Young children acquiring second language vocabulary in preschool group-time: does amount, diversity, and discourse complexity of teacher talk matter?
How do young language minority immigrant children learn words in spontaneous everyday talk in preschool and school group settings? This is the central research question to be examined in this article. We know that vocabulary development is highly variable among children, both first and second language learners. For first language learners, this variation has been strongly related to parental socioeconomic status and education (Hart & Risley, 1995; Hoff, 2003) as well as to quantities and qualities of oral language exposure (Bornstein, Haynes, & Painter, 1998; Childers & Tomasello, 2002; Hart & Risley, 1995; National Institute of Child Health and Human Development, 2000; Pan, Rowe, Spier, & Tamis-Lemonda, 2004; Pan, Rowe, Singer, & Snow, 2005; Tabors, Snow, & Dickinson, 2001; Weizman & Snow, 2001). Much less attention has been paid so far to the role of lexical input in spontaneous talk in second language vocabulary acquisition (for a discussion, see Ellis, 1999). To explore this role is the motivation behind the study presented here, which addresses relationships between qualities of the talk young second language learners hear in preschool and school and their vocabulary acquisition. The impact of second language vocabulary exposure in teacher-led multiparty settings on young children's vocabulary acquisition is examined, focusing on Turkish-speaking children in Norwegian preschools and schools. Immigrant children in Scandinavia have been found to have limited second language vocabularies, even when their teachers regarded them as almost fluent second language speakers (Jorgensen, 2003). Given the importance of vocabulary for later academic achievement and reading comprehension for both first language (Dickinson, McCabe, Anastasopoulos, Peisner-Feinberg, & Poe, 2003; Scarborough, 2001; Storch & Whitehurst, 2002) and second language learners (Carlisle, Beeman, Davis, & Spharim, 1999; Droop & Verhoeven, 2003; Garcia, 2003; Proctor, August, Carlo, & Snow, 2006), identifying qualities of spontaneous talk exposure that support second language vocabulary development in young immigrant learners is of interest to both researchers and practitioners.
Most studies that explore relationships between qualities of exposure to spontaneous everyday talk and word learning have focused on first language learners. Due to the lack of studies examining second language vocabulary learning in settings of spontaneous talk, previous research on relationships between talk exposure and vocabulary learning in monolingual children will be included in the review, while the author acknowledges that significant differences may exist between preschool children's first and second language word learning.
This study is anchored in a socio-pragmatic approach to word learning (Nelson, 1996; Tomasello, 1999; Weizman & Snow, 2001). The proposal within this approach is that word learning begins when children start figuring out the meaning of the words they hear by inferring the speaker's attentional focus. In young children, this process may be scaffolded by the jointly engaged nature of conversations they take part in. Older preschool children may find support for word meaning interpretation by attending to the discourse context in which new words appear. Children who are offered pragmatic support when exposed to new words seem to acquire larger vocabularies (Weizman & Snow, 2001). Children's capacities for social analysis may be part of the reason why word learning is rapid in the preschool years (Beals, 1997; Tomasello, 1999; Weizman & Snow, 2001).
Most studies of word learning in young children have examined caregivers' speech to children in dyadic or small-group interactions, in which children can be addressed in ways that are attuned to their comprehension skills. Research in a variety of cultural and interactional contexts has, however, indicated that children can and do learn language as well as other socio-cognitive skills from keenly observing the interactions of others and listening in on talk (Crago, 1992; Forrester, 1988; Lieven, 1994; Rogoff, Paradise, Mejia-Arauz, Correa-Chavez, & Angelillo, 2003; Saffran, Newport, Aslin, Tunick, & Barrueco, 1997). Akhtar, Jipson, and Callanan (2001) found that 2-year-olds were equally good at learning novel words (words that they had never heard in speech directed to them) when they overheard them as when they were directly addressed. The findings from an even more recent study (Akhtar, 2005) highlight the active role played by children in vocabulary acquisition when listening in on others talking. Thus, although there is robust evidence suggesting that word learning is supported in situations of interaction, participation, and negotiation, the socio-pragmatic line of research offers a broadening conception of ways of vocabulary learning by including the learning that may take place when children overhear others talking.
Lexical Input and Vocabulary Acquisition This section will review three dimensions of talk input that have been found to have an impact on children's word learning. First, research on monolingual children's vocabulary has suggested that the sheer amount of talk children hear may promote acquisition. Purely quantitative factors, such as the frequency with which children hear a new word, have been found to affect children's learning (Childers & Tomasello, 2002). A major factor explaining differences in vocabulary size for young monolingual children (prior to the acquisition of literacy) is the number of words to which they have been exposed during the first years of life. The more language that children hear, the larger their vocabularies are (Hart & Risley, 1995; Hoff & Naigles, 2002).
Second, in addition to the number of words heard, the nature of them appears to be an important determinant of children's vocabularies. Studies have found an association between preschoolers' vocabulary knowledge and their exposure to low-frequency vocabulary in everyday mother-child conversations (Beals, 1997; Weizman & Snow, 2001) and in teacher-child conversations in the preschool classroom (Dickinson, 2001; Han, Roskos, Christie, Mandzuk, & Vukelich, 2005), suggesting that children acquire larger vocabularies as a result of exposure to diverse and rare words, not just exposure to a lot of talk. Work with monolingual children from low-income families in the United States suggests that diversity in lexical input was a better predictor of child vocabulary outcome than was the number of words addressed to the child per se (Weizman & Snow, 2001), or was the only predictor, as opposed to number of words (Pan et al., 2005).
Third, central to Nelson's approach to word learning is the claim that word learning is embedded in discourse contexts that support children's inferences about relevant word meaning: "Thus the general solution to the puzzle of how words are learned at any age lies in the fact that people--children as well as adults--use discourse context to interpret language" (Nelson, 1996, p. 139). With a few noticeable exceptions, questions of word learning in light of word exposure within varying discourse complexity have not been examined. Beals (1997) and Weizman and Snow (2001) pointed to the importance of the transparency of the pragmatic contexts in which children heard new words being used, and to the engagement with which the topic was talked about. Beals (1997) found positive relationships between mothers' informative use of rare words in discourse and children's later vocabulary size, suggesting that discourse supported word learning: "Words are not stand-alone mental objects that have some abstract semantic content. It is in conversations with others that children are exposed to words embedded in the messiness of everyday life" (p. 692).
Talk in group settings cannot be attuned to individual children's comprehension needs to the same extent as in the mother-child interaction studied by Weizman and Snow and by Beals. A handful of studies have examined relationships between talk exposure in preschool and children's vocabulary acquisition. McCartney (1984) found that children's language growth was strongly associated with the amount of time they spent talking with and listening to adults. The most beneficial type of talk was representational talk, which is talk that communicates information and is not used to control children's behavior. Similarly, Dickinson (2001) found that group-time conversations that were cognitively challenging and included a nonpresent topic were associated with stronger vocabulary scores. Further evidence of the potential of group times for supporting children's development came from the finding that the percentage of talk coded as explanatory in preschool was related to several child language measures in kindergarten, with the result being strongest for the children's receptive vocabulary. This finding is in accordance with various other studies revealing relationships between children's exposure to and participation in extended discourse inside and outside of classrooms (explanatory talk being one such type of talk) and children's language growth (Aukrust, 2004; Beals, 1991; Blum-Kulka, 2002; Snow, Cancino, De Temple, & Schley, 1991; Tabors et al., 2001).
This review has suggested three dimensions of talk exposure that seem to impact children's first language vocabulary acquisition: amount of talk heard, vocabulary diversity, and discourse complexity. Few studies have examined longitudinal relations between talk exposure in preschool and vocabulary acquisition in kindergarten and school. Dickinson (2001, see above) found associations between cognitively challenging preschool group-time conversations and kindergarten vocabulary scores. Furthermore, Tabors, Snow, and Dickinson (2001) reported on positive correlations between scores on children's kindergarten extended discourse participation (participation in talk that developed understandings beyond the here and now and that required the use of several utterances to build a linguistic structure) and 4th- and 7th-grade children's receptive vocabulary. They concluded that extended teacher discourse in preschool was a unique predictor of the children's later receptive vocabulary. Weizman and Snow (2001) similarly found a unique long-term impact on vocabulary acquisition of being exposed, in a home setting, to rare words in conversational contexts offering much interpretive support (also when controlling for vocabulary skills at the time of exposure). These studies suggest that acquiring a vocabulary is hard work for young children. A rich language environment in preschool may set off a vocabulary learning process that over time may result in a larger vocabulary, pointing to the importance of studying associations between rich language environments and vocabulary acquisition in both immediate and long-term perspectives.
Lexical Input and Second Language Vocabulary Acquisition: Research Questions
The studies on first language learning reviewed above suggest that words heard in semantically rich contexts were learned better. The assumption is that children are offered more opportunities to develop a rich understanding of the various meaning aspects of words when they appear within diverse and complex discourse contexts. Children who speak one language at home and another in school depend, to a large extent, on the language input they receive at school in their second language vocabulary acquisition. While children using the same language at home and at school continuously encounter the same words in a variety of situations and contexts in and out of school, second language learners encounter the challenge of acquiring the meaning of words solely heard in the context of school. Moreover, because new words often appear in multiparty school contexts in which talk is less tailored to the second language learner's level of understanding, word meaning learning may rely on the interpretive support offered in the discourse contexts in which new words are embedded.
Bilingual children may use their first language as an interpretive resource when trying to figure out the meaning of new words in another language, particularly when the two languages have many cognates. In a study of Spanish-speaking children acquiring English as a second language, Ordonez, Carlo, Snow, and McLaughlin (2002) examined bilingual 5th-graders' ability to produce word definitions and object descriptions in the two languages, and found that skills in word definition production in the first language were a reliable predictor of the same skill in the second language, but only when controlling for the number of words children knew in each language. Some intervention studies indeed suggest that when children receive support from their teachers in discovering how they may use transfer to infer word meaning in cognate words, their vocabulary improves (for a review, see Ordonez et al). Leseman (2000), on the other hand, found either weak (for receptive vocabulary) or absent (for productive vocabulary) cross-language transfer in a group of Turkish immigrant children learning Dutch. When examining the unique contribution of lexical input in preschool on children's second language vocabulary acquisition, relations between children's first and second language vocabulary should be considered.
This study explores relationships between lexical input in teacher talk in preschool and 1st-grade circle-time and Turkish-speaking second language learners' vocabulary acquisition in Norway. Norwegian preschools are known to value children's free play and peer interaction and de-emphasize school preparation per se (for Norwegian preschool ideology compared to various other countries, see Aukrust et al., 2003; Edwards et al., 2006). Along a scale of incidental, contextualized language learning versus learning through direct instruction, Norwegian preschool and elementary school teaching is close to the incidental end of the scale (Bezemer, Kroon, Pastor, Ryen, & Wold, 2004; Einarsdottir & Wagner, 2006). Relations between Turkish-speaking second language learners' peer play interaction and their vocabulary acquisition in Norwegian preschools have been examined elsewhere (Aukrust, 2004; Rydland & Aukrust, 2005; Rydland, in press). Preschools typically also include one daily teacher-led full-group gathering, considered important for creating a sense of group belonging and community in the children. Circle time, moreover, is the most school-like situation taking place in Norwegian preschool everyday life, as it is characterized by group interaction and attendance to decontextualized topics (topics beyond the here and now). I decided to explore relations between the lexical input the children received in this distinct situation and their second language vocabulary acquisition.
Extending the need for more knowledge of how aspects of talk exposure in spontaneous preschool discourse over time may impact young second language learners' vocabulary acquisition, the study raises the following more specific research questions:
1. Are there relationships between teacher talk lexical input (amount, diversity, and complexity of talk) and young children's concurrent second language vocabulary skills in preschool and 1st grade?
2. Does lexical input in preschool teacher talk (amount, diversity, and complexity of talk) predict variance in children's 1st-grade vocabulary skills?
Participants, Schools, and Sites
Participants were children speaking Turkish as their home language and Norwegian as their school language. Of the 27 children participating, 18 (9 boys and 9 girls) were recruited into the project in their next-to-last year of preschool attendance, while 9 children (6 boys and 3 girls) were recruited in their last preschool year. The children who were recruited when they were 4 were visited for the first time during the spring term (March-May) of their next-to-last year of preschool attendance, and were visited again 12 months later when they were 5, as well as visited a third time during the spring term of 1st grade. The children who were recruited when they were 5 were visited for the first time during the spring break (March-May) of their last preschool year and a second time one year later in 1st grade. Table 1 offers an overview of the number of children for which data is available at ages 4, 5, and 6, respectively.
The children were recruited through their preschools and lived in two large cities in Norway. Preschools in areas that were known to have many immigrants were contacted and their teachers were asked to distribute information about the study to parents of Turkish origin. Parental consent was based on information offered orally by teachers as well as in written form, in Turkish and Norwegian.
The children attended a variety of preschool and 1st-grade classrooms, all public and located in the multicultural neighborhoods of the children's homes. All lead teachers were fluent speakers of Norwegian, while some of the teacher assistants were native speakers of minority languages, in addition to speaking Norwegian as their second language. The common language in all preschools and schools was Norwegian, while the children spoke a variety of languages at home, with Urdu, Punjabi, Somali, Arabian, and various Eastern European languages being the most used languages in the children's homes, in addition to Turkish. While all the preschools and schools followed immersion programs, the specific form of language immersion that was practiced varied, and some of the children received some additional literacy teaching in their mother tongue in 1st grade.
Parental Education. Except for two children, who lived with their mothers, all children lived together with both parents. Of the 27 mothers, 18 had received all of their education in Turkey, while 9 received all or some of their education in Norway. Most mothers (14) had completed five years of school, 8 mothers had middle school, and 5 had high school (one of these also had some college education).
Just less than half of the fathers had earned a middle school degree, while nine fathers had completed high school (one of them had, in addition, a college degree). The remaining fathers had five years of elementary school. Most parents who were schooled in Turkey had participated in a "Norwegian for immigrants" course. While parental education varied somewhat, the sample mostly included children who had parents with little formal education.
Information on home language was gathered in parental interviews and in interviews with the target children when they were 6 years old. Interviews confirmed that all of the children spoke Turkish at home with their parents; some of the children reported that they switched between Norwegian and Turkish when speaking with siblings, particularly when they met in the context of school.
Age of Preschool Entrance. The target children had a mean age of 2;9 years when they started in preschool (preschool entrance age varying between 1;5 years and 4 years). All children attended preschool the two last years before entering school, whether they were recruited into the project as 4- or 5-year-olds. All 4-year-olds had a minimum of eight months of preschool attendance at the first time of data collection.
Talk Exposure: Videotaping Circle Time Talk. Circle time was a daily recurrent language event in all the target children's classrooms, combining some routinized aspects (talk about presence and absence, day, week, month, schedule of the day) with opportunities for talking about specific topics brought up by either children or teachers. Each target child was videotaped in circle time once a year, the camera focusing on the entire group, including the teacher. Preschool groups were age-heterogeneous, including children from 3 to 6 years of age. Circle time in preschool was typically led by the lead teacher, but sometimes other adults (teacher assistants, co-teachers) participated. The children would be seated in a circle, sitting on pillows on the floor or on low benches. In some groups, the children were invited to present a story of their liking or show an object they had brought from home. The 1st-graders attended age-homogeneous classrooms, varying in class size from 15 to 27 children, and with two teachers. Most often, circle time was led by one of them; occasionally, both teachers shared the responsibility for circle time. The children were seated on the floor (as in preschool) or at their desks. Teachers in the Norwegian school system postpone systematic teaching of reading skills until the 2nd grade, but preliteracy games and tasks were occasionally part of circle-time gatherings. Monological book reading by teachers was not included in the recordings.
Some classrooms included a couple of target children. When both target children were present during videotaping, that same circle time was sampled as a basis for identifying the qualities of teacher talk these children were exposed to. For the 4-year-olds, the analysis builds on 13 circle-time recordings, on 20 recordings for the 5-year-olds, and on 26 recordings for the 1st-graders. The videotaped circle times had a mean duration of 17.13 minutes for the 4-year-olds, 18.45 minutes for the 5-year-olds, and 19.51 minutes for the 6-year-olds. The total set of videotaped circle times included 19,664 transcribed utterances and 117,500 tokens (words used). As circle time duration varied from classroom to classroom, measures of talk exposure are offered per minute.
Videotaped circle-time conversations were transcribed into computer files using the transcription conventions (CHAT) of the Child Language Data Exchange System (MacWhinney, 1995) (1). Utterance boundaries were based primarily on intonation contour, and secondarily on pause duration. Contextual information needed to understand the interaction was included in the transcripts. Transcription was limited to utterances in Norwegian. Utterances in other languages were marked, but not transcribed.
Identifying Vocabulary Skills. Words can be learned in many different ways. Researchers have made major distinctions between competence in comprehending versus producing words, in paradigmatic versus syntagmatic word knowledge, and in breadth versus depth of children's vocabularies (Nation, 2001; Read, 2004). Bilingual children have been found to have difficulties developing superordinate meaning relations, as they are less exposed to words in a variety of settings compared to monolingual children (Ordonez et al., 2002; Verhallen & Schoonen, 1998). In this study, vocabulary was measured as receptive vocabulary and as word definition skills and was assessed in both languages, using translated versions of the following instruments (2). Receptive vocabulary skills were measured with the Peabody Picture Vocabulary Test--III, form B (PPVT; Dunn & Dunn, 1997). For this task, the child was shown panels of four pictures, the assessor named one of the pictures, and the child was asked to point to the picture that matched the word said by the assessor. A total of 57 items out of the first 62 were administered to the children. Five items were omitted either because of their cultural inappropriateness or dialectical variation. Raw scores were used, as the PPVT has not been standardized for Norwegian. Word definitions were of interest because they constitute one example of what has been referred to as decontextualized language use--"language used in ways that eschew reliance on shared social and physical context in favour of reliance on a context created through the language itself" (Snow et al., 1991, p. 90). To assess word definition skills, the children were given the Word Definition subtask from Reynell Developmental Language Scales, administered and scored according to the manual (Hagtvet & Lillestolen, 1985). The child was asked to define eight familiar words (book, dress, towel, wash, sleep, hungry, cold, last), using the prompts of "What is--" or "What does--mean?" Each word definition was coded on a scale of 0 (no or irrelevant definition) to 2 (definition focuses on important function or reason, offers a superordinate concept, etc.), with a maximum possible score of 16 (3).
All children were tested individually in both Turkish and Norwegian at school. The tasks were presented by a native speaker in each language, and sessions in the two languages were separated by at least a day. Some children insisted on responding to the Turkish word definition task as a request for a Norwegian translation, even when tested by a native Turkish speaker, illustrating the extent to which testing is a situated activity. The test results of these particular children were not included, reducing somewhat the number of children with a complete data set in Turkish word definitions.
Data Analysis and Treatment
Three subsets of exposure measures were computed that reflected the amount of input, the diversity of words used, and discourse complexity.
Identifying Amount of Talk. Automated computer analysis of the transcripts using the facilities of the Computerized Child Language Analysis (CLAN, see MacWhinney, 1995) procedures yielded the number of word tokens (total number of words) produced during circle time. Amount of talk was identified as number of word tokens per minute.
Identifying Vocabulary Diversity. For ease in automatic searching of different words, the CLAN program known as FREQ was used. As the researchers were focused on vocabulary diversity, transcripts were screened against a list of exclusion words, which included: 1) the 150 most common words in Norwegian newspapers (4); 2) the following words, which were high-frequency in circle time conversations, but not among the most frequently used newspaper words: "ja" (yes), "nei" (no), "du" (you), "dere" (you), numbers from 1-12, all personal names; and 3) vocal gestures with discursive significance (e.g., "oj" (wow!)). Morphological variants of a single word root were considered to constitute a single word type. Different dialectical pronunciations of the same word were reduced ("dere"-"dokker," both versions of the plural "you"), as long as they were not semantically distinct. When two items were derived from the same root, but differed significantly in their semantics, they were both included (e.g. "troll"-"trollete," meaning "troll" and "troll-like"). The final word frequency lists that were computed for each circle time were examined visually to eliminate any inconsistencies in spelling/transcription. Vocabulary diversity was identified as number of word types (different words) per minute.
Identifying Discourse Complexity. Spontaneously occurring explanations were selected as a useful speech event for identifying discourse complexity. Explanations emerged in a variety of formats, topics, and participation structures in circle time, but shared the feature that they resulted from expressed or assumed problems of understanding. In extending the work of Blum-Kulka (2002), explanations were defined as exchanges characterized by one or more conversational partners making something plain or clear, following from a partner explicitly or implicitly expressing a lack of understanding of the ongoing interaction or from the speaker assuming such a lack of understanding.
Aukrust (2004) identified various types of explanatory utterances appearing in second language learners' peer play conversations; explanations addressing the immediate activity, principle-based explanations (referring to a principle, rule, or concept), and frame-based explanations (referring to the framing of conversations per se). While Aukrust found that principle-based and frame-based explanations showed associations with learner vocabulary acquisition in peer play, explanations addressing the immediate activity did not. For this reason, immediate activity explanations were not included in the present study. The following criteria applied to the identification of types and subtypes of explanatory talk in the teacher-student conversations: Principle-based explanations: 1) offered premises for or consequences of something; 2) addressed rules, principles, or social generalizations with the purpose of arguing for something; and 3) explored something conceptual (e.g., what something is like). Frame-based explanations, which were much less frequent than principle-based explanations, appeared as negotiations of the interaction per se (for examples, see Table 2). Explanatory types and subtypes were specified to support identification of explanatory talk, but no attempt was made at distinguishing between types or subtypes in the final analysis.
The following types of exchanges were not included as explanatory talk: naming of objects without any further conceptual exploration, simple descriptive sequences about future events ("og sa skal vi, og sa skal vi ..."/ "and then we are going to, and then we are going to ..."), simple teacher rephrasings, even though these might reflect the teacher's awareness of possible lack of student understanding ("skal de av dere som er storst # de store barna"/ "then those of you who are big # the big children"), and translations (e.g. from Norwegian to Turkish).
Explanations were first identified as segments of text, according to the criteria above (text-level). Second were those utterances that had explanatory power within such segments identified, not including utterances that did not contribute to the explanation and move it forward (utterance-level). The preliminary analysis identified discourse complexity in circle time conversations in three different ways: 1) as relative frequency of explanatory utterances (utterance-level) out of total number of utterances in circle-time, 2) as percentage of word types that appeared within explanatory talk out of total number of word types (word types identified according to the criteria specified in the section "Identifying vocabulary diversity"), and 3) as density of word types appearing within explanatory talk (calculated as number of word types within explanatory segments per minute). As all three ways of calculating discourse complexity resulted in similar findings regarding relationships between discourse complexity and target child vocabulary acquisition, alternative 3 (density of word types within explanatory talk) was used as a discourse complexity index.
Data Treatment. The analysis explored three predictors of vocabulary skills in 1st grade: amount of talk (number of word tokens per minute), vocabulary diversity (number of word types per minute), and discourse complexity (number of word types within explanatory talk per minute). Theoretical considerations as well as correlational analysis of relationships between background variables and outcome measures suggested that use of maternal education and age of preschool entrance should be used as control variables in the regression analysis to ensure that effects found were not caused by these dimensions. (Paternal education did not show associations to the children's vocabulary skills as did maternal education.) Maternal education was expressed as a continuous variable, where 1 = 5 years of schooling, 2 = middle school, and 3 = high school or more. Age of preschool entrance was expressed as a continuous variable, with 1 = preschool entrance at age 1, 2 = at age 2, 3 = at age 3, and 4 = at age 4.
Interrater reliability was estimated for explanations based on 20 percent of the circle times. The reliability was computed using Cohen's kappa, a measure of reliability corrected for chance occurrence. The Cohen's kappa statistics for the coding of explanatory talk were .77, indicating "substantial agreement" (Landis & Koch, 1977).
The reported interrater reliability in the scoring of Reynell word definition scale, Norwegian version, varies from .93 to .97 (Hagtvet & Lillestolen, 1985). The reported median reliability coefficient for the PPVT-III (English version) is .94 (Dunn & Dunn, 1997). To get a check on the validity of the translation of PPVT, test results in Norwegian PPVT (henceforth referred to as PPVT-NO) were correlated against the receptive language scales of Reynell Developmental Language Scales, a standardized instrument with Norwegian norms, for the 4- and 5-year-olds. Correlations between the two tests were high (r = 0.73, p < .001 for the 4-year-olds, and r = 0.74, p < .001 for the 5-year-olds), suggesting that the translated version of PPVT-NO measured skills similar to the receptive language scale test that was standardized for Norwegian.
The central motivation behind this work was to examine whether exposure to word tokens, to word types, and to word types within complex discourse contributed to the explanation of variance in vocabulary outcomes, both immediately (research question 1) and in a long-term perspective from preschool to 1st grade (research question 2). To address these questions, descriptive information on the children's vocabulary outcomes and associations between these measures and demographic variables is provided first. Second, variation in teacher talk lexical input at ages 4, 5, and 6 is presented. Third, I turn to research question i and the analysis of correlations between lexical input and vocabulary outcomes to detect possible immediate predictor-outcome relationships. Finally, in response to research question 2, the results of the regression analysis of long-term predictor-outcome relationships are reported.
Outcome Measures: How Did Children's Vocabulary Skills Vary Across Ages, Vocabulary Types, and Languages?
Across ages 4, 5, and 6, the children displayed more developed vocabulary skills (both receptive vocabulary skills and word definition skills) in their home language (Turkish) than in their school language (Norwegian). Test results thus confirmed information offered by children as well as parents in interviews. Table 3 reports on the mean raw scores of the PPVT test and the word definition test in both Norwegian and Turkish at ages 4, 5, and 6, showing an increase in mean raw scores in both languages across age groups. The standard deviations for both PPVT scores and word definition scores, furthermore, indicated a wide variation among the children.
The children's PPVT-NO scores and word definition scores in Norwegian (WD-NO) correlated for each age group as well as across age groups (see Table 4). Relations between vocabulary measures showed a more mixed pattern in Turkish, with weaker or absent correlations between children's scores on, respectively, the PPVT test and the word definition test at ages 4, 5, and 6 (for PPVT-TU and WD-TU at age 4, r = .65, p < .05; at age 6, r = .41, p < .05; at age 5, non-significant). The sole cross-language correlation that appeared was between the children's PPVT scores in Norwegian and Turkish at age 5 (r = 0.43, p < .05).
Maternal education correlated with most second language vocabulary measures; with PPVT-NO at ages 4 and 6 and with word definition skills in Norwegian at all three ages. In spite of the fact that Turkish first and foremost was learned at home, maternal education did not show a similar significant association with the children's PPVT-TU scores or their word definition skills in Turkish. Moreover, even though Norwegian was learned in preschool, the duration of preschool attendance per se did not relate to outcome measures. Maternal education correlated with the children's age of preschool entrance (r = -.50, p < .01), reflecting that mothers with more education tended to put their children into preschool at an earlier age.
Input Measures: How Did Dimensions of Teacher Talk Lexical Input Vary Across Classrooms for 4-, 5-, and 6-year-old Children?
No significant differences appeared between preschool and school in amount of teacher talk (density of word tokens), in vocabulary diversity (density of word types), and in discourse complexity (density of word types within explanatory talk). The mean numbers were slightly higher in 1st grade for all dimensions of teacher talk lexical input, but due to high standard deviations, these differences between preschool and 1st grade were not significant. Overall, there were large differences from one classroom to the other in teacher lexical input.
The teachers were the main contributors of talk in this specific setting. Teachers offered approximately 60 percent of all utterances during circle time (61.6 percent for the 4-year-olds, 59.2 percent for the 5-year-olds, 61.4 percent for the 6-year-olds). When two teachers participated in circle time, their shared contribution was included in the count. These findings regarding participation structure are in line with other studies, which have found that teachers in group settings offer approximately two-thirds of all utterances (for a review, see Cazden, 2001). Teachers furthermore offered between 73.5 percent and 79 percent of all word tokens, and between 84.7 and 87.5 percent of all word types. Conversely, the target children participated sparsely in circle time, typically offering some few utterances on their own. The teachers generally minimized the demands on linguistic contribution offered by single students in the group setting. The second language learners did, however, join into various forms of collective participation, which the teachers invited now and then.
The various measures of teacher talk complexity showed high correlations for each age group. At age 4, density measures of teacher tokens correlated with word types (r = .93, p < .001) and discourse complexity (r = .83, p < .001); at age 5, teacher tokens correlated with word types (r = .91, p < .001) and with discourse complexity (r = .69, p < .001). In 1st grade, the interrelation between word tokens and word types was r = .67 (p < .001) and between word tokens and discourse complexity r = .56 (p < .01).
The data presented so far have demonstrated that the children in the study were provided with a broad range of words that may have supported their learning, and have also revealed differences in circle-time vocabulary exposure across classrooms. Was exposure to rich teacher talk associated with simultaneous and subsequent target child vocabulary acquisition? In order to investigate these questions, correlations were first computed between teacher talk lexical input and child vocabulary at ages 4, 5, and 6, respectively. Second, a regression was undertaken to consider the long-term impact of preschool talk exposure to target children's school vocabulary.
Were There Relationships Between Teacher Talk Lexical Input and Target Children's Simultaneous Second Language Vocabulary Skills?
At age 4, correlations were found between the number of teacher word types and the target children's simultaneous PPVT-NO, and between teacher discourse complexity and PPVT-NO. When controlling for background factors (maternal education and age of preschool entrance), correlations between teacher talk and PPVT-NO at age 4 lost significance. No correlations were found between teacher talk exposure and the children's word definition skills at 4.
At 5 years, no dimension of teacher talk showed associations with the target children's simultaneous vocabulary skills (neither with PPVT-NO nor with word definition skills in Norwegian). Controlling for background factors did not change this picture.
At 6 years, teacher talk discourse complexity (word types within explanatory talk) correlated with PPVT-NO (r = .44, p < .05), and this correlation kept significance when controlling for background variables (r = .49, p < .05). Other dimensions of lexical input in teacher talk did not correlate with PPVT-NO. No associations were found between teacher talk and the target children's word definition skills at 6. These infrequent correlations all applied to simultaneous relations between teacher talk exposure and target children's vocabulary skills. One important research question remains--that is, whether teacher talk exposure in preschool predicted later vocabulary skills in 1st grade.
Did Teacher Talk Lexical Input in Preschool Predict Variation in Later School Vocabulary Skills?
Children's scores on subsequent vocabulary outcomes were regressed on measures of teacher talk input in order to address the research question concerning the impact of lexical input on vocabulary outcomes. As noted above, I included as controls factors that could introduce confounds into the analysis: number of years of maternal education was introduced as the first step and age of preschool entrance as the second. The child's PPVT-NO score or word definition skills score at ages 4 and 5, respectively, was introduced singly as a third control, to assess the degree to which differences in quality of teacher talk input were related to 1st-grade outcomes, independent of the child's earlier vocabulary skills.
First, effects of talk exposure at age 4 on word definition skills and PPVT at age 6 was examined. Teacher talk input at age 4 predicted significantly word definition skills at age 6. The result of this regression is summarized in Table 6. Maternal education predicted 37 percent of the variation in word definition skills at age 6; age of preschool entrance and word definition skills at age 4 did not add significantly to this prediction. Between 15 and 18 percent of the additional variance in word definition skills in 1st grade was predicted by teacher talk lexical input at age 4. Maternal education and teacher talk exposure at age 4 thus predicted approximately 60 percent of all variation in word definition skills in 1st grade.
Also, much of the variation in PPVT-NO scores at age 6 was predicted by maternal education (applied to 31 percent of the variance, p < .05). Age of preschool entrance did not add significantly to this prediction, while 28 percent (p < .05) of the additional variation in PPVT-NO scores in 1st grade was predicted by the child's PPVT-NO score at 4. The first three steps in this regression model thus accounted for 64 percent of the variation in children's PPVT-NO scores in 1st grade, while none of the three dimensions of teacher talk predicted additional variation. When the children's PPVT-NO scores at the age of 4 were removed from the model, with teacher talk exposure entered as the third step in the regression, teacher discourse complexity (word types within explanatory talk) predicted 17 percent (p < .05) of the variation in 1st-grade PPVT-NO.
Second, effects of talk exposure at age 5 were examined. Concerning word-definition skills, maternal education accounted for 32.6 percent of the variation in 1st-grade skills. Age of preschool entrance did not add to the prediction of variation, while word definition skills at age 5 did and predicted 19 percent of the additional variation in word definition skills at age 6 (p < .01). Altogether, 53 percent of the variation in 1st-grade word definition skills was predicted by the three first steps in the regression. Teacher talk lexical input at age 5 did not significantly predict any additional variation in word definition scores at age 6.
Finally, the regression testing the impact of various aspects of teacher talk at age 5 on PPVT-NO scores at age 6 accounted for as much as 64-65 percent of the age 6 variation in PPVT-NO scores (Table 7). Maternal education accounted for 23 percent of the variation, while age of preschool entrance did not reach significance. PPVT-NO scores at age 5 accounted for 20 percent of the additional variation in 1st-grade PPVT-NO scores (p < .05). The various dimensions of teacher talk exposure at age 5 predicted significantly between 18 and 19 percent (p < .01) of the variation in PPVT-NO scores in 1st grade.
As noted, there was considerable variation in teacher talkativeness (density of word tokens), in the number of different words used by teachers (density of word types), and in the extent to which teachers embedded words in explanatory talk. Teachers differed in the extent to which they used occasions that spontaneously occurred in the classrooms to introduce new words and to embed them in discourse contexts that supported understanding of word meaning. The following examples may serve to illustrate the types of teacher talk that predicted vocabulary skills acquisition in this study. In the first example, the teacher had just struck a match to light a candle, when a child, Ali, commented on the smell. The teacher used the opportunity to introduce the word "sulfur." The way she framed the word in the conversation suggested that she did not expect the children to know the meaning of the word:
Example 1: Circle time with 4-year-old target child
Ali: Det lukter royk--flamme (It smells of smoke--flame)
Teacher: Ja det lukter--svovel heter det--fyrstikken har svovel slik at den kan tenne og lyse--det er svovel. (Yes--something smells--sulfur is the name--the match has sulfur on the top so that it can catch fire and light--the smell is from sulfur).
In the next example, one of the children, Miyase, took the floor in circle time to tell everyone that she had seen slippers similar to Emre's in the newspaper. The teacher seemed uncertain about what Miyase was trying to say before suggesting that Miyase was perhaps referring to a newspaper advertisement.
Example 2: Circle time with target child Emre, 5 years old
Miyase: Vet dere hva jeg sa i avisen? (Do you know what I saw in the newspaper?)
Teacher: Nei. (No.)
Miyase: Jeg sa sanne tofler som Emre har (I saw slippers such as Emre has)
Teacher: Sa du sanne tofler som Emre har i avisen? (Did you see slippers like those Emre has in the newspaper?) (pause)
Teacher: Det var kanskje en reklame for en butikk hvor man fikk kjopt sanne tofler tror du det? (Maybe it was an advertisement for a shop where you could buy slippers like Emre's--do you think so?)
The third example refers to a teacher, who during circle time, observed a price tag on the t-shirt of a child and commented that "det var et roverkjop--et bra kjop--du var heldig" ("that was a great bargain--a good buy--you have been lucky"). The teacher, apparently using a word she was not certain that the children knew ("bargain"), supported their word understanding by offering an immediate explanation of the word.
While these excerpts are examples of words appearing within explanatory talk (subcategory "conceptual issues," Table 2), teachers also varied in the extent to which they exposed children to different word types outside of explanatory talk. An illustrative example of teacher talk with high word type density was found in the preschool classroom of a teacher who, during circle time, told and dramatized a well-known fairytale about three goats crossing a bridge with a troll below. She introduced a variety of different animals (lions, tigers, hyenas, alligators) and nature formations (desert, jungle, thicket) in the fairytale script as she narrated, exposing children to a lot of new words within a well-known format. Children who attended preschool classrooms in which teachers exposed them to a variety of different words and often did so within explanatory contexts had more developed vocabularies in 1st grade.
This study generated three sets of findings. The first set of findings demonstrated a linkage between teacher talk exposure during circle time in preschool and 1st-grade vocabulary performance. While only a few associations were found between teacher talk lexical input and the children's simultaneous receptive vocabulary skills and word definition skills (research question 1), teacher talk appeared to have a long-term impact on children's vocabulary acquisition (research question 2).
The second and third set of findings resulted from examining other factors that were expected to have an impact on associations between teacher talk exposure and vocabulary acquisition. The second set of findings applied to relations between background variables and vocabulary skills. Maternal education correlated highly with second language vocabulary skills, while age of preschool entrance, per se, was not important. Apparently, it was not the quantity of preschool attendance that mattered for vocabulary acquisition, but the quality of the talk addressed to the learners (given that all these preschoolers had at minimum eight months of preschool attendance when recruited into the study).
The third set of findings applied to cross-skill, cross-age, and cross-language associations. While second language receptive vocabulary skills and word definition skills correlated highly with each other and across ages, only one weak cross-language association was found (PPVT in Turkish and Norwegian children at age 5). The findings are in accordance with those of Leseman (2000), who found only modest cross-language correlations between receptive vocabulary scores and no cross-language correlations between productive vocabulary scores in a group of Turkish immigrant children acquiring Dutch. They are different, however, from the results of a study by Ordonez et al. (2002), who did find transfer of word knowledge (as displayed in word definitions) from Spanish to English. Spanish and English are languages with more cognates than Turkish and Norwegian or Turkish and Dutch, and transferring knowledge of word meaning in one of the languages to cognates in the other may therefore be a more useful strategy for bilingual Spanish- and English-speaking children than for children speaking Turkish and Norwegian. Methodological issues also should be considered, as the items used in the word definition test (in Turkish and Norwegian) were all of a fairly everyday type. It is possible that testing the children's understanding of more academic, technical, or low-frequency words would have resulted in other cross-language relationship findings.
The remaining discussion will first attend to the impact of maternal education on second language vocabulary acquisition, and finally on aspects of talk exposure that predicted second language vocabulary acquisition. Relationships between maternal education and vocabulary development, robustly documented in former research on monolingual acquisition and reviewed in the introductory section, were also confirmed in this study. The associations between maternal education and second language vocabulary (which was not learned from the mothers) also point to the importance of framing studies of second language acquisition in motivational and identity formation processes. Children with better educated mothers may have received more motivational support from their mothers in acquiring their second language (for a discussion of relationships between marginalization and psychological adaptation in the Turkish adolescent population in Norway, see Virta, Sam, and Westin, 2004). Teacher talk lexical input in preschool thus did not work outside of maternal closeness or distance to the education system, as reflected in maternal level of education. Moreover, the children's second language vocabularies were not related to years of preschool attendance per se. It is, however, important to recognize that none of the children were newcomers in preschool. It is very possible that below a certain level, the duration of preschool attendance per se may be an influential factor in second language acquisition.
All the three dimensions of talk exposure that were identified contributed to the prediction of 1st-grade receptive vocabulary and word definition skills. Teacher talk exposure at age 4 predicted word definition skills at age 6, while teacher talk exposure at age 5 predicted receptive vocabulary at age 6. The fact that the details of the associations varied somewhat across age groups and vocabulary types, as well as the small number of participants in the study, suggest that caution is warranted in drawing conclusions. In spite of these limitations, the results indicate that teacher talk lexical input, as identified in the study, was beneficial to the children's vocabulary acquisition. The same talk exposure dimensions that have been found to have predictive power in studies of first language vocabulary acquisition (e.g., Tabors et al., 2001; Weizman & Snow, 2001) also appeared related to second language acquisition.
The teachers who offered more word types offered them more often within complex discourse contexts and also offered more words overall (tokens). The strong interrelationships between these dimensions of talk exposure are also apparent in the fact that combining them in a composite variable did not improve the prediction. In a recent study involving a sample of young low-income children (Pan et al., 2005), maternal talkativeness (total number of words produced, or word tokens) was not related to vocabulary growth, while the diversity of vocabulary input (number of different words produced, or word types) was. The question raised by Pan et al. is whether quantity and quality of lexical input to children may have different effects in low-income samples, characterized by other forms of communicativeness (e.g., more use of nonverbal means). While this question is clearly relevant for young second language learners who, more so than native speakers, may rely on nonverbal communication and on an otherwise interpretable discourse context to infer the meaning of new words, the separate impact of quantity (density of word tokens) versus quality (density of word types and of word types within explanatory discourse) cannot be distinguished in this study.
Teachers in circle time generally minimized the demands on linguistic contributions that they required from the second language learners. They responded with acceptance to children's unwillingness to talk in the large-group situation, treated students as collective participants, and invited second language learners to participate with simple responses. Opportunities to engage in such non-threatening participation, in combination with opportunities to learn by observing and listening in, may constitute beneficial learning experiences for young second language learners with limited capacities to use language-based negotiation strategies. It is, of course, very possible that teachers who offered a lot of words, and a variety of different words appearing in cognitively challenging discourse contexts, in this specific circle-time setting also did so in everyday preschool situations, which, to a larger extent, invited the second language learners' initiative and participation (e.g., in dyadic narrative productions about events of interest to the child and in dialogic bookreading). Therefore, we cannot exclude the possibility that teacher talk lexical input in more interactive and scaffolded activity initiated by the child was the real mechanism behind the associations between teacher talk lexical input and vocabulary acquisition revealed in this study. Recent research (e.g., Akhtar, 2005; Crago, 1992; Saffran et al., 1997) has, however, pointed to the importance of word learning through overhearing others talking. Extending this perspective, the findings presented here suggest that opportunities to listen to words within rich and complex exposure also support children's second language word learning. Akhtar (2005) concluded that as children participate in everyday multiparty interactions, they are likely to overhear conversations that include unfamiliar words. "While learning through overhearing may play a particular important role in communities in which children are relatively infrequently directly addressed, and in communities in which they hear few child-focused conversations, it is probably a context in which all children everywhere acquire some of their early vocabulary" (Akhtar, 2005, p. 208). Opportunities to listen to and overhear talk about topics that were interesting enough and complex enough to need a large repertoire of words and explanations to be covered seemed to support second language vocabulary acquisition, suggesting that there are many pathways to learning the words of a new language.
Finally, this study shows that a rich language environment in preschool may set off a learning process that over time may result in larger vocabularies, pointing to the importance of studying associations between rich language environments in preschool and vocabulary acquisition in a long-term perspective. While the children's second language vocabulary showed considerable stability from one year to another, teacher talk lexical input in preschool predicted uniquely second language vocabulary one and two years later, in 1st grade, when controlling for vocabulary at the time of exposure. While a few other studies also have reported on the long-term effects of rich lexical input on children's vocabulary years later (e.g., Weizman & Snow, 2001), the mechanisms behind the various word learning processes that may be triggered by more or less rich language environments are not fully understood and should be examined further in future research.
This research was supported by a grant from the Norwegian Research Council. The author gratefully acknowledges the support of Veslemoy Rydland and Kamil Ozerk in data collection. Thanks also to the children, parents, and teachers who contributed their time and conversations.
Akhtar, N. (2005). The robustness of learning through overhearing. Developmental Science, 8, 199-209.
Akhtar, N., Jipson, J., & Callanan, M.A. (2001). Learning words through overhearing. Child Development, 72, 416-430.
Aukrust, V. G. (2004). Explanatory discourse in young second language learners' peer play. Discourse Studies, 6, 393-412.
Aukrust, V. G., Edwards, C. P., Kumru, A., Knoche, L., & Kim, M. (2003). Young children's close relationships outside of the family: Parental ethnotheories in four communities in Norway, United States, Turkey, and Korea. International Journal of Behavioral Development, 27, 481-494.
Beals, D. (1991). "I know who makes ice-cream." Explanations in mealtime conversations of low-income families of preschoolers. Unpublished doctoral dissertation, Harvard University, Cambridge, MA.
Beals, D. (1997). Sources of support for learning words in conversation: Evidence from mealtimes. Journal of Child Language, 24, 673-694.
Bezemer, J., Kroon, S., Pastor, L., Ryen, E., & Wold, A. H. (2004). Language teaching and learning in a multicultural context. Case studies from primary education in the Netherlands and Norway. Oslo, Norway: Novus Press.
Blum-Kulka, S. (2002). "Do you believe that Lot's wife is blocking the road (to Jericho)?": Co-constructing theories about the world with adults. In S. Blum-Kulka & C. E. Snow (Eds.), Talking to adults: The contribution of multiparty discourse to language acquisition (pp. 85-116). Mahwah, NJ: Lawrence Erlbaum.
Bornstein, M. H., Haynes, M. O., & Painter, K. M. (1998). Sources of child vocabulary competence: A multivariate model. Journal of Child Language, 25, 367-393.
Carlisle, J. F., Beeman, M. M., Davis, L. H., & Spharim, G. (1999). Relationship of metalinguistic capabilities and reading achievement for children who are becoming bilingual. Applied Psycholinguistics, 20, 459-478.
Cazden, C. B. (2001). Classroom discourse: The language of teaching and learning (2nd ed.). Portsmouth, NH: Heineman.
Childers, J. B., & Tomasello, M. (2002). Two-year-olds learn novel nouns, verbs, and conventional actions from massed or distributed exposures. Developmental Psychology, 38, 967-978.
Crago, M. B. (1992). Communicative interaction and second language acquisition: An Inuit example. TESOL Quarterly, 26, 487-505.
Dickinson, D. (2001). Large-group and free-play times. Conversational settings supporting language and literacy development. In D. Dickinson & P. Tabors (Eds.), Beginning literacy with language (pp. 223-256). Baltimore: Paul Brookes.
Dickinson, D., McCabe, A., Anastasopoulos, L., Peisner-Feinberg, E., & Poe, M. (2003). The comprehensive language approach to early literacy: The interrelationships among vocabulary, phonological sensitivity, and print knowledge among preschool-aged children. Journal of Educational Psychology, 95, 465-481.
Droop, M., & Verhoeven, L. (2003). Language proficiency and reading ability in first- and second-language learners. Reading Research Quarterly, 38, 78-103.
Dunn, L. M., & Dunn, L. M. (1997). Peabody Picture Vocabulary Test, 3rd edition (PPVTIII). Circle Pines, MN: American Guidance Service.
Edwards, C. P., Knoche, L., Aukrust, V., Kumru, A., & Kim, M. (2006). Parental ethnotheories of child development: Looking beyond independence and individualism in American belief systems. In U. Kim, K. Yang, & K. Hwang (Eds.), Indigenous and cultural psychology. Understanding people in context (pp. 141-163). New York: Springer.
Einarsdottir, J., & Wagner, J. T. (Eds.). (2006). Nordic childhoods in early education. Philosophy, research, policy, and practice in Denmark, Finland, Iceland, Norway, and Sweden. Greenwich, CT: Information Age Publishing.
Ellis, R. (1999). Factors in the incidental acquisition of second language vocabulary from oral input. In R. Ellis (Ed.), Learning a second language through interaction (pp. 35-62). Amsterdam: John Benjamins Publishing Company.
Forrester, M.A. (1988). Young children's polyadic conversation monitoring skills. First Language, 8, 201-225.
Garcia, G. E. (2003). The reading comprehension development and instruction of English-language learners. In A. P. Sweet & C. E. Snow (Eds.), Rethinking reading comprehension (pp. 30-50). New York: The Guilford Press.
Hagtvet, B., & Lillestolen, R. (1985). Reynells spraktest. Oslo: Universitetsforlaget.
Han, M., Roskos, K., Christie, J., Mandzuk, S., & Vukelich, C. (2005). Learning words: Large group time as a vocabulary development opportunity. Journal of Research in Childhood Education, 19, 333-345.
Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday life of young American children. Baltimore: Brookes.
Hoff, E. (2003). The specificity of environmental influence: Socioeconomic status affects early vocabulary development via maternal speech. Child Development, 74, 1368-78.
Hoff, E., & Naigles, L. (2002). How children use input to acquire a lexicon. Child Development, 73, 418-433.
Jorgensen, J. N. (2003). Bilingualism in the Koge project. International Journal of Bilingualism, 7, 333-352.
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159-167.
Leseman, P. P. M. (2000). Bilingual vocabulary development of Turkish preschoolers in the Netherlands. Journal of Multilingual and Multicultural Development, 21, 93-112.
Lieven, E. V. M. (1994). Crosslinguistic and crosscultural aspects of language addressed to children. In C. Gallaway & B. J. Richards (Eds.), Input and interaction in language acquisition (pp. 56-73). Cambridge, UK: Cambridge University Press.
MacWhinney, B. (1995). The CHILDES project: Tools for analyzing talk. Hillsdale, NJ: Erlbaum.
McCartney, K. (1984). Effect of quality of day care environment on children's language development. Developmental Psychology, 20, 244-260.
Nation, I. S. P. (2001). Learning vocabulary in another language. Cambridge, UK: Cambridge University Press.
National Institute of Child Health and Human Development Early Child Care Research Network. (2000). The relation of child care to cognitive and language development. Child Development, 71, 960-980.
Nelson, K. (1996). Language in cognitive development: Emergence of the mediated mind. Cambridge, UK: Cambridge University Press.
Ordonez, C. L., Carlo, M., Snow, C. E., & McLaughlin, B. (2002). Depth and breadth of vocabulary in two languages: Which vocabulary skills transfer? Journal of Educational Psychology, 94(4), 719-724.
Pan, B. A., Rowe, M., Spier, E., & Tamis-Lemonda, C. (2004). Measuring productive vocabulary of toddlers in low-income families: Concurrent and predictive validity of three sources of data. Journal of Child Language, 31,587-608.
Pan, B. A., Rowe, M. L., Singer, J. D., & Snow, C. E. (2005). Maternal correlates of growth in toddler vocabulary production in low-income families. Child Development, 76, 763-782.
Proctor, P., August, D., Carlo, M., & Snow, C. (2006). The intriguing role of Spanish vocabulary knowledge in predicting English reading comprehension. Journal of Educational Psychology, 98, 159-169.
Read, J. (2004). Plumbing the depths: How should the construct of vocabulary be defined? In P. Bogaards & B. Laufer (Eds.), Vocabulary in a second language (pp. 209-227). Amsterdam: John Benjamins.
Rogoff, B., Paradise, R., Mejia-Arauz, R., Correa-Chavez, M., & Angelillo, C. (2003). Firsthand learning through intent participation. Annual Review of Psychology, 54, 175-203.
Rydland, V. (in press). "Whow--when I was going to pretend drinking it tasted coke for real!" Second-language learners' out-of-frame talk in peer pretend play. A developmental study from preschool to first grade. European Journal of Developmental Psychology.
Rydland, V., & Aukrust, V. G. (2005). Lexical repetition in second language learners' peer play interaction. Language Learning, 55, 229-274.
Saffran, J. R., Newport, E. L., Aslin, R. N., Tunick, R. A., & Bar rueco, S. (1997). Incidental language learning: Listening (and learning) out of the corner of your ear. Psychological Science, 8, 101-105.
Scarborough, H. S. (2001). Connecting early language and literacy to later reading (dis)abilities: Evidence, theory, and practice. In S. B. Neuman & D. K. Dickinson (Eds.), Handbook of early literacy research (pp. 97-110). New York: Guilford Press.
Snow, C. E., Cancino, H., De Temple, J., & Schley, S. (1991). Giving formal definitions. In E. Bialystok (Ed.), Language processing in bilingual children (pp. 90-112). Cambridge, UK: Cambridge University Press.
Storch, S. A., & Whitehurst, G. J. (2002). Oral language and code-related precursors to reading: Evidence from a longitudinal model. Developmental Psychology, 38, 934-947.
Tabors, P., Snow, C. E., & Dickinson, D. (2001). Homes and schools together: Supporting language and literacy development. In D. Dickinson & P. Tabors (Eds.), Beginning literacy with language (pp. 313-334). Baltimore: Paul Brookes.
Tomasello, M. (1999). The cultural origins of human cognition. Cambridge, MA: Harvard University Press.
Verhallen, M., & Schoonen, R. (1998). Lexical knowledge in L1 and L2 third and fifth graders. Applied Linguistics, 19, 452-470.
Virta, E., Sam, D., & Westin, C. (2004). Adolescents with Turkish background in Norway and Sweden: A comparative study of their psychological adaptation. Scandinavian Journal of Psychology, 45, 15-25.
Weizman, Z. O., & Snow, C. E. (2001). Lexical input as related to children's vocabulary acquisition: Effects of sophisticated exposure and support for meaning. Developmental Psychology, 37, 265-279.
Vibeke Grover Aukrust
Institute for Educational Research
University of Oslo
(1) The transcripts used the following CHAT conventions: < > = overlap; # = noticeable pause (other than following utterance terminators), +/. = interrupted utterance, + ... = trailing off, %com = contextual information, xxx = inaudible utterance(s), yyy = utterance(s) offered in a non-Scandinavian language.
(2) The Norwegian version of the Word Definition subscale of Reynell Developmental Language Scales was taken from Hagtvet & Lillestolen (1985); the Turkish version of the same subscale was translated from Norwegian to Turkish by Kamil Ozerk. The Norwegian version of Peabody Picture Vocabulary Test was translated from English to Norwegian by Astri Heen Wold. The Turkish version was translated from English and Norwegian by Kamil and Meral Ozerk.
(3) Following Hagtvet and Lillestolen (1985), score 2 was used for word definitions, focusing on an important purpose or function (TOWEL: to dry someone with), giving a cause or reason (WASH: because you are dirty), offering a synonym (SLEEP: rest), offering a superordinate concept (DRESS: piece of clothing), a negation (LAST: not first) or offering two or more relevant associations (SLEEP: lie in the bed in the night). Score 1 was used for word definitions focusing on a more distant purpose or function (TOWEL: for washing), offering one relevant association (COLD: need to get in), or an example relying on the test word (SLEEP: sleep in the bed). Score 0 was used when the child responded with gestures, repeated the test word (DRESS: is a dress), offered a private association (TOWEL: with teddy bears on).
(4) The list of Norwegian words, in descending frequency based on newspaper journals, can be found on www.tekstlab.uio.no/norsk/bokmaal/#tekstene
Table 1 Sample Information Data collection points Number of participants Mean age Next-to-last 18 (9 boys, 9 girls) 4 years; 10 months year of preschool (range 4;3-5;4) "4-year-olds" Last year of 26 (15 boys, 11 girls) 5;11 (range preschool (17 initially recruited 5;3-6;4) "5-year-olds" at age 4, plus 9 recruited at age 5) First grade 27 (15 boys, 12 girls) 6;10 (range 6;1- "6-year-olds" (18 initially recruited 7;2 months) at age 4, plus 9 recruited at age 5) Note. Two of the children recruited when they were in their next-to-last year of preschool attendance did not attend preschool the following year. For one of them, vocabulary tests were administered at home. Thus, for the group of "5-year-olds," vocabulary test data were available for 26 children, while group time recordings were available for 25 children. Table 2 Criteria for Identifying Explanatory Talk Explanatory Explanatory Example types subtypes Premises and/or Teacher: "Nei jeg er ikke tyrkisk, men consequences jeg laerte fra mammaen til Fatima" ("No, I am not Turkish, but I learned this from Fatima's mum") Principle- Rules and Teacher: "Men det vil jeg at dere based principles skal, for det er hyggelig nar noen bar explanations bursdag, da er det de som far lov til a bestemme det" ("But I want you to do that, because it is nice when someone has his birthday that he is allowed to decide on that") Conceptual Child: "Masse drager jeg vil kjope." issues Teacher: "Sanne som flyr opp i lufta?" (Child: "I want to buy lots of kites." Teacher: "Ones that fly up in the air?") Frame-based Negotiating Target child asking his preschool explanations the ongoing teacher (who was a native speaker of interaction Swedish) to give him some "glass" (the word "glass" means "glass" in Norwegian and "ice cream" in Swedish and was used by the child in a Norwegian utterance context, but with a smile). The preschool teacher offered a framing explanation of the interaction by asking if the child was practicing his Swedish. Table 3 Means and Standard Deviations for the Vocabulary Measures Used in the Study 4 years (n=18) 5 years (n=26) 6 years (n=27) M SD M SD M SD PPVT-NO 29.9 10.0 43.5 7.7 50.4 7.0 WD-NO 2.9 3.0 7.7 4.0 10.6 3.2 PPVT-TU 40.9 6.7 53.3 6.1 57.6 4.0 WD-TU 7.7 3.9 10.4 2.3 12.0 1.8 Note. For 4-year-olds, n = 18, with the exception of the Turkish word definition task (WD-TU), for which n = 14. Table 4 Intercorrelations Between PPVT-NO and Word Definition Skills in Norwegian (WD-NO) Across Ages and Demographic Variables 1 2 3 4 1.PPVT-NO 4 -- 2.PPVT-NO 5 .76 *** -- 3.PPVT-NO 6 .70 *** .54 ** -- 4.WD-NO 4 .58 * .68 ** .67 ** -- 5.WD-NO 5 .70 *** .67 *** .50 ** .54 * 6.WD-NO 6 .69 ** .45 * .71 *** .55 * 7. Maternal .58 * ns .48 * .78 *** education 8. Age of pre- ns ns ns ns school entrance 5 6 7 8 1.PPVT-NO 4 2.PPVT-NO 5 3.PPVT-NO 6 4.WD-NO 4 5.WD-NO 5 -- 6.WD-NO 6 69 *** -- 7. Maternal .50 ** .58 *** -- education 8. Age of pre- ns ns -.50 ** -- school entrance Note. For 4-year-olds, n = 18; 5-year-olds, n = 26; 6-year-olds, n = 27. For correlations between demographic variables, n = 27; * p < .05, ** p < .01, *** p < .001 Table 5 Teacher Talk Density of Word Tokens, of Word Types, and of Word Types Within Discourse Complexity 4 years (n=18) 5 years (n=25) M SD M SD Word tokens 69.3 18.5 65.5 26.9 per minute Word types 8.2 2.2 7.8 2.9 per minute Word types 2.8 2.5 3.1 2.5 within discourse complexity per minute 6 years (n=26) M SD Word tokens 72.7 22.3 per minute Word types 9.2 2.6 per minute Word types 3.9 2.4 within discourse complexity per minute Table 6 Predicting Word Definition Skills in Norwegian in 1st Grade From Teacher Talk Lexical Input at 4 Years of Age: Summarizing the Results of Hierarchical Regression Analysis Controlling for Maternal Education, Age of Preschool Entrance, and Word Definition Skills at Age 4. N = 18 R Square (step 1) (step 4) Step Change R square [beta] [beta] 1. Maternal education .37 ** .37 .60 2. Age of preschool entrance .08 .44 3. Previous skills .01 .45 4. Teacher word tokens .15 * .60 .46 4. Teacher word types .15 * .60 .47 4. Teacher word types .18 * .63 .49 within discourse complexity * p <.05, ** p < .01 Table 7 Predicting PPVT-NO Scores in 1st Grade From Teacher Talk Lexical Input at Age 5. Summarizing the Results of Hierarchical Regression Analysis While Controlling for Maternal Education, Age of Preschool Entrance, and PPVT-NO Scores at Age 5. N = 25 R Square (step 1) (step 5) Step Change R square [beta] [beta] 1. Maternal education .23 * .23 .48 2. Age of preschool entrance .03 .26 3. Previous skills .20 * .46 4. Teacher word tokens .18 ** .64 .46 4. Teacher word types .19 ** .65 .48 4. Teacher word types .19 ** .65 .47 within discourse complexity * p <.05, ** p < .01
|Printer friendly Cite/link Email Feedback|
|Author:||Aukrust, Vibeke Grover|
|Publication:||Journal of Research in Childhood Education|
|Date:||Sep 22, 2007|
|Previous Article:||Increasing the literacy behaviors of preschool children through environmental modification and teacher mediation.|
|Next Article:||Special education referral, evaluation, and placement practices for preschool English language learners.|