Semantic Development: Learning the meanings of words
Semantic Development
[During the course of this process, which is usually called semantic development] Children’s
strategies for learning word meanings and relating them to one another change as their
internal representation of language constantly grows and becomes reorganized. [Very young
children understand the pragmatic intent of adults’ utterances before they can understand the
words themselves. This earliest comprehension is at the emotional, social, and contextual levels.
The way parents speak, including their prosodic contours, can convey different emotions or
messages. For example, when a parent speaks in a soothing tone with gentle rhythms, it may
convey comfort. Conversely, raised pitch or sharp intonation might indicate excitement or
displeasure respectively.]
THE RELATION BETWEEN WORDS AND THEIR REFERENTS
The word is a sign that signifies a referent, but the referent is not the meaning of the word.
Explanation: In this context, the referent is the actual object or concept in the world that a word
or phrase refers to.
For example: [You say to a child] “Look at the doggie.” [the dog is the referent, but not the
meaning of doggie—if the dog ran away or were run over by a truck— the word would still have
meaning because meaning is a cognitive construct. It's not just about the object itself, but also
encompasses the associations, ideas, and concepts that are connected to that word in our
minds.]
[Let us assume that the child learns that the word doggie refers to her dog. What is the relation
between the word and the dog? Dogs can be called doggie, Hund, perro, gou, or pek depending
on whether one is speaking English, German, Spanish, Mandarin, or Yucatec Mayan. There is
nothing intrinsic to dogs that makes one name or another more appropriate or fitting: The
relation between the name and the thing is thus arbitrary, and it is by social convention in a
particular language that speakers agree to call the animal by a particular word (Morris, 1946).
This arbitrary relation between the referent (the dog) and the sign for it (the word dog) is
symbolic. Nonverbal signs can also share this symbolic nature; the red light that means stop, for
instance, is purely symbolic because there is no obvious connection between the color red and
the action of stopping. We could agree to have blue lights or even green lights mean stop, as
long as we all agreed on the meaning of the light.]
[Why is it symbolic? The connection between the word "dog" and the actual animal it represents
is arbitrary. The meaning of the word "dog" is determined by convention and agreement within
a linguistic community.]
[Imagine you have a picture of a cat. The picture is like the word, and the actual cat is like the
referent.
The picture (word) represents the cat (referent). When you see the picture, you think of the cat.
But the picture itself isn't the cat. It's just a representation, a symbol, or a sign of the cat.
Similarly, words represent things in the world, but they aren't the things themselves. They're
like labels or symbols we use to understand and talk about the world around us.]
For a few words, the relation between word and referent is not arbitrary.
For example: “The book fell with a thud.“ the relation between the word thud and the actual
sound referred to is not arbitrary, since the word is an attempt to resemble the real sound. Nor
is the name of the cuckoo bird arbitrary: It represents the sound that the bird actually makes.
Use of words that resemble real sounds is called onomatopoeia.
Many of children’s earliest words or protowords have a less-than-arbitrary relation to their
referents; trains are called choo-choos and dogs are bow-wows.
[Some of these words are in the baby-talk lexicon that adults use when attempting to communicate with
babies and others are the children’s own creations. It is probably easier for children to learn a word that
is more directly related to its referent than one that is totally arbitrary and symbolic, and as some
research has shown, young children believe that the name and the referent are intrinsically related.
They think that one cannot change the name of something without changing its nature as well; for
instance, years ago, Vygotsky (1962) noted that many children believe that, if we decided to call a dog a
cow, it would begin to moo.]
Plato, writing in the fourth century b.c.e., discussed the question of whether there is a natural
relation between names and referents in his Cratylus dialogue.
The Anomalists of Plato’s day believed that the relation was inexplicable, but the Analogists
believed that through careful etymology the essential nature of words could be revealed
(Bloomfield, 1933).
For example: Blueberry and Bedroom
[It is called blueberry because it is a berry that is blue. It is called bedroom because it is a room
containing a bed.]
[The "Cratylus" dialogue is deep philosophical inquiries about the nature of language and its role in
human understanding and communication.]
MENTAL IMAGES
Although meaning is a mental representation, or concept, that is not to say that meaning is a
mental picture.
[Even though it is true that many people are able to visualize words, many words, such as happy or
jealousy, do not have picturable referents, and still we know their meanings. Even if one has an image
for a word, it is likely to be quite individualized: Dog, for instance, might evoke a picture of a brown
cocker spaniel for one person and a golden retriever for another. Furthermore, images tend to be quite
idiosyncratic; speakers who share meaning may hold very different internal images. One speaker’s
mental image of a house may look like a mansion, whereas another’s may be a simple cottage, yet both
speakers recognize new instances of houses when they encounter them. Additionally, to be useful for
communication, meaning cannot reside solely in the mind of the individual, but must be shared by a
speech community.]
CONCEPTS
One of the child’s primary tasks in semantic development is to acquire categorical concepts
(e.g., to learn that the word dog refers to a whole class of animals) and to be able to extend
the word to appropriate new instances of the category.
Theorists differ as to how to characterize the nature of children’s categorical concept
acquisition.
One view is that children acquire categories by learning the essential semantic features of the
category.
[This view is called the semantic feature view where children learn a set of distinguishing features for
each categorical concept. For example, at first the word dog may be understood to apply only to the
child’s own dog, but the child soon comes to understand that other creatures may also be called dog as
long as they share a small set of critical features: Dogs are animate, are warm-blooded, have four legs,
and bark. Other theorists propose that categories are defined by a set of weighted, rather than equally
critical, features. The child’s task, then, is to sort out which features are most important for membership
in a particular category. For example, the feature “bark” might be weighted relatively heavily for the
concept dog, because most dogs do bark, whereas other warm-blooded, four-legged mammals do not.
Second, they first learn prototypical examples of a category.
[According to prototype theory (see Figure 4.1), children acquire prototypes, or very good examples of
concepts, when they acquire meaning and only later come to recognize category members that are
distant from the prototypes. These right here are the examples of prototypical vegetables, dog, fruit,
bird, flower, and chair. So, why are they a prototypical member of the said categories? Well, they are
considered as such because they possess a combination of features or characteristics that are most
commonly associated with that category. These features are typically perceived as central or
representative of the category as a whole. For example, when children first learn about fruits, they
may encounter apples as one of the earliest examples. Apples possess many characteristics that are
typical of fruits—they are usually sweet, grow on trees, and have seeds inside. Therefore, children may
initially form their concept of a fruit based on their experiences with apples. So, as children gain more
experience and exposure to different examples within each category, their understanding becomes
more refined, and they can recognize that there is diversity within categories. This process reflects how
cognition develops and how humans organize information to make sense of the world around them.]
Third, they use a probabilistic strategy in assigning category membership.
[Upon first seeing a penguin, children (and adults) would decide that it is probably a bird, because it has
many birdlike features, such as a beak and wings. Thus, even though it does not fly or chirp, it still
qualifies for membership in the bird category. Some researchers, notably Smith and Medin (1981), have
pointed out that even if children are acquiring their concepts as categories, there are differences in the
nature of the concepts themselves. For instance, there are classical concepts, such as triangle, which
can be unambiguously defined: All triangles must have three angles, or they are simply not triangles.
Bird, on the other hand, is an example of a probabilistic concept. Most, but not all, birds have many
features in common, but there is not a single set of essential features. Furthermore, some concepts
have fairly sharp boundaries and are hierarchically organized, while others are not; for instance, most
adults can agree on what is and is not a dog and know that dogs belong to the superordinate category of
animals.]
THEORITICAL PERSPECTIVE ON SEMANTIC DEVELOPMENT
[One of the simplest explanations of how children acquire the meanings of their first words is that they
do so through associative learning.]
1. Learning Theory
2. predicts that repeated exposure to a stimulus (e.g., hearing the parent say the word kitty) paired
with a particular experience (seeing the family cat appear) will result in the child associating the
sound of the word kitty with the family cat.
3. Learning theory may explain the earliest and simplest kinds of linking between words and
objects.
[Eventually, the infant will react to the word alone as if the cat were there—looking around for it or
getting excited and ready for play. For learning to have taken place, it suffices that the word kitty and
the actual cat have been associated, so that they evoke at least some of the same responses. Children
are especially sensitive to novelty in their environment and predisposed to apply new words to new
objects (Smith, 1999). Thus, it is likely that many of their earliest words, such as bottle and blanket,
which have concrete referents, could be learned through association. Exclusive reliance on associative
learning, however, would be slow, effortful, and idiosyncratic and result in many errors.]
DEVELOPMENTAL THEORIES
consider semantic development within the wider context of the child’s unfolding social,
cognitive, and linguistic skills.
[During the first few months of life, before they actually begin producing words, infants are laying the
foundation for language development. Clark (1993) theorized that by the time they start learning
language, all children have developed a set of ontological categories; these are concepts about how the
world is organized. Ontological categories include objects, actions, events, relations, states, and
properties. These are the basic categories in all languages that speakers refer to when they use
language.
Ontological Categories
are concepts about how the world is organized.
include objects, actions, events, relations, states, and properties.
Even equipped with a set of ontological categories, the infant’s task is still quite daunting.
[This challenge likely stems from the complexity of the world around them and the
vast amount of information they need to process and make sense of. In other
words, while infants have some foundational knowledge, they must still navigate a
wide array of stimuli, experiences, and interactions, which can be overwhelming
and difficult to fully comprehend.]
Developmental theories attempt to explain how the child acquires first words, why the scope of
reference of children’s early words may not match that of adults, and how children’s semantic
systems become more adult-like over time.
[Let us consider what an infant must understand about verbal communication in order to begin mapping
words she hears to referents.]
The infant is in her home, and the family dog, Rufus, is lying nearby on a rug with a bone. The
baby hears her mother say words such as Rufus, dog, bone, and look.
[An infant may initially assume that the word dog applies only to the family dog. Eventually, however,
young children must come to understand that a single label can be applied to more than one specific
case (i.e., dog refers not only to their own Rufus, but to many different dogs, seen in the park, pictured
in books, on dog food boxes). Without this insight, infants cannot begin to understand the nature of
reference or to communicate about objects, actions, and properties (Clark, 1993; 2009).]
Note: However, this understanding is only one step in cracking the mapping puzzle. Not only does the
label dog refer to many different dogs, a particular dog may be labeled in many different ways (Rufus,
dog, setter, pet, puppy).
One way that young children may avoid this mapping nightmare is to rely on their rudimentary
understanding of other people’s attentional and intentional states and how those states relate to
what is likely to be communicated (Tomasello, 2003).
[It is because in order to become efficient word learners, young children must come to understand, for
example, that a novel word they hear probably relates to an object or event that the speaker is paying
attention to. If the infant simply assumes that the word she hears relates to whatever is present or that
it refers to whatever she herself is attending to, she will be relying solely on associative learning and no
doubt make many mismappings.
Graham, Nilsen, Collins, and Olineck (2010) have shown that by 24 months of age, infants hearing an
adult produce an unfamiliar label check to see whether the adult is attending to the same object or
event they themselves are and, if not, adjust their own focus of attention to match the adult’s.
[The ability to establish and maintain joint focus of attention with those around them, as well as a basic
understanding of others’ intentions and goal-directed actions, is crucial for children’s efficient word
learning (L. Bloom, 2000; Tomasello, 2003).
Other theorists have suggested that children are aided by a number of lexical principles that constrain
the number of possible word–referent mappings. For example, young children may tend to assume
that a new word they hear refers to an object (Golinkoff, Mervis, & Hirsh-Pasek, 1994) and, further,
that the word refers to the whole object rather than to its parts (Markman & Wachtel, 1988).
[So, these two tendencies together may predispose the child to eliminate the family dog’s floppy ears or
the way he runs around the living room as likely referents for the label dog.]
Other lexical principles suggest that children tend to avoid two labels for one referent (Hansen &
Markman, 2009).
Principle of mutual exclusivity - is a cognitive bias observed in language acquisition, particularly in
young children, where they tend to assume that each object or concept has only one label or name.
This assumption leads children to avoid assigning multiple labels to the same object, preferring instead
to assign a single label to each object they encounter. [the child in our example will be inclined to
eliminate Rufus as a possible referent for bone, because Rufus already has a name.]
[However, Clark (2007) proposes a slightly different child principle that words contrast in meaning.]
Principle of contrast - refers to the tendency to assign distinct labels to objects or concepts based on
their perceived differences. [In this principle, the child doesn’t necessarily eliminate existing labels
when encountering new ones. Instead, the child assumes that the new label ("bone" in this case) doesn't
perfectly overlap in meaning with the existing label ("Rufus" representing the dog's name). This
assumption allows them to understand that the new word refers to something different from what they
already know, in this case, a bone instead of the dog named Rufus.]
Bélanger and Hall (2006) found that, by 20 months, most toddlers can use the presence or absence of
an article to distinguish between proper names (“that’s Daxy”) and count nouns (“that’s a Daxy”).
[And as they get older, Hall (1994a) showed that children used their world knowledge that dogs often
have proper names, whereas caterpillars do not, to interpret zav as a proper name in a sentence such as
“This dog is Zav,” but as an adjective in the sentence, “This caterpillar is zav.]
According to Hollich, Hirsh-Pasek, and Golinkoff (2000), children take advantage of multiple cues in
learning words and weight those cues differently at different points in development.
Beginning of word learning
children may give more weight to perceptual information, such as the concreteness of an action
or the visible shape of an object, and only later draw more heavily on social and linguistic cues.
[they initially base their understanding of words on what they can see or experience directly. However,
as they continue to develop and interact with others, they start to incorporate social and linguistic cues,
such as context, gestures, and language use, into their understanding of words.]
Preschool Years
children adhere less strictly to lexical principles such as the whole-object bias.
[It means that children grow older and their language skills develop, they may become more flexible in
their interpretations of new words. They may start to understand that words can refer to parts of
objects, attributes, actions, or even abstract concepts, rather than just the whole object itself. For
example, they may be particularly likely to learn the word fur when hearing it used in conjunction with a
familiar object and marked with possessive syntax: “Look at the doggie’s fur!”]
Fast Mapping
18-months-old
can make an initial word–referent mapping after only a few exposures to a new word, often also
without explicit instruction by an adult.
[Through natural exposure to language in their environment, coupled with their innate cognitive
abilities, such as pattern recognition and associative learning, toddlers begin to form connections
between words and their meanings even without the guidance or instructions by an adult.]
Carey and Bartlett (1978) first demonstrated fast mapping by providing 3- and 4-yearolds with
exposure to unfamiliar words in the course of classroom activities.
[So, these children were not taught the words explicitly but were simply asked, for example, “Bring me
the chromium tray, not the blue one, the chromium one.” The researchers found that most children
remembered something about the sound and meaning of the target word (such as that it was a color
word) a week later. Later research showed that fast-mapped labels are remembered by preschoolers for
at least a month (Markson & Bloom, 1997), a capacity that probably helps ensure that new words will
not be forgotten quickly if they are encountered infrequently.]
[Although children’s memory for nonlinguistic facts is inferior to adults’, Markson and Bloom found that
children remembered fast-mapped words over several weeks’ time as well as adults. As with most other
kinds of learning, exposure distributed over several days makes for more successful word learning than
the same number of exposures concentrated in a single day (Childers & Tomasello, 2002). In this
respect, 2-year-olds learning novel words and college students studying for exams seem to adhere to
similar learning principles. ]
Children ages 2 and older may learn nouns as effectively through incidental learning as through
ostension (i.e., when objects are labeled explicitly)
For example, [upon hearing a sentence like] “Mother is feeding the ferret.” [Children may use what
they know about the meaning of the word feed (that one only feeds animate objects) to seek out an
animate referent for the word ferret]
THE STUDY OF EARLY SEMANTIC DEVELOPMENT
By early in their second year, most children have begun to produce some words themselves.
Such as mommy, daddy, doggie, and water tend to be common early words across children;
while vase, and policeman are not.
[These patterns indicate that children begin with words related to what is socially and
intellectually most meaningful to them (Anglin, 1995).]
Early studies on language development, dating back to the 18th century, often relied on
observations of authors' own children recorded in diaries. This method continued into the
19th and 20th centuries, with many psychologists keeping diary records of their children's
development.
[However, even though diaries are valuable for tracing individual children's language
development, they can be biased because authors may be inclined to record unusual or
interesting events rather than daily occurrences. And so, researchers have developed ways to
enhance diary studies by providing parents with checklists of words their children are likely to
acquire. This helps parents focus on ordinary language development milestones that they might
otherwise overlook.]
One commonly used tool for tracking early language development is the MacArthur-Bates
Communicative Development Inventories (CDI)
[It provides checklists for parents to report their children's language skills, allowing researchers
to validate these reports against the children's actual language use. Additionally, the
MacArthur-Bates CDI has been translated and standardized across various communities and
languages, allowing for the comparison of vocabulary development norms across different
linguistic backgrounds.]
These norms indicate general developmental progression patterns for typically developing
children throughout their first 30 months of life for both (a) productive vocabulary (words
children have learned to produce within appropriate contexts) and (b) receptive vocabulary
(words children have learned to understand to varying degrees)
WHAT ARE EARLY WORDS LIKE?
children’s early words tend to fulfill a social purpose.
[Early words like "mama," "dada," or "hi" help children establish connections with their caregivers and
other individuals in their environment. These words are often among the first learned because they are
essential for initiating and maintaining social relationships.
Children are exposed to a significant amount of language before they begin acquiring their
initial vocabulary.
[This exposure occurs through interactions with caregivers, family members, and the surrounding
community. Even before they begin speaking, children listen to the language around them, absorbing
words, phrases, and intonations. This early exposure lays the foundation for language acquisition by
familiarizing children with the sounds, rhythms, and patterns of their native language. It provides the
necessary input for their developing language skills and sets the stage for the emergence of their first
words and phrases.]
Children's early vocabulary tends to consist of words that label concrete objects in their
environment.
[These words are tangible and observable, making them easier for children to understand and learn.
Examples of such concrete object labels include "ball," "car," "dog," "cup," "book," and "banana." These
words represent physical entities that children can see, touch, or interact with directly. By learning these
concrete object labels, children begin to build their vocabulary and develop the ability to communicate
about the objects and experiences in their world.]
Vocabulary acquisition in early childhood follows a trajectory of gradual increase, starting
with learning approximately one word per week during the second year of life.
[This slow but steady increase in vocabulary reflects the natural progression of language development as
children become more exposed to and engaged with their linguistic environment.]
By the end of the second year and throughout the first 5 years of life, the rate of word
learning intensifies, with children learning an average of one new word every two waking
hours.
Some studies identify a vocabulary spurt around 18 months, characterized by a rapid increase
in the number of words learned. However, not all children display this abrupt increase.
[Due to various reasons such as individual differences, environmental factors, cognitive abilities,
language input, and developmental context.]
Vocabulary acquisition is generally described as a gradual process, with children becoming
more skilled word learners over time.
[It is a gradual process because children must be exposed first to a lot of processes such as they acquire
new words through exposure to language, contextual learning, semantic mapping, social interaction,
cognitive development, and metalinguistic awareness.]
The words that children acquire in their early productive vocabularies are influenced by many
factors. Early words tend to share phonetic features, occur frequently in the speech that
children hear, and be shorter in length than later-acquired words (Storkel, 2004).
[Firstly, early words often share phonetic features that make them easier for children to pronounce.
These words typically have simpler syllable structures or commonly occurring sounds, facilitating their
production by young children. Secondly, words that occur frequently in the speech children hear are
more likely to be included in their early vocabularies. These high-frequency words become familiar to
children through repeated exposure in their linguistic environment. Additionally, early-acquired words
tend to be shorter in length compared to later-acquired words because shorter words are easier for
children to pronounce and remember.
Researchers have analyzed the phonology of children’s first 50 words, studied children’s
imitations of words, and tried to teach new words to 1-year-olds (Schwartz & Leonard, 1982).
[The results of these studies show that words that are easier for children to pronounce are more
likely to be included in their early productive vocabularies but also that favored sound patterns
vary greatly across children.]
Research suggests that at around 20 months of age, children can use specific phonetic
information in learning similar words.
[An example of this is when a child learns the word "cat" and then successfully applies the phonetic
pattern of that word to learn similar words like "bat" or "hat."]
Nespor, Peña, and Mehler (2003) have proposed that vowels and consonants play different
roles in speech processing and language acquisition, arguing that consonants are more salient
in word identification. Before the moment of the first reliable adult-like production of a new
word, young children attempt over time a sequence of sounds that increasingly resembles the
adult word, and they tend to use this form consistently in the appropriate context (Roy, 2011).
[For example, consider a child learning the word "ball." Initially, the child may produce a simplified
version of the word, such as "ba" or "baw." As the child's language skills develop, they refine their
pronunciation, eventually producing the word "ball" accurately. Throughout this process, the consonant
sounds, such as "b" and "l," may be more salient and easier for the child to identify and reproduce
compared to the vowel sound "a."]
Among nouns, those that are the easiest to distinguish from the surroundings, such as
animate beings or things that move, are the earliest learned (Gentner, 2006).
[It is because children are naturally drawn to things that are dynamic and catch their attention,
making them more likely to focus on and learn the names of these entities. Additionally, nouns
are often highlighted in the first or last position of mothers’ utterances to children, making them
more salient. Also, nouns are favored over verbs in acquisition because nouns tend to refer to
clearer, more concrete, and more readily identifiable referents than verbs.]
The linguistic and conceptual complexity of verbs may be one reason that children initially rely
on general-purpose verbs such as do, go, make, and get (Clark, 1993).
[It is because they are simpler and more versatile in their applications. And apparently, adults
also encounter difficulties in mapping and understanding the verbs, indicating that verb
acquisition presents challenges across different stages of language development.]
A recent proposal for a “unified developmental theory of word learning” argues that
grammatical categories do not offer the best explanation to understand early word learning.
Maguire, Hirsh Pasek, and Golinkoff have proposed the “SICI continuum” of Shape,
Individuability, Concreteness, and Imageability.
[So, instead of debating whether children first learn nouns or verbs, this theory proposes a more
comprehensive explanation, arguing that, regardless of grammatical form, children’s early words
tend to refer to concepts with a reliable and consistent shape, that can more easily be
distinguished from others (individuability), that are more perceptually salient and concrete
(concreteness), and that can easily yield a mental image (imageability).]
[Let us take into consideration the many different types of nouns and verbs in a language, this
theory predicts why nouns that are highly abstract (such as idea or justice) tend to be learned
much later in life than other nouns (such as dog or cup) as well as why many concrete verbs,
such as dance or eat, are learned early in development.]
Children often add several new words for one semantic domain all at once.
For example, when 1-year-old Damon learned ant, bug, and ladybug all in 1 week, and frog,
snake, and alligator the next.
[Semantic networking can also be observed using] Preferential looking paradigms – are
experimental techniques used in developmental psychology to study the visual attention and
perceptual preferences of a child.
[In these experiments, a child will be presented with visual stimuli, typically displayed
simultaneously on two or more screens or areas of interest. By observing which stimulus a child
preferentially looks at or spend more time attending to, researchers can infer the child’s
perceptual abilities, preferences, and cognitive processes.]
Children seek links, relations, and conceptual wholes in everything they experience, including
language.
[As such, it is not surprising that children’s earliest words can be predicted by the numbers of
word associates that they have in the input; that is, the more semantic neighbors that a word
has in the parent’s input to a child, the faster such items are learned.]
Semantic links are also evident in young children’s inappropriate use of certain words after
they have learned the appropriate use.
In such cases, Bowerman (1978), who thoroughly detailed early stages of one child’s speech,
observed that there was some semantic overlap between the word used incorrectly and the
correct word.
[For example, consider a child who has learned the words "big" and "small." A child might
initially learn these words to describe size differences, such as big and small toys. However, due
to the semantic overlap between these words in terms of comparative size, the child
might incorrectly use "big" to describe something that is actually small.]
