Conference PaperPDF Available

Environmental Regularities Shape Semantic Organization throughout Development

Authors:

Abstract and Figures

Our knowledge of the world is an organized lexico-semantic network in which concepts can be linked by relations, such as "taxonomic" relations between members of the same stable category (e.g., cat and sheep), or association between entities that occur together or in the same context (e.g., sock and foot). Prior research has focused on the emergence of knowledge about taxonomic relations, whereas association has received little attention. The goal of the present research was to investigate how semantic organization development is shaped by both taxonomic relatedness and associations based on co-occurrence between labels for concepts in language. Using a Cued Recall paradigm, we found a substantial influence of co-occurrence in both 4-5-year-olds and adults, whereas taxonomic relatedness only influenced adults. These results demonstrate a critical and persistent influence of co-occurrence associations on semantic organization. We discuss these findings in relation to theories of semantic development.
Content may be subject to copyright.
Environmental Regularities Shape Semantic Organization throughout Development
Abstract
Our knowledge of the world is an organized lexico-semantic
network in which concepts can be linked by relations, such as
“taxonomic” relations between members of the same stable
category (e.g., cat and sheep), or association between entities
that occur together or in the same context (e.g., sock and
foot). Prior research has focused on the emergence of
knowledge about taxonomic relations, whereas association
has received little attention. The goal of the present research
was to investigate how semantic organization development is
shaped by both taxonomic relatedness and associations based
on co-occurrence between labels for concepts in language.
Using a Cued Recall paradigm, we found a substantial
influence of co-occurrence in both 4-5-year-olds and adults,
whereas taxonomic relatedness only influenced adults. These
results demonstrate a critical and persistent influence of co-
occurrence associations on semantic organization. We discuss
these findings in relation to theories of semantic development.
Keywords: semantic development; semantic organization;
categories
Introduction
Our knowledge about the world is fundamental to many of
the cognitive feats we accomplish on an everyday basis,
including applying what we know to new situations,
retrieving knowledge from memory, and incorporating new
information into existing knowledge (Bower, Clark,
Lesgold, & Winzenz, 1969; Heit, 2000; Tse, Langston,
Kakeyama et al., 2007). These feats are possible due to the
organization of our knowledge into an interconnected
lexico-semantic network of related concepts (Cree &
Armstrong, 2012; McClelland & Rogers, 2003). For
example, our knowledge of dogs is often connected to our
knowledge of other similar animals (e.g., cats), as well as to
our knowledge about the contexts in which dogs appear,
such as with leashes and doghouses.
Although the fact that our concepts are organized is
hardly controversial (e.g., McClelland & Rogers, 2003), the
processes that drive the development of semantic
organization are a topic of considerable debate. To date, this
debate has focused on how connections between concepts
from the same stable, “taxonomic” category (e.g., animals,
foods) are formed, in spite of the fact that they may be
difficult to observe: Members of the same (especially
superordinate) taxonomic category do not necessarily look
similar, or occur together. Some have proposed that
semantic development begins with easy to observe relations
that are then used to bootstrap taxonomic knowledge
(Lucariello, Kyratzis, & Nelson, 1992). Alternately, others
have proposed that we are endowed with early-emerging
biases towards learning taxonomic relations (e.g., Gelman &
Markman, 1986).
The goal of this research is to investigate another
possibility: That easy to observe relations specifically, co-
occurrence play a fundamental role in shaping knowledge
organization from early in development through adulthood.
In this paper, we first review traditional theoretical accounts
that have focused on taxonomic relations, then highlight key
findings suggestive of a role for co-occurrence that these
accounts fail to capture, and an alternate perspective that we
test in the present experiment.
Traditional Accounts of Semantic Development
Most extant accounts of the development of semantic
organization have focused on how semantic knowledge
becomes organized according to membership in taxonomic
categories, such as foods. According to some accounts,
referred to here as restructuring accounts, taxonomic
relations are the endpoint of development. Critical to these
accounts is the idea that the order in which relations
between concepts are acquired is dictated by how
observable they are. For example, it is easy to observe that
cups have the same shape, or reliably co-occur with juice or
milk, whereas membership in the same superordinate
taxonomic category is more difficult (if not impossible) to
observe. Restructuring accounts propose that early
organization is shaped by information readily available in
the environment, and that taxonomic knowledge comes to
replace this (more rudimentary) organization.
An early restructuring account was proposed by Inhelder
and Piaget (1964), in which the transition to taxonomic
organization is driven by experiences that highlight the
inadequacy of earlier modes of organization (although the
mechanisms by which this transition occurs are not clear).
Another, more specified restructuring account is Nelson and
Lucariello’s (1992) slot-filler account, which highlights
environmental input in which some members of the same
taxonomic category play the same role in the same context,
such as some members of the taxonomic category of foods
(e.g., eggs and bacon) reliably being eaten in a breakfast
context. According to this account, young children are
sensitive to these regularities, such that semantic knowledge
is first organized into contextually-constrained taxonomic
groups, which are gradually integrated together as children
recognize when entities play the same role in different
contexts (e.g., foods being eaten in different meal contexts).
According to another set of accounts, referred to here as
taxonomic bias accounts, taxonomic relations predominate
semantic organization from early in development due to
early-emerging (possibly innate) biases towards learning
which entities are members of the same taxonomic category.
These biases include beliefs that entities in the world belong
to taxonomic categories, and that labels are indicative of
category membership (e.g., Gelman & Coley, 1990). A role
for other types of environmental input, such as the regularity
with which entities co-occur, is not specified.
A final type of account reviewed here, which we refer to
as featural learning, posits that the development of semantic
organization is driven by detecting clusters of features
1156
whose appearance in entities is reliably correlated, and
which are often associated with taxonomic category
membership (Rosch, 1975). For example, membership in
the category of birds is associated with possessing wings,
feathers, and a beak. Featural learning accounts propose that
sensitivity to these correlations yields taxonomic
organization (e.g., McClelland & Rogers, 2003). In contrast
with taxonomic bias accounts, featural learning accounts
argue in favor of the gradual emergence of taxonomic
organization over the course of development. However,
featural learning accounts do not consider spatial or
temporal co-occurrence of items in the world (or language)
as contributors to semantic organization.
Environmental Regularities Overlooked by
Traditional Theoretical Accounts
Of the influential accounts reviewed in the previous section,
only some restructuring accounts posit any role in semantic
development for environmental regularities with which
entities and their labels co-occur. Even in these accounts,
these regularities are ultimately overwritten. However,
several findings highlight a potential importance of co-
occurrence regularities throughout development.
First, statistical learning studies suggest that sensitivity to
the regularity with which different entities co-occur is
apparent from very early in development (Bulf, Johnson, &
Valenza, 2011). Moreover, numerous findings attest to the
influence on children’s reasoning of semantic relations that
may be derived from co-occurrence, such as schematic
relations between entities that occur in the same context
(e.g., cow and barn) and thematic relations between entities
that play complementary roles (e.g., nail and hammer)
(Blaye, Bernard-Peyron, Paour, & Bonthoux, 2006; Fenson,
Vella, & Kennedy, 1989; Lucariello et al., 1992; Walsh,
Richardson, & Faulkner, 1993). Additionally, a handful of
studies conducted by Fisher, Godwin and Matlen (Fisher,
Matlen, & Godwin, 2011; Matlen, Fisher, & Godwin, 2015)
point more directly towards an influence of co-occurrence
on childrens semantic reasoning. In these studies,
participants were asked to infer whether a property (e.g.,
has blicket inside) attributed to a target (e.g., glove) was
shared by either a strongly taxonomically related item (e.g.,
mitten) or a more weakly taxonomically related item (e.g.,
sweater). These studies revealed that four year old children
only reliably chose the strongly taxonomically related item
when its label co-occurred with the target either in corpora
of childrens speech input (e.g., bunny-rabbit, Fisher et al.,
2011) or an empirically manipulated speech stream (Matlen
et al., 2015). These findings suggest that accounts of
semantic development that do not posit any role for co-
occurrence are at best incomplete.
Second, a handful of findings suggest that semantic
relations that may be derived from co-occurrence continue
to shape semantic organization into adulthood. For example,
Lin and Murphy (2001) found that relations between entities
that adult raters judged as associated in scenes or events
(which likely co-occur) had a pervasive influence on adults’
categorization and reasoning that was frequently greater
than the influence of taxonomic relations. This evidence is
inconsistent with restructuring accounts, in which an early
influence of co-occurrence is eventually overwritten.
Finally, the potential contributions of co-occurrence
regularities are highlighted by a mechanistic account and
corroborating behavioral evidence presented by Sloutsky,
Yim, Yao, and Dennis (2017). According to this account,
exposure to co-occurrence regularities in language fosters
both the learning of associations between concepts whose
labels directly co-occur in sentences (e.g., fork and
spaghetti), and between taxonomically related concepts
whose labels share patterns of co-occurrence (e.g., spaghetti
and pie). However, whereas co-occurrence in a sentence can
be directly gleaned from input and therefore rapidly learned,
shared patterns of co-occurrence that often link members of
the same taxonomic category are learned more slowly
because they can only be derived from multiple instances of
direct co-occurrence. This account predicts both that (1)
direct co-occurrence should contribute to semantic
organization throughout development, and (2) the
contributions of direct co-occurrence to semantic
organization should be evident earlier in development than
the contributions of taxonomic relatedness. Initial evidence
for this account comes from a series of experiments
presented in Sloutsky et al. (2017) in which children and
adults were asked to infer the category membership of a
novel word (e.g., whether it was an animal or a machine)
that was presented within a list of familiar words. Both
children and adults readily inferred the category
membership of the novel word when it appeared in a list of
words that are associated (and therefore likely to co-occur)
with the same category. For example, participants inferred
that the novel word referred to an animal when it appeared
in a list of words including “furry” and “zoo”. However,
only adults inferred this meaning when the novel word
appeared in a list of words referring to members of the
category, such as “lion” and “bunny”.
Together, these prior findings suggest that co-occurrence
regularities may shape semantic development. However, in
addition to being overlooked in traditional theoretical
accounts of the development of semantic organization, this
possibility has received only limited empirical investigation
to date, and the way in which it has been investigated has
not been designed to assess relational knowledge for items
that actually co-occur in the environment. Critically, this
research has instead investigated knowledge for relations
between items either judged by researchers or participants
as co-occurring according to researcher-specified criteria, or
produced in free association tasks. Neither ratings nor free
associations are inputs from the environment from which
semantic relations can be learned: They are outcomes of
relations already learned and present in semantic knowledge
(Hofmann, Biemann, Westbury et al., 2018). A more direct
investigation of the role of co-occurrence in shaping
semantic development could be accomplished by assessing
1157
the contributions of co-occurrence regularities present in
actual environmental input.
Current Study
The overall purpose of the current study was to investigate
the contributions of co-occurrence regularities and
taxonomic relatedness to the organization of lexico-
semantic knowledge from early childhood to adulthood.
This investigation was designed to arbitrate between
competing theoretical accounts of the development of
knowledge organization. Specifically, restructuring accounts
predict that co-occurrence should contribute to knowledge
organization in childhood, but be replaced by taxonomic
relations in adulthood. Both taxonomic bias and featural
learning accounts are agnostic about the contributions of co-
occurrence, but whereas the former predict that taxonomic
relations should contribute from childhood through to
adulthood, the latter predict that the contributions of
taxonomic relations should substantially increase with age.
A different developmental pattern is predicted by recent
proposals that highlight a key role throughout development
for co-occurrence in which it both directly fosters relations
between concepts, and indirectly fosters relations between
concepts that share patterns of co-occurrence and are often
taxonomically related (e.g., Sloutsky et al., 2017).
Specifically, such proposals predict that the contributions of
co-occurrence should be evident in both children and adults,
whereas contributions of taxonomic relatedness should be
evident only later in development.
We accomplished this investigation by measuring the
degree to which familiar concepts were related in young
children (4-year-olds) and adults’ semantic knowledge when
either the concepts’ labels reliably co-occur in linguistic
input, or when they are members of the same taxonomic
category. To target actual experienced co-occurrence, we
identified pairs of words familiar to young children that co-
occurred more reliably with each other than with other
words in corpora of child-directed speech.
To measure the contributions of co-occurrence and
taxonomic relations to children and adults lexico-semantic
knowledge, we used a Cued Recall paradigm to measure the
effects of co-occurrence and taxonomic relatedness on
memory retrieval. We selected this paradigm for two
reasons. First, the sensitivity of this task to semantic
relatedness is attested by numerous findings that semantic
relatedness influences the accuracy with which people
(including children) recall word pairs and lists (Bjorklund &
Jacobs, 1985; Blewitt & Toppino, 1991). Second, this task
facilitates a comparison between children and adults
because it measures contributions to lexico-semantic
knowledge without requiring participants to reason about
relations, which adults may more easily.
Method
Participants
The sample included 30 4-5 year old children (Mage=4.50
years, SD=1.62 years), and 29 Adults (Mage=20.16 years,
SD=3.66 years). The child age group was selected because
the 4-5 year period is one during which the nature of
relations that organize lexico-semantic knowledge has been
the subject of active debate (Lucariello et al., 1992; Nguyen
& Murphy, 2003; Waxman & Namy, 1997). Children were
recruited from families, daycares, and preschools in a
metropolitan area in a Midwestern US city. Adults were
undergraduates from a public university in the same city and
participated in exchange for partial course credit.
Stimuli and Design
The primary stimuli used in this experiment were word pairs
that belonged to one of three Semantic Relatedness
conditions: Co-Occur (pairs that reliably co-occurred with
each other more often than with other words in child speech
input), Taxonomic (words close in meaning from the same
taxonomic category) or Unrelated. (words that neither
reliably co-occur nor are similar in meaning).
Co-Occurrence Criteria. The first step taken to select
pairs in each condition was to identify a set of words for
which lexical norms collected using the MacArthur-Bates
Communicative Development Inventory (MB-CDI) were
available from WordBank (an open database of children's
vocabulary development, Frank, Braginsky, Yurovsky, &
Marchman, 2016), and measure their rates of co-occurrence
in 25 child speech input corpora from the CHILDES
database (MacWhinney, 2000). To reduce the computational
expense of measuring word co-occurrence rates, some
classes of words that would a priori not be used as stimuli
were removed, such as sounds (e.g., “moo”), leaving a list
of 538 words. Additionally, to ensure that co-occurrences
were measured from speech input, CHILDES corpora were
pre-processed to remove speech produced by children. Co-
occurrences between these words were then calculated by
taking all possible pairs of words in this set, and calculating
how frequently they co-occurred with each other within a 7-
word window across 25 CHILDES corpora. Finally, to
account for the fact that more frequent words co-occur with
other words simply by chance, t-scores (Evert, 2008) were
calculated for each word pair using the formula below based
on their measured co-occurrence frequencies (O), adjusted
for the frequency of co-occurrence expected by chance
based on their respective frequencies across the corpora and
the size of the corpora (E):
Table 1: Pairs of words used in the Co-Occur, Taxonomic,
and Unrelated conditions
Co-Occur
Taxonomic
Unrelated
bottle
baby
ball
crayon
frog
foot
shoe
pig
towel
bread
brush
hair
horse
blocks
cereal
cup
juice
carrot
balloon
tree
cheese
mouse
fork
sheep
pancake
car
street
popcorn
pizza
lion
soup
spoon
airplane
fish
bed
milk
cow
sock
duck
swing
1158
Word pairs for use in the Co-Occur condition were then
selected as pairs of nouns with t-scores > 2.5 (following
Baayen, Davidson, & Bates, 2008) in which, according to
lexical norms accessed from WordBank, both words were
produced by >80% of 36-month-old children (one year
younger than children in our sample).
Taxonomic Criteria. Taxonomic relatedness was
determined based on both the membership of concepts in
the same taxonomic category (e.g., clothing, foods, animals)
and similarity in meaning between their labels. Similarity in
meaning was measured as similarity between the definitions
of candidate words from WordNet (a database of word
definitions composed by lexicographers). This measure
captures the essence of taxonomic relatedness i.e., close
similarity in meaning without relying on participant
judgments that may be influenced by non-taxonomic
relations (Wisniewski & Bassok, 1999). In WordNet, nouns
are first grouped into sets of synonyms, which are in turn
linked into a hierarchy according to “IS A” and part-whole
relations. Similarity in meaning between word pairs was
measured using Resnik similarity, i.e., the information
content (specificity) of the word lowest in the WordNet
hierarchy within which the pair of words is subsumed. For
example, dog and cat are subsumed within carnivore,
whereas dog and kangaroo are subsumed within mammal;
because the information content of carnivore is greater than
the information content of mammal, Resnik similarity is
higher between dog and cat versus dog and kangaroo.
Candidate Taxonomic pairs nouns with Resnik
similarities of > 5 and t-scores < 1.5 in which both were
produced by at least 80% of 36-month-old children
according to WordBank norms. The rationale of the Resnik
similarity criterion of > 5 is illustrated in Fig. 1, which
shows that this value distinguished between same- vs.
different-category items.
Unrelated Criteria. Candidate Unrelated word pairs
were noun pairs that met the WordBank production norm
criterion with t-scores and Resnik similarities of < 1.5.
Composition of Full Set. From the sets of candidate
pairs, eight pairs were selected for each of the Relation
conditions (Co-Occur, Taxonomic, and Unrelated) such
that: 1) The mean percentage of 36-month-olds who
produced the words in the pairs according to Wordbank
norms was equated across conditions, and 2) No words
appeared in more than one condition (Table 1). An
additional 4 nouns that met the WordBank production norm
criterion were selected to construct pairs used for
demonstration and practice (see Procedure below). All
words were recorded by both a male and a female speaker
using an engaging, child-friendly intonation.
The eight pairs in each Relation condition were divided
into two Stimulus Sets, each with four pairs in each
condition, because pilot testing indicated that 12 pairs was
the maximum number that could be presented to children
without producing floor effects. Within each Stimulus Set,
each word in a pair was randomly assigned to be either the
Cue or Target. In the experiment, Cue words were presented
using the male speaker’s voice, and Targets using the
female’s voice. Additionally, the 12 word pairs were
pseudorandomized into three blocks, such that each block
contained 1-2 pairs from each condition. The order of these
blocks was counterbalanced across participants.
Procedure. Adult participants were tested in a quiet
space in the lab, and children were tested either in a quiet
space in the lab, or at their preschool or daycare. The
procedure was identical for adults and children (including
the auditory presentation of the same recorded Cue-Target
pairs), with the exceptions that: 1) Instructions were
conveyed by an experimenter for children, and as text on a
computer screen for adults, and 2) Children made verbal
responses recorded by the experimenter, whereas adults
typed responses.
To start, participants were informed that they were
going to play a game with two sock puppets depicted on the
computer, Izzy and Ozzy, in which Izzy and Ozzy would
say pairs of words. The two demonstration/practice
unrelated Cue-Target spoken word pairs were then played,
while animations depicted one puppet “saying” the Cue
word, and the other saying the Target word. Participants
Figure 1: Graphs depicting Resnik similarity between one item from a Taxonomic pair and: (1) The other item from the
pair (highlighted), (2) Other items from the same taxonomic category, and (3) Items from other categories.
1159
then completed two practice rounds with the same Cue-
Target pairs consisting of a Study Phase, in which
participants were instructed to remember the words that
went together in pairs, and a Test phase, in which only the
Cue in each pair was presented and participants were
prompted to either say or type the Target that had been
spoken by Ozzy. Participants received corrective feedback
after each practice trial, and completed up to three practice
rounds until they either responded with the correct Target
for both Cues within around, or the experiment was
terminated.
Participants then proceeded to complete the three
blocks of Cue-Target pairs in the Stimulus Set to which they
had been randomly assigned. Each block followed the same
Study and Test phase format as the practice rounds, with the
exception that participants did not receive feedback.
Results
The primary outcome measure of interest for this study was
the accuracy with which participants recalled Target words
paired with Cues in each of the three Relation conditions:
Co-Occurrence, Taxonomic, and Unrelated
1
. Responses
were scored as accurate when participants made responses
identical to the Target, morphological variants of the Target
(e.g., “spoons” instead of “spoon”), or close synonyms to
the Target (e.g., “road” instead of “street”).
1
We also analyzed participants errors to test the frequency with
which the incorrect responses participants in each age group
produced either co-occurred with or were taxonomically related to
the Cue. However, these analyses did not contribute meaningfully
to our results. The majority of incorrect responses in both age
groups were other words from the set of word pairs the participant
heard (64% in children, 82% in adults). Of these responses, only a
small minority (7-14%) were either co-occurring with or
taxonomically related to the Cue, which was likely the result of the
random chance with which some words from the list, when
randomly recombined with Cues, happen to be related to them in
some way. Of responses not drawn from the list of word pairs, the
only detectable pattern was a tendency for children to respond with
incorrect words that co-occurred with the Cue (52%) more often
than words that were taxonomically related to the Cue (6%). This
pattern mirrors the results of analyses of childrens accuracy.
All analyses were conducted in the R environment.
Mixed effects models were generated using the lme4 (Bates,
Maechler, Bolker, & Walker, 2015) package, and
corresponding 2 or F-statistics for main effects and
interactions were generated using the car package (Fox &
Weisberg, 2011).
Preliminary Analyses: Stimulus Set Comparison
We first tested whether any effect of condition varied across
the two Stimulus Sets in children and adults. For data from
each age group, we generated a binomial generalized linear
mixed effects model with Accuracy (0 or 1) as the outcome
variable, Relation condition (Co-Occurrence, Taxonomic,
and Unrelated) and Stimulus Set (1 vs. 2) as fixed effects,
and participant and item as random effects. This analysis
revealed no significant interaction between Relation
condition and Stimulus Set (ps > .23). For all subsequent
analyses, we therefore collapsed across Stimulus Sets.
Primary Analyses
Accuracy by age and condition is presented in Figure 2. To
test the relative influences of Relatedness conditions (Co-
Occurrence, Taxonomic, and Unrelated) on accuracy, we
generated an omnibus binomial generalized linear mixed
effects model with Accuracy (0 or 1) as the outcome
variable, Relatedness condition and Age group (children and
adults) as fixed effects, and participant and item as random
effects. This analysis yielded main effects of Relatedness
condition (2(2)=25.26, p<.001) and Age group
(2(1)=10.36, p=.001) that were qualified by an interaction
(2(2)=7.87, p=.02).
To investigate the interaction between Relatedness
condition and Age group, we conducted two sets of
analyses: A first set in which we compared the effects of the
different Relatedness conditions in each Age group, and a
second set in which we compared the effects of each
Relatedness condition in children versus adults.
Relation Conditions in Each Age Group. In these
analyses, we generated for each age group a binomial
generalized linear mixed effects model with Accuracy as the
outcome variable, Relatedness condition as a fixed effect,
and participant and item as random effects. These models
Figure 2: Accuracy in children (left) and adults (right) in the Relation Conditions. Error bars represent standard errors.
1160
revealed significant effects of Relatedness condition in each
age group (ps < .001) (Figure 3). To conduct pairwise
comparisons of the Relatedness conditions in each age
group, we re-generated the model for each age with each of
the Relatedness conditions as the reference level, and
applied Bonferroni-adjustments to the resulting p-values. In
children, these analyses revealed significant differences
between the Co-Occurrence (M=0.60, SD=0.49) and both
Unrelated (M=0.25, SD=0.43) and Taxonomic conditions
(M=0.29, SD=0.45) (ps < .001), but no difference between
the Taxonomic and Unrelated conditions (p > .99). In adults,
these analyses revealed a significant difference between the
Co-Occurrence (M=0.71, SD=0.46) and Unrelated
conditions (M=0.34, SD=0.48) (p < .0001), the Taxonomic
(M=0.59, SD=0.49) and Unrelated conditions (p=.033), and
no significant difference between Co-Occurrence and
Taxonomic conditions (p=.237).
Comparison of Children and Adults. To compare the
accuracy of children versus adults in each Relatedness
condition, we generated a binomial generalized linear mixed
effect model for each Relatedness condition, each with Age
Group as a fixed effect, and participant and item as random
effects. Additionally, we applied Bonferroni-adjustments to
all p-values to correct for multiple comparisons. These
analyses revealed only a significant difference between
children and adults in accuracy in the Taxonomic condition
(p<.001). In comparison, there was no significant difference
in accuracy between children and adults in either the Co-
Occur or Unrelated conditions (ps>.2).
General Discussion
The purpose of the present experiment was twofold: (1) To
investigate how semantic development is shaped by co-
occurrence regularities and taxonomic relatedness, and (2)
More broadly, to investigate whether the development of
semantic organization involves the maintenance of early-
emerging taxonomic organization throughout development
(as in taxonomic bias accounts), the restructuring of
semantic organization (as in restructuring accounts), or the
addition of new semantic knowledge that does not replace
earlier-emerging knowledge.
In this experiment, we observed substantial effects of co-
occurrence in both young children and adults. In contrast, an
influence of taxonomic relatedness was only apparent in
adults. Importantly, due to our use of an implicit measure of
semantic knowledge, this developmental pattern is unlikely
to be attributable to developmental improvements in
reasoning. These findings therefore support a key role for
co-occurrence in semantic development, and are consistent
with an overall developmental trajectory in which some
types of semantic knowledge (such as taxonomic) tend to
supplement rather than supplant earlier-emerging
knowledge.
Generalizability of Findings
In order to evaluate the support for a key role for co-
occurrence in lexico-semantic development, it is important
to consider the possibility that the cued recall paradigm used
in this experiment biased the results in favor of this
outcome. Specifically, accurately recalling pairs of words
may better evoke participants prior knowledge of word
pairs that they have experienced occurring together than
their knowledge of taxonomically related words.
However, this possibility is undermined by corroborating
evidence from very different paradigms that do not involve
recalling word pairs. First, as described in the introduction,
findings from studies conducted by Fisher, Godwin, and
Matlen (Fisher et al., 2011; Matlen et al., 2015) have
provided evidence for the contribution of co-occurrence to
semantic reasoning. Specifically, these studies found that
young children only reliably infer that an item shares a
property with another, strongly taxonomically related item
when their labels co-occur (e.g., bunny-rabbit). Moreover,
the pattern of results in adults and children has recently been
replicated using another, very different paradigm in which
the contribution of a given form of relatedness is measured
based on the degree to which it interferes with participants
ability to identify when a picture (e.g., of a baby) does not
depict the same thing as a preceding word (e.g., bottle)
(Unger & Sloutsky, Under Review). Taken together, these
findings suggest a general contribution of co-occurrence to
lexico-semantic knowledge that is not dependent upon the
use of a cued recall-based assessment.
References
Baayen, R. H., Davidson, D. J., & Bates, D. M. (2008).
Mixed-effects modeling with crossed random effects for
subjects and items. Journal of Memory and Language, 59,
390-412.
Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015).
Fitting Linear Mixed-Effects Models Using lme4. Journal
of Statistical Software, 67, 1-48.
Bjorklund, D. F., & Jacobs, J. W. (1985). Associative and
categorical processes in children's memory: The role of
automaticity in the development of organization in free
recall. Journal of Experimental Child Psychology, 39,
599-617.
Blaye, A., Bernard-Peyron, V., Paour, J.-L., & Bonthoux, F.
(2006). Categorical flexibility in children: Distinguishing
response flexibility from conceptual flexibility. European
Journal of Developmental Psychology, 3, 163-188.
Blewitt, P., & Toppino, T. C. (1991). The development of
taxonomic structure in lexical memory. Journal of
Experimental Child Psychology, 51, 296-319.
Bower, G. H., Clark, M. C., Lesgold, A. M., & Winzenz, D.
(1969). Hierarchical retrieval schemes in recall of
categorized word lists. Journal of Verbal Learning and
Verbal Behavior, 8, 323-343.
Bulf, H., Johnson, S. P., & Valenza, E. (2011). Visual
statistical learning in the newborn infant. Cognition, 121,
127-132.
Cree, G. S., & Armstrong, B. C. (2012). Computational
models of semantic memory The Cambridge Handbook of
1161
Psycholinguistics (pp. 259-282). Cambridge: Cambridge
University Press.
Evert, S. (2008). Corpora and collocations. Corpus
linguistics. An international handbook, 2, 1212-1248.
Fenson, L., Vella, D., & Kennedy, M. (1989). Children's
knowledge of thematic and taxonomic relations at two
years of age. Child Development, 60, 911-919.
Fisher, A. V., Matlen, B. J., & Godwin, K. E. (2011).
Semantic similarity of labels and inductive generalization:
Taking a second look. Cognition, 118, 432-438.
Fox, J., & Weisberg, S. (2011). An R Companion to Applied
Regression (Second Edition ed.). Thousand Oaks, CA:
Sage.
Frank, M. C., Braginsky, M., Yurovsky, D., & Marchman,
V. A. (2016). Wordbank: An open repository for
developmental vocabulary data. Journal of Child
Language.
Gelman, S. A., & Coley, J. D. (1990). The importance of
knowing a dodo is a bird: Categories and inferences in 2-
year-old children. Developmental Psychology, 26, 796.
Gelman, S. A., & Markman, E. M. (1986). Categories and
induction in young children. Cognition, 23, 183-209.
Heit, E. (2000). Properties of inductive reasoning.
Psychonomic Bulletin & Review, 7, 569-592.
Hofmann, M. J., Biemann, C., Westbury, C., Murusidze,
M., Conrad, M., & Jacobs, A. M. (2018). Simple
CoOccurrence Statistics Reproducibly Predict
Association Ratings. Cognitive Science, 42, 2287-2312.
Inhelder, B., & Piaget, J. (1964). The early growth of logic
in the child. New York: Norton.
Lin, E. L., & Murphy, G. L. (2001). Thematic relations in
adults' concepts. Journal of Experimental Psychology:
General, 130, 3-28.
Lucariello, J., Kyratzis, A., & Nelson, K. (1992).
Taxonomic knowledge: What kind and when? Child
development, 63, 978-998.
MacWhinney, B. (2000). The CHILDES project: The
database (Vol. 2): Psychology Press.
Matlen, B. J., Fisher, A. V., & Godwin, K. E. (2015). The
influence of label co-occurrence and semantic similarity
on children’s inductive generalization. Frontiers in
Psychology, 6, 1146.
McClelland, J. L., & Rogers, T. T. (2003). The parallel
distributed processing approach to semantic cognition.
Nature Reviews Neuroscience, 4, 310-322.
Nguyen, S. P., & Murphy, G. L. (2003). An Apple is More
Than Just a Fruit: CrossClassification in Children's
Concepts. Child development, 74, 1783-1806.
Rosch, E. (1975). Basic objects in natural categories:
Language Behavior Research Laboratory, University of
California.
Sloutsky, V. M., Yim, H., Yao, X., & Dennis, S. (2017). An
associative account of the development of word learning.
Cognitive Psychology, 97, 1-30.
Tse, D., Langston, R. F., Kakeyama, M., Bethus, I.,
Spooner, P. A., Wood, E. R., . . . Morris, R. G. (2007).
Schemas and memory consolidation. Science, 316, 76-82.
Unger, L., & Sloutsky, V. M. (Under Review).
Environmental Regularities Influence Semantic
Organization throughout Development.
Walsh, M., Richardson, K., & Faulkner, D. (1993).
Perceptual, thematic and taxonomic relations in children’s
mental representations: Responses to triads. European
Journal of Psychology of Education, 8, 85-102.
Waxman, S. R., & Namy, L. L. (1997). Challenging the
notion of a thematic preference in young children.
Developmental Psychology, 33, 555-567.
Wisniewski, E. J., & Bassok, M. (1999). What makes a man
similar to a tie? Stimulus compatibility with comparison
and integration. Cognitive Psychology, 39, 208-238.
1162
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
What determines human ratings of association? We planned this paper as a test for association strength (AS) that is derived from the log likelihood that two words co‐occur significantly more often together in sentences than is expected from their single word frequencies. We also investigated the moderately correlated interactions of word frequency, emotional valence, arousal, and imageability of both words (r's ≤ .3). In three studies, linear mixed effects models revealed that AS and valence reproducibly account for variance in the human ratings. To understand further correlated predictors, we conducted a hierarchical cluster analysis and examined the predictors of four clusters in competitive analyses: Only AS and word2vec skip‐gram cosine distances reproducibly accounted for variance in all three studies. The other predictors of the first cluster (number of common associates, (positive) point‐wise mutual information, and word2vec CBOW cosine) did not reproducibly explain further variance. The same was true for the second cluster (word frequency and arousal); the third cluster (emotional valence and imageability); and the fourth cluster (consisting of joint frequency only). Finally, we discuss emotional valence as an important dimension of semantic space. Our results suggest that a simple definition of syntagmatic word contiguity (AS) and a paradigmatic measure of semantic similarity (skip‐gram cosine) provide the most general performance‐independent explanation of association ratings.
Article
Full-text available
The MacArthur-Bates Communicative Development Inventories (CDIs) are a widely used family of parent-report instruments for easy and inexpensive data-gathering about early language acquisition. CDI data have been used to explore a variety of theoretically important topics, but, with few exceptions, researchers have had to rely on data collected in their own lab. In this paper, we remedy this issue by presenting Wordbank, a structured database of CDI data combined with a browsable web interface. Wordbank archives CDI data across languages and labs, providing a resource for researchers interested in early language, as well as a platform for novel analyses. The site allows interactive exploration of patterns of vocabulary growth at the level of both individual children and particular words. We also introduce wordbankr , a software package for connecting to the database directly. Together, these tools extend the abilities of students and researchers to explore quantitative trends in vocabulary development.
Article
Full-text available
Semantically-similar labels that co-occur in child-directed speech (e.g., bunny-rabbit) are more likely to promote inductive generalization in preschoolers than non-co-occurring labels (e.g., lamb-sheep). However, it remains unclear whether this effect stems from co-occurrence or other factors, and how co-occurrence contributes to generalization. To address these issues, preschoolers were exposed to a stream of semantically-similar labels that don't co-occur in natural language, but were arranged to co-occur in the experimental setting. In Experiment 1, children exposed to the co-occurring stream were more likely to make category-consistent inferences than children in two control conditions. Experiment 2 replicated this effect and provided evidence that co-occurrence training influenced generalization only when the trained labels were categorically-similar. These findings suggest that both co-occurrence information and semantic representations contribute to preschool-age children's inductive generalization. The findings are discussed in relation to the developmental accounts of inductive generalization.
Article
Full-text available
This study examined 2-year-old childrern's ability to make category-based inferences. Subjects were asked a series of questions that they could answer based on category membership, appearances, or both. In one condition, all pictures were named; in a second condition, none were named. Children performed well on prototypical pictures regardless of whether they were named; on atypical pictures, they performed better when category labels were provided. A control study demonstrated that children ignored the label when it named a transient property rather than a stable category. Contrary to standard views of young children, these results indicate an early-emerging capacity to overlook salient appearances. However, one important development still to take place is the ability to use subtle perceptual cues to determine category membership in the absence of language.
Article
Full-text available
Maximum likelihood or restricted maximum likelihood (REML) estimates of the parameters in linear mixed-effects models can be determined using the lmer function in the lme4 package for R. As for most model-fitting functions in R, the model is described in an lmer call by a formula, in this case including both fixed- and random-effects terms. The formula and data together determine a numerical representation of the model from which the profiled deviance or the profiled REML criterion can be evaluated as a function of some of the model parameters. The appropriate criterion is optimized, using one of the constrained optimization functions in R, to provide the parameter estimates. We describe the structure of the model, the steps in evaluating the profiled deviance or REML criterion, and the structure of classes or types that represents such a model. Sufficient detail is included to allow specialization of these structures by users who wish to write functions to fit specialized linear mixed models, such as models incorporating pedigrees or smoothing splines, that are not easily expressible in the formula language used by lmer.
Article
Full-text available
This research explored the development of children's use of multiple conceptual organizations (thematic, taxonomic) in sorting sets of pictures. Experiment 1 revealed that between 5 and 9 years, two forms of categorical flexibility can be distinguished: response and conceptual flexibility. It appeared that children's multiple sorts do not necessarily reflect the use of different conceptual organizations. Such a lag was mainly due to a difficulty in accessing taxonomic representations, specifically in the younger age groups. Therefore, Experiment 2 investigated the development of taxonomic representations using an original approach requiring participants to decide whether new items could be included into an existing taxonomic sort. This approach showed that taxonomic representations were only gradually differentiated from thematic and perceptual ones over the 5 to 10 years period. The discussion raises new hypotheses about the interaction between developing executive control (specifically, increasing resistance to interference of irrelevant information) and increasing conceptual knowledge in accounting for the development of conceptual flexibility.
Article
Full-text available
Categorizations which humans make of the concrete world are not arbitrary but highly determined. In taxonomies of concrete objects, there is one level of abstraction at which the most basic category cuts are made. Basic categories are those which carry the most information, possess the highest category cue validity, and are, thus, the most differentiated from one another. The four experiments of Part I define basic objects by demonstrating that in taxonomies of common concrete nouns in English based on class inclusion, basic objects are the most inclusive categories whose members: (a) possess significant numbers of attributes in common, (b) have motor programs which are similar to one another, (c) have similar shapes, and (d) can be identified from averaged shapes of members of the class. The eight experiments of Part II explore implications of the structure of categories. Basic objects are shown to be the most inclusive categories for which a concrete image of the category as a whole can be formed, to be the first categorizations made during perception of the environment, to be the earliest categories sorted and earliest named by children, and to be the categories most codable, most coded, and most necessary in language.
Article
Word learning is a notoriously difficult induction problem because meaning is underdetermined by positive examples. How do children solve this problem? Some have argued that word learning is achieved by means of inference: young word learners rely on a number of assumptions that reduce the overall hypothesis space by favoring some meanings over others. However, these approaches have difficulty explaining how words are learned from conversations or text, without pointing or explicit instruction. In this research, we propose an associative mechanism that can account for such learning. In a series of experiments, 4-year-olds and adults were presented with sets of words that included a single nonsense word (e.g. dax). Some lists were taxonomic (i.,e., all items were members of a given category), some were associative (i.e., all items were associates of a given category, but not members), and some were mixed. Participants were asked to indicate whether the nonsense word was an animal or an artifact. Adults exhibited evidence of learning when lists consisted of either associatively or taxonomically related items. In contrast, children exhibited evidence of word learning only when lists consisted of associatively related items. These results present challenges to several extant models of word learning, and a new model based on the distinction between syntagmatic and paradigmatic associations is proposed.