In this unit you will revisit two renowned theorists in the field of Psychology: Jean Piaget and Lev Vygotsky. Each of these innovative thinkers has put forth ideas about how children learn language in relation to their views on learning and intelligences

Early Semantic Development: The Developing Lexicon Nina C. Capone, PhD, William O. Haynes, PhD, and Kristy Grohne-Riley, MA

OBJECTIVES ____________________________________________________ To learn the milestones of early lexical–semantic development To understand the course of word learning over time in development To understand factors that influence word learning in children

INTRODUCTION _________________________________________________ The child spends one year preparing for his first word with motor, pragmatic, cognitive, and phonological developments. For example, the child’s ability to develop fine motor movements of the hand and mouth depends on the gross motor development of the body. Stability in the trunk muscles supports sitting at approximately six months of age. This development provides a stable base for mobility in the hands to explore objects and mobility for the muscles of the mouth for babbling and later word productions.

At the onset of first words, the child’s utterances are one word in length until the child has accumulated an expressive vocabulary of approximately 50 words (Nelson, 1973). It is then, at the 50-word milestone, that these utterances are expanded into early word combinations of mostly two- and then some three-word utterances. It is also at this time that children begin to accrue new vocabulary at an exponential rate. This transition from word learning being a slow process to a rapid fire is termed the word spurt (Bates et al., 1994; Goldfield & Reznick, 1990). The single-word period typically extends from about 12 months of age until the child is nearly 2 years old.

Children, of course, have not read the developmental literature indicating exactly when they should enter and leave the single-word stage, so there is considerable variability among youngsters in traversing the single-word period. This chapter discusses the developmental

CHAPTER 7

198 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

course of early semantic milestones, the course of word learning, and factors that influence the child in learning a word.

PREPARING THE FIRST YEAR: PERLOCUTIONARY STAGE _________________ Bates (1976) defined the development of communicative functions as consisting of three stages: perlocutionary, illocutionary, and locutionary. From birth to approximately 8 to 10 months of age, babies are in the perlocutionary stage of communication development. In this stage, the infant produces vegetative sounds (e.g., burping), sound play (e.g., coo-goo), and other behaviors (e.g., eye gaze) to which the adult infers a communicative intent. For instance, a child looks at a toy and the adult infers that the child wants it. Communication has taken place even though the child has not necessarily intended a specific message to the adult. Vegetative sounds such as cries and burps are also interpreted by the adult. The adult often responds to these with lan- guage (Oh, you’re hungry), and so begins the child’s journey onto a road of linking the intention (pragmatics), the meaning (semantics), and the form (phonology, morphology, and syntax) of language. Three important behaviors that are necessary for successful communication through- out the life span are already observed during the perlocutionary stage—namely, eye contact, joint attention, and turn-taking. These behaviors were discussed in Chapter 5.

Prelinguistic speech behaviors are also observed during this stage of development, includ- ing vocal play and babbling. These speech behaviors do not have meaning or intention to com- municate. Nevertheless, a number of studies have shown that certain types of babbling appear to have some connection to later lexical development (e.g., de Boysson-Bardes & Vihman, 1991; Stoel-Gammon & Cooper, 1984; Vihman, 1996; Vihman, Macken, Miller, Simmons, & Miller, 1985). Chapter 8 discusses the relationship between babbling and first words.

INTENT TO COMMUNICATE: ILLOCUTIONARY STAGE ____________________ The second stage of communication is illocution. In this period the child displays the inten- tion to communicate with gestures and nonlinguistic vocalizations. A major difference be- tween the perlocutionary stage and the illocutionary stage is that the child has clear intention to communicate in the latter stage—it is not merely an adult inferring a communicative intent. The illocutionary stage is typically characterized by the child using gestural communication or using gestures accompanied by vocalizations (e.g., jargon, protowords) that are not words. Primarily, children attempt to communicate at this time for two purposes: (1) to regulate joint attention with the adult and (2) to regulate joint action with the adult (Bruner, 1975). For example, a child will repeatedly show an object to an adult to gain attention; when a child wants to have a wind-up toy activated, she will physically manipulate the adult by moving

The First Word: Locutionary Stage 199

the adult’s hand to the toy. The latter behavior is referred to as a ritual request gesture, as dis- cussed in Chapter 6.

Children progress from a period of babbling to a point at which they begin to stabilize cer- tain vocalizations around specific situations, events, and objects. These vocalizations have been called various names by researchers, including protowords (Halliday, 1975), prelexical forms (Bates, 1976), and phonetically consistent forms (PCF; Dore, Franklin, Miller, & Ramer, 1976). These vocalizations do not meet the criteria for classification as true words because they do not resemble adult productions. Prelexical transition utterances tend to be associated with a situa- tion or event rather than with a particular physical referent. They are regarded by many as sim- ply a part of an activity rather than language that is referential in nature. Dore and colleagues (1976) indicate that these forms have the following characteristics:

They are independent units bounded by pauses.• They are produced repeatedly, at different times.• They are correlated with recurring conditions.• They are phonetically more stable than babbling but not as stable as word productions.•

These PCFs are not necessarily attempts at word approximations and are highly individual to specific children. One child might say /na na na/ every time he protests; another child may produce a different vocalization for protest. The key point is that the vocalization will be simi- lar for each protest situation for an individual child.

One example of a PCF is an affect expression, in which a child stabilizes a vocalization around an emotional reaction such as anger, frustration, joy, protest, or some other affective situation.

A second PCF is an instrumental expression, in which a child vocalizes a consistent pattern whenever he is attempting to regulate adult behavior to obtain goods or joint activity. An exam- ple would be a child who vocalizes /uh/ whenever he is attempting to regulate an adult. Again, a child who says /uh/ to be lifted up is not necessarily saying up because the same vocalization is produced when the adult is enlisted to open a jar or wind a toy.

A final example of a PCF is an indicating expression. The indicating expression is produced in concert with pointing, where the goal is the direction of adult attention and not necessarily the regulation of action. The child may say /ba/ every time he points to an object to direct adult attention.

THE FIRST WORD: LOCUTIONARY STAGE _____________________________ Most speech-language pathologists wish they had a dime for every parent who reported that their child spoke his first word at the age of six months. Parents typically scan the output of

200 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

their infants with great interest early on for evidence of the first word, because this event marks the beginning of verbal communication. There is an important difference, however, between a six-month-old uttering /mama/ randomly and an adult using many words in specific, consis- tent, and appropriate contexts. The parent’s criterion for the first word is probably phonologi- cal in nature, not semantic (Kamhi, 1988). Thus, if a word sounds like an adult word, the child is given credit for using it, even though it may have been void of semantic or pragmatic intent.

The notion of a true word should be separate from the mere uttering of a bisyllable by an infant. In babbling, infants often approximate an adult word such as /di/ and parents might say “Didn’t she just say dish?”, when actually it was just a phonetic accident that the child could never duplicate. A true word has to have “a phonetic relationship to some adult word” and the child must “use the word consistently to mark a particular situation or object” (Owens, 1988, p. 199). These criteria rule out babbling or jargon as true words.

The locutionary stage of communication begins at approximately 12 months, when the hearing child utters his first word. Twelve months is only the average age, however: Children typically vary in this milestone from 11 to 13 months. The child might say ba to indicate the bottle, say up and extend the arms toward the adult in an attempt to be picked up, or say doggie to direct the adult’s attention to an oncoming canine. Nelson (1973) reported that the mean age of a 10-word vocabulary was 15 months, with a range of 13 to 19 months of age.

During the time that children are acquiring these first 10 words, their vocabulary is unsta- ble. That is, they tend to have appearance and then disappearance of words. This phenomenon should not be confused with vocabulary regression, which is reported by parents of children with language impairments. In typical development, it is not uncommon for children to use a word for a few weeks and then stop using it. However, this disappearance is accompanied by new words being used. For children with language impairments, new words do not appear; rather, they experience a loss of using words more generally.

Nelson (1973) reported the mean age for acquisition of the 50-word lexicon was 19.75 months—though children vary widely around that milestone, with a range of 15 to 24 months. Word learning is fairly slow in the first half of the second year as compared to the second half of that same year and beyond. The 50-word lexicon milestone is important, however, because two key milestones follow it: Around the time the child accumulates 50 words, he enters a rapid period of word acquisition (i.e., the word spurt) and begins to combine words.

Between 18 and 24 months of age, most children will exhibit what has been variously re- ferred to by researchers as vocabulary spurt, naming insight, nominal insight, or naming explosion (Bates et al., 1994; Bloom & Capatides, 1987; Goldfield & Reznick, 1990; Gopnik & Meltzoff, 1986). All of these terms refer to a discernable rapid increase in the number of words a child is learning and using. Various changes in development of related skills have been investigated to account for this sudden and rapid increase in word learning (e.g., object permanence; the ability

The First Word: Locutionary Stage 201

to categorize objects—Gopnik & Meltzoff, 1986; temperament—Bloom & Capatides, 1987; connectionist learning algorithm—Plunkett, Sinha, Moller, & Strandsby, 1992). Findings are inconsistent in terms of whether these variables relate to the word spurt, however, and there re- mains no definitive explanation as to why or how these factors contribute to this rapid increase in the rate of new word learning.

Bloom (2004) has argued that the word spurt is a “myth” (p. 205) because other factors could account for the findings of an increase in rate of word learning. Two possible reasons Bloom suggests for the observed word spurt are the child’s general talkativeness and the way in which the researcher of a study defines a word spurt (e.g., a child may be defined as having a word spurt when she acquires 12 new words in a 2-week period). On the first point, it is pos- sible that the number of words children produce is simply a function of how much they talk, not how many new words they know. Put another way, the more one talks, the more words a researcher is likely to find (p. 213). On the second point, Bloom argues that researchers may define a child as having had a word spurt when 12 new words were acquired within a 2-week period but do not reference how many words were acquired the previous 2 weeks. The child may have learned 10 new words and then 12 new words in each respective time period, indicat- ing a gradual increase in vocabulary rather than a sudden shift in word learning. These ideas are controversial, so it is wise to remember that children start out learning words slowly and then get better at it. It is unclear whether there needs to be a separate cognitive or social mechanism to explain this phenomenon.

The timely acquisition of early vocabulary milestones is important for continued lan- guage development. Marchman and Bates (1994) have suggested that the size of the lexicon must reach a necessary threshold if other language domains are to fully develop, particu- larly morpho-syntax. For example, the word spurt (if one exists) and two-word combination milestones tend to emerge after children have a threshold of at least 50 words in their lexicon. Nelson (1973) also found that vocabulary size at 2 years of age was related to mean length of utterance (MLU—a measure of syntactic development) at 30 months of age. Children who do not meet early vocabulary milestones in a timely manner are referred to as late talkers. While many late-talking toddlers outgrow their early delay and are reclassified as late bloomers, a pro- portion of these children persist in language impairments that transcend the domain of vo- cabulary to those of morphology, syntax, and written language (e.g., Rescorla, Dahlsgaard, & Roberts, 2000; Thal & Tobias, 1994; Thal, Tobias, & Morrison, 1991).

To see how this development is assessed, consider the early semantic milestones of the three case studies from Chapter 1. Johnathon was reported to speak his first word at 12 months, Josephine did not speak her first word until 15 months, and Robert’s first word emerged at 18 months. Therefore, at the time of the evaluation, Josephine and Robert al- ready had a history of delayed semantic development in expressive vocabulary. Josephine was

202 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

22 months at evaluation and had only 7 words in her vocabulary. Robert was 27 months at evaluation and had only 14 words in his vocabulary, some of which were sound effects. Only Johnathon was using new words and combining words. Josephine and Robert were both de- layed for learning new words regularly and for combining words. Josephine and Robert have a history of delay that is persisting into toddlerhood. Their semantic delay could now be char- acterized as having (1) a small expressive vocabulary size for their age, (2) too few new words being added to their expressive vocabulary, and (3) a delay in combining words. Finally, while Johnathon relies on words to communicate, Josephine and Robert still rely on gestures and nonlinguistic vocalizations. Johnathon is evaluated as typical for his age, whereas Josephine and Robert show delays in vocabulary development that place them in a category of language impairment. During this time in development, we refer to them as late talkers.

A PREPONDERANCE OF NOUNS ____________________________________ Nelson (1973) analyzed the word classes present in the first 50-word lexicon using the classifi- cation system of specific nominals, general nominals, action words, modifiers, personal–social words, and function words. (See Table 7-1 for definitions of these terms.) The proportion of total vo- cabulary that each word class (e.g., noun, action) represents in the early lexicon is remarkably

TABLE 7˜1 Word Classifications

Word Class Definition

Specific Nominals These words refer to a specific exemplar of a category whether or not it is a proper name (e.g., mommy, daddy, pet’s name).

General Nominals These words refer to all members of a category and include classes such as objects, substances, animals, people, letters, numbers, pronouns, and abstractions.

Action Words These words are used to describe or demand an action.

Modifiers These words refer to properties or qualities of things or events, such as attributes, states, locations, or possessives.

Personal–Social Words

These words express affective states and social relationships such as assertions (e.g., no, yes, want) and social expressive words (e.g., please, ouch).

Function Words These words refer to items that serve a grammatical function in relating to other words (e.g., what, is, for, to).

Source: Nelson (1973).

A Preponderance of Nouns 203

similar across children. General nominals (i.e., common nouns) make up the largest proportion of the lexicon, accounting for 51% of vocabulary. Action words and specific nominals each ac- count for 14% of the lexicon, and modifiers and personal–social words each represent 9% of the lexicon. Only 4% of the vocabulary is made up of function words. This same classification system is used today in clinical analyses of early language (e.g., Retherford, 2000).

Despite significant differences in language and culture, children learning Mandarin Chinese, Japanese, Kaluli, German, Italian, Hebrew, and Turkish are also observed to have more nouns than other word classes in their early lexicon (e.g., Caselli et al., 1995; Gentner, 1982; Goldfield, 1993). The universal nature of noun preference in early lexicons has garnered some attention by language researchers, and several hypotheses about the early predominance of nouns have been put forth.

For example, is there a frequency effect? That is, are nouns used more often than verbs? This is unlikely because even though adults use more nouns than other word classes, there is a larger variety of nouns. This means that individual nouns are used less frequently than the smaller lexicon of verbs in a language (Goldfield, 1993). With fewer verbs to use, adults will use each more often. Put another way, the pattern should be reversed (more verbs than nouns in the early lexicon) because children actually hear each of the smaller number of verbs more often than each of the larger number of nouns.

Another explanation might lie in the way parents teach language to their children. When speaking to their child, do adults organize their language around object naming? American par- ents largely focus heavily on object naming with their children (Nelson, Hampson, & Shaw, 1993). However, in Kaluli, parents are not particularly interested in teaching object names to their children, and Korean-speaking parents tend to focus their interactions around actions and activity (Kim, McGregor, & Thompson, 2000). With little difference in the early lexicons of children from these disparate cultures, it would appear that this is not a salient factor in noun acquisition.

Two factors that have received more attention for their part in this phenomenon are (1) the ordering of word classes in a sentence of a language and (2) conceptual differences between nouns and other word classes. We address each of these issues in turn.

The ordering of word classes in a sentence generally refers to the order of nouns, in subject (S) and object (O) position, and verbs (V). In English, we might say The girl is reading a book, where The girl is the noun phrase that occupies the subject position, a book is the noun phrase that occupies the object position, and is reading is the verb phrase that occupies the verb posi- tion. In English, nouns occupy the most salient positions in sentences. For example, the speech signal tends to change at the end of sentences and clauses (e.g., greater stress, elongated vowel, or fricative duration), making the words and morphemes that occur in this position more sa- lient for the listener. It is possible that children acquire more nouns early on because they occur

204 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

in salient sentence positions, and particularly in final sentence positions. In contrast, languages such as Korean, German, Kaluli, and Turkish have verbs in the final positions of sentence (S-O-V; O-S-V)—yet children learning these languages have lexicons that are predominantly made up of nominals as well.

Gentner (1982, 2006) proposes that concrete nouns promote more rapid learning than other word classes—particularly verbs—because they allow for greater transparency of the map- ping between lexical and semantic information (natural partitions hypothesis; Gentner, 1982, p. 327). Nouns represent more perceptually stable entities than other word classes. Other word classes (verbs, prepositions, modifiers) are slower to be acquired because relationships among entities are less spatially cohesive and concrete than they are for objects. Predicates, for ex- ample, have more ambiguous relations to the perceptual world, and their meanings are harder to glean in a single exposure. Finally, because object concepts are concrete, they allow children to bootstrap into the language system. Bootstrapping is a term that refers to the child’s use of known information to infer unknown information. In essence, knowing object concepts gives children a foothold in the speech stream, and it creates a scaffold from which to learn other word class–semantic relations. Simply put, learning nouns helps to you learn verbs, and so on.

It is likely that both language input and conceptual factors contribute to the universal pattern of noun learning. Kim, McGregor, & Thompson (2000) set out to compare early vocabularies of English- and Korean-speaking children to confirm or refute the claim that nouns dominate early vocabularies universally, and to explore why that might be. Specifically, their study followed eight English-speaking and eight Korean-speaking mother–child dyads to examine the influence of ma- ternal language on the child’s early vocabulary. Children were followed from 16 months of age to 21 months of age, just before or around their 50-word milestone. Kim and colleagues (2000) found that both Korean- and English-learning children acquired significantly more nouns than verbs, despite the differences the researchers observed between the two cultures in lan- guage input and parent activity. This dominance of nouns over verbs was found for Korean- speaking children even though their caregivers tended to emphasize verbs in their interactions with them. As expected, English-speaking caregivers tended to emphasize nouns in their talk to the children. These results suggest that all children come to the task of early word learning with a strong predisposition toward linking nouns to objects.

That being said, there was also an influence of the frequency and saliency of verbs in the Korean-speaking input. Specifically, even though Korean-speaking children had more nouns than verbs in their vocabulary, they had a greater proportion of verbs in their vocabulary as compared to the number of verbs in the English-speaking children’s vocabulary. Therefore, while children appear to have some universal predisposition for noun learning, their environment also shapes the developing lexicon to some extent. In this case, the environment includes social activity and surface structure of the grammar that the child hears.

Expressive versus Referential Word Learners 205

This pattern of noun learning over other word-class learning continues to be observed in the development of preschool and early school-age children, with and without language impair- ments. For example, two author teams—Rice, Buhr, and Nemeth (1990) and Oetting, Rice, and Swank (1995)—examined word learning after just a brief exposure to vocabulary items from a cartoon video. Both studies found that typically developing children and children with language impairments showed a preference for learning object terms over other word-class items, includ- ing action, attribute, and affective terms. Because non-object labels are difficult to map in a single exposure, more experience and contexts may be needed to map the meanings and linguistic spec- ifications of these word classes (Oetting, Rice, & Swank, 1995; Waxman, 1994). We discuss the idea that richness of experience influences acquisition of words later in the chapter. First, however, we return to the early lexicon and the importance of noun learning.

EXPRESSIVE VERSUS REFERENTIAL WORD LEARNERS ___________________ When Nelson (1973) analyzed the first 50-word lexicon, she found that although the majority of children had a preponderance of general nominals, there was a group of children for which this was not the case. Nelson classified children as referential if general nominals accounted for more than 50% of their total vocabulary. Those children with general nominals accounting for less than 50% of their lexicon were referred to as expressive. The expressive children used many personal–social words, whereas the referential children used many nouns (see also Bates, Bretherton, & Snyder, 1988; Bloom, 1973; Goldfield, 1986, 1987; Snyder, Bates, & Bretherton, 1981). Snyder, Bates, & Bretherton (1981) found the expressive–referential distinction in children as young as 13 months, when the children had only 10 to 12 words in their lexicons. Note that expressive children and ref- erential children should not be viewed dichotomously. Indeed, most authorities indicate that chil- dren who are categorized as expressive versus referential probably represent extreme points on a continuum, rather than separate classifications or typologies of language-developing children (Bates, Bretherton, & Snyder, 1988; Nelson, 1981; Nelson & Lucarielly, 1985).

As research began to focus on the differences between expressive and referential children, it was found that certain characteristics, in addition to lexical composition, were often observed in each group. For instance, referential children were reported to develop language earlier and more rap- idly than expressive children did (Bates, Bretherton, & Snyder, 1988; Horgan, 1978, 1979; Nelson, 1973; Ramer, 1976).

Specifically, Bates and colleagues (1988) noted that referential children had larger vocabular- ies and reached morpho-syntactic milestones sooner than expressive children. Referential chil- dren had a word spurt, whereas expressive children tended to learn words at a slow and steady pace without a word spurt. Referential children showed greater growth of verb vocabulary at 20 months and more productive control over function words (e.g., determiners, prepositions)

206 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

by 28 months. Initially, expressive children had more function words in their lexicon because they used phrases as holistic chunks. That is, they did not analyze the words of a phrase as in- dividual words but rather as one lexical item. A U-shaped curve in function word development was observed: Once expressive children began analyzing individual words, these functional items dropped out and reappeared later as productive forms. Also, referential children tended to use words in context-flexible ways (e.g., saying cup to refer to many cups, not just their cup). They would decontextualize their word use by talking about absent objects. Decontextualization refers to the gradual distancing of a symbol from the original referent or learning context (Werner & Kaplan, 1963). The ability to decontextualize language is important for spoken and written language development as the child ages, particularly for social, emotional, and academic success.

The relationship between size of object vocabulary and vocabulary growth was also ob- served in the late talkers studied by Rescorla, Mirak, and Singh (2000). These authors found that even among children who were delayed in language development, if they had larger vo- cabularies (although smaller than is typical), the children tended to have greater growth in total vocabulary size over time than did the children who had the smallest vocabularies.

INNATE BIASES MAKE WORD LEARNING EFFICIENT _____________________ If children relied on explicit teaching (i.e., “this is a cup,” “this is another cup”) to learn each word of their language, language learning would be a laborious, effortful, and inefficient task. Luckily, children come to the task of word learning with some innate biases (also referred to as constraints or principles of word learning) that help them narrow down the many possible refer- ents that could be paired with a word that they hear. For example, if mother says, “Oh, I see the dax,” where dax is a novel word, the child must figure out which of the many objects, parts of objects, actions, and events going on in the room that dax is referencing.

Several biases can help the child achieve this feat. First, the child must differentiate between spoken language and other sounds as labels. The principle of reference states that words—but not other sounds—label objects, actions, and events (e.g., Hollich et al., 2000). When Balaban and Waxman (1997) presented nine-month-olds with objects paired with either a tone or a word, the infants showed a preference to link a word—but not a tone—with an object. This occurred even though both tones and words captured their attention in the learning phase.

Three related biases that help the child map a new word to the right referent are the novel name–nameless category principle (N3C; Golinkoff, Mervis, & Hirsh-Pasek, 1994), the prin- ciple of mutual exclusivity, and the whole-object bias (Markman, 1989). The principle of mutual exclusivity states that if the child already has a name for an object (cup, comb), it cannot receive another name. The flip side of this bias is the N3C principle, which states that a novel word will be taken as the name for a previously unnamed object. For example, in experiments ex-

Innate Biases Make Word Learning Efficient 207

amining how children map novel words to unfamiliar objects, an array of three objects may be presented to the child: a cup, a comb, and a novel object previously unseen or unnamed. When the experimenter states, “Give me the dax,” the child will hand over the novel object because she already has names for the other objects. Children choose objects in this way even though they have never been explicitly taught, “This is a dax.” Finally, the whole-object bias guides the child to infer that the word label refers to the entire object and not just a part, an attribute, or its mo- tion. For example, the child knows that the word car refers to the whole entity, and not just the wheels, windows, or steering wheel of the vehicle.

Children also come to language learning armed with a principle of conventionality (Clark, 1993). That is, children know that there are culturally agreed-upon names for things and that these names do not change. Our use and understanding of language would be quite chaotic if the names given to things were ever changing. Children make this assumption early on.

Two additional biases that make word learning efficient for the child are the principle of ex- tendibility (Hollich et al., 2000) and the shape bias (e.g., Clark, 1973; Landau, Smith, & Jones, 1988). These issues are addressed again later in the chapter, so they are mentioned only briefly here. Put simply, the principle of extendibility states that a word does not refer to only one object, but rather refers to a category of objects, events, or actions if they share similar proper- ties. Therefore, a word will label all instances of an object if all of those instances have the same shape and/or function. The shape bias constrains word extension based on shared perceptual features of the original referent and the novel exemplar.

We can readily see the shape bias at work in some of children’s naming errors. For ex- ample, a child may look at the moon and say ball because both moon and ball are round. The shape bias has been extensively studied in terms of its relationship with word learning (e.g., Booth & Waxman, 2002; Gershkoff-Stowe & Smith, 2004; Jones, 2003). Gershkoff-Stowe and Smith (2004) found a positive parallel relationship in toddlers’ increasing use of the shape bias and an increase in their expressive noun vocabulary. Further, Jones (2003) found that late talkers—children defined by their small vocabulary size—do not demonstrate a shape bias. It would seem, then, that the shape bias is particularly important for vocabulary growth.

Even though a shared shape is one basis for word extension, a shared function between objects is also a salient feature that organizes a category of objects by one name (e.g., Booth & Waxman, 2002; Clark, 1973; Kemler Nelson, 1999). For example, Kemler Nelson (1999) assessed the role of shared function in word extensions made by 28-month-olds. These toddlers heard the names of objects and had opportunities to enact their functions. They were then tested on word extension to four types of objects: similar shape–similar function, similar shape–dissimilar function, dissimilar shape–similar function, and dissimilar shape–dissimilar function. In this study, toddlers were able to transcend shape similarities and extend words based on shared function. That is, they used the same name for two objects if the function was the same, even for objects that did not look alike.

208 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

Other object features, such as size of an object, have not been shown to influence extension de- cisions (e.g., Jones, Smith, & Landau, 1991). For example, children label a real car and a toy car both as car despite the obvious difference in size (and texture). Even though the size difference between their parent’s car and their Fisher-Price miniature car is quite large, children call both car because the items share shape properties that include four wheels, doors, seats, windshield, steering wheel, and other salient features.

Toddlers are also quite smart about adjusting their shape bias. When Jones, Smith, and Landau, (1991) placed eyes on their objects, children accepted shape changes and extended a word label to another object that shared texture (e.g., fur). In this case, the presence of eyes signaled that the object was an animate entity, and animate entities tend to share material (e.g., bears have fur, humans have skin, birds have feathers) and change shape (i.e., we change shape as we move). Therefore, children appear to adapt to and integrate multiple perceptual cues to make inferences about what something is or is not.

THE EMERGENT COALITION MODEL OF WORD LEARNING ________________ As stated earlier, little of language learning is accomplished by direct and explicit teaching. Instead, much of what the child learns is a result of what he infers about the language that he hears and what is going on around him. The previous section discussed the innate principles or biases that the child brings to the word learning task. Other cues in the environment also scaffold word learning for the child—for example, a parent’s point or eye gaze, the grammatical structure of the language model, the perceptual salience of the object. The emergentist coalition model (ECM; Hollich et al., 2000) describes how children coalesce environmental cues and in- nate biases to learn new words. In addition to the interaction between cues and strategies, the ECM posits that children calculate their success and failure rate of mapping words to referents. This error signal feeds back into the learning system to improve the reliability of coalescing these cues and strategies as the child develops.

A full range of cues is always available to the child, though he weighs each cue differently over time. With development and experience, the child learns which cues are more reliable indicators of word–referent pairings. In turn, the child’s accuracy in linking words to referents becomes more efficient. When a parent produces a word label, the child recruits attentional cues (e.g., perceptual salience of objects), social cues (e.g., eye gaze, pointing), and linguistic cues (e.g., regularities in syntax) in the environment to further constrain the possibilities of which referent that word is labeling. Infants aged 12 to 18 months initially consider percep- tual salience (e.g., a moving object) to be a more important cue about what is being labeled than a social cue (e.g., eye gaze or pointing to it). Therefore, the child infers that the inter- esting moving object is what is being labeled, even though the parent may be looking at the

Learning a Word 209

boring object. By ages 18 to 24 months, children have learned that the social cue is a more reli- able indicator of the word to referent mapping. Now when they hear a word, they pay atten- tion to the adult’s gesture and eye gaze to ensure a reliable word-to-referent mapping occurs. Put most simply, by toddlerhood children can accurately map the names of the boring objects as well as the interesting ones (Golinkoff & Hirsh-Pasek, 2006).

LEARNING A WORD ______________________________________________ Learning a word is not an all-or-nothing phenomenon. Instead, as Carey (1978) described, word learning is a gradual and long-term process. Learning a word grossly encompasses learning the lexeme (word form or label), the semantic representation (word meaning), and grammatical specifications (e.g., word-class information), plus making connections between the various rep- resentations. As children encounter words in a variety of contexts, over time their knowledge of individual words is enriched.

Carey (1978) delineated two phases of word learning: fast mapping and slow mapping. A word is considered fast mapped if there is an initial association made between word and refer- ent. A related phenomenon has been termed quick incidental learning (QUIL; Oetting, Rice, & Swank, 1995; Rice, Buhr, & Oetting, 1992). QUIL reflects more naturally occurring word learning situations—that is, situations that offer minimal environmental support with multiple cues in ongoing scenes.

In experimental studies, many new words may be presented as part of a cartoon program prior to the test of word learning. In contrast, the fast mapping phenomenon is often measured after a more explicit or structured task with fewer word–referent pairs needing to be inferred. The former scenario better matches the child’s natural word learning experiences. What is important for the speech-language pathologist to keep in mind is that a child will not know much about a fast-mapped word–referent pair. Lexical and semantic information are weakly represented in memory after such a brief exposure. Therefore, the child needs several to many exposures to a word, depending on the word, to truly discern the subtleties of its meaning. This happens during the slow mapping phase of word learning.

Slow mapping is the process of enriching lexical–semantic representations after a word is fast mapped into memory. The child’s representations are enriched through increased fre- quency of exposure and/or richer quality of exposure. This process has significant implications for the clinical practice of speech-language pathologists, because intervention provides the child with increased frequency of exposure to language and a richer quality of language learning. Quality of learning is enriched through the therapeutic scaffolds used by the clinician. Such therapeutic scaffolds could include the explicit use of verbal models of the word label being repeated with greater frequency for the child, or visual cues such as gesturing about the object

210 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

while labeling it, or organizing objects by category to help the child build connections between related vocabulary items.

At any given time, words may vary in the richness of their lexical–semantic representations simply because some words are encountered more often than others. For example, as you are learning the vocabulary of speech-language pathology, you are mapping many new terms into memory. Your knowledge of these terms is still weak as compared to that of the master-level student who is about to enter the profession. The master-level student has had many more exposures through additional coursework and richer quality of each experience, and through advanced reading and clinical practicum. The richness of these professional terms in your own vocabulary will evolve over time as you move through undergraduate and graduate study. Conversely, if this is the last class you ever take in speech-language pathology, then your lexi- cal–semantic knowledge of these terms will remain sparse and perhaps weaken as other words in your lexicon are continually updated and strengthened with repeated exposure.

AN ASSOCIATIONISTIC ACCOUNT OF LEXICAL° SEMANTIC REPRESENTATIONS

According to associationistic accounts, lexical and semantic information are stored and linked within a distributed neural network in the brain (e.g., Barsalou, 1999a, 1999b; Plunkett, Karmiloff-Smith, Bates, Elman, & Johnson, 1997). For example, the concept of bone com- prises visual information (shape, color), thematic associates (dogs chew bones), actions (chew- ing), the proprioceptive–tactile experience of feeling its weight and rough texture, and its lexeme (/bon/). Lexemes are further divisible into phonemes (/b/, /o/, and /n/). Each of these is a node of information (semantic, phonological, lexical) in the network; connections within and be- tween representations, in turn, allow for spreading of neural activation among semantic and lexi- cal nodes. For example, when you see a bone, your shape node is activated. Through connections in the network, there is activation of the other nodes as well. This includes the lexeme /bon/.

The strength of those nodes and connections’ activation weights influences whether nodes will be activated and information can be recalled from memory. When a semantic representa- tion is stronger, it has many connected nodes of information, which leads to a greater number of connections to the lexeme. This richer quality and quantity of connections provides an ac- tivation strength that will reach the activation threshold of the lexeme such that it, too, will be activated (i.e., recalled).

Each node and connection pair carries an activation weight (excitatory or inhibitory). Fast- mapped words or infrequently encountered words tend to have few nodes with weak activation weights. These lexical–semantic representations are not distinct from other weakly represented words that share some, but not all, features. It is through the slow mapping phase that lexical– semantic representations are enriched. Richer lexical–semantic representations are distinct

Naming Errors of Overextension and Underextension 211

because a greater number of nodes and connections are present within and between lexical– semantic representations, and because their activation weights are stronger. When semantic ac- tivation weights and connections to the lexeme are strong, the likelihood of recalling the lexeme (i.e., activating it in memory) is quite good.

For example, upon seeing an image of a pig, the child may activate several lexical–semantic representations that are weak if he has little experience with farm animals. These representations perhaps all have two ears, four legs, and a tail. Several lexemes may exist to name this weak repre- sentation the child has of farm animals in memory (e.g., pig, horse, and cow). However, as the child enriches his knowledge of pigs, horses, and cows, only the relevant lexeme will be activated when he sees one of these, because the distinct semantic features are being activated for one of them. As part of this process, the unique feature nodes activate the correct lexeme for retrieval. At the same time, inhibitory signals are sent to the horse and cow lexeme nodes to forestall their activation.

Before the child develops a richer representation, she is more likely to fail to retrieve the name of an intended word (e.g., Capone & McGregor, 2005; McGregor & Appel, 2002; McGregor, Freidman, Reilly, & Newman, 2002; McGregor, Newman, Reilly, & Capone, 2002). For example, the child may say doggie instead of cow, where cow is the intended target word. The next sections of the chapter deal with the issue of naming errors in children.

NAMING ERRORS OF OVEREXTENSION AND UNDEREXTENSION ___________ Parents and researchers have noted for years that children often misuse the words in their vo- cabularies. The early lexicon is in an almost continual state of flux and begins to stabilize only after the child’s word spurt (Gershkoff-Stowe, 2001; Nelson, 1973). When a child calls a cow doggie, he is exhibiting the limitations of his vocabulary. Limitations can take the form of either weakness of knowledge or an immature retrieval process (Gershkoff-Stowe, 2002).

When a child uses a word too broadly to refer to referents that may be similar in percep- tual feature or function, the error is referred to as an overextension. Some examples might be calling the moon a ball or calling a strange man daddy. Conversely, the child may also produce underextensions. Underextended words have too narrow a meaning. An example is the use of the word dog only when referring to the child’s pet, and not when referring to other dogs. Clark (1973) has estimated that overextensions and underextensions occur frequently and can repre- sent as much as one-third of a child’s early vocabulary between the ages of one and two years. Two traditional theories attempt to explain extension errors.

First, the semantic feature hypothesis (Clark, 1973, 1975) states that children classify and organize referents in terms of perceptual features such as size, shape, animacy, and texture. This phenomenon could explain some overextensions in which a child generalizes a word based on perceptual similarity (e.g., ball–moon).

212 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

Second, the functional core hypothesis (Nelson, 1974) states that words are overextended because of the actions or functions performed on objects rather than the perceptual features of the referents. Thus a child may say the word rake when a person is sweeping because the actions are similar.

Aspects of each of these traditional theories are present in the current theories of word retrieval in children. While it is true that children organize categories around perceptual (e.g., shape) or functional features, several other factors also influence word learning and retrieval. These factors include the phonological make-up of the word (e.g., high versus low phonotactic probability; Storkel, 2001), the frequency with which the word is encountered and/or practiced (Dell, 1990; Gershkoff-Stowe, 2002), and the richness of the word’s semantic representation (e.g., Capone & McGregor, 2005; McGregor, Friedman, Reilly, & Newman, 2002). For ex- ample, Storkel (2001) showed that preschoolers (ages three to six years) take advantage of how frequently phonological sequences is found in the language (phonotactic probability). When taught words made up of low-phonotactic-probability sequences and others consisting of high- phonotactic-probability sequences, preschoolers learned the high-phonotactic words faster.

The next section of the chapter delves into the nature of word retrieval/naming errors and explores how error and target are related in systematic ways.

NAMING ERRORS ________________________________________________ The retrieval errors that children (and adults) make are often logically related to the word that was targeted for expression. Therefore, errors in word retrieval reflect a speaker’s knowledge (Dell, 1990; Dell, Reed, Adams, & Meyer, 2000; McGregor, 1997).

For example, as the child’s vocabulary grows in size, the semantic system begins to take on a hierarchical organization. Children acquire basic-level (or ordinate terms) lexical items first (e.g., dog) and later develop a hierarchy of superordinate (e.g., animal) and subordinate (e.g., col- lie) terms. If a target word is dog, the child’s error is more likely to be animal than spoon, because dog and animal are more closely related (or connected) in the lexicon than are dog and spoon. These items are rarely, if ever, associated together in our experiences.

Word retrieval errors can actually relate to their targets in several ways:

Phonologically—for example, saying • chicken instead of kitchen or miracleride for merry- go-round Semantically—for example, saying • key for door or skating for skiing Phonologically and semantically—for example, saying • elevator for escalator An indeterminate response—for example, saying • thing or I don’t know A perseverative response—that is, using the same word to label different objects within • a set time interval A visual misperception—for example, saying • lollipop for balloon

Naming Errors 213

McGregor (1997) found that the most prevalent type of word retrieval errors were se- mantic errors and indeterminate errors. These errors indicate that retrieval failure is likely related to a weak or missing semantic representation of the target word or weak links between semantic knowledge and lexical labels (Gershkoff-Stowe, 2001; Lahey & Edwards, 1999; McGregor, 1997; McGregor, Newman, Reilly, & Capone, 2002; Plunkett, Karmiloff-Smith, Bates, Elman, & Johnston, 1997). McGregor showed that preschoolers with and without lan- guage impairments make the same types of word retrieval errors (predominately semantic er- rors), but children with language impairments make more errors.

During the word spurt, the toddler begins mapping many new words into memory. Given that the child has so many weak representations at this age, the word spurt is a good develop- mental time for examining the nature of word retrieval errors because the child makes so many during this time (Gershkoff-Stowe, 2002; Gershkoff-Stowe & Smith, 1997). Gershkoff-Stowe and Smith (1997) found that during this shift in rapidly learning new words, toddlers show a parallel trend in increasing naming errors. As their word acquisition increases so rapidly, they also have a sudden but temporary increase in retrieving words for the naming of known refer- ents. Word retrieval errors during the word spurt period may occur related to words that tod- dlers had named accurately during the pre-spurt period. Consistent with the findings of other studies, these naming errors are predominately semantically related to the target word, although perseverative errors are also prevalent. Phonological errors were significantly less common.

Perseverative errors subsequently decline in the post-spurt period (Gershkoff-Stowe, 2002). Gershkoff-Stowe and Smith (1997) suggest that such errors reflect a general fragility in the retrieval process and/or a weakness of word representations, reflecting the fact that children of this age are new word learners. Semantic errors persist in the post-spurt period. This rela- tionship is to be expected because the retrieval process itself has stabilized, yet new words are continually fast mapped. As a consequence, the child will always have a continuum of weak to richer representations. Weaker representations will be more prone to error in retrieving them for naming purposes (Capone & McGregor, 2005). As the child practices saying words and slow-mapping their semantics, the child may make increasingly fewer errors on those words that are repeatedly encountered and enriched (Gershkoff-Stowe, 2002).

Recall the associationistic theory of the lexical–semantic system. We can think of lexical– semantic representations as existing along a continuum of representations from weak to richer forms, depending on the quantity and quality of experience with a particular word (Capone & McGregor, 2005). Richer representations are distinct in that they have more unique nodes of information and stronger activation weights. This summed excitatory activation at the tar- get lexical node reduces the likelihood that the wrong lexeme (or no lexeme) will be retrieved. Storkel (2001) found low phonotactic probability words are prone to error more than high phonotactic probability words. She hypothesized that high phonotactic probability words are

214 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

learned faster, which most likely allows the child to map and integrate semantic information better for these words than for low phonotactic probability words. That is, high phonotactic probability words are less prone to error because they are more richly represented in memory than low phonotactic probability words.

In another study, Gershkoff-Stowe (2002) manipulated toddlers’ frequency of experience in saying words via two conditions of naming practice. In one condition, toddlers had practice naming pictures from a picture book; in the second condition, toddlers had extra practice nam- ing the same pictures (i.e., book, picture cards). Toddlers in the extra-practice condition pro- duced fewer naming errors than the low-practice condition.

Richness of semantic representation also influences word retrieval (e.g., Capone & McGregor, 2005; McGregor, Friedman, Reilly, & Newman, 2002; McGregor, Newman, Reilly, & Capone, 2002). McGregor and colleagues have studied the richness of semantic rep- resentation associated with both accurate naming and naming errors by typically developing children and those with specific language impairment (SLI). Children with SLI are known for having difficulty with word retrieval (e.g., McGregor, 1997). In the studies conducted by McGregor, Friedman, Reilly, and Newman (2002) and McGregor, Newman, Reilly, and Capone (2002), children were asked to name a set of pictures to assess naming accuracy and then were subsequently asked to draw pictures of the target words and to define the target words. A group of adults rated pictures and definitions for quality and quantity of informa- tion, respectively. The results showed that for both groups of children, when they produced naming errors on target words, the corresponding drawings and definitions of those words were weak. Specifically, drawings were rated as poor in quality and definitions contained fewer pieces of semantic information than did drawings and definitions of accurately named targets. Drawings and definitions of accurately named targets contained more information in both drawings and definitions, indicating richer representations of these words. In summary, richer semantic knowledge was associated with accurate word retrieval for naming, while weaker se- mantic knowledge was associated with naming errors.

Experience with words is a key factor in creating a richer lexical–semantic network. Having more experience with certain words can lead to retrieval of them despite other factors that are often associated with expressive word use, such as older age and higher IQ (Bjorklund, 1987). Experience can be enriched via quantity (or frequency) of experience or via quality of each expe- rience (e.g., through scaffolding). For example, in the Gershkoff-Stowe (2002) study discussed earlier, frequency of experience was manipulated by increasing practice in saying words; this extra practice had a positive effect on retrieving words for naming.

Capone and McGregor (2005) examined the influence of enriched quality of semantic learning for its effect on word retrieval. In their study, the frequency of word exposure was controlled but the semantic representation was enriched via representational gesture cues.

Working Memory 215

Frequency of word exposure was controlled by ensuring that the training objects and words were novel and the objects were labeled the same number of times in each condition. Semantic enrichment varied by three conditions: shape, function, and control. In the shape condition, tod- dlers heard the word label and also saw a gesture that highlighted the shape of the object. In the function condition, toddlers heard the word label and also saw a gesture that highlighted the function of the object. In the control condition, no gesture cues accompanied the word label. By the study’s end, toddlers had learned the same number of words under all three learning condi- tions, but the quality of learning differed by condition. Toddlers retrieved more words for nam- ing when semantic enrichment (shape, function) was provided than when words were learned without it (control). Indeed, when semantic learning was assessed, toddlers knew more about the objects in the shape and function conditions than in the control condition.

WORKING MEMORY ______________________________________________ Our discussion thus far has largely referred to learning and retrieving words within the long- term memory system. Long-term memory is where information is stored after learning. Another memory system, known as working memory (Baddeley, 2000; Baddeley & Hitch, 1974), also plays a key role in word learning. The working memory system is involved in active, online pro- cessing of information; it allows for temporary storage of information while it is being manipu- lated or processed. Working memory has a limited capacity of resources to process information, so there is always a trade-off in terms of where those resources are being directed. For example, when the child sees an unknown referent and hears a novel word, she engages in a process of linking the word to the referent, but also fast-maps phonological and semantic information and integrates the entire picture with what she already knows (i.e., information from long-term memory). This kind of processing is termed online processing. Working memory is the system that allows us to process information online. It is the system used to make sense of new informa- tion and to integrate new information with known information stored in long-term memory.

Several distinct components constitute the working memory system. First, two workspaces process visual and verbal information. The workspace for verbal information, which is called the phonological loop, encodes, maintains, and manipulates speech-based input. The second work- space, called the visuo-spatial sketchpad, manipulates visual information for visual recognition and orientation of stimuli. These two subsystems compete for the limited processing resources available in the working memory system. The third component of the system, known as the cen- tral executor, acts as the overseer. It maximizes the processing of visuo-spatial and phonological information by allocating processing resources to one or the other of the two workspaces or by splitting the resources between the two. The central executor modulates attention to each type of information. A fourth component of the working memory system, the episodic buffer, provides

216 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

a place for integration of information to occur after the initial processing. It allows temporary representations to be integrated (new and old information) prior to that final integrated repre- sentation being sent on for storage in long-term memory (Baddeley, 2000).

The integrity of the phonological loop is critical for vocabulary development (Gathercole & Baddeley, 1989, 1990; Gathercole, Willis, Emslie, & Baddeley, 1992; Jarrod & Baddeley, 1997). For example, Gathercole and Baddeley (1989) showed that the size of a child’s receptive vocabulary was positively related to his ability to remember phonological information. Four- and five-year-olds were better at repeating nonwords when they had larger, rather than smaller, vocabularies. A child’s performance accuracy in nonword repetition is a common measure of integrity of the phonological loop. Of course, all new words to the child are nonwords initially until they make a connection with their referents; therefore, nonword repetition is a good mea- sure of the phonological loop’s capacity for processing new words. Gathercole, Willis, Emslie, and Baddely (1992) identified the same relationship between the integrity of the phonological loop and vocabulary development in children four to eight years of age. Conversely, children with language learning impairments (e.g., SLI, Down syndrome) have been found to have poor nonword repetition and smaller vocabularies than typically developing children (Gathercole & Baddeley, 1990; Jarrod & Baddeley (1997). Therefore, the relationship is the same: Good non- word repeaters have larger vocabularies and poor nonword repeaters have smaller vocabularies.

To date, little research has focused on the integrity of the visuo-spatial sketchpad and vocabu- lary development. Jarrod and Baddeley (1997) found that children with Down syndrome showed a relative strength in tasks that required processing of visual information (visuo-spatial sketchpad measure) when compared to tasks requiring processing of verbal information (phonological loop).

LATER LEXICAL DEVELOPMENT _____________________________________ Word learning continues across the life span (Nippold, 2007). During the first half of the second year, the infant’s vocabulary consists largely of nouns. As he passes into toddlerhood, the child’s verb vocabulary expands. The size of other word classes grows as the child moves through the preschool years and beyond (e.g., spatial terms, temporal terms, pronouns, con- junctions). As the child enrolls in school, word learning opportunities shift from oral to writ- ten. The school-age child begins to learn many new vocabulary items via academic and social readings. Again, the child relies largely on inferences from the context that surrounds new words to glean their meanings. There is also some more direct teaching through instructors and text reading. For example, definitions are stated explicitly for some words.

One of the most significant accomplishments for the child from preschool to school age and into adolescence and adulthood is the ability to learn and use more abstract language. The

Summary 217

child develops a flexibility in understanding that words can have multiple meanings and that sometimes these meanings are concrete (e.g., cold refers to temperature), and sometimes they are more abstract (e.g., cold refers to a person’s temperament). More concrete meanings are learned first, in the preschool years; later, the psychological and abstract meanings are acquired. Also, the child expands the type of lexical items learned by including figurative language (e.g., meta- phors, idioms, slang terms). Metaphors, idioms, and slang terms are extensions of lexical devel- opment. For example, idioms (e.g., It’s raining cats and dogs) are believed to function as a single word. These constructs convey a single meaning that differs from the meaning that the words convey in isolation. Children’s understanding of figurative language follows its own developmen- tal course, with literal meanings being learned earlier (preschool and early school age) and figu- rative meanings being acquired later as the child moves through middle school, adolescence, and adulthood. Mastery of figurative language is essential for social and academic success. For exam- ple, humorous play on words is often a device used in advertising. Even in adulthood, however, performance on figurative language tasks does not reach a ceiling of 100% accuracy. Therefore, lexical acquisition and figurative language use are life-long in their development.

SUMMARY _____________________________________________________ Word learning is a lifelong, complex, and dynamic process. The child comes equipped with certain tools (i.e., biases) to help him jump into the word learning system; many scaffolds are also available for the child in the environment (e.g., gestures). The child readily takes advantage of these aids. The child spends a full year preparing for his first word. Although word learning is slow to start, by toddlerhood the child is a word-learning machine.

The early lexicon largely consists of nouns. As the child moves through toddlerhood and into the preschool years, however, vocabulary expands to include many other word classes, in- cluding verbs, prepositions, pronouns, conjunctions, and adjectives. The lexicon also expands to include words with multiple and figurative meanings, as well as larger language units that function as a single lexical item.

The process of learning a single word is an extended process that moves from an initial fast mapping to a longer period of slow mapping that enriches the lexical and semantic rep- resentation of the word. The richness of lexical–semantic representation is related to the likelihood of whether the child will retrieve that word from memory. Speech-language in- tervention provides the child with increased frequency and richer quality of word learning experiences, thereby narrowing the gap between a child with language impairments and his typically developing peers.

218 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

KEY TERMS ____________________________________________________ Action words Basic-level terms Central executor Decontextualize Emergentist coalition model Expressive children Distributed neural network Episodic buffer Fast mapping Function words Functional core hypothesis General nominals Hierarchical organization of the lexicon Indeterminate errors/responses Innate biases Late bloomers Late talkers Lexeme Lexical representation Lexical–semantic network Modifiers Naming explosion Naming insight Natural partitions hypothesis Nominal insight Novel name–nameless category principle (N3C) Ordering of word classes Ordinate terms Overextension Perseverative errors Personal–social words

Phonetically consistent form (PCF) Phonological errors Phonological loop Prelexical forms Principle of conventionality Principle of extendibility Principle of mutual exclusivity Principles of word learning Principle of reference Protowords Quick incidental learning (QUIL) Referential children Semantic errors Semantic representation Slow mapping Semantic feature hypothesis Sensorimotor morphemes Shape bias Specific nominals Subordinate terms Superordinate terms Underextension Visual misperception errors Visuo-spatial sketchpad Vocabulary spurt Whole-object bias Word learning constraints Word retrieval errors Word spurt Working memory

References 219

STUDY QUESTIONS ______________________________________________ What are the precursor speech-language behaviors to first words observed in the • infant? Describe lexical development from first words to two-word combinations, including the • word class that predominates in this early lexicon. What is the word spurt? How does naming accuracy (word retrieval) change from pre- • to post-word spurt? Define the principles of word learning, and explain how they aid the young child in • learning words. Compare and contrast fast mapping and slow mapping of words. • What is the effect of how richly a word is represented on retrieval of the lexeme?• Describe the emergentist coalition model of word learning.• Describe working memory, and explain how it relates to word learning.•

REFERENCES ___________________________________________________ Baddeley, A. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences,

4(11), 417–423. Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. Bower (Ed.), The psychology of learning and

motivation (Vol. 8, pp. 47–90). New York: Academic Press. Balaban, M. T., & Waxman, S. R. (1997). Do words facilitate object categorization in 9-month-old infants?

Journal of Experimental Child Psychology, 64(1), 3–26. Barsalou, L. W. (1999a). Perceptual symbol systems. Behavioral and Brain Sciences, 22, 577–660. Barsalou, L. W. (1999b). Language comprehension: archival memory or preparation for situated action?

Discourse Processes, 28(1), 61–80. Bates, E. (1976). Language in context: The acquisition of pragmatics. New York: Academic Press. Bates, E., Bretherton, I., & Snyder, L. (1988). From first words to grammar: Individual differences and dissociable

mechanisms. New York: Cambridge University Press. Bates, E., Marchman, V., Thal, D., et al. (1994). Developmental and stylistic variation in the coposition of

early vocabulary. Journal of Child Language, 21, 85–123. Bjorklund, D. F. (1987). How age changes in knowledge base contribute to the development of children’s

memory: An interpretive review. Developmental Review, 7, 93–130. Bloom, L. (1973). One word at a time. The Hague: Mouton. Bloom, L., & Capatides, B. (1987). Expression of affect and the emergence of language. Child Development, 58,

1513–1522. Bloom, P. (2004). Myths of word learning. In G. Hall & S. Waxman (Eds.), Weaving a lexicon (pp. 205–224).

Cambridge, MA: MIT Press. Booth, A. E., & Waxman, S. R. (2002). Object names and object functions serve as cues to categories for in-

fants. Developmental Psychology, 38(6), 948–957. Bruner, J. (1975). The ontogenesis of speech acts. Journal of Child Language, 2, 1–19.

220 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

Capone, N. C., & McGregor, K. K. (2005). The effect of semantic representation on toddlers’ word retrieval. Journal of Speech, Language, and Hearing Research, 48, 1468–1480.

Carey, S. (1978). The child as word learner. In M. Halle, J. Bresnan, & G. A. Miller (Eds.), Linguistic theory and psychological reality (pp. 264–293). Cambridge, MA: MIT Press.

Caselli, M. C., Bates, E., Casadio, P., et al. (1995). A cross-linguistic study of early lexical development. Cognitive Development, 10(2), 159–199.

Clark, E. (1973). What’s in a word? On the child’s acquisition of semantics in his first language. In T. Moore (Ed.), Cognitive development and the acquisition of language (pp. 65–110). New York: Academic Press.

Clark, E. (1975). Knowledge, context and strategy in the acquisition of meaning. In D. Dato (Ed.), Developmental psycholinguistics: Theory and application. Washington, DC: Georgetown University Press.

Clark, E. (1993). The lexicon in acquisition. Cambridge, England: Cambridge University Press. de Boysson-Bardes, B., & Vihman, M. (1991). Adaptation to language: Evidence from babbling and first words

in four languages. Language, 67, 297–319. Dell, G. S. (1990). Effects of frequency and vocabulary type on phonological speech errors. Language and

Cognitive Processes, 5, 313–349. Dell, G. S., Reed, K. D., Adams, D. R., & Meyer, A. S. (2000). Speech errors, phonotactic constraints, and im-

plicit learning: A study of the role of experience in language production. Journal of Experimental Psychology, 26(6), 1355–1367.

Dore, J., Franklin, M., Miller, R., & Ramer, A. (1976). Transitional phenomena in early language acquisition. Journal of Child Language, 3, 13–28.

Gathercole, S. E., & Baddeley, A. D. (1989). Evaluation of the role of phonological STM in the development of vocabulary in children: A longitudinal study. Journal of Memory and Language, 28, 200–213.

Gathercole, S. E., & Baddeley, A. D. (1990). Phonological memory deficits in language disordered children: Is there a causal connection. Journal of Memory and Language, 29, 336–360.

Gathercole, S. E., Willis, C. S., Emslie, H., & Baddeley, A. D. (1992). Phonological memory and vocabulary de- velopment during the early school years: A longitudinal study. Developmental Psychology, 28(5), 887–898.

Gentner, D. (1982). Why nouns are learned before verbs: Linguistic relativity versus natural partitioning. In S. Kuczaf (Ed.), Language development: Volume 2: Language, thought and culture. (pp. 301–334) Hillsdale, NJ: Lawrence Erlbaum.

Gentner, D. (2006). Why verbs are hard to learn. In K. Hirsh-Pasek & R. M. Golinkoff (Eds.), Action meets word: How children learn verbs (pp. 544–564). New York: Oxford University Press.

Gershkoff-Stowe, L. (2001). The course of children’s naming errors in early word learning. Journal of Cognition and Development, 2(2), 131–155.

Gershkoff-Stowe, L. (2002). Object naming, vocabulary growth, and the development of word retrieval abili- ties. Journal of Memory and Language, 46, 665–687.

Gershkoff-Stowe, L., & Smith, L. B. (1997). A curvilinear trend in naming errors as a function of early vocabu- lary growth. Cognitive Psychology, 34, 37–71.

Gershkoff-Stowe, L., & Smith, L. B. (2004). Shape and the first hundred nouns. Child Development, 75(4), 1098–1174.

Goldfield, B. (1986). Referential and expressive language: A study of two mother–child dyads. First Language, 6, 119–131.

Goldfield, B. (1987). The contributions of child and caregiver to referential and expressive language. Applied Psycholinguistics, 8, 267–280.

References 221

Goldfield, B. (1993). Noun bias in maternal speech to one year olds. Journal of Child Language, 20, 85–99. Goldfield, B. A., & Reznick, J. S. (1990). Early lexical acquisition: Rate, content and the vocabulary spurt.

Journal of Child Language, 17, 171–283. Golinkoff, R. M., & Hirsh-Pasek, K. (2006). Baby wordsmith: From associationist to social sophisticate.

Current Directions in Psychological Science, 15(1), 30–33. Golinkoff, R. M., Mervis, C. V., & Hirsh-Pasek, K. (1994). Early object labels: The case for a developmental

lexical principles framework. Journal of Child Language, 21, 125–155. Gopnik, A., & Meltzoff, A. (1986). Relations between semantic and cognitive development in the one word

stage: The specificity hypothesis. Child Development, 57, 1040–1053. Halliday, M. (1975). Learning how to mean. New York: Elsevier. Hollich, G. J., Hirsh-Pasek, K., Golinkoff, R. M., et al. (2000). Breaking the language barrier: An emer-

gentist coalition model for the origins of word learning. Monographs of the Society for Research in Child Development, 65(3).

Horgan, D. (1978). How to answer questions when you’ve got nothing to say. Journal of Child Language, 5, 159–165.

Horgan, D. (1979). Nouns: Love ’em or leave ’em. Address to the New York Academy of Sciences. Jarrod, C., & Baddeley, A. D. (1997). Short-term memory for verbal and visuospatial information in Down’s

syndrome. Cognitive Neuropsychiatry, 2(2), 101–122. Jones, S. S. (2003). Late talkers show no shape bias in a novel name extension task. Developmental Science, 6(5),

477–483. Jones, S. S., Smith, L. B., & Landau, B. (1991). Object properties and knowledge in early lexical learning. Child

Development, 62, 499–516. Kamhi, A. (1988). Three popular myths about language development. Child Language Teaching and Therapy,

4(1), 1–12. Kemler Nelson, D. G. (1999). Attention to functional properties in toddlers’ naming and problem solving.

Cognitive Development, 14, 77–100. Kim, M., McGregor, K. K., & Thompson, C. K. (2000). Early lexical development in English- and Korean-

speaking children: Language-general and language-specific patterns. Journal of Child Language, 27(2), 225–254.

Lahey, M., & Edwards, J. (1999). Naming errors of children with specific language impairment. Journal of Speech, Language, and Hearing Research, 42(1), 195–205.

Landau, B., Smith, L. B., & Jones, S. S. (1988). The importance of shape in early lexical learning. Cognitive Development, 3, 299–321.

Marchman, V. A., & Bates, E. (1994). Continuity in lexical and morphological development: A test of the critical mass hypothesis. Journal of Child Language, 21(2), 339–366.

Markman, E. M. (1989). Categorization and naming in children: Problems in induction. Cambridge, MA: MIT Press.

McGregor, K. K. (1997). The nature of word-finding errors of preschoolers with and without word-finding deficits. Journal of Speech and Hearing Research, 40(6), 1232–1244.

McGregor, K. K., & Appel, A. (2002). On the relation between mental representation and naming in a child with specific language impairment. Clinical Linguistics and Phonetics, 16(1), 1–20.

McGregor, K. K., Friedman, R. M., Reilly, R. M., & Newman, R. M. (2002) Semantic representation and naming in young children. Journal of Speech, Language, and Hearing Research, 45(2), 332–346.

222 CHAPTER 7 / Early Semantic Development: The Developing Lexicon

McGregor, K. K., Newman, R. M., Reilly, R. M., & Capone, N. C. (2002) Semantic representation and nam- ing in children with specific language impairment. Journal of Speech, Language, and Hearing Research, 45(5), 998–1014.

Nelson, K. (1973). Structure and strategy in learning to talk. Monographs of the Society for Research in Child Development, 143(38).

Nelson, K. (1974). Concept, word, and sentence: Interrelations in acquisition and development. Psychological Review, 81, 267–285.

Nelson, K. (1981). Individual differences in language development: Implications for development and lan- guage. Developmental Psychology, 17, 170–187.

Nelson, K., Hampson, J., & Shaw, L. K. (1993). Nouns in early lexicons: Evidence, explanations and implica- tions: Erratum. Journal of Child Language, 20(2), 228.

Nelson, K., & Lucarielly, J. (1985). The development of meaning in first words. In M. Barrett (Ed.), Children’s single word speech (p. 13). Chichester, UK: John Wiley and Sons.

Nippold, M. A. (2007). Later language development: School-age children, adolescents, and young adults, 3rd ed. Austin, TX: Pro-Ed.

Oetting, J., Rice, M., & Swank, L. (1995). Quick incidental learning (QUIL) of words by school-age children with and without SLI. Journal of Speech and Hearing Research, 38, 434–445.

Owens, R. (1988). Language development: An introduction. Boston: Allyn and Bacon. Plunkett, K., Karmiloff-Smith, A., Bates, E., Elman, J. L., & Johnson, M. H. (1997). Connectionism and de-

velopmental psychology. Journal of Child Psychological Psychiatry, 38(1), 53–80. Plunkett, K., Sinha, C., Moller, M. F., & Strandsby, O. (1992). Symbol grounding or the emergence of sym-

bols? Vocabulary growth in children and a connectionist net. Connection Science, Special Issue: Philosophical Issues in Connectionist Modeling, 293–312.

Ramer, A. (1976). Syntactic styles in emerging language. Journal of Child Language, 3, 49–62. Rescorla, L., Dahlsgaard, K., & Roberts, J. (2000). Late-talking toddlers: MLU and IPSyn outcomes at 3;0

and 4;0. Journal of Child Language, 27(3), 643–664. Rescorla, L, Mirak, J., & Singh, L. (2000). Vocabulary growth in late talkers: Lexical development from 2;0 to

3;0. Journal of Child Language, 27(2), 293–311. Retherford, K. (2000). Guide to analysis of language transcripts, 3rd ed. Eau Claire, WI: Thinking Publications. Rice, M. L., Buhr, J. C., & Nemeth, M. (1990). Fast mapping word learning abilities of language-delayed pre-

schoolers. Journal of Speech and Hearing Disorders, 55, 33–42. Rice, M. L., Buhr, J. C., & Oetting, J. B. (1992). Specific-language-impaired children’s quick incidental learn-

ing of words: The effect of a pause. Journal of Speech and Hearing Research, 35, 1040–1048. Snyder, L., Bates, E., & Bretherton, I. (1981). Content and context in early lexical development. Journal of

Child Language, 8, 565–582. Stoel-Gammon, C., & Cooper, J. (1984). Patterns of early lexical and phonological development. Journal of

Child Language, 11, 247–271. Storkel, H. (2001) Learning new words: Phonotactic probability in language development. Journal of Speech,

Language, and Hearing Research, 44(6), 1321–1337. Thal, D., & Tobias, S. (1994). Relationships between language and gesture in normally developing and late-

talking toddlers. Journal of Speech and Hearing Research, 37, 157–170. Thal, D., Tobias, S., & Morrison, D. (1991). Language and gesture in late talkers: A one-year follow-up.

Journal of Speech and Hearing Research, 34, 604–612.

References 223

Vihman, M. (1996). Phonological development: The origins of language in the child. Oxford, UK: Basil Blackwell. Vihman, M., Macken, M., Miller, R., Simmmons, H., & Miller, J. (1985). From babbling to speech: A reas-

sessment of the continuity issue. Language, 61, 397–445. Waxman, S. R. (1994). The development of an appreciation of specific linkages between linguistic and concep-

tual organization. Lingua, 92, 229–257. Werner, H., & Kaplan, B. (1963). Symbol formation. New York: John Wiley & Sons.