Handbook of Language and Literacy Development - a Roadmap from 0 to 60 Months

children image

research

Vocabulary Development (0-60 Months)click to print Print
Research Review / Parent

Written by: Andrew Biemiller, Ontario Institute for Studies in Education, University of Toronto

Introduction: The Importance of Vocabulary Developed By Kindergarten for Later School Achievement

Children with relatively small vocabularies by the end of kindergarten generally show reduced reading comprehension by grades three or four and continue to show delays through high school (Chall, Jacobs & Baldwin, 1990; Cunningham & Stanovich, 1997; Nation & Snowling, 2004; Scarborough, 2001; Snow, Porche, Tabors, & Harris, 2007). Kindergarten vocabulary is less predictive of grade one reading. In the primary grades (kindergarten to grade three inclusive), children's texts often have restricted vocabulary. However, if vocabulary is developing slowly by kindergarten, children will be disadvantaged by the middle elementary school years.

By the time children begin kindergarten, they have already acquired much of their language. They speak in sentences, and they understand simple narratives and simple expository language. By 60 months of age, most children probably know more than one or two thousand root word meanings.

However, there are substantial differences in language achievement among children at the beginning of kindergarten. How do these differences come about? Age differences alone have a substantial impact. The youngest kindergartners are 20 percent younger than the oldest kindergartners. There are also substantial differences in language and vocabulary associated with social class, and children growing up in a non-English-speaking family are at a disadvantage. Let us consider briefly the range of language differences at the beginning of school, the developmental nature of language growth, and some of the early influences on language growth.

An Overview of Vocabulary Development in the First Five Years

I estimate that the vocabulary size by the beginning of kindergarten ranges from 2300 root word meanings (average for lowest quartile) to 4700 root word meanings (average for the highest quartile). This estimate is based on a projection from vocabulary size at the end of grade two. The estimate is discussed below in the section on Words Developed in the First Five Years. During the subsequent primary grades, the gap between the lowest and highest quartiles continues to expand (Biemiller, 2005). By the end of grade two, children in the lowest quartile average 4000 root word meanings, average children have about 6000 meanings, and the highest quartile children average 8000 meanings. Note that these large vocabulary differences have developed before children have had much of an opportunity to build vocabulary in conjunction with their own reading. Young children mainly read "primer" texts with restricted reading vocabulary.

Thus, some children begin school with reduced vocabulary, and their vocabulary disadvantage becomes greater during the primary grades. Reducing or eliminating this early vocabulary gap between low- and high-vocabulary children should go some distance towards improving disadvantaged children's chances of achieving well in school.

We might hope that once children enter school at age 4 or 5, there would be more opportunities for less advantaged children to build vocabulary—even to "catch up" with more advantaged children. Unfortunately, the limited available data on the vocabulary effects of school attendance in primary grades is discouraging—kindergarten, grade one and grade two children appear to gain no vocabulary as a result of a year in a primary grade. The evidence is simple—on average, the youngest first graders have just one month's more vocabulary than the oldest kindergartners (Christian, Morrison, Frazier, & Massetti, 2000; Morrison, Smith, & Dow-Ehrensberger, 1995). Similarly, the youngest second graders average just one month's vocabulary more than the oldest first graders (Cantalini, 1987). This lack of school effect on vocabulary probably reflects insufficient attention given to vocabulary, compared to early reading and math skills. The average month older first grader does have markedly better reading than math skills than a month younger kindergartner, albeit not as strong as the skills of the oldest first graders.

In what follows, I will discuss how word meanings are acquired in the early years, individual differences in word acquisition, and how home and group care/school influences affect the number of word meanings acquired. In addition, I will discuss evidence that interventions might increase vocabulary acquisition in the early years. I will also provide some preliminary discussion of what word meanings might be targeted for attention by parents and teachers, particularly for children whose language development is at risk. I conclude that more could (and should) be done to facilitate language (and particularly vocabulary) development during the first five years of development.

How Word Forms and Meanings are Acquired and "Mapped"

In this section, I discuss processes and individual differences affecting word acquisition, as well as home differences and child care interventions affecting word acquisition, and conclusions.

Processes

Word form and word meaning. A "word form" is the spoken (and heard and remembered) word. "Word meaning" is the knowledge (stored experience, cognitive organization of information received) "mapped" or accessed by a word. The word meaning is sometimes called the "referent" of the word. A "word" is the combined word form and meaning. About a third of the root word forms likely to be known by age 5 are mapped to more than one word meaning. This estimate is based on an examination of a sample of root word meanings commonly known by age 5—see section on Words Developed During the First Five Years.

Fast-mapping meanings. Some years ago, Susan Carey (1978) introduced the concept of "fast-mapping" word acquisition. She described fast-mapping as when "One, or a very few, experiences with a new word can suffice for the child to enter it into his mental (list of words) and to represent some of its syntactic and semantic features" (p. 291). She observed that preschool students could very quickly acquire a conceptual referent for a new word but took much longer to have a relatively complete understanding of the word. By "conceptual referent", Carey meant the "relevant part of the internal representational system in terms of which the person or animal describes and understands the world and his own actions in it.... Take for example, the lexical domain (or mental list) of color words, and the conceptual domain of mental representation of colors" (p. 269). While I doubt that many root meanings are simply "inferred", I suspect that given the simultaneous presence of a novel word form and conceptual referent, children can sometimes "map" or establish a connection between the form and referent very quickly. This situation—simultaneous presence of word form and conceptual referent—can occur either as a result of direct explanation by an older speaker or as a result of naturally occurring dialogue or story reading.

Under what circumstances are meanings mapped? To what extent are words acquired as the result of intentional explanations, to what extent do acquisitions follow from explicit child efforts to determine word meaning, and to what extent are meanings acquired without intentional effort to teach or to infer meaning?

It has been asserted that vocabulary acquisition is largely a matter of unintended inference (Landauer & Dumais, 1997; Nagy, Anderson, & Herman, 1987). Acquisition of vocabulary by inference is argued because "...it is evident that most children must be acquiring the vast bulk of their vocabulary knowledge apart from instruction specifically devoted to vocabulary learning." (Nagy & Herman, 1987, p. 23 ). It is possible that this claim is true for elementary (grades four to six inclusion) or secondary school students, although I doubt it. There are differences among students in their ability to learn new words incidentally; these differences relate to their concurrent vocabulary levels (McKeown, 1985) and to their comprehension levels (Swanborn & de Glopper, 1999). However, accounts of word acquisition in the preschool period in fact describe mainly explicit adult:child language episodes, in which specific meanings are either explicitly provided or a very salient situation provides an opportunity to map a word form-word meaning association (Bloom, 1998; Woodward & Markman, 1998).

Explicit adult:child language episodes include pointing to referent objects (Tomasello, Carpenter, & Liszkowski, 2007); highlighting action (Woodward & Markman, 1998); and adult supply of words when the child is attending to a referent (Nelson, 1988) among others. Nelson wrote:

"The typical way children acquire words in their first language is almost completely the reverse of the Quinean paradigm...Children do not try to guess what it is that the adult intends to refer to; rather they have certain conceptions of those aspects of the world they find interesting and, in successful cases of word acquisition, it is the adult (at least in western middle-class societies) who guesses what the child is focused on and supplies an appropriate word." (1988, as cited in Woodward & Markman, 1998, p. 410).

"Salient situations" occur when a parent labels an object while handing the object to a child (e.g. "Johnny, have a banana"), and "overhearing" explanations given to another (e.g., a lemon is yellow and an orange is orange, the lemon is smaller than an orange, both are citrus fruits). Readers can find further examples in Woodward and Markman (1998).

Sources of Differences in Word Acquisition

Is it simply the case that some children have been advantaged by heredity and that this accounts for observed differences? Alternatively, is it simply the case that some children have been advantaged by circumstances—by the family, community, and education system in which they grow up—and that this advantage accounts for observed differences? It is clear that both constitutional differences in children (nature) and differences in children's experience (nurture) have substantial impacts on how rapidly a child's language develops (Olson & Gayan, 2002). Indeed, the relationship between listening vocabulary and many of the other skills measured on "intelligence tests" is very strong (Carroll, 1993). On the other hand, there is clear evidence that variations in children's environment are associated with marked differences in language and vocabulary development, and that fairly simple interventions can double normal vocabulary growth rates (to be discussed in further on, Does Early Intervention Make a Difference?). Language development does not simply follow a biologically programmed schedule. Let us consider briefly some of the factors influencing language and vocabulary growth.

Person Characteristics (Talent)

Constitutional Factors. To learn words easily, one must be able to form "phonological files" that is, easily-stored information that allows a person to readily distinguish, remember, and reproduce the sound of a novel word. In addition, one must develop "referent connections" to which the word-sound refers. These referent concepts (or "collections of connections" as Miller, 1991, describes them) must either form easily or already be extant when a "word" is learned. For example, an infant may already have a mental "file" about cups when she first learns the word, "cup". E. Gibson (1968) and J. Gibson (1962) cite evidence that something like files are established for objects within the first month of life. Eleanor Gibson described most perceptual input as organized into objects, space, and events with the later addition of iconic and symbolic information. Learning the meaning of the word, "cup" simply involves connecting the word-form with the pre-existing non-verbal "file" on cups. On the other hand, when the word "fair" is learned, a complex referent concept regarding "fair" and "unfair" social transactions is probably gradually constructed. It is unlikely that a non-verbal "file" on fairness exists prior to encountering "fair" in verbal interaction. A completely different referent concept regarding a "fair" as a social setting is also learned at some point.

There are probably constitutional factors affecting both the ease of forming phonological files and the ease of developing referent files and connections. Thus, it is not surprising that young children with a strong command of word sounds develop larger vocabularies than comparable children with less strong phonological skills (Gathercole, Service, Hitch, Adams & Martin, 1999; Gathercole, 2007). In addition, "working memory", a general cognitive capacity necessary for word formation, is also implicated in language development. One must be able to attend to a word form and its meaning simultaneously if word form and word meaning are to be mapped (Case, 1985; Gathercole et al, 1999). More generally, the growth of working memory appears to underlie basic conceptual development (Case, 1985; Case & Okamoto, 1998). Thus, progress through the various "stages" of cognitive development has been shown to be closely associated with the growth of working memory. This cognitive development is likely to influence the mental or cognitive "level" of words learned, in addition to the ease of learning specific words.

Rose, Feldman, and Jankowski (2009) have recently reviewed research on cognitive-developmental abilities that have a bearing on vocabulary acquisition as early as 12 months of age and reported results of a new study on this subject. They report that three processes assessed at 12 months of age —visual recognition (evidence: distinguishing between familiar and novel stimuli), recall (evidence: repeating observed acts), and representational competence (evidence: cross-modal transfer of shape perceived tactually and then visually) were associated over 30 percent of variance in vocabulary assessed at 36 months on the Peabody Picture Vocabulary Test (PPVT) (Dunn & Dunn, 1997). These processes when combined with observed vocabulary at 12 months of age (as reported by parents) accounted for 44 percent of vocabulary at 36 months of age.

Much more could be said of constitutional factors influencing language growth, but these examples should be sufficient to establish that such factors have an effect. To a large extent, the effect of constitutional factors is to increase or decrease the rate or ease of word learning. In short, there is reasonable evidence for the existence of a constitutionally-based "talent" for learning words and language. However, the number of words learned, (and other aspects of language structures developed) will be influenced by a combination of this potential learning rate and the opportunities children encounter for learning language.

Vocabulary Acquisition at Home

It is not news that children who grow up in low income and poorly educated families are much more likely to have smaller vocabularies and less advanced language development than their more advantaged peers. Research on family and social class factors was extensively summarized in McLloyd (1998); Duncan, Brooks-Gunn, and Klebanov (1994); and Landry and Smith (2006). These differences are clearly apparent by the time children start school. However, social class, as indexed by family income and education, is a gross index for what must be a combination of specific family and community child rearing practices plus some possible group differences in constitutional cognitive potential. I have already discussed constitutional differences that may affect the rate of language development. Let us now turn to child-rearing practices and opportunities that may also affect language development.

It is clear that root words (word form+meaning) can only be acquired when (a) a word novel to the child is used with or around a child and (b) the meaning is in some way introduced to the child while the child is attending to the word form. In most cases, not only must the word be used, but the context must be clear. Beals (1997), Tabors, Beals and Weizman, (2001) and Weizman and Snow (2001), have shown that greater parent use of "rare words" (words not commonly known by grade four, Chall & Dale, 1995) and the presence of more children's books in the home are associated with increased vocabulary by age five. Tabors' and Beals' studies of home and school language involved a sample of disadvantaged homes whose children were attending Headstart—a preschool program for disadvantaged children.

Betty Hart and Todd Risley (1995, 1999) have written what may be the clearest book describing language and achievement differences between advantaged and disadvantaged children, and differences in their early experiences. They provide correlational data showing that parents who use a lot of language (e.g., many different words) with and around their child have a child who uses many more words (and almost undoubtedly knows many more words). By three years of age, advantaged children were found to have twice the vocabulary of welfare children, and were adding productive vocabulary at twice the rate (6 words/day vs. 3 words/day). Vocabulary was strongly correlated with a measure of parent interaction combining (a) different words per hour, (b) feedback tone or warmth on interaction, (c) "symbolic emphasis", (d) suggestive rather than directive "guidance style", and (e) "responsiveness" - the proportion of parent responses to child initiated talk (r = .78). Hart and Risley (1995) summarized their overall observations of differences in children's language-related experience:

"To illustrate the differences in the amount of children's language experience using numbers, rather than just "more" and "less", we can derive an estimate based simply on words heard per hour. The longitudinal data showed that in the everyday interactions at home, the average (rounded) number of words children heard per hour was 2,150 in the professional families, 1,250 in the working-class families, and 620 in the welfare families. .... Given the consistency we saw in the data, we might venture to extrapolate to the first 3 years of life. By age 3 the children in professional families would have heard more than 30 million words, the children in working class families 20 million words, and the children in welfare families 10 million" (pp. 131-132).


Similar findings exist in other observational studies. In Gordon Wells' study of language in the home, a significant correlation was found between total adult talk to children and the quality and gains over time in children's language. Larger correlations were found between gains in child language and the percentage of adult talk that involved extensions of child language (Wells, 1985). As in the Nelson quotation earlier, this finding implies that vocabulary grows faster when parents actively help children build vocabulary. Burton White's famous study contrasting the home environment of advantaged and disadvantaged children similarly indicated marked differences in the quantity and quality of dialogue directed to children (White & Watts, 1973). In general, children develop language more rapidly in homes where parents (usually mothers) seriously engage in dialogue with their language-learning preschoolers. There is some evidence that this variation occurs independently of socio-economic (SES) differences. In other words, children develop stronger language when being raised in lower SES circumstances but interacting with their mothers in ways similar to the way high SES children interact with their mothers (Hart & Risley, 1995).

Hoff (2003) investigated "high SES" (college-educated parents) and "mid SES" (high school graduate parents, mainly working in low-skill occupations). She studied mother-child talk at an average child age of 21 months and again at 31 months. Both groups of children produced similar numbers of different words at 21 months of age —about 36 in 90 utterances. By 31 months, the high SES children were producing an average of 51 words in 90 utterances, significantly more than the mid SES children who produced an average of 46 different words in 90 utterances. (We don't have this information for seriously disadvantaged children.) All of the observed SES difference in individual children's vocabulary was accounted for by the fact that high SES mothers spoke to their children in longer utterances (Mean length of utterance - MLU), thereby providing more words to be learned, and more elaborate sentences providing context for the words. Once mother's mean length of utterances was entered as a predictor, SES added no other variance. This estimate is based on an examination of a sample of root word meanings commonly known by age 5—see section on Words Developed During the First Five Years. Note that in Hoff's 2006 study, parental variables accounted for about 25% of observed variance in vocabulary gains during the study. Other factors including measurement problems and child cognitive process differences clearly also played a role in observed vocabulary gains.

Vocabulary: Effects of Motivation and Parental Encouragement.

It is clear that some homes provide more language support (word explanations, praise for using new words, etc.) (Hart & Risley, 1995). Very young children are observed to inquire about unknown word meanings (Brown, 1965). However, by age 4 or 5 years, Beals (1997) suggested that children ask about very few meanings, at least during family meals. Note that Beals' study was conducted with disadvantaged Headstart children in their homes. It is easy to visualize such interactions. "Mommy, what does think mean?" If a child is discouraged by an answer like "Don't bother me" or equivalent just a few times, I suspect that the odds of asking about unknown words becomes rare, as was reported by Beals. In my personal experience, many questions about unknown words certainly come from four- or five-year old children in advantaged homes. Landry and Smith's (2006) review noted that the importance of parental responsiveness to children's talk, including maintaining attention, responding to children's vocalizations, and emotionally warm support were all associated with advancing verbal skills.

Vocabulary: Acquisition in Group Care or School Setting

The preceding studies were all "observational". In other words, they describe relationships between observed parent activities and child speech. They do not prove that parent behaviours cause more or less child language development. It could be possible that parents with high language abilities have children with high language abilities and just happen to talk more to their children.

However, a very important study by Barbara Tizard and her colleagues (Tizard, Cooperman, Joseph, & Tizard, 1972) provides strong evidence that what parents and caretakers "do" matters, perhaps as much or more as the genetic potential passed on to these children (Biemiller, 1999). Tizard and her colleagues studied young children in English residential institutions -- orphans or children who had been removed from their parents. They found that the organization of care (whether care was centrally controlled or largely the responsibility of a single care-giving nurse) was strongly associated with the level of language development. When an individual nurse was largely responsible for the activities she carried out with a group of 6 children, she spent more time talking, reading, and playing with her charges, was more "accepting" of children, took the children on more trips, etc., than when nurses had less responsibility for a specific group, and followed centrally-determined procedures. In terms of language, nurses with greater responsibility "offered more informative talk, spoke in longer sentences, gave fewer negative commands, and were more likely to explain themselves when they told the child to do something, than were those in the institutionally oriented groups."(Tizard & Tizard, 1974, p. 143). Dramatic differences were associated with these rearing conditions. While the non-verbal intelligence of the children in various groups did not differ, mean verbal comprehension among the 18 children in the three most verbally-supportive groups was 1.5 standard deviations higher than the three most institutionally oriented groups -- a difference identical to that found between the children of professional families and children of manual workers.

In a large-scale study of daycare organization and quality, Ruopp, Travers, Glantz, and Coelen (1979) similarly found that organization of child care arrangements had a large effect on child language and other outcomes. Child care programs organized into small groups (8-10 children), with one staff person responsible for the children in their group, resulted in significantly greater language gains. Ruopp and colleagues also reported differences in language used with children with fewer negative commands and more verbal explanations. Note that the "ratio" of staff to children was much less critical than the organization of child care programs (within a range of 1:6 to 1:12).

Effects of Reading to Children.

Beyond studies of parent-child dialogue are studies of the specific effects of reading to children. In a major review, Bus, Van IJzendoorn, and Pellegrini (1995) concluded that reading to children has an average "effect size" of .67 on "language variables" (vocabulary, comprehension). However, the effect of reading to children under 4.5 years of age was relatively small (5 studies, mean effect size of .34). Both Bus et al (1995) and another review (Scarborough & Dobrich, 1994) reported that amounts of story reading to preschoolers accounted for just 8 percent of child vocabulary variance. It is very likely that simply reading stories to children is not enough—children need more general supportive verbal interaction during reading and at other times.

There are practices that could be included in story reading with preschool children (and even with toddlers and infants). With the youngest children (1 and 2 years of age), effective story reading probably deals more with words shown in pictures and focuses on text within single pages. Unfortunately, I know of no studies of "reading" or presenting word books and simple story books to children at these ages, much less what the consequences for vocabulary development might be.

At ages 3 and 4, there have been more studies of reading and vocabulary/language development (Phillips, Hayden & Norris, 2006). Several reviewers (Bus, 2002; Phillips, Norris, & Anderson, 2008; Sénéchal et al, 2001) have concluded that differences in the quality (such as the engagement of the child with books by an adult by asking questions, pointing out parts of an illustration, and prompting the child to "read") as well as quantity of reading to children affects children's vocabulary. Bus also suggested that adult reading competence and interest in reading is associated with the impact of storybook reading on children's vocabulary.

At the kindergarten level (children just over 5 years of age), Robbins and Ehri (1994) have shown that some vocabulary is acquired simply as a result of hearing a story read twice. Words which occurred more frequently in the texts read were more likely to be learned. Children acquired an average of 1.5 new meanings (out of about 10 unknown prior to the reading—note that there was a 25% random chance of passing a word). However, children with vocabularies below 100 (i.e., below average) on the PPVT, acquired no new words as a result of hearing stories.

Whitehurst et al (1994) introduced "dialogic reading" as a program used by Headstart teachers and by parents of the same Headstart children. Dialogic reading involved various pauses to discuss specific words and the general meaning of the story (comprehension). No significant effect was found for "general vocabulary" (as assessed by the PPVT). Note that Whitehurst et al did not assess a sample of words addressed directly in the program. The failure to have much effect on "general vocabulary" (as assessed with the PPVT or similar measures) may not be important. To impact 'general vocabulary" (without simply teaching the words on the PPVT) over a year, a child would have to successfully learn several hundred word meanings beyond those being learned without instruction. Existing intervention studies have not been long enough to have such an effect.

In another application of dialogic reading, Hargrave and Sénéchal (2000) also investigate dialogic reading's effectiveness. They read stories to 3 or 4 year-old children individually. Each story was read twice. In one condition, there was further "dialogic" discussion of the books and words based on the work of Whitehurst et al., (1994). Reading without dialogic discussion resulted in gains of one word. On average, dialogic discussion increased this by one more word. Overall, we would have to conclude that dialogic reading may not be the most effective way of supporting vocabulary with disadvantaged 3 and 4 year old children.

Sénéchal (1997) read one of two stories individually to 3 or 4 year old children. A single reading condition was contrasted to 3 readings, and 3 readings with questions about target words. The single reading resulted in acquiring 0.4 words (not significant). Simple repeated reading led to adding 1.3 words out of 10 words (receptive, using pictures) while repeated reading with questioning led to adding 2.9 words. These results—gaining 13% of pretested words learned from hearing a story 3 times vs. 29% from added word discussion—are typical of findings for repeated reading plus instruction with primary grade children (reviewed in Biemiller & Boote, 2006).

Coyne, McCoach, and Kapp (2007) favour "extended instruction". They taught 6 words in some depth to kindergarten children. This led to 1.5 more words being learned with instruction than from simple reading without instruction. Further "extending" instruction (more attention to both context and definitions) than from "embedded" instruction (briefly explaining the word's meaning when it was encountered in a story) added another 0.5 words learned (on average). These results actually suggest that teaching more words less intensively may result in greater overall gains than teaching a few words in depth. Similar results were true for the work of Beck and McKeown (2007).

Roberts and Neal (2004) reported success with a 16 week California-supported preschool program for English Language Learners (Hmong and Hispanic). They also taught 6 words each week. Children receiving a comprehension-focus story-based approach added about 2 word meanings per week. On average about a third of the words proved to be known prior to instruction. Another third were learned, and a third of the words were not learned. In other words, they learned half of the words not previously known. These results are typical of language instruction with young children and are very similar to findings with primary grade children (Biemiller & Boote, 2006). Maintaining this level of word acquisition over 16 weeks is very encouraging. It appears that such methods can profitably be used over long enough periods of instruction to have a meaningful effect on vocabulary.

In general, it appears that embedding word meaning instruction in narrative contexts can result in adding enough word meanings to significantly increase vocabulary — if enough words are taught for enough weeks (Biemiller, 2006; Biemiller & Boote, 2006). This finding apparently holds at least for 4 year-old children (Hargrave & Sénéchal, 2000; Sénéchal, 1997). I have not found studies of reading and vocabulary development for younger children.

Does Early Intervention Make a Difference?

We do not know to what extent introducing kindergarten and primary grade children to more vocabulary can compensate for the lack of earlier intensive language experiences of the types described by Hart and Risley (1995, 1999), and by Whitehurst and his colleagues (1994). Certainly a more proactive approach to vocabulary learning in the primary grades will help. (Note that the California State Board of Education is now requiring substantial vocabulary components in published elementary school curricula. This specification has resulted in such components being added to new curricula coming into use at present (Examples are SRA/McGraw-Hill's Imagine It! and Houghton-Mifflin's Houghton-Mifflin Reading programs). However, the gap children have when kindergarten begins will be difficult to "fix" in 3 years. Success would mean that disadvantaged children were adding new word meanings faster than advantaged children. Reducing the gap before students arrive in kindergarten would improve the chances of eliminating vocabulary as a source of reading difficulty.

Vocabulary: Cumulative Vocabulary Effects

The acquisition of new vocabulary at any given point will be a result of environmental opportunities (presence of words to be learned, direct meaning support, plus some encouragement for the effort), vocabulary talent (ease of encoding phonological forms, ease of constructing "meanings" from instruction and inference), and the child's pre-existing vocabulary at that point. This cumulative vocabulary will be important both for understanding explanations and for any inference or construction of meanings.

Vocabulary is largely developed in a predictable sequence (Biemiller & Slonim, 2001). We are not certain why the sequence is so robust. However, it seems likely that some or much of the prior vocabulary is necessary for learning subsequent vocabulary. If a child is familiar with most of the words in surrounding language, the chances of being able to learn or infer a new meaning are probably much greater than if a child is lacking a number of surrounding word meanings.

This is true not only for generating new lexical or meaning representations, but also for interpreting affixed or compound words. My hypothesis is that young children rarely infer meanings of words that are completely unknown, but that they may often infer meanings of affixed, inflected, or compounded words when the components are known, and especially when most other words in the surrounding language are known. They may also consolidate meanings of words that have been briefly explained before.

Experience Factors and Language Development: Summary

Although it is clear that there are some constitutional differences among children in the rate at which they learn language, it is equally clear that there are large differences due to the opportunities presented to learn and use language. The dramatic differences in language experienced by advantaged and disadvantaged children in the first five years of life lead to marked differences in their vocabulary and available language structures by the beginning of kindergarten. There is compelling experimental evidence that much of this difference is due to experience, rather than a simple consequence of inherited ability.

Words Developed During the First Five Years

Some receptive vocabulary is acquired by 12 months of age (Nelson & Lucariello, 1985) while 50 words are produced on average by 18 months of age (Nelson, 1973). More importantly, substantial receptive vocabulary begins to be acquired around 12 months of age. Bates, Dale, and Thal (1995) suggest that an average of 200 word meanings are understood by 16 months of age. There is discussion of rapid word acquisition towards the end of the "one-word" infant period (around 70 words per month, Dromi, 1987). We will see that this is similar to the rate of vocabulary gains from age 2 to age 8 years. Similarly, there are reports of rapid word acquisition of new words in the third year (from 2 to 3 years of age) (Bloom, 1998). However, both Bloom (1998) and Bates, Dale, and Thal (1995) suggest that observed rapid acquisition of words produced may be more a case of becoming able to produce words that are already comprehended, rather than suddenly learning 10 or 20 root words per day. Researchers do not have a clear outline of vocabulary acquisition after that age, but they can report that since an average of 6000 root word meanings are known by the end of grade two (approximately 8 years of age), and very few at 12 months of age, an average of 860 root word meanings are acquired annually after 12 months of age or about 70 word meanings per month or about 2.5 per day. After second grade, (possibly starting earlier) about 1000 root meanings are learned per year or 2.7 per day (Anglin, 1993; Biemiller, 2005; Nagy & Scott, 2001; Nation, 2001).

There is a consensus in the work of Anglin (1993), Biemiller (2005), and Nagy and Scott (2001) that by the end of second grade, average children know about 6000 root word meanings plus another 18000 derived word meanings, including affixed words (both syntactic and semantic variants) as well as compound words and idiomatic expressions. Remember that there are more meanings than word forms (e.g., lean—lean - slant to side (gr. 2); lean - without fat (grade 4); lean - thin (grade 6); lean (on) - to count on (grade 8); and two others listed in Dale and O'Rourke (1981). The 6000 meanings known at grade two include a lot of multiple meanings for some word forms.

One of the major issues concerning vocabulary size is to what extent we should count root meanings or root plus derived meanings? In my view, we should emphasize root forms. There is clear evidence that children grasp syntactic and semantic variations at young ages (Bloom, 1998; Brown, 1957). Children are applying some semantic affixes before age 5 (Woodward & Markman, 1998)).Examination of word meanings known by children in grade 4 or later grades (Dale & O'Rourke, 1981) very frequently show roots and derived forms appearing at the same or adjacent grade levels.

Can we similarly expect derived forms to be mastered near the same time as root words at preschool ages? We do not know. My present practical solution is to treat irregular syntactic forms as separate word meanings (e.g., ride, rode), while regular syntactic forms and simple affixed semantic forms (e.g., uneaten) are probably mastered or inferred at an early age.

Estimating Vocabulary Size

To estimate vocabulary size at 60 months (roughly the beginning of kindergarten), I adjusted average vocabulary of 6000 root word meanings at the end of grade two down by 4/7ths, yielding an average estimated vocabulary size of approximately 3400 meanings. Note that there will be a wide variation among children in vocabulary, both due to environment and also to age. The youngest children in a single year will average 800 fewer meanings than the oldest children within the same grade or year, simply due to age (Cantalini, 1987; Christian, Morrison, Frazier, & Massetti, 2000; Morrison, Smith & Dow-Ehrensberger, 1995). As noted previously, a year of school experience has not been shown to have any measurable impact on vocabulary development in the primary grades. The same researchers do report that a year of primary school experience certainly does have an effect on word recognition and early arithmetic skills.

Estimating Kinds of Words Known

My final database list of root word meanings likely known at age five or the beginning of kindergarten includes some 3500 meanings. I estimate that such a child might know 4000 or more word meanings, including some at higher grade levels than the grade two meanings included here. I arrived at this list using a list of word meanings that are known about 80% by the end of grade two (Biemiller & Slonim, 2001). Another 1200 meanings reported known at grade 2 were omitted for various reasons. Among reasons for omitting meanings were derived (e.g., brownish), slang (cream - to beat badly), "sounds" (e.g., boom), archaic (e.g., britches), or redundant (e.g., carry—two similar meanings with reference to computation-one was retained).

I expect that a child 5 years of age with an above-average vocabulary would know 80% or more of these words. By grade two, more than 80% of these 3500 word meanings will be known by most children. However, the children who learn these words during the primary grades are already behind their classmates in vocabulary development. More advanced children are acquiring additional words at this point. Details of this estimation are given in Appendix A.

Word Categories

Word meanings were classified by concrete vs. not, and by word type category.

Concrete vs. Verbal Referents. These were classified as "concrete" (meanings which could be indicated by pointing or acting out, e.g., shower, wrestling) or "verbally-defined" (meanings requiring some verbal explanation of the meaning, e.g., think, secret). About 70% of meanings probably or possibly known at 60 months were classified as "concrete". Note that knowledge of concrete referents is fairly easy to test using pictures as in the Peabody Picture Vocabulary Test (PPVT) (Dunn & Dunn, 1997). However, verbal referents are difficult to picture, and in fact are not often assessed on the PPVT. Testing these meanings with children below grade two can be difficult (see Biemiller, 2005; Biemiller & Slonim, 2001).

Age and Type of Referent. In the second year of life (first year of acquiring words) children learn almost exclusively concrete meanings (Booth & Waxman, 2009). These researchers also report no acquisition of adjectives in the 14 - 18 month age-range. During the third year (24 - 36 months), some "relational" terms are added: prepositions and locational terms (e.g., under, there) and also some "action/event" terms (e.g., bring, drop, listen, remember, breakfast) and "description" terms or adjectives (e.g., asleep, bad, heavy, long) (Bloom, 1998; Hart & Risley, 1999; Nelson, 1974). Many adjectives are relational, for example big vs. little.

However, by 36 to 60 months, many children are beginning to acquire verbally-defined meanings as well (e.g., clever, weather, amaze, meaning) in addition to adding more of the kinds of words learned earlier. Knowledge of verbally-defined meanings distinguishes children making average progress compared to disadvantaged children (Biemiller & Boote, 2006).

Word Type Categories. For this review, I categorize word meanings five categories into:


Roughly half of these words were nouns. A quarter of the words were verbs.

Estimating Which Word Meanings are Worth Teaching

The list of 3500 root meanings includes some which are probably known by most (80% plus) children at age 5 years (60 months), and some root meanings which are probably known by some children at this age (estimated 40 to 80 percent). Meanings predicted to be known by relatively few children at 60 months (or less than 40%) have been excluded, but are doubtless known by some children at 60 months.

I prioritized the 3500 meanings into three categories to determine which meanings might be desirable to use and teach. Categories included:

More examples are given in Table 1. The complete list of "Teach" words is attached to the Parent Narrative.

I suggest that the teach words are those meanings which children who have large vocabularies at 60 months probably already know, while the teach words will frequently be the meanings not known by those with small vocabularies at 60 months. This method of selecting meanings for increased attention by parents, teachers, and child-care staff is based on the observation that words are acquired in a predictable sequence. This has been shown for children from kindergarten up (Biemiller & Slonim, 2001). The same logic is plausible for children under five years of age.

The "Teach" words that I have identified for attention or explanation were much more likely to be "verbally defined" (46%) than the "Easy" words (14% verbally defined) or "Low priority" (20% verbally defined). This identification is similar to findings for words known by average second grader children compared to words known by disadvantaged second grade children (Biemiller & Boote, 2006).

Let me be clear that this is a very preliminary effort at identifying words for attention with preschool children. Readers of this handbook will doubtless disagree with many of my choices. Good! But we have to start somewhere. Such a preliminary list is provided here. My hope is that in the not too distant future, knowledge of at least some of these words by young children can be assessed empirically. If the words assessed are selected randomly from the various categories of words, we should be able to generalize results to other words in the same categories.

Table 2 shows the breakdown of words by priority and word category.

Some 1500 meanings were identified as potentially worth introducing to children ages 3 to 5. They are listed in the Parent Narrative. By age 5 or 6, I estimate that these are meanings that are known by children with relatively large vocabularies, while children with relatively small vocabularies will not know many of these priority words. In the present listing of meanings worth teaching to preschool children, there are almost 700 verbally-defined meanings—or about half of the total recommended meanings.

Note that the "definitions" provided here, which were taken from Dale and O'Rourke's Living Word Vocabulary (1981), are brief. Teaching or explaining these meanings should generally be done when the words are encountered in context, and the meaning explained with reference to the context. These brief meanings as supplied by Dale and O'Rourke help to distinguish different meanings of the same word form (e.g., sample - to taste, and sample - an example).

Conclusions

We have seen that by 60 months, average children have acquired more than 3000 root word meanings. However, substantial differences have already developed, with children in the lowest quartile having at least 1000 fewer meanings than average children.

There is evidence that much of this range in vocabulary acquisition reflects home circumstances. These circumstances have not, to date, been much affected by child care or school practices. However, there is also research showing that both group care and school settings could help to reverse early gaps in vocabulary. Changes in home could also improve vocabulary development, if parents and caregivers were willing and able to make the effort. Really, "home" often means home or group day care, since the large majority of children are not cared for by parents fulltime. Relatives, paid home care providers, and others provide care for roughly half of many preschool children's waking hours.

At present, by grade two children in the lowest vocabulary quartile have half the vocabulary that children in the largest vocabulary quartile have. During both the preschool period and the primary school period, vocabulary continues to diverge for the least advantaged children. As long as children arrive at grade two already at least two years and 2000 root word meanings behind average children in vocabulary, children will continue to experience serious reading comprehension problems throughout the elementary and high school years. Vocabulary is not the whole solution, but without roughly grade-level vocabulary, success in school achievement is highly unlikely. What about English Language Learners—children who speak another language at home? I have not addressed this issue here. However, I will note that Cummins (1985) and others have found that those who have relatively large vocabularies in their native language typically become fully literate in English. Those with smaller vocabularies are at some risk.

The earlier we can support needed vocabulary development, the easier the continuing maintenance of vocabulary development will be and the most children's language and literacy development will flourish.

Table 1: Sample Preschool Meanings by Priority

Easy Meaning Harder but Low Priority Teach/Explain
about on the point of acorn oak tree nut ABC’s the alphabet
bad naughty drinking place amaze to surprise
block child’s toy blend mix battle part of war
button to fasten brownie chocolate nut cake burglar robber
cider apple juice captain in charge of ship chose picked out or selected
cries yells chubby fat and round crouch bend low
drum musical instrument darling sweetheart ever at any time
easy not hard ear muffs ear warmers flap move back and forth
fire a flame fingernail nail on finger Friday day of week
follow to go along (after another) gang group of people grain oats, corn, etc.
goes on keeps on going grown-up older person hero very brave man
her belonging to a girl house to give shelter to jet kind of airplane
into toward the inside lad young boy letter a written message
liar person who tells lies mane horse’s neck hair letter one of the abc’s
mix stir together necklace worn around neck meaning the sense of words
o.k. alright pack fill tightly nightfall coming of night
out not in pine a tree part a piece of the whole
peas small green vegetables queen king’s wife prepare to get ready
puppy a young dog religion belief in god repeat do over
rush to hurry rowboat oar-driven boat season spring or summer, etc.
show movie or tv program sill lower part of window then next
speed quickness of moving stalk stem of plant tough hard to take
sweep brush off a floor take off remove (clothes) vitamin needed for health
toilet stool in the bathroom (?) trunk auto compartment west where the sun sets
watch stool in the bathroom (?) wheat a plant wire metal thread
yours belonging to you zebra striped animal zoo wild animal park

Table 2: Word Categories by Priority

Word Category Easy Low Priority Teach
Nouns 535 756 712
Verbs 327 183 399
Modifiers 144 65 276
Prepositions 54 3 40
Other (function or interrogation) 23 5 13
Total 1083 1012 1440

Biemiller, A. (2009). Research Review: Vocabulary Development (0 – 60 Months). In L.M. Phillips (Ed.), Handbook of language and literacy development: A Roadmap from 0 – 60 Months. [online], pp. 1 - 51. London, ON: Canadian Language and Literacy Research Network. Available at: Handbook of language and literacy development