University of Southern California Dissertations and Theses

Toddlers’ interpretation of 'thing' and 'one'

(USC Thesis Other)

Toddlers’ interpretation of 'thing' and 'one'

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Copyright 2023 Olesia Bokhanovich

Toddlers’ Interpretation of Thing and One

by

Olesia Bokhanovich

A Thesis Presented to the
FACULTY OF THE USC DORNSIFE COLLEGE OF LETTERS, ARTS, AND SCIENCES
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF ARTS
(PSYCHOLOGY)

August 2023

ii
TABLE OF CONTENTS
List of Figures………………………………………………………………………………...iii
Abstract……………………………………………………………………………………….iv
Chapter One: Introduction……………….………….………………………………………....1
Chapter Two: Method…………………….…………………………….…………….……...10
Chapter Three: Results……………………………...…………………….……….…………15
Chapter Four: Discussion…………………….………………………………….…………...23
Chapter Five: References………………….………………….………………………...……30
Chapter Six: Appendix A…………………………………………………………………….31

iii
LIST OF FIGURES
1 Figure 1:A set of objects of 1 trial…………………………………………………....10
2 Figure 2: Child’s point of view during experimental trial………...…………………12
3 Figure 3: The percentages of all choices in the first and second runs of
3-year-olds in condition One……………………….……………………………...…16
4 Figure 4: The percentages of all choices in the first and second runs of
3-year-olds in condition Thing……………………………………………….….…...16
5 Figure 5: The percentages of all choices in the first and second runs of
24-30-months-olds in condition One………………………….………….…….…….18
6 Figure 6: The percentages of all choices in the first and second runs of
24-30-months-olds in condition Thing………………………………………………18
7 Figure 7: Summed basic-level matches in the first run across the ages in months......20
8 Figure A1: Full set of objects………………………………………...………………31

iv
Abstract
Two phrases, “Look at this one!” and “Look at this thing!” might appear very similar at first
glance as both contain taxonomically underspecified referents one and thing, meaning that
they do not reveal a particular category. However, using the word one implies that the
speaker has a particular shared category in mind, while using the word thing does not warrant
such presupposition. That is, these phrases differ in presuppositional content or the
knowledge shared between the speaker and the listener, which is important for making and
understanding references. This study was motivated by research by Mintz (2006)
demonstrating evidence that presuppositional content plays a role in language development.
The present study explores 2-and-3-year-old children’s ability to understand the differences
in the presuppositional content entailed in the word one and the presuppositional content
entailed in the word thing. To answer this question, an online study was conducted where
toddlers (n=77) were presented with a target object and asked to show another from the four
test objects displayed: a basic-level match, a superordinate-level match, and two unrelated
distractor objects. Some children heard the pronoun one, while other children heard the word
thing when being presented with unfamiliar objects (e.g., clay shapes). The results indicated
that the basic-level match was the most prominent choice across both 3-year-old and 2-year-
old participants, with a marginal difference between the selection rate of the basic-level
match across conditions. Additionally, older children were more likely to select a basic-level
match compared to younger children. The findings indicate that toddlers showed sensitivity
to presuppositional content contained in the phrases with underspecified reference terms
thing and one. Understanding this sensitivity further is important for exploring children's
pragmatics skills development as well as for creating research designs and educational
materials with an interchangeable use of the terms thing and one.

1
Chapter One: Introduction
The use of language, fundamental to human communication, is a multifaceted process
where slight variations in wording can have significant implications for interpretation. This
paper explores the presuppositional content that can be seen in two seemingly similar
phrases, “Look at this one!” and “Look at this thing!” While both are taxonomically
unspecified, indicating a lack of specificity in a category, in contrast to “Look at that car!”
which activates the speaker's construal of a specific conceptual category (car), the use of the
pronoun one in the former phrase suggests a presupposition of a shared category between
speaker and listener (Mintz, 2006). In contrast, the use of the noun thing in the latter phrase
lacks this presupposition and might indicate that the speaker lacks knowledge about the exact
category or label for the object of their reference (Mintz, 2006). Therefore, the interpretation
of thing and one is based on their use in context, and they vary in how they should be
interpreted. The purpose of the study is to investigate when and how children start spotting
this linguistic distinction, with the goal of expanding our knowledge of language
development.
Presuppositional Content and Word Learning

Past studies indicate that presuppositional content plays a significant role in word
learning. Mintz and Gleitman (2002) showed that toddlers' learning of adjectives depended
on whether the noun that it modified was a specific or vague reference. In this study, toddlers
were presented with familiar toy objects of known categories (cars) with some adjustments to
their property (being covered in star stickers). Among specific references used were basic-
level nouns (e.g., car) and superordinate-level nouns (e.g., toy) The authors analyzed
toddlers’ behavior when definitive and underspecified reference terms one and thing were
used. As children were presented with an object, they also heard a phrase containing a novel
adjective (e.g., “…stoof car”). They were then asked, "Which is the stoof car/toy/thing/one?"

2
and given two choices to pick from (1) a kind match, a toy matching the category that was
shown before or (2) a property match, a toy not matching the category shown before but
matching by the property. The reference terms at the end of the phrase were nouns that
allowed children to access the category (basic-level names, superordinate-level names) or
nouns that were vague references (one, thing). If the child had learned the meaning of the
novel adjective stoof, they were expected to match it with the property (e.g., being covered in
star stickers). Both 24-month-olds and 36-month-olds were able to select the property match
above the chance level when they heard sentences containing specific nouns (e.g., “stoof
car”). When hearing sentences containing vague-referent nouns (e.g., “stoof one” or “stoof
thing”), only 36-month-olds were able to select the property and only in the condition where
the pronoun one was used.
This finding of Mintz and Gleitman (2002) sparked more questions regarding the
pragmatic distinction in the interpretation of the reference words one and thing and warranted
looking at them separately. Mintz (2005) further explored how word-learning can be
influenced by presuppositional content of underspecified reference words of one and thing
with both familiar and, importantly, unfamiliar objects, a category for which cannot be as
easily computed as with a known object (e.g., toy car). The study included 240 children who
were presented with unfamiliar adjectives (stoof) paired with unfamiliar (e.g., shapes) and
familiar (e.g., toy animals) objects. All presented objects within one training trial shared a
common property (e.g., covered in felt). This study found that adjective acquisition was
conditioned on the interaction between modified nouns and the familiarity of the object
referenced. Specifically, 36-month-olds mapped adjectives to the common property only
when the word thing was used. The author explains that this might be because, with
unfamiliar objects, there is no prior knowledge of the category, therefore, thing will be a
logical stand-in for the basic-level category. Interestingly, one and thing elicited different

3
responses in 36-month-olds depending on the context of familiar and unfamiliar objects. On
the other hand, 24-month-olds were at chance for both thing and one when unfamiliar objects
were used. At this age participants were hypothesized to have some limitations in retrieving
category information and understanding presuppositions of the discourse. Overall, this study
provided further evidence that toddlers’ ability to map novel adjectives depends on the noun
it modifies (specific or vague) as well as on the familiarity with the category. That is,
presuppositional content within the category construal might guide the learner toward the
interpretation and therefore serve as a facilitator in learning adjectives and other lexical
categories (Mintz, 2005).
Pragmatic Oddness
As previously discussed, presuppositional content might have a facilitatory function
in word learning among 2-and-3-year-old children. However, some studies indicate that
young children experience difficulties in mapping novel adjectives to object characteristics.
In the study by Klibanoff & Waxman (2000), toddlers were presented with two target objects:
the first represented a novel property (e.g., transparency), and the second represented a
contrastive property (e.g., opacity). In within-basic condition, the targets were of the same
basic-level category (e.g., transparent plate, opaque plate). In across-basic condition, the
targets were of the different basic-level categories (e.g., transparent plate, opaque
toothbrush). The experimenters referred to the targets using a phrase with a novel adjective
(e.g., “Gogi says this is a very blickish one. He says this one is not blickish. Can you give
Gogi another one that is blickish?”). The choices the children could make were both from
another basic-level category (e.g., bottles) and were between a matching object (e.g., a
transparent bottle) and foil (e.g., an opaque bottle). The study found that only 4-year-old, but
not 3-year-old participants, were able to map novel adjectives to objects of the same and
basic-level category. A similar pattern was observed in other studies which indicate that

4
younger learners might be limited in novel adjectives to object characteristics mappings
(Klibanoff & Waxman, 2000; Taylor & Gelman, 1988; Hall, Waxman, & Hurwitz, 1993).
A potential explanation for these limitations was provided by Mintz (2005), who
suggested that failure in mapping might occur not due to the failure in the identification of
the category but rather due to the pragmatic oddness created by the violation of pragmatic
discourse in tasks used. For example, a violation of this kind occurs when one labels an
unfamiliar object, or thing labels a familiar object. In this case, a participant might be
confused and limited in arriving at the meaning.
Additionally, support for this hypothesis was also brought up by Hiramatsu, Rulf, &
Epstein (2010), who indicated that previous failure of younger children to extend novel
adjectives to the same and different basic-level categories as compared to older children
might be attributed to the complexity and oddness of the tasks used. The researchers explored
how both adults and 3–4-year-old children interpret the pronoun one. They found that
children, as well as adults, were able to extend novel adjectives across categories when the
task was designed in a way that would not be pragmatically odd for subjects. In the first
experiment, subjects were shown a target object (e.g., yellow toothbrush) described as “This
is a yellow one” and asked, “Is there another yellow one?”. In this instance, the oddness is in
a potentially dual interpretation of one as it could be interpreted as referring to a kind
(another yellow toothbrush) or as a property (another yellow object). In this pragmatically
odd situation, it was found that adults preferred the kind-match. However, in the second
experiment, pragmatic oddness was removed by substituting the use of one with a different
linguistic cue all over and presented objects were characterized as “stoof all over”. In this
case, adults and toddlers were able to arrive at a property-match interpretation. The authors
thus concluded that the language abilities of younger children in the past research might have

5
been underestimated due to a potential hindering by the confusion created by pragmatically
odd tasks (Hiramatsu, Rulf, & Epstein, 2010).
Similarly, Akhtar (2002) found success in mapping novel adjectives by even younger
participants (i.e., 2-year-olds). In this study, when presented with a novel object, children
were provided with a clear contrast by hearing the phrase "This is a dacky one" in the
sequence with the other two objects' descriptions, "This is a round one", and "This is a square
one". In this case, children were able to map a novel adjective to the shape. Moreover, when
the known adjectives "smooth" and "fuzzy" were used to describe other objects instead,
children then mapped novel adjectives to the texture.
Lidz, Waxman & Freedman (2003) provide further evidence for very young toddlers’
competency in interpreting the referents when presented with familiar objects. The study
examined how children interpret the sentences with a referent one. In the familiarization
phase, participants were shown a target object (e.g., a yellow bottle) and told “Look! A
yellow bottle!”. In the test phase, participants were shown two objects: one matched the
target by kind and color, while the second one matched only by kind (e.g., a yellow bottle and
a blue bottle). The toddlers were exposed to either a control condition in which they heard
“Now look. What do you see now?” or an anaphoric condition in which they heard “Now
look. Do you see another one?”. The authors found that toddlers in anaphoric condition
looked longer to the familiar object, which matched in kind and color, compared to the novel
object, which matched only by kind. The pattern was reversed for the control condition, in
which toddlers looked longer at the novel object. These results indicate toddlers as young as
18 months old possess a linguistic structure aiding the acquisition of syntax and are able to
interpret the word one as a reference to the category of the presented familiar object. Overall,
when using simple tasks, explicit directive cues, and explicit contrasts which help in

6
constraining the interpretations, 2-year-old children might also be successful in adjective
mapping (Akhtar, 2002; Hiramatsu, Rulf, & Epstein, 2010; Waxman & Klibanoff, 2000).
Support for the effect of the cues used in a given task was also demonstrated by
Sandhofer and Smith (2007), who taught 2-year-olds color words by presenting them with 3
toys of red, yellow, and green colors and indicating their color out loud by saying “This is a
red one.” in an absent syntactic cue case and “This is a red dinosaur.” in a strong syntactic
cue case. Children were then asked to match the presented 3 dinosaur toys with other 3 cup
toys of the same colors. The authors found that younger children learned more adjectives in
the absence of strong syntactic cues, when only asked. “Show me where red is”, while older
children learned more with strong syntactic cues provided, when asked, “Can you give me
the red cup?”. The authors suggest that noun specification and syntactic cues affect children’s
word learning and that adjective learning, specifically, is a function of children’s noun
knowledge as it provides additional information about the property. In sum, the studies above
indicate that presuppositions of the speaker affect the young learner's ability to arrive at the
meaning of the words; however, more research is needed to better understand children’s
sensitivity to presuppositions contained within specific linguistic units as well as to
disentangle a potential effect of pragmatic coherence or incoherence.
The reviewed literature provides indirect evidence that nouns and, specifically, their
category construal facilitate word learning. In the present study, we will take a closer look
into the presuppositional content alone to examine toddlers’ interpretation of thing and one
more directly using an easy matching task, potentially lessening the effect of possible
pragmatic confusion based on Mintz’s (2006) research design conducted in-person which
provided additional evidence that 3-year-olds can be sensitive to differences in
presuppositional content. In Mintz (2006), 24 and 36-month-olds were presented with a
target object and asked to show another using various phrases that either used the pronoun

7
one or the word thing with both familiar and unfamiliar objects. There were 4 choices the
children could make: (1) basic-level match which shares a category and a kind to the target
object, (2) superordinate level match, which shares the category of the target object but not
the kind; and (3 and 4) two unrelated objects, which share neither the category nor the kind
with the target object. The results indicated that 36-month-olds, when shown familiar objects
and heard one, selected the basic-level match above the chance level, potentially interpreting
one as a stand-in for the basic-level noun, whereas when they heard thing, their selections
were random, potentially interpreting thing as unconstrained. In the case of unfamiliar
objects, the pattern was reversed; when heard the word thing, the basic-level match was the
most prominent choice, and when heard one, the selections were random. Providing strong
evidence that 3-year-olds are sensitive to presuppositions of thing and one in different
contexts, 2-year-olds' performance in this paradigm remains to be further investigated.
Research Question and Hypotheses
The present study takes a closer look at 2-year-olds’ performance in the task design
by Mintz (2006). It adds to the existing findings by examining continuous data from 2-and-3
year-olds, which includes a wide variety of ages, analyzing a larger sample size between age
groups. Additionally, children’s null responses were recorded such as not selecting an object
or saying “no” which was not the focus of the earlier studies. The study is also conducted in a
completely online format. Though having drawbacks, the online format was innovative as it
allowed us to access a diverse sample and better control the stimuli as pre-recorded videos
were used; therefore, all subjects were exposed to the same intonation and articulation of the
speaker.
The questions of the current study are as follows: Are children between 24- and 36-
months of age sensitive to the pragmatic differences between thing and one, and if so, what is
the developmental age-related of this sensitivity? To answer this question, we will look at

8
what children’s choices of the objects would be given the presentation of the phrases
containing the associated category (using one) versus not (using thing) with unfamiliar
objects. Using unfamiliar objects allows us to look at children’s choices of objects that do not
have a known category. Moreover, familiarity with certain objects, such as knowing what a
giraffe is and how it looks, might benefit children’s learning of properties associated with that
object (Hall, Waxman, & Hurwitz, 1993; Mintz, 2005). Given this, it is interesting how
children would react when we are referring to a category using the pronoun one when no
category can be attributed to the objects displayed.
Considering earlier findings of Mintz (2005), the first hypothesis is that in the case of
unfamiliar objects, 3-year-olds, but not 2-year-olds, will be able to compute presuppositional
content upon hearing thing as they will interpret it as a stand-in category for a basic-level
noun due to a lack of prior knowledge about the categories (Mintz, 2005; 2006). If this
hypothesis is true, 3-year-olds are expected to select the basic-level match above chance upon
hearing the referent thing labeling unfamiliar objects.
The second hypothesis is that the referent one will not confuse subjects with
unfamiliar objects, as it was believed before, and would still allow for presuppositional
content computation among 3-year-olds potentially due to using a simpler task (Mintz, 2006).
If this hypothesis is true, 3-year-olds will select the basic-level match above chance upon
hearing the referent one labeling unfamiliar objects. As for 2-year-olds, they are predicted to
remain at chance level for both conditions due to having potential limitations in category
retrieval abilities as well as limitations in presuppositional content computation in a given
discourse, as shown by previous studies (Mintz & Gleitman, 2002; Mintz, 2005).
Moreover, the study explores whether age and condition can predict selecting a
superordinate-level match when asked to choose again from the remaining objects. Such
analysis is interesting as it allows us to explore toddlers’ understanding of taxonomic levels

9
when underspecified referents are used. In parallel to the familiar taxonomies such as dogs
(type), mammals (class), and animals (kingdom), the stimuli used in the present study call
upon a similar and perhaps more obvious taxonomy due to specific visual features of the
objects used: sharing the exact shapes and textures. Therefore, the use of referents thing and
one in this context might result in a different choice switch: from the basic-level match at
first to another category or to a higher level within the same taxonomy when asked to select
again. Furthermore, it allows examining whether age might play a role in this pattern. The
hypothesis A in this analysis would be that older participants should be more likely to select
the superordinate-level match. The rationale behind this is that older participants would be
more likely to create a category construal as shown by older toddlers’ successful category
interpretation and use in word learning (Mintz,2005). Therefore, it could be expected that
presuppositional content computation development might help them in selecting the
superordinate-level match with more likelihood as compared to younger participants. The
hypothesis B in this analysis would be that participants who heard the referent one would be
more likely to make this basic-level to superordinate-level match jump compared to
participants who heard the referent thing. The rationale behind this hypothesis is that one
given its ability to reference a specific category as opposed to thing where no specific
category is made available would prompt more subjects to select an object in the second run
that still shares the same category or features as the target instead of making a selection
among distractors.

10
Chapter Two: Method
Participants
The participants of the study (N=77) were typically developing 2- and 3-year-olds
coming from predominantly English-speaking families. Participants were recruited using an
online platform called “Lookit”. Parents gave informed consent prior to the study sessions.
Fifteen subjects were excluded because they did not complete 6 valid trials in the first run (6),
were exposed to English less than 80% of the time (2), were fussy (1), reported atypical
development (1), or because significant parent intervention occurred (5). Parents received an
Amazon card of a $5 value as a token of appreciation within 2 weeks of their participation.
Materials and Design

The study included 2 conditions, One and Thing, using unfamiliar objects. The
objects were presented in a slightly shifted diamond shape where one object is presented at
the top, one at the bottom, and two other objects on the sides. For one trial, a full set of
objects included a target, a basic-level match that shared a kind and features with the target, a
superordinate-level match that shared features but not the kind with the target, and two
distractor objects. A full set of objects for one trial is presented in Figure 1 below.
Figure 1
A Set of Objects of 1 Trial

11
Note. This figure depicts a complete set of objects for one trial: a target object, basic-level
match, superordinate-level match and two distractor objects.
The positions of the objects were counterbalanced across subjects. For each trial of
the 6 trials, subjects were presented with 1 target object. The study used a total of 6 target
objects of different shapes and colors. The full set of objects is displayed in Figure A1 in the
appendix section. The objects among which the subjects were to make their selection were
presented on the side simultaneously with the target object. Underneath each object, except
the target object, was placed a number from 1-4 in order to label objects for caregivers to
simplify the process of reporting which object was selected.
In the One condition, the subjects were presented with the phrase: “Look at this! Can
you find another one?” for 6 consecutive trials asking them to select one from another 4
displayed on the screen. In the Thing condition, the phrase “See this thing? Can you find
another?” was used instead. The phrases are not uniform due to the effort to make the phrases
as natural as possible while still using the words one and thing (Mintz, 2006). If participants
hesitated to choose an object, another video was played, encouraging them to choose again
and directing their attention to the object display panel. The phrases were presented in a

12
video format featuring a female American-English speaker. This method ensured the
consistency of the stimulus across all trials and minimized possible pronunciation and
production differences or errors.
Procedure
The study was conducted online using the conferencing platform Zoom. The
caregiver was asked to hold the child on their lap or to sit beside their child. Both the parent
and the child were looking at and attending to the computer screen. The experimenter first
went over the informed consent process, an information sheet describing the procedures of
the study, and time commitments (3 min). Parents were explained how to report which
objects were selected and other instructions of the study. They were also asked to minimize
distractions, remain silent, and only report the object selected. Simple cue words “Look!” or
“Point!” were allowed to pronounce to get the child’s attention and clarify that they needed to
point to the object to select it. Such cues did not hint at any of the object or objects category.
Before beginning the experiment, children were shown a practice panel with familiar objects
(e.g., truck) and asked to point to any object they liked, this was done to ensure that the child
understood that they needed to point when asked to find another object on the screen (5 min).
During the experiment, the child watched pre-recorded videos that featured a person
who asked to focus on the target object and used the phrase “Look at this! Can you find
another one?” or “See this thing? Can you find another?” prompting a child to select an
object out of the 4 objects presented on the other side. This continued for a total of 6 trials,
each about 10-20 seconds long. An example of a trial setup can be seen in Figure 2 below.
Figure 2
Child’s Point of View during Experimental Trial

13

Note. This figure represents the exact trail set-up presented to participants. The target object
is presented on the left and the objects to choose from with numbers labeling each are
presented on the right.
During the first run of 6 trials total, each time after the child selected an object, it
disappeared, and upon completion of the 6th trial, the second run started without the
previously selected objects. The subject then went over the same 6 trials without the
previously selected objects (10 seconds each). In the case when a child did not make a
choice, they were played another video featuring the same person repeating the phrase and
pointing to the panel to encourage them to make a choice. All obtained data is anonymous.
The total time for the experiment was around 15 minutes.
Coding
After each trial, the experimenter coded the child’s choices into a digital sheet after
obtaining the number that indicated whether the object selected was a basic-level match,
superordinate-level match, distractor 1, or distractor 2 directly from the caregivers’ report.
This method was used due to the difficulties of observing the children’s responses online
using a web camera. Not selecting an object was coded as “-” and a clear indication of

14
unwillingness to select as “N”. The unwillingness was recorded when the subject pronounced
such phrases as “No”, “I can’t”, “There is no other one!” and other phrases that indicated a
negative answer to the prompt to select an object. The difference between not selecting an
object and unwillingness to select was in the presence of a negative verbal indication in the
latter. All responses were recoded by a second experimenter for reliability checks, with the
k= .99 for the first run and k= .95 for the second run of the study. The disagreements were
resolved by replaying the session recording to determine whether the disagreement was
caused by (1) an accident in the labeling of the object, in which case it was corrected to the
label coded by the second experimenter, (2) sequential choice of the objects by a subject in
which case their first choice was recorded, or (3) unclear whether the subject didn’t select an
object or said “No”, in which case if a vocalization similar to “No” or “No other one” was
spotted, the selection was coded as “N”.

15
Chapter Three: Results
The analyzed sample after exclusions contained 62 participants; 28 were 3-year-olds
(15 females, M= 40.96 months, SD= 3.16 months, range: 36-46 months), and the rest 34 were
2-year-olds (14 females, M= 29.21 months, SD= 4.05 months, range: 24-35 months). Half of
the participants in each age group were in condition One while the other half were in
condition Thing.
3-Year-Olds’ Performance
Subjects' first run performance was recorded as selecting a basic-level match or not (1
and 0 respectively). We used a one-sample Wilcoxon signed rank test with continuity
correction to evaluate the basic-level match selection rates in the first run. We calculated the
chance level for this analysis by multiplying the chance level of selecting a basic-level match
(0.25) by the number of trials (6), which resulted in a value of 1.5 for the first run. The results
indicated that in the first run, subjects selected basic-level match greater than the chance level
of 1.5 in both condition One (M= 0.8, SD= 0.4, Selection Tally: 67 basic-level choices, 17
non-basic-level choices) V= 103.5, p= 0.001, and condition Thing (M= 0.63, SD= 0.48,
Selection Tally: 53 basic-level choices, 31 non-basic-level choices), V= 94.5, p = 0.004. This
finding indicates that 3-year-olds were able to compute that both referent one and thing
referred to a basic-level category.
Similarly, the superordinate-level match selection rates were evaluated in the second
run of the study among those who selected the basic-level match in the first run. The chance
level for this analysis was calculated by multiplying the chance level of selecting a
superordinate-level match (0.33) by the number of trials (6), which resulted in a value of 2
for the second run. The results indicated that in the second run, there was no evidence
suggesting that subjects selected the superordinate-level match above chance level in both
condition One (M= 0.51, SD= 0.5, Selection Tally: 34 superordinate-level choices, 33 non-

16
superordinate-level choices), V= 65.5, p= 0.212, and condition Thing (M= 0.47, SD= 0.5,
Selection Tally: 25 superordinate-level choices, 28 non-superordinate-level choices), V= 33,
p= 0.7. Overall selections of participants in both the first and second runs can be seen in
Figure 3 and Figure 4, displaying conditions One and Thing, respectively, and a red line
indicating the chance level.
Figure 3
The Percentages of All Choices in the First and Second Runs of 3-Year-Olds in Condition
One

Note. This figure depicts 3-year-olds’ choices in condition One in a percentage format. The
red line indicates the chance level. The object choices include basic-level match,
superordinate-level match, and distractors. Additionally, the percentage of cases when a
participant did not choose an object is presented as “No Choice” and “Said No” when a
participant indicated an unwillingness to select an object.
Figure 4
The Percentages of All Choices in the First and Second Runs of 3-Year-Olds in Condition
Thing

17

Note. This figure depicts 3-year-olds’ choices in condition Thing in a percentage format. The
red line indicates the chance level. The object choices include basic-level match,
superordinate-level match, and distractors. Additionally, the percentage of cases when a
participant did not choose an object is presented as “No Choice” and “Said No” when a
participant indicated an unwillingness to select an object.
24-30-Month-Olds’ Performance
To take a closer look at the performance of our youngest participants, the children 24-
30 months of age were selected from the original dataset. The sample included 17 subjects (7
females, M= 25.65 months, SD= 1.8 months, range: 24-29 months), 8 assigned to condition
One, and 9 assigned to condition Thing. We performed a Wilcoxon signed rank test, the
results of which indicated that, in the first run, subjects selected basic-level match greater
than the chance level of 1.5 in both conditions One (M= 0.6, SD= 0.49, Selection Tally: 29
basic-level choices, 19 non-basic-level choices) V= 34.5, p= 0.01, and condition Thing (M=
0.5, SD= 0.5, Selection Tally: 27 basic-level choices, 27 non-basic-level choices), V= 39, p=
0.028. This finding indicates that 24-30-month-olds were able to compute that both referent
one and thing referred to a basic-level category.
Similarly, the superordinate-level match selection rates were compared to the chance
level of 2 across conditions in the second run of the study for this age range among
participants who selected the basic-level match in the first run. For this analysis, one trial was
removed, due to the absence of a selection in the first run, therefore having a different chance

18
level from the rest of the second run data, of which the chance level is 0.33 after one of the
objects is removed. The results indicated that in the second run, subjects selected the
superordinate-level match with no significant difference compared to the chance level of 2 in
both condition One (M= 0.41, SD= 0.5, Selection Tally: 12 superordinate-level choices, 17
non-superordinate-level choices), V= 13, p= 0.785, and condition Thing (M= 0.48, SD= 0.51,
Selection Tally: 13 superordinate-level choices, 14 non-superordinate-level choices), V= 8,
p= 0.748. Overall selections of participants in both the first and second runs can be seen in
Figure 5 and Figure 6, displaying conditions One and Thing, respectively, and a red line
indicating the chance level.
Figure 5
The Percentages of All Choices in the First and Second Runs of 24-30-Months-Olds in
Condition One

Note. This figure depicts 24-30-month-old’s choices in condition One in a percentage format.
The red line indicates the chance level. The object choices include basic-level match,
superordinate-level match, and distractors. Additionally, the percentage of cases when a
participant did not choose an object is presented as “No Choice” and “Said No” when a
participant indicated an unwillingness to select an object.
Figure 6
The Percentages of All Choices in the First and Second Runs of 24-30-Months-Olds in
Condition Thing

19

Note. This figure depicts 24-30-month-old’s choices in condition Thing in a percentage
format. The red line indicates the chance level. The object choices include basic-level match,
superordinate-level match, and distractors. Additionally, the percentage of cases when a
participant did not choose an object is presented as “No Choice” and “Said No” when a
participant indicated an unwillingness to select an object.
Logistic Regression: Basic-Level Choice by Binary Condition and Continuous Age
Multilevel logistic regression to account for the clustering structure of the repeated
measures nested within a participant was performed to evaluate age as a continuous variable
among all participants. The function glmer from “lme4” package (version 1.1.31; Bates et al.,
2015) was used to perform the logistic regression with random intercept for participants and
sets of objects. This analysis modeled the relationship between condition and age on the odds
of selecting a basic-level match in the first run of the experiment. The equation for the model
is as follows:
𝐿𝑒𝑣𝑒𝑙 1: 𝑌
!(#,%)
~𝐵𝑒𝑟𝑛𝑜𝑢𝑙𝑙𝑖0𝜇
!(#,%)
2
𝜂
!(#,%)
= 𝑙𝑜𝑔𝑖𝑡(𝜇
!(#,%)
)
𝜂
!(#,%)
= 𝛽
'(#,%)

𝐿𝑒𝑣𝑒𝑙 2:𝛽
'(#,%)
= 𝛾
''
+𝛾
'(
𝐶𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛
#
+𝛾
')
𝐴𝑔𝑒
#
+𝛾
'*
𝐶𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛
#
∗𝐴𝑔𝑒
#
+𝑢
'#
+𝑣
'%

where 𝑌
!(#,%)
denotes the binary outcome of choosing either basic-level match in the first run
or not for participant j at trial i with sets of object k, 𝜇
!(#,%)
denotes the probability of

20
choosing the basic-level match in the first run, 𝜂
!(#,%)
denotes the log odds of choosing the
basic-level match in the first run, 𝛾
''
denotes the grand mean of the log odds, 𝛾
'(
denotes
the effect of condition on the log odds, 𝛾
')
denotes the effect of age on the log odds, 𝛾
'*
the
effect of the interaction of age and condition on the log odds, 𝑢
'#
denotes the participant-
level random effect, and 𝑣
'%
denotes random effect for the sets of items. The variable
condition was contrast-coded where participants in condition One were assigned a value of 1
and participants in condition Thing were assigned a value of -1. The variable age was mean
centered. A significant main effect of age in predicting the probability of selecting a basic-
level match in the first run was observed (p= 0.015). When controlling for condition, the odds
of selecting a basic-level match in the first run increased by 1.08 (95% CI [1.01, 1.15]) for
every one month increase in age. This finding indicates that as older children were more
likely to compute presuppositional content and interpret underspecified referents as labeling a
basic-level category.
A marginal main effect of condition was also observed (p= 0.097). While controlling
for the age, the odds of selecting a basic-level match in the first run increased by 1.41 (95%
CI [0.94, 2.13]) for those who were in condition One. This finding indicates that there might
be a potential difference in how children interpret vague referents one and thing when
labeling unfamiliar objects, where one marginally eliciting more basic level interpretations as
compared to thing. The condition by age interaction was not significant (p= 0.669). The
demonstration of the relationships described above can be seen in Figure 7.
Figure 7
Summed Basic-Level Matches in the First Run across the Ages in Months

21

Note. This figure shows data from all participants (n=62), that is, both 2-and-3-year-olds in
both conditions One (blue) and Thing (purple). The y-axis represents the summed basic-level
match responses in the first run for each participant across all six trials. The x-axis represents
age in months.
Logistic Regression: Superordinate-Level Choice by Condition and Age
Another analysis was whether selecting a superordinate-level match in the second run
of the study was different depending on the condition and age among subjects who selected a
basic-level match in the first run. To answer this question, a multilevel logistic regression
accounting for the repeated measures was performed subsetting participants who selected a
basic-level match in the first run of the study, otherwise the equation of this model would be
identical to the previously described:
𝐿𝑒𝑣𝑒𝑙 1: 𝑌
!(#,%)
~𝐵𝑒𝑟𝑛𝑜𝑢𝑙𝑙𝑖0𝜇
!(#,%)
2
𝜂
!(#,%)
= 𝑙𝑜𝑔𝑖𝑡(𝜇
!(#,%)
)
𝜂
!(#,%)
= 𝛽
'(#,%)

22
𝐿𝑒𝑣𝑒𝑙 2:𝛽
'(#,%)
= 𝛾
''
+𝛾
'(
𝐶𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛
#
+𝛾
')
𝐴𝑔𝑒
#
+𝛾
'*
𝐶𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛
#
∗𝐴𝑔𝑒
#
+𝑢
'#
+𝑣
'%

where 𝑌
!(#,%)
denotes the binary outcome of choosing either superordinate-level match in the
second run or not for participant j at trial i with sets of object k, 𝜇
!(#,%)
denotes the
probability of choosing the superordinate-level match in the second run, 𝜂
!(#,%)
denotes the log
odds of choosing the superordinate-level match in the second run, 𝛾
''
denotes the grand
mean of the log odds, 𝛾
'(
denotes the effect of condition on the log odds, 𝛾
')
denotes the
effect of age on the log odds, 𝛾
'*
the effect of the interaction of age and condition on the log
odds, 𝑢
'#
denotes the participant-level random effect, and 𝑣
'%
denotes random effect for the
sets of items. The variable condition was contrast coded where participants in condition One
were assigned a value of 1 and participants in condition Thing were assigned a value of -1.
The variable age was mean centered. The results of this model showed that condition was not
a significant predictor of selecting a superordinate-level match (p=0.951). The age also did
not significantly predict the outcome variable (p=0.986).

23
Chapter Four: Discussion

The basic-level match was the most prominent choice as the first selection in a trial
and was significantly above the chance level for both condition One and Thing and for both
3-year-olds and 24–30-month-olds. Moreover, the logistic regression analysis on the whole
dataset revealed a significant effect of age, where the older the participants were, the more
likely they were to select a basic level match. This analysis also showed a marginal effect of
condition, where subjects in condition One were more likely to select the basic-level match.
Therefore, the first hypothesis which states that 3-year-olds, but not 2-year-olds, will be able
to compute presuppositional content upon hearing thing labeling unfamiliar objects was
partially supported, since participants in condition Thing did, in fact, select the basic-level
match above chance; however, this pattern was not unique to 3-year-olds. The second
hypothesis which stated that participants will not be confused upon hearing one labeling
unfamiliar objects was also partially supported since participants also selected the basic-level
match above chance in condition One; however contrary to this hypothesis, this pattern was
also not unique to 3-year-olds. Additionally, contrary to hypothesis A and B which stated that
older children would be more likely to select the superordinate-level match and that
participants in condition One would be more likely to do so were not supported. Neither age
nor subjects’ condition placement significantly influenced superordinate-level match
selection in the second run of the study.
Implications

In the case of unfamiliar objects, 3-year-olds did, in fact, make presuppositional
content computation in the Thing condition where thing could be interpreted as a stand-in for
a basic-level noun. As outlined by previous literature, a noun used with unfamiliar objects
that carry little taxonomic information with no prior labels or knowledge available to the
participant becomes mapped to some shape-based property of the presented object (Mintz,

24
2005). Moreover, this was not unique to the condition Thing. In condition One, subjects also
made presuppositional content computation without being confused by the fact that the word
one is used with objects that do not have a particular known category.
A finding that goes against the posed hypotheses is that 2-year-olds, 24-30-months-
olds specifically, could also succeed in our task and select the basic-level match above the
chance level despite findings that suggested otherwise (Mintz & Gleitman, 2002; Mintz,
2005). This shows that toddlers of this age do understand that vague reference terms such as
one and thing can stand in for basic-level category labels.
Task demands are a possible explanation for why this study observed a different
effect from the earlier studies on this matter. The present study’s design might have required
lower task demands in comparison to earlier adjective mapping designs since subjects have
all the exemplars (basic-level match, superordinate-level match) right in front of them at all
times. This might make the basic-level match more salient for our younger subjects which in
turn made it easier for them to select it; therefore, possibly resulting in better performance at
this task. The low demands of this task might have also made it easier to perform and still
select a basic-level match when a confusing referent one was used with unfamiliar objects for
3-and-2-year-olds. Besides, the adjective mapping task might have increased the difficulty of
the task, since subjects also had to learn a novel adjective with little guidance or cues.
Moreover, the category being referenced with underspecified terms changed during the same
trial. For example, in Mintz (2005), participants were presented with three objects of three
different categories, and all of them were referenced with the same nominal, one, or thing.
Such continuous changes of the category paired with a challenge to learn a novel adjective
might also contribute to the additional cognitive load necessary to process all the information.
Another implication of the present research is the exploration of the developmental
shift. Earlier studies suggest that there is an interesting change occurring somewhere from 24

25
months to 36 months of age related to the ability to use underspecified noun labels and
understand the context in which they are used to arrive at an interpretation of the category
they refer to (Mintz 2005). The present study might have confirmed this shift by observing a
significant effect of age, where older participants were more likely to perform the task
successfully. However, it is also important to note that, in general, older toddlers show better
performance on experimental tasks which makes it difficult to state that the observed age
effect is task dependent.
Furthermore, this study allowed us to look at participants' on-the-spot categorization
processing in a virtual environment. Interestingly, in this study, we also recorded responses
that indicated that the participants did not want to select any objects. This usually occurred in
the second run of the study, when the basic-level match was already selected and taken away.
In this case, participants sometimes said, "There is no other one!" or otherwise indicated an
unwillingness to choose again. Although not significant, this difference visually seems to be
greater in condition One than Thing for both of the age groups (Figures 3,4,5,6). This might
suggest that even in the absence of an explicit category, the referent one might make subjects
more restrictive with their choices. A possible rationale for this is that one presupposes a
category in the context, while thing does not. This might allow participants to be more
focused on categories or more unequivocal when one is used as a referent while being more
open to their own interpretations or their own object preferences when thing is used. Future
studies can further examine this phenomenon by implementing measures of the "No"
response in toddlers in the context of categories understanding.
Another implication of the present study, as well as the earlier work, is the suggestion
that toddlers might have a differential understanding of the vague reference terms one and
thing (Mintz & Gleitman, 2002; Mintz, 2005). This raises possible concerns about
interchangeable and oblivious use of such terms in diverse study paradigms, designs, and

26
educational materials for young learners. Despite the fact that both thing and one are vague
references, the observed marginal difference between basic-level choice preference might
indicate that there might be a difference in children's understanding of these terms with a
subsequent difference in behavior in the context of unfamiliar objects. More specifically,
participants in condition One were marginally more likely to select the basic-level match as
compared to participants in condition Thing. A possible difference in interpretations brings us
back to the discussion on whether one could be more restrictive even in the context of objects
with no known category. Perhaps, to answer this question further, one can look at whether
there is a difference between one and thing interpretation in a within subjects task, although
this would increase the task demands.
Limitations

One possible limitation of the study is that it was conducted entirely online. Since all
our participants were required to have a stable internet connection, we are leaving the part of
the population outside of this requirement unexamined. Similarly, conducting studies online
and testing children specifically presents challenges such as low control over distractions in
the home settings, variable toddler positioning, and parental interferences (Zaadnoordijk et
al., 2021).
Another potential limitation of this study is lower external validity compared to
higher internal validity due to the experimental trade-off that one can suspect has occurred
due to high control. The high control is reflected in using pre-recorded material, which would
give us uniform pronunciation of the phrases reducing the noise and experimental
manipulations.
In this way, this study may have more difficulty suggesting generalizability outside the
experimental setting. The generalizability might also be low due to a relatively small to
medium sample size. In terms of generalizability across languages, it might be difficult to

27
find the exact same terms as one and thing and expect to get similar results in a different
language. However, this might be an interesting future direction. It is also unknown how
bilingual toddlers might respond to this task; this would be another factor that is curious to
examine. One possible influence of bilingualism on this experiment is that given potentially
rich exposure to referring expressions in more than one language might introduce
possibilities of mixing referents and increased subsequent confusion. Alternatively, from
another view, exposure to referent expressions and categories might inform a clear
understanding and correct use of such referents as one and thing.
Future Directions
The present work examined presuppositional content understanding of 3-and 2-year-
olds providing a focus on 24-30-month-olds performance. It has also demonstrated how this
could be conducted entirely online, which has both advantages and disadvantages to the
research. Future work could focus on conducting this study in the in-person setting to
minimize the potential negative effects of the online environment. Moreover, as we
discussed, the present task appeared to have low task demands, which might have allowed 2-
year-olds to perform the task successfully when having the basic-level match always
displayed as a selection option in comparison to the failure on the previous task in adjective
mapping where subjects had to create their own construal with little supportive cues (Mintz
& Gleitman, 2002). Therefore, it could be interesting to see if using the present task as
training and subsequently performing the novel adjectives task might result in a different
performance of our youngest participants.
Conclusion

28
In summary, the present study explored toddlers’ ability in object category
computation when labeled by different taxonomically underspecified referents thing and one
carrying different presuppositional assumptions. To do so, we presented 2-and-3-year-olds
with a target object and phrases that contained the words thing and one, asking them to select
another object among a basic-level match, superordinate match, and two distractors.
Supporting the posed hypothesis and prior work (Mintz; 2005; 2006), it was found
that when presented with unfamiliar objects, 3-year-olds interpret thing as a stand-in for
basic-level nouns and select a basic-level match above the chance level. A similar pattern of
results was seen in the One condition, where children were not confused by the referent one
used with unfamiliar objects that lack a specific category.
Contrary to our hypothesis and prior work, which indicated that 2-year-olds have
limited abilities in understanding vague reference terms in the case of unfamiliar objects
(Mintz & Gleitman, 2002; Mintz, 2006). The present findings show it might not be the case,
as 24-30-month-olds in both conditions selected the basic-level match significantly above the
chance level. This finding might be attributed to lower task demands associated with the
present design, which features all exemplars in front of the subject and does not add
complexity by introducing novel words. On the same note, older children were more likely to
select a basic-level match, indicating an effect of age on the presupposition sensitivity. The
results also indicate a marginal difference between the selection rate of the basic-level match
across conditions, where children in condition One were marginally more likely to select a

29
basic-level match than children in condition Thing. This difference might be important for
research designs that use both terms interchangeably as vague references. It could also help
develop educational materials for young language learners. More research is needed to tease
apart the exact differences between toddlers’ interpretations of thing and one.

30
Chapter Five: References

Akhtar, N. (2002). Relevance and early word learning. Journal of Child Language, 29(3),
677-686.

Douglas Bates, Martin Maechler, Ben Bolker, Steve Walker (2015). Fitting
Linear Mixed-Effects Models Using lme4. Journal of Statistical Software,
67(1), 1-48. doi:10.18637/jss.v067.i01.

Hall, D. G., Waxman, S. R., & Hurwitz, W. M. (1993). How two‐and four‐year‐old children
interpret adjectives and count nouns. Child Development, 64(6), 1651-1664.

Hiramatsu, K., Rulf, K. E., & Epstein, S. D. (2010). When knowledge causes failure:
Children's extension of novel adjectives and the interpretation of one. Lingua, 120(5),
1209-1218.

Lidz, J., Waxman, S., & Freedman, J. (2003). What infants know about syntax but couldn't
have learned: experimental evidence for syntactic structure at 18 months. Cognition,
89(3), 295-303.

Mintz, T. H. (2005). Linguistic and conceptual influences on adjective acquisition in 24-
and 36-month-olds. Developmental Psychology, 41, 17-29.

Mintz, T. H. (2006, November 3). Toddlers' sensitivity to the presuppositions of ‘one’ and
‘thing’ [Conference session]. Boston University Conference on Language
Development, Boston University, Boston, MA, United States.

Mintz, T. H., & Gleitman, L. R. (2002). Adjectives really do modify nouns: The incremental
and restricted nature of early adjective acquisition. Cognition, 84(3), 267-293.

Sandhofer, C., & Smith, L. B. (2007). Learning adjectives in the real world: How learning
nouns impedes learning adjectives. Language Learning and Development, 3(3), 233-
267.

Taylor, M., & Gelman, S. A. (1988). Adjectives and nouns: Children's strategies for learning
new words. Child Development, 411-419.

Waxman, S. R., & Klibanoff, R. S. (2000). The role of comparison in the extension of novel
adjectives. Developmental Psychology, 36(5), 571.

Zaadnoordijk, L., Buckler, H., Cusack, R., Tsuji, S., & Bergmann, C. (2021). A global
perspective on testing infants online: introducing ManyBabies-AtHome. Frontiers in
Psychology, 12, 703234.

31
Chapter Six: Appendix A
Figure A1
Full Set of Objects

Abstract (if available)

Abstract Two phrases, “Look at this one!” and “Look at this thing!” might appear very similar at first glance as both contain taxonomically underspecified referents one and thing, meaning that they do not reveal a particular category. However, using the word one implies that the speaker has a particular shared category in mind, while using the word thing does not warrant such presupposition. That is, these phrases differ in presuppositional content or the knowledge shared between the speaker and the listener, which is important for making and understanding references. This study was motivated by research by Mintz (2006) demonstrating evidence that presuppositional content plays a role in language development. The present study explores 2-and-3-year-old children’s ability to understand the differences in the presuppositional content entailed in the word one and the presuppositional content entailed in the word thing. To answer this question, an online study was conducted where toddlers (n=77) were presented with a target object and asked to show another from the four test objects displayed: a basic-level match, a superordinate-level match, and two unrelated distractor objects. Some children heard the pronoun one, while other children heard the word thing when being presented with unfamiliar objects (e.g., clay shapes). The results indicated that the basic-level match was the most prominent choice across both 3-year-old and 2-year-old participants, with a marginal difference between the selection rate of the basic-level match across conditions. Additionally, older children were more likely to select a basic-level match compared to younger children. The findings indicate that toddlers showed sensitivity to presuppositional content contained in the phrases with underspecified reference terms thing and one. Understanding this sensitivity further is important for exploring children's pragmatics skills development as well as for creating research designs and educational materials with an interchangeable use of the terms thing and one.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Considering the effects of disfluent speech on children’s sentence processing capabilities and language development

PDF

Do 3-year-olds understand that appearances can ‘deceive’ and lead to false beliefs?

PDF

Examining the impact of prenatal maternal internalizing symptoms and socioeconomic status on children’s frontal alpha asymmetry and psychopathology

PDF

Academic achievement, same- and cross-ethnic positive peer regard among Asian American and Latinx adolescents

PDF

Learning repetition rules and non-adjacent dependencies from human actions in nine-month-old infants

PDF

Adolescents who have never dated: friendship attributes as predictors of prolonged romantic development among Asian American and Latine youths

PDF

Children's ""mutuality demand"" in seeing, hearing, and speech

PDF

Young children selectively transmit true information when teaching

PDF

Subjectivity, commitments and degrees: on Mandarin hen

PDF

When and how infants discriminate between declaratives and interrogatives

PDF

Learning about word learning in ambiguous situations

PDF

Young children’s demand for reciprocal engagement: evidence from typically-developing and autistic preschoolers

PDF

Children's use of social and visual perspective-taking in reference resolution

PDF

Couple conflict during pregnancy: Do early family adversity and oxytocin play a role?

PDF

Why is it so difficult to recognize faces differing only moderately in orientation?

PDF

Development of error monitoring in preschool to 12th-grade students and relations in late childhood and adolescence to social-affective processing and emotions about math

PDF

Can I make the time or is time running out? That depends in part on how I think about difficulty

PDF

Sensitive, specific, and generative face recognition in a newborn visual system

PDF

The development of object recognition in the newborn brain

PDF

Evaluating the effects of virtual reality and acceptance and commitment therapy on music performance anxiety

Asset Metadata

Creator Bokhanovich, Olesia (author)

Core Title Toddlers’ interpretation of 'thing' and 'one'

School College of Letters, Arts and Sciences

Degree Master of Arts

Degree Program Psychology

Degree Conferral Date 2023-08

Publication Date 01/10/2024

Defense Date 07/10/2023

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,semantics,Toddlers,vocabulary

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Mintz, Toben (committee chair), Manis, Frank (committee member), Morales, Santiago (committee member)

Creator Email bokhanov@usc.edu,olesia@bokhanovich.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113263438

Unique identifier UC113263438

Identifier etd-Bokhanovic-12069.pdf (filename)

Legacy Identifier etd-Bokhanovic-12069

Document Type Thesis

Format theses (aat)

Rights Bokhanovich, Olesia

Internet Media Type application/pdf

Type texts

Source 20230711-usctheses-batch-1066 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu