present progressive (i.e., “Point to where the alien is eating
pizza!”). To continue, children had to correctly answer at least two
trials (i.e., point to the picture of the alien eating pizza and not the
distractor pictures), demonstrating that they could watch a brief
video accompanied by a verbal description and point to a picture
that matched the description.
In the subsequent experimental trials, children watched the hand
puppet videos with puppets conversing and introducing a novel
word (binty, daxy, strupy, moky, joomy, gorpy, reksy, tropy) four
times using one of the following sentence frames:
Is: The alien is [novel word]
Feels: The alien feels [novel word].
Feels about: The alien feels [novel word] about something.5
Each child was presented with the same sentence frame for all
experimental trials and no novel word was repeated across trials.
Each experimental video and novel word was presented in random
order.
5
In between each experimental trial was a filler trial containing a
novel verb in the present progressive (ending in -ing: piffing,
tayving, serding) to indicate an action. The filler trials were included to give children a break from the experimental trials and
were not intended as control trials. Nonetheless, child performance
on these trials was included as a covariate in the event that
children’s filler performance significantly impacted their performance on the trials of interest. See online supplemental materials
for analyses of filler trials and main analyses without filler trial
performance as a covariate. Over the course of the session, each
child saw 10 videos (three screening, four experimental, three
filler).
Picture pointing task. Following each video, a screen containing auditory instructions with accompanying text instructed the
child to “Point to where the alien [sentence frame] [novel word].”
The voice from a single individual was used in the auditory
instruction. Children then saw three images of a cartoon alien
presented in a random array, with each image depicting the alien
expressing an emotion, a physical state, or engaging in an action
(see Table 3 for list of image types and Figure 1 for sample trial).
Children were instructed to point to the image that corresponded to
the novel word. Their choice served as the dependent variable.
Three alien identities were used in screening trials. Four different
alien identities were used for the experimental and filler trials, with
the limitation that each experimental trial employed a different
alien (thus, the same alien may have been seen in both an experimental trial and a filler, but not in two experimental trials). Images
were randomized across all trials to appear once (i.e., children saw
all seven possible emotion images appearing in either an experimental or filler trial). All alien stimuli were validated in a separate
sample of 3- to 5-year-olds, in which children were better than
chance at associating the images with the intended physical state,
emotion or action (see online supplemental materials).
Parent questionnaire. While the child completed the computerized task, parents and/or legal guardians completed a voluntary
paper questionnaire. Information was gathered on the child’s birth,
including date, location and whether the mother’s pregnancy was
normal. Children’s general communicative and linguistic development was measured including any information on whether the
child had been evaluated for speech problems or learning disabilities, and a rudimentary number of spontaneous word production
of a subset of word categories. Information on familial language
was gathered, as well as who the child lived with. Parental race/
ethnicity, place of birth and native language, education, career, and
household income were also gathered. The first was used as a
measure of child race/ethnicity.
Procedure. This study was approved by the University of
North Carolina Institutional Review Board. Parents of children
who looked to be 3–5 years of age were approached at the museum
and asked if their child would like to participate in a short video
task on word learning. Only children who were 3–5 years of age
were subsequently enrolled (with the exception of children who
were accidentally enrolled due to a miscommunication about their
5
For the feels about conditions, story endings were: “about brushing his
teeth”; “about cleaning her room”; “about eating cookies”; and “about
playing games.” The picture pointing task instructed kids to “point to
where the alien feels [novel word] about something.”
SHABLACK, BECKER, AND LINDQUIST
6
Table 3
Image Stimuli for Study 1 and Study 2
Emotions
Actions
Physical states
happy, excited, sad, mad, scared, disgusted,
surprised
sleeping, jumping, sitting, falling, cartwheeling,
walking, running, eating pizza/fruitⴱ,
swimmingⴱ
itchy, hot, cold, sick, burned, hungry, hurt
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Note. Four cartoon alien identities exhibited the listed emotion, action, or
physical state. Screening trials included an emotion image and two action
images.
ⴱ
denotes action images that were only included in screening trials.
age or allowed to participate because a sibling participated; these
children were not included in analyses). Children were told that
they would play a game involving aliens. Following the parent/
legal guardian’s consent, the experimenter spent the first few
minutes getting to know the child and obtaining verbal assent. The
experimental task was then administered online on a laptop computer via Qualtrics.
Children who failed two or more of the three screening trials
were thanked for their time and told that the game was over. If
children passed the screening trials, the experimenter then introduced the experimental task by saying “You’re doing great! Let’s
keep playing this game! Now we’re going to watch some more
videos. These videos are going to have a special alien word and I
need help to figure out what that word means! Do you want to
play/Are you ready?” If the child wanted to continue, the experimenter then played the first experimental video containing a novel
“alien” word followed by a screen with visual and auditory instructions for the picture pointing task. Once an image was chosen,
the experimenter confirmed the choice and made the selection by
clicking the radio button beneath the image the child chose. If a
child was hesitant, the experimenter encouraged them up to three
times before continuing to the next trial. Following the video task,
children had the option to complete an additional task, not discussed here. For their participation, children received a hand stamp
and temporary tattoo.
Results
To examine the impact of sentence frame and age on image
choice, mean proportions of each image choice type were created
across trial types (e.g., mean proportion of emotion images chosen
within each experimental condition).6 We opted to treat age as a
categorical variable rather than a continuous variable due to the
ease of interpreting mixed model ANOVAs over regression models with categorical outcomes. We nonetheless also computed
multilevel multinomial logistic regression analyses in which age
was a continuous predictor and sentence frame was a categorical
predictor of choice outcomes (see online supplemental materials).
These findings replicated the mixed model ANOVA findings so
we report the ANOVA findings in the main text for ease of
interpretation.
To examine our main hypothesis that sentence frame and age
would interact to influence children’s image choice, we conducted
a 3 (image choice: emotion, action, physical state) ⫻ 3 (sentence
frame: is, feels, feels about) ⫻ 3 (age: 3, 4, 5) mixed model
ANOVA with mean proportions of Image choice as a within
subjects factor and sentence frame and age as between subjects
factors. To control for the effects of participant gender and screening trial performance, we also conducted a 3 (image choice:
emotion, physical state, action) ⫻ 3 (sentence frame: is, feels, feels
about) ⫻ 3 (age: 3, 4, 5) mixed model ANCOVA with gender and
performance on the screening trials as covariates. Lastly, to account for filler trial performance, we conducted a 3 (image choice:
emotion, physical state, action) ⫻ 3 (sentence frame: is, feels, feels
about) ⫻ 3 (age: 3, 4, 5) mixed model ANCOVA with gender,
performance on the screening trials, and performance on filler
trials as covariates. Filler trial performance was computed as the
proportion of trials in which an action image was chosen. Findings
were largely identical across the three analyses, so we report the
most conservative ANCOVA findings controlling for gender, performance on screening and filler trials. Findings from the other
ANOVA and ANCOVA are available in the online supplemental
materials.
See Table 4 for all effects; we explicitly discuss only significant
effects and predicted effects. First, we found a main effect of
image choice, F(2, 246) ⫽ 3.10, p ⫽ .05, 2 ⫽ .03.7 Children
chose physical state (p ⬍ .001) and emotion (p ⬍ .001) images
significantly more than action images. Children chose physical
state images significantly more than emotion images over all (p ⫽
.002; see Figure 2).
The main effect of image choice was qualified by a significant
interaction between image choice and age, F(4, 246) ⫽ 2.53, p ⫽
.04, 2 ⫽ .04. Simple effects show a significant effect of age on
action images, F(2, 123) ⫽ 4.69, p ⫽ .01, 2 ⫽ .07 and a
significant effect of age on physical state images, F(2, 123) ⫽
3.65, p ⫽ .03, 2 ⫽ .06. Pairwise comparisons reveal that 3-yearolds chose action images (p ⫽ .003) significantly more than
5-year-olds. Five-year-olds chose physical state images (p ⫽ .008)
significantly more than 3-year-olds did. Emotion images were
chosen equally between all ages (see Figure 3).
As predicted, there was a marginal interaction between image
choice and sentence frame, F(4, 246) ⫽ 2.28, p ⫽ .06, 2 ⫽ .04,
suggesting that children relied on more restrictive sentence frames
to inform image choice. Simple effects reveal that the proportion
of trials in which emotion images were chosen is marginally
influenced by the sentence frame, F(2, 123) ⫽ 2.92, p ⫽ .06, 2 ⫽
.05. Pairwise comparisons demonstrated that emotion images were
chosen significantly more in the feels about (p ⫽ .02) than the feels
condition, but that the is condition (p ⫽ .57) did not differ from
feels about condition. Emotion images were chosen marginally
more during is trials than during feels trials (p ⫽ .09; see Figure 4).
We did not find a predicted three-way interaction between
image choice, age, and sentence frame, suggesting that children
did not rely more on sentence frame with increasing age. Instead,
age and sentence frame separately moderated image choice.
6
Eight children had technical difficulties during one or two trials. In
these cases, mean proportions were weighted by the number of completed
trials.
7
Mauchly’s test indicated that the assumption of sphericity was violated, 2 (2) ⫽ 6.36, p ⫽ .04, therefore degrees of freedom were corrected
using Huynh-Feldt corrections follow Girden (1992) suggestion based on
Greenhouse-Geisser estimate, ε ⫽ 0.95, being greater than .75.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
CHILDREN’S EMOTION ACQUISITION
7
Figure 1. Sample trials for Study 1 and Study 2. For the text of the Study 1 video, see Table 2. In Study 2,
children heard a two to three sentence story, see Table 4 for examples. Each video was played on a screen by
itself, and the next screen presented three randomized images of the alien character, accompanied by audio
instructions for the child to point to the image they believe depicts the novel word. See the online article for the
color version of this figure.
Discussion
Results provide preliminary support for our hypotheses in Study
1, which is a conservative test of the syntactic bootstrapping
hypothesis in which children received no other context for the
Table 4
Study 1 Within Subjects Main Effects and Interactions for 3
(Image Choice: Emotion, State, Action) ⫻ 3 (Sentence Frame:
Is, Feels, Feels About) ⫻ 3 (Age: 3, 4, 5) Mixed Model
ANCOVA With Participant Gender, Screening Trial
Performance, and Filler Trial Performance as Covariates
Variable
Image
Image
Image
Image
Image
Image
Image
choice
Choice
Choice
Choice
Choice
Choice
Choice
⫻
⫻
⫻
⫻
⫻
⫻
Screening Trial Performance
Gender
Filler Performance
Age
Sentence Frame
Age ⫻ Sentence Frame
df
F
/2
p
2
2
2
2
4
4
8
3.10
4.23
2.65
3.55
2.53
2.28
1.28
.03
.03
.02
.03
.04
.04
.04
.05
.02
.07
.03
.04
.06
.25
Note. Mauchly’s test indicated that the assumption of sphericity was
violated, |2 (2) ⫽ 6.36, p ⫽ .04, therefore degrees of freedom were
corrected using Huynh-Feldt corrections follow Girden (1992) suggestion
based on Greenhouse-Geisser estimate, 冱 ⫽ .95, being greater than .75,
dferror ⫽ 246.
meaning of a novel word except the sentence frame it was heard in.
We found that age influenced children’s tendency to perceive a
novel word as denoting an internal state (physical state or emotion)
over an action. Three-, 4-, and 5-year-olds were equally likely to
choose emotion images, with the relative proportion of physical
state to action choices increasing over age. Three-year-olds chose
more action items than 5-year-olds, whereas 5-year-olds were
more likely than 3-year-olds to choose physical state images.
Because all of our novel words ended in -y (daxy, joomy, binty),
and words ending in this sound are prototypically adjectives
(happy, bouncy, hungry, thirsty, rosy, furry, smelly), this feature
alone may have been a cue to 5-year-old children that novel words
labeled attributes or states. The fact that 3-year-olds chose the
action images the most out of the age groups suggests that 3-yearolds may not yet understand that words ending in -y are likely to
be adjectives.
We also found suggestive evidence that more restrictive sentence frames impact children’s understanding that novel words
refer to emotions, although this effect was only marginal. Children
were more likely to choose emotions in the most restrictive sentence frame (feels about) as compared with the feels condition.
Children were equally likely to choose emotions in is and feels
about conditions, but recall that is [emotion adjective] is also a
grammatically correct choice.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
8
SHABLACK, BECKER, AND LINDQUIST
Figure 2. Study 1 estimated marginal means of each image choice. Bars represent 95% confidence intervals.
See the online article for the color version of this figure.
We note that we did not find the predicted three-way interaction
between age, sentence frame, and image choice. This null finding
could in part be a result of the conservative study design we used,
in which children received no other context but the linguistic
context. If younger children cannot use sentence frames effectively
and thus choose randomly, whereas older children are reasonably
concluding that is, feels, and feels about are each grammatically
appropriate for a third of the trials, then it may have been difficult
to observe a three-way interaction. This interaction was perhaps
even more difficult to observe in light of all children’s bias toward
choosing physical state images.
There are two potential interpretations of children’s bias toward
physical state images. First, a bias toward choosing the physical
state (as opposed to the emotion) may merely reflect the relative
salience of these images to children, as physical state images in our
study often contained other potentially interesting details such as
areas of differently colored skin, bandages, and so forth, which
were necessary to convey meanings such as cold, hot, hurt, and
sick. This bias might have especially occurred in Study 1 because
there was no other situational information to drive children’s
attention to the other images. Without situational information, the
older children may have used the linguistic cue of the novel word
ending in -y (suggestive of an adjective, rather than the verbindicating -ing) to narrow their choices down to the physical state
and emotion images, and then fixated on the physical state images
due to their additional intrigue. Indeed a few children did note that
they chose images based on color (i.e., “that one because it’s blue”
and “I chose that one cause it’s blue and blue is my favorite
color”).
This interpretation is consistent with the fact that children even
showed a bias toward physical state images in the filler condition
(see online supplemental materials). We included filler items because it is common to include filler items to break up the experimental trials on this type of task (Gerken & Shady, 1996). Note
that accounting for children’s performance on the filler trials did
not alter our results, meaning that children’s ability to perform on
Figure 3. Study 1 estimated marginal means of each image choice by age
in years. Bars represent 95% confidence intervals. See the online article for
the color version of this figure.
Figure 4. Study 1 estimated marginal means of each image choice by
sentence frame. Bars represent 95% confidence intervals. See the online
article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
CHILDREN’S EMOTION ACQUISITION
filler trials did not affect our findings. Nonetheless, we cannot be
sure why children chose the physical state images on the filler
trials. One possibility is that in the absence of any situational
context, children inferred that the images had multiple meanings.
That is, an image of a physical state in which an alien was sick
could have been interpreted as also “disgusted” or “sad” by children, which would have been technically correct. Our validation
data are somewhat consistent with this interpretation; although
children performed greater than chance on all stimuli, the stimuli
that they performed the least accurately on were the physical states
of “sick” and “hurt.” In both cases, children were most likely to
misinterpret these states as emotional (e.g., disgusted, sad). This
fact is not likely to just be a limitation of our stimuli. Mental states
have multiple levels of meaning and individuals differ in the
complexity of the inferences drawn about those states—a person
who has tears coming from their eyes and is emitting sound from
their mouth could be “crying,” “sad,” or “experiencing grief at a
loss” (Vallacher & Wegner, 1987, 1989; Wegner & Vallacher,
1987). Even highly caricatured facial expressions of emotion are
perceived as indicative of mental states in some cultures (e.g., sad)
but as actions (e.g., crying) in others (Gendron, Roberson, van der
Vyver, & Barrett, 2014).
Nonetheless, observations of children’s spontaneous comments,
in combination with our validation data, suggests that on average
children understood the intended meaning of the alien images.
Children were more accurate than chance at identifying the correct
emotion, physical state, and action image in our validation study
(see online supplemental materials). Additionally, children’s spontaneous utterances during our experimental trials suggested that
they understood the meaning of the stimuli. For instance, for action
images children stated “this is running,” “[s/he is] falling down,”
“the one who is walking is moky,” or imitated running. For
physical state images children proclaimed things such as “this
one’s hot,” “sweaty!,” “he has two Band Aids, he hurts,” “maybe
he broke his leg,” and “he’s ouchy.” Finally, for emotion images
children made statements such as “this means mad,” “he’s
scared!,” and “that one is sad.”
Second, it is possible that children had a bias toward the physical state images because the feels versus feels about sentence
frame was not sufficiently helpful across the age range to help
children move beyond their bias toward physical state images. This
interpretation is consistent with our observation of an interaction
between image choice and sentence frame, but lack of a three-way
interaction between sentence frame, age, and image choice. The
marginal interaction between image choice and sentence frame
suggests that children may be able to use the sentence frame to
begin to home in on the correct word meaning, supporting prior
research (Becker, 2015; Gleitman et al., 2005; Papafragou et al.,
2007). However, the fact that older children were not significantly
more likely to choose the emotion image over physical state
images in the feels about condition suggests that even 4- and
5-year-olds are not yet able to use feels about to restrict an
adjective meaning to an emotion rather than a physical state.
Taken together, the findings of Study 1 suggest that although
linguistic cues may be meaningful to children in this age range,
children may not be able to use sentence-level cues alone to
understand emotion adjectives. These findings are consistent with
recent work showing that the same sentence-level cues children are
known to exploit for learning verbs may not be usable for adjec-
9
tives at this stage of development (Booth & Waxman, 2009; Syrett,
Latourrette, Ferguson, & Waxman, 2018). However, as we note,
Study 1 is a particularly conservative test of the syntactic bootstrapping hypothesis, as it is atypical for children to experience
emotion concepts and their words in the absence of an emotional
situation. We thus included situational context in Study 2 to more
clearly test the role of sentence frame in children’s understanding
of novel emotion words when an emotional situation was also
present.
Study 2
Because research on emotion concept development demonstrates that children rely on situational context—including the
causes and consequences of an emotion when determining the
meaning of an emotional facial expression (Kayyal et al., 2015;
Widen & Russell, 2010, 2011)—we modified Study 1 to include
information about the situational context a novel adjective was
occurring in. Study 1 was a conservative test of the syntactic
bootstrapping hypothesis, but we reasoned that including situational context would ultimately be a more ecologically valid test of
the types of contexts in which children learn about emotions (e.g.,
when observing an event and hearing an adult talk about it).
To achieve this end, children in Study 2 viewed videos of a
cartoon alien experiencing emotional situations. We predicted that
with this additional situational context present, we would observe
a three-way interaction such that older children would be more
likely to choose emotion images when the sentence frame was
maximally restrictive. As the procedure is largely identical to
Study 1, we only describe the differences between them below.
Method
Participants. One hundred forty-seven children participated
in the study at the Museum of Life and Science in Durham, North
Carolina. As in Study 1 children who did not meet our a priori
inclusion criteria were excluded from further analysis. Nine were
not within our desired age range (had mistakenly been enrolled or
allowed to participate because a sibling had done so). Seven were
removed from analysis because they lived in households where
languages other than English were spoken greater than 50% of the
time, and one was later found to have a learning disability as
reported by their parent/legal guardian in the parent questionnaires.
Eighteen children did not complete the video task either due to
screening trials failure (n ⫽ 9) or opting not to continue in the
midst of the task (n ⫽ 9). The final sample consisted of 113
(Mage ⫽ 4.04, SDage ⫽ .77; 45 females) children: 31 3-year-olds
(eight female), 46 4-year-olds (26 female), and 36 5-year-olds (11
female).
Parents and/or legal guardians completed a voluntary packet of
questionnaires identical to that in Study 1. Twelve individuals did
not report family income; of the 101 who did, 20.8% reported a
household income ⬍$70,000; 28.7% between $70,000 and
$100,000; 26.7% between $100,000 and $150,000; and 23.8%
reported $150,000 or higher. Sixteen parents/guardians did not fill
out the race/ethnicity information for one or both parents. Of the
97 who did, 82.5% of the children were Caucasian/White, 14.4%
10
SHABLACK, BECKER, AND LINDQUIST
Table 5
Sample Stories for Each Emotion Type of Study 2 and Novel Words
Emotion
Happy
Sad
Mad
Surprised
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Disgusted
Afraid
Excited
Story
Palooza ran a race at school. She was the fastest alien in the race, so she won first place! Now, Palooza [sentence frame]
[novel word]. What do you think [novel word] means?
Xylobean’s best friend moved away to a different planet, and they won’t see each other again. Now, Xylobean [sentence frame]
[novel word]. What do you think [novel word] means?
Chromia was reading her favorite book. Then, another alien took it and tore out a page! Now, Chromia [sentence frame] [novel
word]. What do you think [novel word] means?
One day Wazu came home and all his furniture was turned upside-down. He just stared at his furniture and couldn’t figure out
how that happened. Now, Wazu [sentence frame] [novel word]. What do you think [novel word] means?
Palooza took a bite of an apple. As soon as she bit into it, she realized it was rotten inside. She didn’t want to eat the rest of it.
She threw it in the trash. Now, Palooza [sentence frame] [novel word]. What do you think [novel word] means?
Wazu heard a loud crashing noise in the distance. Then, the sound started getting closer and closer! Now, Wazu [sentence
frame] [novel word]. What do you think [novel word] means?
Chromia always wanted to fly in a spaceship. Now she was going to get a chance to do it! Now, Chromia [sentence frame]
[novel word]. What do you think [novel word] means?
multiracial, 1.03% African American, 1.03% Asian, and 1.03%
Hispanic.
Materials.
Novel word videos. Children viewed videos of an alien cartoon
character with a narrator describing a short story (see Table 5 for
sample stories). Videos were created using GoAnimate (http://www
.goanimate.com) and were on average 23.85 s. A single individual
narrated the story for all videos. The screening video scripts were
similar to Study 1, but now had a single narrator, a name associated
with the alien, and a cartoon accompanying the auditory stimuli.
Children who failed two or more of the three screening trials were
thanked for their time and told that the game was over.
During the experimental trials, children watched seven cartoon
videos with accompanying narration: Each video employed a
different novel word (daxy, moky, reksy, binty, gorpy, joomy,
tropy) in one of three between-subjects experimental sentence
frames (is, feels, feels about) as in Study 1. Each alien character in
the video had a neutral expression. Aliens, video stories, and novel
words were fully randomized such that no single alien, story, or
word were consistently paired together. In contrast to Study 1, the
novel word was presented twice (rather than four times) in the
video and a third time (rather than fifth time) in the picture
pointing task instructions (see Table 5 for sample stories and
Figure 1 for a sample trial). Emotional situations were created by
providing a brief story that highlighted a positive or negative
discrete emotional experience for the alien (the alien was happy,
sad, afraid, mad, excited, disgusted, or surprised), prior to the
introduction of the novel word in the target sentence frame. Stories
were kept short to limit cognitive burden and to keep the child’s
attention. Stories were developed based on prior work (see Widen
& Russell, 2010) drawing from the prototypical causes of emotions
in a North American setting. A total of seven experimental trials of
a single alien identity were presented. We reasoned that variation
across videos (i.e., seeing aliens and stories) would be sufficiently
attention-capturing for children so we did not include filler trials to
maintain their attention. All sessions began with the happy story
followed by the remaining six emotions in randomized order. We
modeled this method after other developmental research on emotion (Russell & Widen, 2002a, 2002b; Widen & Russell, 2010). In
these studies, happy trials are presented first because happiness is
a well-understood emotion concept for 3- to 5-year-olds; it is
assumed that receiving a more difficult emotion concept first
might discourage children.
Picture pointing task. Cartoon alien images were identical to
Study 1 (see Table 3), however, each matched the alien seen in the
video, where s/he was given an identifying name (screening trials:
Chrysanthemum, Magenta, Frebedo; experimental trials: Palooza,
Chromia, Wazu, Xylobean). Female participants viewed either
Palooza or Chromia (the “female” aliens) and male participants
viewed either Wazu or Xylobean (the “male” aliens). We matched
the gender of the aliens and children because there is some evidence that interpersonal similarity facilitates mental state inference
(Ames, 2004). All children saw Chrysanthemum, Magenta, and
Frebedo for the screening trials. During the picture pointing task,
images of the aliens displaying a particular emotion, action, and
physical state image appeared only once throughout the seven
trials such that no image was repeated (e.g., children saw a happy
alien on only one trial throughout the experiment). The emotional
image depicted always matched the story (e.g., children saw a
happy alien for a story describing happiness) but the particular
action and physical state seen were randomly displayed.
Procedure. This study was approved by the University of North
Carolina Institutional Review Board. The procedure was identical to
Study 1, save that participants saw videos of the alien scenarios, rather
than videos of puppets holding short conversations.
Results
Analysis procedures are identical to that in Study 1 with mean
proportions of each image choice type calculated across trial types
(e.g., mean proportion of emotion images chosen within each
experimental condition)8 entered in mixed model ANOVAs. To
examine our main hypothesis, a 3 (image choice: emotion, physical state, action) ⫻ 3 (sentence frame: is, feels, feels about) ⫻ 3
(age: 3, 4, 5) mixed model ANOVA was conducted with image
choice as a within subjects factor and sentence frame and age as
between subjects factors. As in Study 1, we also computed a 3
(image choice: emotion, physical state, action) ⫻ 3 (sentence
8
One child experienced technical difficulties during one trial. As in
Study 1, their mean proportions were adjusted to be out of the number of
completed trials.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
CHILDREN’S EMOTION ACQUISITION
frame: is, feels, feels about) ⫻ 3 (age: 3, 4, 5) mixed model
ANCOVA with Image choice as a within subjects factor and
sentence frame and age as between subjects factors, including
participant gender and performance on the screening trials as
covariates. Findings were largely identical across the two analyses
so we report the more conservative ANCOVA findings. See online
supplemental materials for the ANOVA and multilevel multinomial logistic regression analyses.
See Table 6 for all effects; we discuss only the significant
effects and predicted effects here. As predicted, and as in Study 1,
we found a significant interaction between image choice and age,
F(3.73, 190.38) ⫽ 8.29, p ⬍ .001, 2 ⫽ .14.9 Emotion images
were chosen more as age increases, and both physical state and
action images were chosen less as age increases. Simple effects
reveal significant age differences in the proportion of trials in
which emotion images were chosen, F(2, 102) ⫽ 12.63, p ⬍ .001,
2 ⫽ .20, action images were chosen, F(2, 102) ⫽ 5.65, p ⫽ .005,
2 ⫽ .10, and physical state images are chosen, F(2, 102) ⫽ 4.59,
p ⫽ .01, 2 ⫽ .08. Pairwise comparisons reveal that 5-year-olds
chose emotion images significantly more than 4- (p ⫽ .01) and 3year-olds (…
Purchase answer to see full
attachment
WE’VE HAD A GOOD SUCCESS RATE ON THIS ASSIGNMENT. PLACE THIS ORDER OR A SIMILAR ORDER AND GET AN AMAZING DISCOUNT
Source link