Spanish Nonword Repetition 1 Running Head: Spanish Nonword Repetition Spanish Nonword Repetition: Stimuli Development and Preliminary Results Kerry Danahy Ebert University of Minnesota, Minneapolis Jocelyne Kalanek San Diego Unified School District, California Kelly Nett Cordero Kathryn Kohnert University of Minnesota, Minneapolis Abstract Purpose: The current study presents preliminary nonword repetition data from Spanish-speaking preschool children using a new set of stimuli. Method: Twenty nonwords were constructed to be phonotactically possible in Spanish and to conform to published guidelines for nonword repetition stimuli. Fourteen Spanish-speaking typically-developing preschool children repeated the nonwords. Results: Both age and word length affected repetition accuracy and there was an age by length interaction. Younger children were less accurate overall and showed steeper decreases in accuracy as length increased. Conclusions: The results provide promising evidence that the stimuli may be developed into a Spanish nonword repetition task for both research and clinical purposes. Spanish Nonword Repetition 3 During the past decade, several investigations have examined the viability of nonword repetition (NWR) tasks for the assessment of primary or “specific” language impairment (LI)1 in English-speaking children. These investigations have employed carefully constructed sets of nonword stimuli designed to assess the integrity of a child’s language-learning system while minimizing the influence of differences in language experience. NWR tasks have been applied to a variety of populations and have been successfully completed by children as young as 3 years old (Gathercole, 1995). The skills measured by NWR, as well as their relationship to language learning, have been debated. Many researchers focus on the construct of phonological working memory as the key component of NWR performance (e.g., Gathercole, Willis, Baddeley, & Emslie, 1994); it is suggested that phonological short-term memory supports early vocabulary growth, and thus children with poor phonological working memory and resulting poor NWR performance are slower to learn new words (Gathercole, Willis, Emslie, & Baddeley, 1991). However, other skills may contribute to NWR performance, including speech perception (Frisch, Large, & Pisoni, 2000), articulatory ability (e.g., Archibald & Gathercole, 1996), and lexical long-term memory (e.g., Dollaghan, Biber, & Campbell, 1995). Investigations into the skills that affect NWR performance have examined English nonwords exclusively, creating an obvious potential for English-oriented bias in the results. Thus, while NWR may lend insight into the language learning process by highlighting specific skills that support language learning, it is important to gather data from a diverse set of language learners to ensure that generalizations made about English hold true for other children. 1We use the term primary language impairment and acronym LI rather than the traditional term “specific language impairment” and acronym SLI due to current literature indicating that these children have deficits outside the linguistic domain. Spanish Nonword Repetition 4 Among monolingual English-speaking children, a strong body of evidence indicates that children with LI demonstrate deficits on NWR tasks. Performance on NWR tasks has been shown to consistently differentiate between typically developing and language-impaired groups, although the magnitude of the effect has varied among studies (see Graf Estes, Evans, & Else- Quest, 2007, for a review). In addition, the meta-analysis of NWR studies conducted by Graf Estes et al. (2007) indicated that the size of the difference between children with LI and their typically developing peers does not vary significantly between the ages of 4 and 12 years of age. Thus, the available evidence suggests that performance on well-constructed NWR tasks may be a useful complement to other language measures in clinical assessments across a broad age range. NWR tasks have also been promoted as potentially less-biased assessment tools. The task is designed to emphasize processing capacity over language experience, particularly when stimuli are constructed to be phonotactically possible for a particular language, but to bear minimal resemblance to actual lexical items. Campbell, Dollaghan, Needleman, and Janosky (1997) found that typically-developing children from minority racial groups performed as well on an NWR task as their typically-developing majority (Caucasian) peers. The same group scored lower than the majority group on a more traditional knowledge-based standardized language measure, suggesting that the NWR task was less-biased for children from racial minority groups. NWR has also been shown to be less-biased for members of racial minority groups when the participants include both children with typical development and children with LI (Ellis Weismer et al., 2000). However, the members of racial minority groups included in previous studies have been monolingual English speakers. Children who speak languages other than English are also central to the need for less-biased assessment tasks, and English-based NWR tasks may not be less- Spanish Nonword Repetition 5 biased for these groups. In other words, although English-based NWR tasks are designed to minimize the effects of English-language experience, they may not successfully do so for children learning English as a second language, due to their decreased experience with the English phonological system or to the potential for within-speaker cross-language transfer effects. Thorn and Gathercole (1999) investigated NWR performance in typically-developing simultaneous French-English bilingual, sequential French-English bilingual, and monolingual English-speaking children between 4 and 8 years of age. Results indicated that experience in the test language impacted performance: the monolingual group scored below both bilingual groups in NWR performance with French-based nonwords. Kohnert, Windsor, and Yim (2006) compared typically developing bilingual children who learned Spanish as a first language and English as a second language to two groups of monolingual English speaking children (one typically-developing and one with LI) on an English-based NWR task. Results indicated that the bilingual children performed significantly better than the monolingual children with LI, but also significantly worse than the typically-developing monolingual children. Monolingual performance standards in either or both of a bilingual child’s languages are insufficient for assessment, due to fundamental differences in language experience between single and dual- language learners (e.g., Pearson, Fernandez, & Oller, 1993; see Kohnert, 2008 for review and discussion). Nonetheless, the results obtained by Thorn and Gathercole (1999) and Kohnert et al. (2006) indicate that learner experience in the language used to design nonword stimuli affects repetition accuracy on NWR tasks and therefore English-based nonwords will not be appropriate for children with limited or no exposure to English. Establishing and testing nonword stimuli in a variety of languages is also important for testing the hypothesized relationship between NWR and language development; linguistic skill Spanish Nonword Repetition 6 and NWR performance may not be as tightly linked in languages other than English. For example, among monolingual Cantonese-speakers, Stokes and colleagues found no performance difference on an NWR task between children with LI and their typically developing age peers (Stokes, Wong, Fletcher, & Leonard, 2006). For a given set of nonword stimuli, patterns of performance in both typically-developing and LI populations need to be established in order to determine the stimuli’s ability to separate the populations. Language-specific constraints on the form of nonwords or the manner in which they are processed may also limit the usefulness of NWR tasks in languages other than English. These patterns cannot be discovered until appropriate stimuli are constructed and tested. For example, recent publication of a set of Swedish nonword stimuli, along with data on preschool children’s performance repeating them, has established a potential Swedish NWR task (Radeborg, Barthelom, Sjöberg, & Sahlén, 2006). Preliminary data suggest that NWR tasks using appropriate Spanish-based stimuli may be able to separate impaired from typical Spanish-speaking populations. Calderón (2003) developed a set of 22 nonwords adhering to the phonotactic constraints of Spanish. Stimuli were two, three, or four syllables in length. The NWR task was administered to 16 typically-developing children and 16 children with LI; the mean age of participants was 5;0. Participants were Spanish-English bilinguals living in Southern California who had greater exposure and proficiency (according to several measures) in Spanish than in English. Results showed that the children with LI were significantly less accurate at repeating the Spanish nonwords than their typically developing peers at all three word lengths. In addition, all children were less accurate at repeating the longer words, but there was no interaction between group and length, which differed from some NWR results in English (e.g., Dollaghan & Campbell, 1998; Kohnert et al., 2006). It is possible that including shorter non-words would have established preliminary common ground between Spanish Nonword Repetition 7 children with and without LI; longer non-words may have provided a critical cut-off point at which children with LI were disproportionately affected. Girbau and Schwartz (2007) also constructed an NWR task following the phonotactic constraints of Spanish and administered it to both typically-developing children and children with LI. Participants in the study were monolingual Spanish speakers living in Spain who ranged in age from 8;3 to 10;11. Twenty nonwords were used; word length ranged from one to five syllables. Nonwords included clusters and word-final consonants. Children with LI scored significantly below the children without LI on the three-, four-, and five-syllable words from this set. Although selected examples from this set of stimuli were included in the publication, the full set of nonword stimuli was not included. Despite these encouraging preliminary data, there does not exist a set of Spanish nonword stimuli in the public domain. Publication of a full set of nonword stimuli is an essential step in developing a tool that can be tested with children in diverse areas and in producing independent corroborating evidence of the tool’s effectiveness as well as its limitations. In English, the Children’s Test of Nonword Repetition (CNRep; Gathercole et al., 1994) and the Nonword Repetition Test (NRT; Dollaghan & Campbell, 1998) have been validated in multiple studies and can be compared (e.g. Graf Estes et al., 2007). Brea-Spahn and Frisch (2006) developed and published a large set of nonwords derived randomly from a Spanish lexicon; although these words were rated for wordlikeness by adult Spanish-English bilinguals, they have not yet been used for an NWR task. Finally, Carreiras and Perea (2004) published a set of Spanish nonwords; however, these stimuli were designed to be read by adults and are uniformly disyllabic. In summary, a set of empirically validated Spanish non-words is needed for both clinical and research purposes. Spanish Nonword Repetition 8 The purpose of this study was to obtain preliminary data for a Spanish NWR task with a group of typically developing preschool-aged children, using a set of nonwords that adhere both to Spanish phonotactic constraints and to general guidelines for less-biased nonword stimuli that have been established in English. Ideally, this preliminary testing will lead to further investigations of the NWR task and to a further refinement of the nonword set into one appropriate for research and clinical use with bilingual and monolingual Spanish-speaking children. Method Stimuli A set of twenty nonword stimuli were constructed to adhere to the phonotactic constraints of Spanish. These stimuli, along with task instructions, are listed in the Appendix. Spanish spoken in the United States contains 17 consonants and 5 vowels (Guirao & García Jurado, 1990); the nonword stimuli included all five of the vowels (/a, e, i, o, u/) and 11 of the consonants (/d, g, p, b, t∫, t, n, ɳ, k, j, f/). Construction of syllables and assignment of primary stress in the nonwords followed typical patterns for Spanish, making the stimuli more “wordlike” and therefore easier to remember and repeat (e.g., Gathercole et al., 1991). Syllables were constructed following the consonant-vowel (CV) pattern that is most frequent in Spanish (Guirao & García Jurado, 1990) as opposed to the CVC pattern of English. Stress also followed the canonical pattern for Spanish, with the penultimate syllable receiving primary stress (Núñez Cedeño & Morales-Front, 1999). Attention was also given to the overall frequency of occurrence of phonemes in Spanish. For example, aside from /s/ (at 9.4%), /n/ is the most frequently occurring Spanish consonant (7.1% of all consonants) (Guirao & García Jurado, 1990). Spanish Nonword Repetition 9 Therefore, /n/ was incorporated frequently in our list of nonwords. In addition, /ɳ/ only represents 0.4% of all consonants produced in Spanish (Guirao & García Jurado, 1990) and is extremely rare in word initial position. As such, this phoneme occurs only once and in word - medial position in our list of nonwords. The shortest nonword stimuli were one syllable in length, consistent both with previous research in English (CNRep; Gathercole et al., 1994) and in Spanish (Girbau & Schwartz, 2007). Although single syllable words are considerably less common in Spanish than English, it was considered important to create a broad range of stimuli in the initial phase of development as opposed to making a priori assumptions regarding the relative ease of various word lengths. In addition, because Spanish words are longer, on average, than English words, stimuli of up to five syllables were included in contrast to the maximum of four syllables often used in English NWR stimuli (e.g., Dollaghan & Campbell, 1998). A total of four nonwords at each syllable length (i.e., 1, 2, 3, 4, and 5 syllables) were constructed. The total number of phonemes in the twenty nonword stimuli was 120. It is anticipated that empirical evidence will be used to further refine and develop this initial set of stimuli. The criteria outlined by Dollaghan and Campbell (1998) were used as a general guideline for nonword stimuli development. Of course, given fundamental cross-linguistic differences between Spanish and English, the robust English NWR criteria could not be adopted without modification. Rather we judiciously altered some of the suggested criteria for English NWR to be consistent with the constraints of Spanish. The first criterion is that none of the individual syllables should correspond to actual words, in order to minimize the influence of lexical experience on task performance. This recommendation was followed in large part. However, due to the large number of one syllable words that result from basic consonant and vowel Spanish Nonword Repetition 10 combinations in Spanish it was not possible to fully meet this criterion. Overall, there were 12 syllables that corresponded to true words in Spanish: no, tu/ti (you), de (of), ni (neither, nor), fe (faith), te (tea), ja (already - ya), di (1st person past tense ‘said’), ba (3rd person singular ‘goes’- va), bi (3rd person past tense ‘saw’- vi), and be (3rd person singular ‘sees’ - ve). Second, later-developing consonants were excluded in order to minimize interference from articulatory limitations on task performance. Across most Spanish dialects, /r/ and /s/ tend to be the last phonemes to develop in typical populations (Bedore, 1999). For the same reason, none of the nonwords contained abutting consonants or consonant clusters. Next, nonwords should contain only tense vowels. This third criterion was easy to meet because all vowels in Spanish are tense and are not diphthongized or reduced to schwa (Barrutia, 1994). The fourth constraint, that a consonant occupy a syllable position in which it typically occurs less than 25% of the time, cannot be applied to Spanish as consonants are nearly always syllable-initial. Finally, for English it was suggested that each phoneme should occur only once in a nonword stimulus (Dollaghan & Campbell, 1998). Spanish has a smaller phonemic inventory as well as longer words and therefore this constraint was not fully applied. Participants Participants were 14 preschoolers between the ages of 3;5 (years; months) and 5;6; group mean age was 4;3. The group was composed of 8 girls and 6 boys. All the children spoke Spanish as a primary language and also had some exposure to English, with varying degrees of skill. Children attended a local bilingual early childhood education program and were exposed to both Spanish and English on a daily basis in their bilingual classrooms. Speech and hearing screenings were also completed for all participants; all children passed the hearing screening2 and demonstrated speech production skills within developmental expectations. They produced a 2 Two children were absent from the preschool on the day of hearing screenings. Spanish Nonword Repetition 11 variety of developmentally-appropriate errors on the speech screening, including cluster reduction and distortions or substitutions for the /s/ and /r/ sounds. Participants also completed the Preschool Language Scale-4 (PLS-4) in both Spanish and English (Zimmerman, Steiner, & Pond, 2002; Zimmerman, Steiner, & Pond, 2002). Both the Auditory Comprehension and Expressive Communication subtests were completed. In Spanish, all participants received a Total Language standard score within one standard deviation of the mean; the mean score was 104. In English, Total Language standard scores ranged from 51 to 96, with a mean score of 73. A rough measure of developmental effects within the group of children was created by dividing the group cross-sectionally and comparing performance between the two groups. “Older” and “younger” participants were separated using overall mean age to divide the total sample into two groups. The older group consisted of seven children ranging in age from 4;3 to 5;6, with a mean age of 4;10. Seven children ranging in age from 3;6 to 4;0, with a mean age of 3;9, composed the younger group. This cross-sectional comparison based on age allowed us to gain some insight into potential developmental effects on task performance. The presence or absence of age-related effects here may serve as a useful guide for future longitudinal studies. Procedures, Scoring, and Reliability A native Spanish speaker recorded task instructions and the 20 nonwords onto an audiotape to be used with each child. Task instructions were recorded in Spanish only. Each child was tested individually in a quiet room at the school. The children listened to the nonwords via headphones and were instructed to repeat the stimulus immediately following presentation. Responses were audio recorded for later broad phonetic transcription and scoring. Spanish Nonword Repetition 12 A native Spanish speaker performed the transcription and scoring tasks. Each target phoneme was scored as correct or incorrect. Following Dollaghan and Campbell (1998), phoneme distortions and additions were not counted as errors, whereas phoneme omissions and substitutions were considered incorrect. It should also be noted that although the voiced stop consonants /b, d, g/ were written as the targets of production for the nonwords, the phonological process of spirantization was accounted for in scoring. These voiced stop phonemes are produced with a short lag voice onset time (VOT) in English and prevoicing (initiation of voicing before release of stop articulation posture) in Spanish. However, in intervocalic position in Spanish, they are typically produced as fricatives, or a category often termed spirants. Spirantization leads to the production of /β, ð, γ/ (Branstine, 1991). Therefore, productions of both the voiced stops and their spirant counterparts were counted as correct for all intervocalic (not word initial) productions. Also following Dollaghan and Campbell (1998), each participant received a percentage phonemes correct (PPC) score at each of the five syllable levels as well as for the total number of phonemes (i.e., 120). Data for seven of the fourteen participants was independently transcribed and scored by an additional coder; inter-coder reliability was 0.982 (p < 0.01). Results Results were explored in three ways. First, a 2 (group: younger vs. older) x 5 (length based on number of syllables) mixed ANOVA was used to examine the effects of age and length of stimulus items on repetition accuracy. Significant effects were found both for age group, F(1, 12) = 4.913, η2 = 0.29, p < 0.05, and for length, F(4, 48) = 17.312, η2 = 0.76, p < 0.001. Length effects are illustrated in Figure 1; accuracy levels remained relatively constant across one-, two-, and three-syllable nonwords (at 89.3%, 91.5%, and 89.5%, respectively) but declined between Spanish Nonword Repetition 13 both three and four syllables (from 89.5% to 76.7%) and between four and five syllables (from 76.7% to 59.8%). Simple contrasts indicated that both decreases were significant, F(1, 12) = 22.456, η2 = 0.652, p < 0.001, and F(1, 12) = 30.729, η2 = 0.719, p < 0.001. The ANOVA also indicated a significant age group by length interaction, F(4, 48) = 2.62, η2 = 0.28, p < 0.05. Post hoc comparisons revealed a significant change in the relationship between the Older and Younger groups at the four-syllable level in comparison to the three- syllable level, F(1, 12) = 11.474, η2 = 0.489, p < 0.01. The younger children demonstrated a drop in accuracy when the words increased from three to four syllables, changing from 88.1% to 63.8% accuracy, whereas the older children maintained accuracy near 90% at both word lengths. This effect is apparent in Figure 2, which shows the repetition accuracy for each age group. The second analysis focused on error patterns. Errors were divided into three general types: consonant, vowel, and syllable errors. Syllable errors occurred when a child made an error on both the consonant and the vowel within a syllable, and thus accounted for two incorrect phonemes. When the child erred on the consonant but not the vowel in a syllable, it was counted as a consonant error, and vice versa for the vowel errors. This analysis indicated that 73.7% of incorrect phonemes occurred as part of a syllable error. Consonant errors accounted for 21.4% of phoneme errors, and vowel errors accounted for the remaining 4.9%. Despite the low percentage of vowel errors, more than half of the participants (8 of 14) made at least one vowel error. The third analysis explored the relationship between language scores, as measured by the PLS-4, and NWR performance. Pearson correlations with Bonferroni correction were calculated for the English PLS-4 Expressive Communication subtest raw score, the Spanish PLS-4 Expressive Communication subtest raw score, and the total PPC score from the NWR task. Age Spanish Nonword Repetition 14 was partialed out of these correlations due to the expected developmental increases in all three measures. The partial correlation between Spanish and English expressive PLS-4 scores was significant, r = 0.70, p < 0.05. The partial correlation between Spanish PLS-4 and NWR score was not significant, r = 0.29, p = 0.39, and neither was the partial correlation between English PLS-4 and NWR score, r = 0.41, p = 0.21. Discussion The purpose of this study was to develop a set of nonword stimuli for a Spanish NWR task and to examine preliminary data from a small group of Spanish-speaking preschoolers who completed the task. Our results indicate that the proposed nonwords and the NWR task itself were broadly appropriate for this group of young children; all children completed the task and the breadth of accuracy scores indicates that the nonwords were neither too difficult nor too easy. Although accuracy decreased as word length increased, at each complexity level there were children who successfully repeated all phonemes. In addition, the lowest PPC score received at any single complexity level was 28%, achieved by one child at the five syllable level, and the lowest total PPC score was 43%. Only 1 of 14 children was able to repeat all phonemes correctly. These results also provide preliminary evidence that our set of nonwords may be sensitive to the effects of age, because the older group of children performed significantly better than the younger group. This developmental effect is both consistent with previous testing with preschool-aged children (e.g., Gathercole, 1995) and expected of a task that taps language processing skills (e.g., Kohnert & Windsor, 2004). The generalization of age effects in this study is somewhat limited by the small number of participants as well as the cross-sectional as opposed to longitudinal data. Spanish Nonword Repetition 15 In addition, the analyses of length effects indicated that on average children were able to repeat words of one, two, and three syllables with high accuracy. However, accuracy was not perfect, even for these short words; all children made at least one error on the one-, two-, or three-syllable words, except the one child who received a PPC score of 100% for the entire task. Mean PPC scores for the younger group of children began to decline at the four-syllable level, dropping from 88.1% accuracy to 63.9%, and continued to decline at the five-syllable level (to 48.6%). The older group of children maintained high accuracy until the words reached five syllables, dropping from 89.3% accuracy on four-syllable words to 71.1% on five-syllable words. This pattern suggests that the shorter nonwords (perhaps those of one and two syllables) were relatively easy even for young children, perhaps because the Spanish CV syllable pattern reduces the number of phonemes to be recalled and repeated in these short words. In some cases it may be desirable to eliminate monosyllabic nonwords from Spanish NWR tasks in order to shorten the task for inclusion as part of a broader screening or assessment battery. Conversely, the inclusion of five syllable words appears useful to prevent the task from being too easy for older or more linguistically adept children. However, it is important to note that this set of Spanish nonwords has not yet been tested with children with LI. Children with LI would be expected to perform more poorly than their typically developing peers on NWR and therefore the shorter stimulus items might prove more informative. At the same time, longer nonwords might be subject to ceiling effects. Data from the Girbau and Schwartz (2007) study, involving older children and more complex words, supports this hypothesis; children with LI began to decline in accuracy on two-syllable words, whereas their typically-developing peers maintained high accuracy levels until the words reached four syllables in length. Similarly, the preschool children with LI in the Calderón (2003) Spanish Nonword Repetition 16 investigation demonstrated less than 80% accuracy on two-syllable nonwords. Overall, the variation in average performance on nonwords across ages and between language-intact and language-impaired groups reinforces the idea that nonword stimuli should be varied enough to capture a wide range of NWR ability levels. Prematurely restricting the length of nonword stimuli, without sufficient preliminary data, may limit the value of the nonword stimuli. The analysis of error types also contributes important information about the performance of typically-developing Spanish-speaking children on NWR. The majority of errors involved the entire syllable unit, rather than only the consonant or the vowel. In part, this may be attributed to expected developmental effects: younger children (such as the participants in this study) would be expected to have more difficulty remembering longer words and perhaps to omit segments. Given the typical syllable structure of Spanish, omitting only the consonant or only the vowel would create a phonotactically unlikely sequence (either CC or VV). The entire syllable would be a more likely deletion. It is also important to note that, while the proportion of vowel errors was low, most children made at least one vowel error. If the vowels involved in syllable errors are also counted, there were many vowel errors (i.e., 4.9% plus one-half of 73.7%, for a total of 41.8%). In contrast, Girbau and Schwartz (2007) reported that typically-developing children were “consistently perfect” (p. 68) at repeating vowels. The cross-study difference may be the result of age, as the children in the Girbau and Schwartz (2007) study were an average of 5 years older than the participants in the current study. However, it would be premature to conclude that only children with LI make vowel errors on Spanish NWR; the current results clearly show that young typically-developing children make vowel errors on Spanish NWR. Further research is needed to Spanish Nonword Repetition 17 determine the differential contributions of age, Spanish language ability and, perhaps, previous exposure to English. Lastly, although correlations between the Spanish NWR task score and traditional measures of skill in either Spanish or English in these developing bilinguals were not significant, the correlation coefficients obtained were relatively high (r = 0.29 and r = 0.41) and statistical significance might be obtained with a larger sample size (n=14). A small sample size is one of the major limitations of the current study. Given the promising nature of the preliminary data obtained in this study, it will be important to complete further testing with larger groups of typically-developing children, and to extend data collection to children of different ages, in different areas, and with different language abilities. In summary, the proposed set of Spanish nonword stimuli generally conforms to accepted standards for NWR tasks, including following language-specific phonotactic constraints; avoiding later-developing consonants and lax vowels; and minimizing the resemblance between real words and nonwords. The current data establishes that this set of words can be used with preschool children and that the children’s performance is affected by both word length and development in the expected directions. Further testing using these stimuli both with larger groups of typically developing Spanish-speaking children and with groups of language-impaired Spanish-speaking children is necessary to explore the relationship between performance on this NWR task and other measures of language skill and language development. The preliminary data shown here suggest that these stimuli are promising for use as one measure of language processing ability in a communication screening or assessment battery. Spanish Nonword Repetition 18 References Archibald, L.M.D., & Gathercole, S.E. (2006). Nonword repetition: A comparison of tests. Journal of Speech, Language, and Hearing Research, 49, 970-983. Barrutia, R. (1994). Fonética y Fonología Española: Teoria y Practica 2nd Edition. New York: John Wiley & Sons, Inc. Bedore, L. (1999). The acquisition of Spanish. In O.L. Taylor. & L.B. Leonard (Eds.), Language acquisition across North America: Cross-cultural and cross-linguistic perspectives (pp. 157-207). San Diego, CA: Singular. Branstine, Z. (1991). Stop/Spirant alternations in Spanish: On the representation of contrast. Studies in the Linguistic Sciences (21), 1-22. Brea-Spahn, M.R., & Frisch, S.A. (2006). Wordlike or not? Segment probability, bilingualism, and word similarity ratings. Poster presentation to the American Speech-Language- Hearing Association Convention, Miami, FL. Calderón, J. (2003). Working memory in Spanish-English bilinguals with language impairment. Unpublished doctoral dissertation, University of California, San Diego, and San Diego State University. Campbell, T., Dollaghan, C., Needleman, H., & Janosky, J. (1997). Reducing bias in language assessment: Processing dependent measures. Journal of Speech, Language, and Hearing Research, 40, 519-525. Carreiras, M., & Perea, M. (2004). Naming pseudowords in Spanish: Effects of syllable frequency. Brain and Language, 90, 393-400. Spanish Nonword Repetition 19 Conti-Ramsden, G. (2003). Processing and linguistic markers in young children with specific language impairment (SLI). Journal of Speech, Language, and Hearing Research, 46, 1029-1037. Dollaghan, C.A., Biber, M.E., & Campbell, T.F. (1995). Lexical influences on nonword repetition. Applied Psycholinguistics, 16, 211-222. Dollaghan, C., & Campbell, T. (1998). Nonword repetition and child language impairment. Journal of Speech, Language, and Hearing Research, 41, 1136-1146. Ellis Weismer, S., Tomblin, J.B., Zhang, X., Buchwalter, P., Chynoweth, J.G., & Jones, M. (2000). Nonword repetition performance in school-age children with and without language impairment. Journal of Speech, Language, and Hearing Research, 43, 865-878. Gathercole, S.E. (1995). The assessment of phonological memory skills in preschool children. British Journal of Educational Psychology, 65, 155-164. Gathercole, S. E., Willis, C. S., Baddeley, A. D., & Emslie, H. (1994). The Children’s Test of Nonword Repetition: A test of phonological working memory. Memory, 2, 103–127. Gathercole, S.E., Willis, C., Emslie, H., & Baddeley, A.D. (1991). The influences of number of syllables and wordlikeness on children’s repetition of nonwords. Applied Psycholinguistics, 12, 349-367. Girbau, D., & Schwartz, R.G. (2007). Nonword repetition in Spanish-speaking children with specific language impairment (SLI). International Journal of Language and Communication Disorders, 42, 59-75. Graf Estes, K., Evans, J.L., & Else-Quest, N.M. (2007). Differences in the nonword repetition performance of children with and without specific language impairment: A meta- analysis. Journal of Speech, Language, and Hearing Research, 50, 177-195. Spanish Nonword Repetition 20 Guirao, M., & García Jurado, M.A. (1990). Frequency of occurrence of phonemes in American Spanish. Revue Quebecoise de Linguistique, 19, 135-150. Frisch, S.A., Large, N.R., & Pisoni, D.B. (2000). Perception of wordlikeness: Effects of segment probability and length on the processing of nonwords. Journal of Memory and Language, 42, 481-496. Kohnert, K. (2008). Language disorders in bilingual children and adults. San Diego: Plural. Kohnert, K., & Windsor, J. (2004). In search of common ground- Part II: Non-linguistic performance by linguistically diverse learners. Journal of Speech, Language, and Hearing Research, 47, 891 – 903. Kohnert, K., Windsor, J., & Yim, D. (2006). Do language-based processing tasks separate children with primary language impairment from typical bilinguals? Learning Disabilities Research & Practice, 21, 19-29. Núñez Cedeño, R.A., & Morales-Front, A. (1999). Fonología generativa contemporánea de la lengua española. Washington, D.C.: Georgetown University Press. Pearson, B.Z., Fernandez, S.C., & Oller, D.K. (1993). Lexical development in bilingual infants and toddlers – Comparison to monolingual norms. Language Learning, 43, 93-120. Radeborg, K., Barthelom, E., Sjöberg, M, & Sahlén, B. (2006). A Swedish non-word repetition test for preschool children. Scandinavian Journal of Psychology, 47, 187-192. Stokes, S.F., Wong, A.M.Y., Fletcher, P., & Leonard, L.B. (2006). Nonword repetition and sentence repetition as clinical markers of specific language impairment: The case of Cantonese. Journal of Speech, Language, and Hearing Research, 49, 219-236. Spanish Nonword Repetition 21 Thorn, A.S.C., & Gathercole, S.E. (1999). Language-specific knowledge and short-term memory in bilingual and non-bilingual children. The Quarterly Journal of Experimental Psychology, 52A, 303-324. Zimmerman, I.L., Steiner, V.G., & Pond, R.E. (2002). Preschool Language Scale, Fourth Edition – English Edition. San Antonio, TX: Harcourt Assessment. Zimmerman, I.L., Steiner, V.G., & Pond, R.E. (2002). Preschool Language Scale, Fourth Edition – Spanish Edition. San Antonio, TX: Harcourt Assessment. Spanish Nonword Repetition 22 Appendix Nonword stimuli and task instructions 1 Syllable 2 Syllables 3 Syllables 4 Syllables 5 Syllables /du/ /'na Fe/ /te'Ba ka/ /t∫i no'te Ba/ /be te t∫o'Du pe/ /gi/ /'bo pa/ /t∫o'Bi Fa/ /no t∫i 'Be ni/ /do Bi nu't∫i Fa/ /pe/ /'tu ɳo/ /be't∫i po/ /di tu'ja Bu/ /ni t∫u to'na Be/ /bo/ /'de fu/ /nu 'ti fe/ /te Do't∫i Fo/ /te Be t∫i 'no Fe/ Note: Word-medial voiced stop consonants have been written as spirants in these stimuli; as noted in the text above, both stop (/b, d, g/) and spirantized (/B, D, F/) versions of these consonants were accepted as correct. Also note that voiceless stops (/p, t, k/) are not aspirated in Spanish. Original task instructions in Spanish: “Voy a decir unas palabras inventadas. No son palabras de verdad y suenan chistosas. Después de que yo diga una palabra, quiero que tu la repitas.” English translation of task instructions: “I am going to say some made-up words. They are not real words, and they sound funny. After I say a word, I want you to repeat it.” Figure Captions Figure 1. Nonword repetition accuracy for all subjects by word length in syllables. Figure 2. Nonword repetition accuracy by word length and age group. Repetition Accuracy by Word Length 0 20 40 60 80 100 1 2 3 4 5 Total Syllables Pe rc en t P ho ne m es C or re ct Repetition Accuracy by Word Length and Age Group 0 20 40 60 80 100 1 2 3 4 5 Tota l Syllables Pe rc en t P ho ne m es C or re ct Younger Older Nonword stimuli and task instructions