INFLUENCES ON VOCABULARY KNOWLEDGE

One line of research on vocabulary development examines the effects of instruction on vocabulary growth (e.g., Biemiller & Slonim, 2001). Direct classroom instruction in word definitions, though effective in promoting vocabulary acquisition overall, is relatively ineffective in narrowing the gap between those with good versus poor vocabulary levels (e.g., Anderson & Nagy, 1991; Baker, Simmons, & Kameenui, 1998). Among older children, direct instruction in vocabulary knowledge is less important relative to reading experiences in developing vocabulary skill (e.g., Nagy & Anderson, 1984). The transition from learning to read in young children to reading to learn among older children is essential for advanced vocabulary development. Reading is strongly associated with vocabulary development (e.g., Baker et al., 1998; Stanovich, 1986), so that those with the sparsest vocabulary levels are often those with the poorest reading skills as well (see Baker et al., 1998, for a review). In the present study, apart from the above-documented effects of teaching, which are involved in both vocabulary and reading skill growth, we were particularly interested in those cognitive skills that are associated with variability in vocabulary knowledge.

Cognitive components related to vocabulary acquisition have been the focus of relatively little research. For example, Gathercole and colleagues (Gathercole et al., 1992) noted, “Although studies of word learning in children have documented the remarkable facility of preschool children to acquire new vocabulary…the factors underpinning the large individual differences in young children's abilities to learn new words are as yet little understood” (p. 887). Several research studies (e.g., Bowey, 2001; Gathercole et al., 1992, 1999) have therefore focused on the importance of phonological skills in promoting vocabulary learning. Tasks of memory, phonological awareness, and articulation have all been used to explain vocabulary growth (e.g., Bowey, 2001; Gathercole et al., 1992, 1999).

Different researchers have argued convincingly that particular different phonological skills are especially important for vocabulary development. For example, Gathercole and colleagues (1992, 1999) demonstrated that phonological memory (as measured by nonword repetition) was strongly associated with vocabulary acquisition in both young children and adolescents. Avons and colleagues (Avons et al., 1998) obtained similar results in their study of preschool children. However, their assessment of what they termed phonological short-term memory also included what others might term phonological sensitivity, a measure of rhyme awareness. These researchers found that, even controlling for initial vocabulary knowledge, both rhyme detection and memory span were significant predictors of preschoolers' vocabulary skill 1 year later (Avons et al., 1998). Metsala (1999), in contrast, showed that both measures of phonological awareness and nonword repetition skills were strongly associated with vocabulary among 3- to 5-year-olds. Bowey (2001) also noted that phonological sensitivity, in addition to nonword repetition, predicted vocabulary development in young children. However, she suggested that perhaps both phonological sensitivity and nonword repetition were representative of general phonological processing. Overall, studies on the associations between phonological processing and vocabulary knowledge have focused on different phonological processing skills and tested these associations in samples of different ages. Given these varied findings, in the present study we sought to measure general phonological processing using tasks of speeded naming, nonword repetition, and phonological awareness. The associations of these with vocabulary knowledge were not the primary focus of the study but, rather, served as measures to be statistically controlled in our exploration of the role of morphological awareness in vocabulary knowledge.

WHAT IS MORPHOLOGICAL AWARENESS?

Our working definition of morphological awareness is awareness of and access to the meaning and structure of morphemes in relation to words. Morphemes are the smallest units of meaning in language. Carlisle (1995, p. 194) similarly defines morphological awareness as, “children's conscious awareness of the morphemic structure of words and their ability to reflect on and manipulate that structure.” Our focus is on children's abilities to distinguish and manipulate morphemes at the word level. This broad definition allows us, theoretically, to consider children's knowledge of both derivations and inflections in language simultaneously. Derivational morphology includes knowledge of prefixes (e.g., the un in undisciplined or the pre in preoperational), suffixes (e.g., the ation in graduation or simulation), and compounding (e.g., cowboy and sunlight are both compound words). Inflectional morphology focuses primarily on indicating grammatical changes in words (e.g., the s in dogs or the ed in acted are both grammatical inflections). Our own crosslinguistic work has convinced us that the concept of morphological awareness must be flexible if it is to be used successfully across languages. For example, inflectional morphology is obviously important in English or Finnish, but is relatively unimportant in Chinese. In contrast, lexical compounding is far more common in Chinese than it is in English.

In the present study, we acknowledge the utility of a connectionist approach in exploring the broad associations among aspects of sound, meaning, and orthography across languages (Gonnerman, Seidenberg, & Andersen, 2005). According to Gonnerman et al. (2005), understanding and speaking any given language involves an interaction of semantic representations and phonological patterns through a weighted distribution of processing. In any language, it is the strength of associations of the connections among processing units (similar to neurons in the brain) that determines how words are learned and accessed. In this view,

We view this perspective as a flexible one, capable of simultaneously accounting for development in both derivational and inflectional (e.g., Plunkett & Marchman, 1991, 1993) morphology. This connectionist perspective on morphological structure can also account for crosslinguistic differences in priming of words. For example, Plaut and Gonnerman (2000) found that in a morphologically rich artificial language, priming of words that shared some morphological association but were semantically opaque (such as miktav [letter] and katava [article] in Hebrew; from Bentin & Feldman, 1990) occurred in their model. In contrast, the same effects did not occur in the model when it was trained in an artificial language created to be relatively morphologically opaque. These effects suggest that languages that are relatively morphologically rich and transparent will tend to preserve the associations of sound and meaning in language, even when the associations of two words share only an apparent similarity in form. Our previous work on morphological awareness in relation to Chinese character recognition (McBride–Chang, Shu, Zhou, Wat, & Wagner, 2003) piqued our interest in studying morphological awareness in English.

Across languages, the boundaries across categories of morphology may become increasingly fuzzy and confusing. For example, in Finnish, inflectional morphology is complex relative to English (Lyytinen & Lyytinen, 2004) and is a prominent early marker of language impairments in childhood, whereas in Chinese, inflectional morphology is limited (Packard, 2000). Thus, the contrast in the importance of inflectional morphology for early language learning across languages can be quite striking, as in this example. Furthermore, in many Indoeuropean languages, as Casalis and Louis–Alexandre (2000) discuss, inflectional morphology tends to develop relatively early, whereas derivational morphology knowledge continues to grow through primary and secondary school and beyond. Inflectional and derivational morphology differ in other ways as well. For example, inflectional morphology is relatively limited and does not change the grammatical category of a word (e.g., from a verb to a noun). In contrast, derivational morphology can change these categories (e.g., manage/manager or overwork/workman).

Despite these contrasts, however, our focus in the present study on awareness of and access to unspecified categories (including both inflectional and derivational, bound and unbound) of morphemes in words lead us to construct a task of morphological structure knowledge that included different aspects of morphological knowledge. Given our previous experiences and definitions by others (Carlisle, 1995; Gonnerman et al., 2005) on the nature of morphology and morphological awareness in language, our measure of morphological awareness was flexible in including both lexical compounding and inflectional grammar in our task. Admittedly, this flexibility may represent a confounding of distinct aspects of morphological awareness, as might be argued by some linguists. On the other hand, this confounding may also be relatively faithful to children's language development, which may include simultaneously developing knowledge of lexical compounding and inflectional grammar in English for children of the age groups included in the present study, kindergarten and second grade.

Previous research has suggested that various aspects of morphological awareness may be particularly useful for vocabulary building. For instance, Wysocki and Jenkins (1987) found that students in fourth to eighth grade were able to learn some new words by generalizing from those sharing a root morpheme. An example might be comprehending the word piety based on previous knowledge of the word pious. Wysocki and Jenkins (1987) termed this ability morphological generalization. Others (e.g., White, Power, & White, 1989) have argued that such generalization does facilitate vocabulary building, provided that students are taught morphemes within an appropriate linguistic context. In young children, Lyytinen and Lyytinen (2004) demonstrated that poor production of inflectional morphology at age 2.5 years was a strong marker of subsequent language learning difficulties at age 5 in children at risk for dyslexia.

MORPHOLOGICAL AWARENESS AND LITERACY

Morphological awareness and vocabulary knowledge are often discussed in the specific context of literacy learning as well. For example, Carlisle (2000) found individual differences in morphological awareness for third- and fifth-grade students to be uniquely predictive of their reading comprehension. Another study of at risk second-grade readers (Nagy, Berninger, Abbott, Vaughan, & Vermeulen, 2003) also demonstrated that morphological awareness uniquely predicted reading comprehension in these children, although not in fourth graders at risk for writing difficulties. Fowler and Liberman (1995) showed that word reading was significantly correlated with tasks of morphological awareness, even controlling for age and vocabulary level, among second to fourth graders. Carlisle and Nomanbhoy (1993) also found that a measure of morphological production significantly predicted word reading in first graders, once phonological awareness was statistically controlled. Morphological production measured in first grade was significantly related to word recognition and reading comprehension in second grade, with phonological awareness controlled as well (Carlisle, 1995).

Different types of morphological awareness measured at different ages appear to have different effects on different aspects of literacy development. For example, measures of inflectional grammar tended to be associated with a stage model of spelling in a study of children ages 6–10 years (Nunes, Bryant, & Bindman, 1997). In that study, it was demonstrated that children's knowledge of spelling of inflectional morphemes tends to progress from having no clear association with grammatical category (e.g., kist for kissed) to an overgeneralization of morphemes across verbs (e.g., sleped for slept) to a clear understanding of regular versus irregular endings. Thus, morphological awareness tended to progress in a clear developmental pattern in relation to spelling of morphemes indicating past tense. Deacon and Kirby (2004) demonstrated that the same measure of inflectional grammar used in that of Nunes et al. (1997) administered in second grade also predicted unique variance in measures of pseudoword reading and reading comprehension but not word reading 4 years later (when the children were in fifth grade), even after statistically controlling for measures of intelligence and phonological awareness. They interpreted these results as indicating a clear distinction between morphological awareness and phonological awareness in reading acquisition.

Indeed, because English morphemes are strongly associated with phonological units such as syllables or phonemes, researchers sometimes wonder about the extent to which tasks of morphological and phonological awareness are distinguishable (e.g., Mann, 2000). Most studies of morphological awareness and English word recognition underscore the strong association of morphological and phonological awareness (Carlisle, 1995; Egan & Pring, 2004; Fowler & Liberman, 1995). For example, in their discussion of factors predicting reading and reading disability, Fowler and Liberman (1995) argued that “…if morphology does play a separate role from phonology…it is a small role indeed” (p. 179). In the present study, we explore the utility of morphological awareness, as overlapping but separable from phonological processing in understanding vocabulary development using somewhat different tasks of morphological awareness than have been used previously.

DISTINGUISHING TWO ASPECTS OF MORPHOLOGICAL AWARENESS

Like phonological processing, morphological awareness is likely comprised of multiple dimensions. We focused on two such dimensions that might ultimately be useful for understanding vocabulary growth because they can be assessed in very young children without using print. These constructs were originally developed in Chinese (McBride–Chang et al., 2003) in relation to early Chinese character recognition. Both are adaptable across languages and may be distinct features of morphological awareness in relation to vocabulary acquisition.

Morpheme identification is the ability to distinguish different meanings across homophones. This skill is demonstrated when a child understands that the flower in flowerpot is represented by a plant with petals as opposed to a sack of white powder (flour). This aspect of morphological awareness might help young language learners to distinguish among meanings of syllables with identical sounds, facilitating language analysis and vocabulary growth. In this case, morphological awareness involves understanding that different meanings can simultaneously be attributable to phonologically identical words. Using oral language measures, it is relatively clear that homophone distinction may involve morpheme knowledge that is separable from phonology, which is held constant in these examples. Morpheme identification is distinct from vocabulary knowledge because it relies on understanding of possible multiple meanings in the same spoken morpheme. Oral vocabulary knowledge requires learning to map a given spoken word to a given meaning. In contrast, morpheme identification requires that the meanings of two or more morphemes that are identically pronounced can be distinguished based on meanings.

A second aspect of morphological awareness, morphological structure awareness is the ability to create new meanings by making use of familiar morphemes. Berko's (1958) study of children's grammatical knowledge illustrates this skill. A child who understands that the famous concept of greater than one wug is represented by the word wugs, involving two morphemes, demonstrates morphological structure awareness skill. This particular example reflects inflectional knowledge of morphology. Morphological structure awareness can also be demonstrated by tapping derivational knowledge of morphology, via compounding. For example, the child who understands the idea of treetop as the highest point in a tree might be encouraged to think of a new term to represent the lowest point in the tree. One who asserts that treebottom is a reasonable term for this low point in the tree would, in our view, demonstrate morphological structure awareness. The core ability demonstrated by both the wugs and treebottom examples is the ability to construct new meanings from knowledge of previously learned morphemes. This morphological structure task requires children to make use of linguistic knowledge to derive new meanings. Skill in manipulating language, variously referred to as generativity, creativity, or productivity of language, may be important in learning new meanings within one's language.

Because morpheme identification and morphological structure awareness require somewhat different types of morphological analysis, we were interested in the unique contributions of each to vocabulary knowledge. In contrast, the extent to which tasks of morphological awareness would uniquely explain vocabulary skill once other phonological processing skills, including word recognition itself, were statistically controlled was unclear because phonological and morphological awareness tend to be strongly associated in English (Mann, 2000). Given that phonological processing skills are strong predictors of vocabulary skill across several studies (e.g., Avons et al., 1998; Bowey, 2001), morphological awareness might be linked to vocabulary skill only through its shared variance with phonological processing skills.

In the present study, we tested the associations of morphological structure awareness and morpheme identification with measures of reading and vocabulary in kindergarten and second-grade American children. In particular, we tested the extent to which each of these measures of morphological awareness might uniquely predict vocabulary across these two ages. We view this as a tentative step in understanding cognitive factors underlying vocabulary. If morphological awareness is strongly predictive of vocabulary knowledge, this might suggest that early tutoring in morphological structure awareness or morpheme identification might facilitate children's reading and/or language comprehension development. This hope was the backdrop for the present correlational study.