LSHSS Clinical Forum

Phonological Skills in Predominantly English-Speaking, Predominantly Spanish-Speaking, and Spanish– English Bilingual Children Brian A. Goldstein Leah Fabiano Patricia Swasey Washington Temple University, Philadelphia, PA

D

espite the recent increase in the number of cross-linguistic studies examining phonological skills in non-English speakers (e.g., Amayreh & Dyson, 1998, for Arabic), there is a paucity of information on the phonological skills of bilingual speakers, particularly Spanish–English bilingual speakers. Existing studies of Spanish–English bilingual children either have not examined their Spanish skills (Gildersleeve, Davis, & ABSTRACT: Purpose: There is a paucity of information detailing the phonological skills of Spanish–English bilingual children and comparing that information to information concerning the phonological skills of predominantly English-speaking (PE) and predominantly Spanish-speaking (PS) children. The purpose of this study was to examine the relationship between amount of output (i.e., percentage of time each language was spoken) in each language and phonological skills in Spanish-English bilingual children and PE and PS children. Method: Fifteen typically developing children, ranging in age from 5;0 (years;months) to 5;5 (mean = 5;2), participated in the study. The participants consisted of 5 PE speakers, 5 PS speakers, and 5 bilingual (Spanish–English) speakers. A single-word assessment was used to gather information on phonological skills (consonant accuracy, type and frequency of substitutions, frequency of occurrence of phonological patterns [e.g., cluster reduction], accuracy of syllable types [e.g., CV, CVC, CCV, etc.]), and type and rate of cross-linguistic effects. Results: The results indicated that there was no significant

LANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

Stubbe, 1996; Gildersleeve-Neumann & Davis, 1998) or have not included matched groups of predominantly monolingual English and/or Spanish speakers (e.g., Goldstein & Washington, 2001). Given that it is unknown whether the phonological skills of bilingual children are either similar to or different from the phonological skills of predominantly English-speaking (PE) and predominantly Spanish-speaking (PS) children, it is crucial to include correlation between amount of output in each language and phonological skills either in the Spanish skills of PS children and Spanish-English bilingual speakers or in the English skills of PE children and Spanish-English bilingual speakers. In addition, there was no significant difference in segmental accuracy, syllabic accuracy, or percentage of occurrence of phonological patterns between either the Spanish skills of PS children and Spanish-English bilingual speakers or the English skills of PE children and SpanishEnglish bilingual speakers. Finally, the children showed a limited number of cross-linguistic effects. Clinical Implications: Results from this study indicate no link between parent estimates of language output and phonological skill and demonstrate that Spanish–English bilingual children will have commensurate, although not identical, phonological skills as compared to age-matched PS and PE children.

KEY WORDS: phonological, bilingual, predominantly Spanish, predominantly English

SCHOOLS • Vol. 36 • 201–218 • July 2005 © American Speech-Language-Hearing Association 0161–1461/05/3603–0201

201

speakers from all three groups in order to begin to understand phonological skills in children with varying linguistic profiles (Watson, 1991).

PHONOLOGICAL SKILLS IN ENGLISH AND SPANISH MONOLINGUAL CHILDREN Compared with the number of studies examining English phonological development, there is a small number of studies examining phonological skills in monolingual Spanish-speaking children (see Goldstein, 1995, for a detailed review). Existing studies suggest that by approximately 5 years of age, typically developing (TD) Spanishspeaking children have mastered (i.e., accuracy greater than 90%) the majority of sounds in the language, with the exceptions of /g/, /f/, /s/, /≠/, /r/1 (trill), and /|/ (flap) (Acevedo, 1993; De la Fuente, 1985; Eblen, 1982; Gonzalez, 1981). At that age, TD monolingual Spanishspeaking children still exhibit moderate amounts (percentages of occurrence greater than 10%) of cluster reduction, final consonant deletion, unstressed syllable deletion, tap/ trill deviation, and fronting (Anderson & Smith, 1987; Goldstein & Iglesias, 1996; Gonzalez, 1981). These general findings from monolingual Spanish speakers are comparable to those for monolingual English speakers. By 6 years of age, TD monolingual English speakers also have mastered the majority of sounds in the language, with the exceptions of /s/, /z/, /T/, /D/, /Z/, and /®/1 (e.g., Smit, Hand, Freilinger, Bernthal, & Bird, 1990; see Vihman, 2004, for a detailed review). By that age, these children still may exhibit the phonological processes of stopping (particularly on interdentals) and gliding of liquids (Grunwell, 1997). This brief description of phonological skills in monolingual Spanish- and English-speaking children suggests that the attainment of phonological skills in TD monolingual Spanish- and English-speaking children is comparable. It is not identical, of course, given the phonotactic distinctions between the languages. For example, because word length in Spanish is relatively longer than that of English (Vihman, 1996), it is not surprising that unstressed syllable deletion is exhibited with a higher frequency of occurrence and for a longer period of time in Spanish speakers than in English speakers. It is still unknown, however, whether the phonological skills of predominantly monolingual speakers (English and Spanish) are commensurate with those of bilingual (Spanish–English) children.

PHONOLOGICAL SKILLS IN BILINGUAL CHILDREN There are comparatively few studies that have examined the phonological skills of bilingual preschool children, 1

The symbol /r/ is used for the Spanish trill; the symbol /®/ is used for the prevocalic, nonlateral liquid in English.

202

L ANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

much less Spanish–English bilingual children. Although these studies have helped to inform our understanding of phonological development in bilingual children, clinical applications of these studies are limited for three main reasons. First, it is primarily the English phonological skills of Spanish–English bilingual children that have been examined. Gildersleeve et al. (1996) examined the English phonological skills (they did not examine the Spanish phonological skills) of 29 TD 3-year-old bilingual (Spanish–English) children and compared them to the phonological skills of 14 TD 3-year-old monolingual English speakers and 6 TD 3-year-old monolingual Spanish speakers. They found that, in English, the bilingual children showed an overall lower intelligibility rating, made more consonant and vowel errors overall, distorted more sounds, and produced more uncommon error patterns (e.g., backing) than either the monolingual English or the monolingual Spanish speakers. In addition, the bilingual children had the highest or second highest percentage of occurrence on every phonological pattern (initial consonant deletion, cluster reduction, final consonant deletion, stopping, gliding, and final consonant devoicing). Gildersleeve-Neumann and Davis (1998) extended the study of Gildersleeve et al. (1996) by examining the English phonological skills of 27 TD 3-year-old Spanish– English bilingual children at the end of the school year and comparing them to the phonological skills of 14 TD 3-yearold monolingual English speakers and 6 TD 3-year-old monolingual Spanish speakers. Again, they did not examine the bilingual children’s Spanish phonological skills. Compared with the monolingual English and monolingual Spanish speakers, the bilingual children demonstrated more phonological processes and exhibited, on average, a higher percentage of occurrence on six of the ten phonological processes analyzed (i.e., cluster reduction, backing, final consonant deletion, final devoicing, initial voicing, and stopping). Gildersleeve-Neumann and Davis concluded that bilingual speakers demonstrated different developmental patterns than their monolingual peers and exhibited more errors initially than monolingual speakers. These differences, however, decreased over time. The second main limitation is that data collected from bilingual children (including Spanish–English bilingual children) typically are compared retrospectively to previously published data from monolingual speakers (e.g., Goldstein & Washington, 2001). Goldstein and Washington examined phonological patterns in twelve TD 4-year-old Spanish–English bilingual children. There were no significant differences between the two languages on percentage of consonants correct, percentage of consonants correct for voicing, place of articulation, manner of articulation, or percentage of occurrence for phonological processes. However, the children exhibited different patterns of production across the two languages (e.g., fricatives were more accurate in the children’s Spanish productions than in their English productions). On the basis of retrospective comparisons, the children also exhibited different patterns compared to those of monolingual children of either language. For example, the bilingual children showed higher consonant accuracy in English than did monolingual

SCHOOLS • Vol. 36 • 201–218 • July 2005

English-speaking children. Consonant accuracy for the bilingual children’s Spanish productions, however, was lower than that of monolingual Spanish-speaking children. In a study of 16 TD Cantonese- and English-speaking preschoolers, Dodd, So, and Li (1996) found that the children’s error patterns were different for their two languages. In a retrospective comparison to monolingual English- and Cantonese-speaking children, they found that the bilingual children’s patterns were atypical for monolingual speakers of either language; that is, the bilingual children exhibited error patterns that the monolingual children did not show (e.g., initial consonant deletion, backing, aspiration, and frication). Differences in results between monolingual and bilingual children may be the result of using retrospective studies for comparison. Retrospective comparisons may be biased because of differences in methodology and analyses. For example, Goldstein and Washington (2001) compared the bilingual children’s English phonological skills to four different studies that also included 4-year-old monolingual English speakers (Austin & Shriberg, 1997; Haelsig & Madison, 1986; Shriberg, Kwiatkowski, Best, Hengst, & Terselic-Weber, 1986; Smit et al., 1990). The number of children in those comparison studies ranged from 20 (Haelsig & Madison, 1986) to 186 (Smit et al., 1990). Given the marked differences between studies, it is necessary to compare data from bilingual children to those collected prospectively from predominantly monolingual children in order to decrease the possibility that the differences between groups are due to differences in method (Smit, 1986). The third limitation of previous studies is that they largely have been case studies rather than group studies (e.g., Johnson & Lancaster, 1998). Paradis (2001) argued for the “need for group studies of children acquiring the same language pair to tease apart the possible trends from individual variation” (p. 20). On the other hand, individual variation in (bilingual) language acquisition is well established (Keshavarz & Ingram, 2002). For example, the differences in the frequency of cross-linguistic effects for Schnitzer and Krasinski’s (1994, 1996) 2 children indicate the need for examining individual variation in the speech of bilingual children. Thus, there seems to be a need for the delineation of both group and individual trends in the study of bilingual children’s phonology. In addition, comparison data from predominantly English and Spanish speakers will allow differentiation of typical developmental patterns and characteristics of crosslinguistic effects (e.g., Vihman & McLaughlin, 1982). Traditionally, cross-linguistic effects (i.e., bidirectional transfer) have been viewed as a fundamental characteristic of bilingual language development. In terms of phonology, these effects may be the result of differences in the repertoires of each language; for example, a Spanish– English bilingual child substituting a trill (a sound not in the English repertoire) for the English prevocalic “r” (a sound not in the Spanish repertoire) (e.g., /b®IdZ/ “bridge” ➝ [brIdZ]) (Goldstein, 2001a). These effects also may be demonstrated because of differences in the distribution of sounds in the two ambient languages (Goldstein, 2001a).

For example, in Spanish, there are only five word-final consonants: /s/, /n/, /|/, /l/, and /d/. Spanish speakers may omit word-final sounds other than these five in their production of English. Finally, cross-linguistic effects may be exhibited because of different phonotactic constraints in each language (Perez, 1994). For example, because wordinitial clusters cannot begin with /s/ in Spanish, Spanish speakers may exhibit epenthesis in attempting word-initial /s/ clusters in English so that /stAmp/ “stomp” would be produced as [estAmp]. Most studies of bilingual phonological acquisition note some cross-linguistic effects. In the speech of a child from ages 8 to 20 months acquiring English and Farsi, Keshavarz and Ingram (2002) noted some examples of bidirectional cross-linguistic effects. That is, the child was using segments specific to Farsi in his English productions and segments specific to English in his Farsi productions. Fantini (1985) found that his preschool son aspirated voiceless stops in Spanish (/tetSo/ “roof” ➝ [tÓetSo]) even though stops in Spanish typically are described as unaspirated (Hammond, 2001). In studies of their two sons, Schnitzer and Krasinski (1994, 1996) found that their older son (followed from ages 1;6 [years;months] to 4;6) showed a low frequency of cross-linguistic effects, but their younger son (followed from ages 1;11 to 2;5) exhibited a high proportion of cross-linguistic effects. In a group of 12 Spanish–English bilingual children aged 5;0– 7;0, Goldstein, Fabiano, and Iglesias (2003) noted a low frequency of cross-linguistic effects. Less than .3% of all segments produced were the result of cross-linguistic influence. Thus, bilingual children do seem to show some evidence of cross-linguistic effects, though typically with a low frequency. Vihman and McLaughlin (1982) argued, however, that so-called cross-linguistic effects may be relatively common substitution patterns that are also exhibited by monolingual speakers. Including control groups of predominantly English and Spanish speakers would help to disentangle true cross-linguistic effects from more common substitution patterns. The purposes of this study were to examine the relationship between amount of output (i.e., parent report of language use) in each language and phonological skills in Spanish-English bilingual children and PE and PS children and to determine phonological skills between those groups. The focus of the study motivated three main research questions: 1. Is there a relationship between language output (i.e., parent report of language use) and phonological skills (consonant accuracy, percentage of occurrence for phonological patterns, and syllable accuracy) in the Spanish of bilingual children and the Spanish of PS children and in the English of bilingual children and the English of PE children? 2. Are there between-group differences in phonological skill between the Spanish of bilingual children and the Spanish of PS children and between the English of bilingual children and the English of PE children? 3. What are the rates and types of cross-linguistic effects exhibited by bilingual and PS and PE children?

Goldstein et al.: Phonological Skills in Bilingual and Predominantly Spanish and English Speakers

203

(Spanish–English) speakers (age range: 5;0–5;5, mean = 5;1) (Table 1). A one-way analysis of variance (ANOVA) did not reveal a significant difference for age between the three groups. The children’s language status (i.e., PS, PE, or bilingual) and language profile in each language (e.g., years of exposure, proficiency, input, and output) were determined by parent report (Gutiérrez-Clellen & Kreiter, 2003; Peña, Bedore, & Rappazzo, 2003) (Table 1). All children in the bilingual group were reported to use each language more than 20% of the time, a criterion based on previous work showing that output in both languages should be at least 20% in order for children to be considered competent speakers in both languages (Pearson, Fernandez, Lewedeg & Oller, 1997). Children assigned to the PS or PE group were reported to use the “predominant” language more than 20% of the time and the nonpredominant language less than 20% of the time. For the bilingual children, language output, on average, was 43% for English (SD = 15.3; range = 25%–63%) and 57% for Spanish (SD = 15.3; range = 37%–75%). For the PS children, language output, on average, was 7% for English (SD = 4.3; range = 0%–10%) and 93% for Spanish (SD = 4.3; range = 90%–100%). For the PE children, language output, on average, was 97% for English (SD = 6.8; range = 85%–100%) and 3% for Spanish (SD = 6.8; range = 0%–15%). Paired t tests (with a corrected alpha level of .01) were computed to discern differences in language output within each group. There

METHOD Participants Fifteen children ranging in age from 5;0 to 5;5 (mean = 5;2) participated in the study. All children were considered to be TD by both parent and teacher report. According to parent report, all children were of low socioeconomic status as indicated by their enrollment in a Head Start program or their receipt of free or reduced-price lunches. There was no parent concern about speech, language, or cognitive development. None of the children had been diagnosed with a communication disorder, and thus none had received previous intervention for a communication disorder. Moreover, according to teacher report, each child exhibited normal functioning in the classroom (e.g., able to follow classroom routines and directions). Finally, each child had passed a hearing screening that was administered by the school. All 15 children were of Latino descent, and although all 15 children were exposed to English and Spanish, according to parent report, their output (i.e., percentage of time each language was spoken) in each language varied. From that report, the children were subdivided into three groups based on amount of output (described below) in each language: 5 PE (age range: 5;0–5;5, mean = 5;3), 5 PS (age range: 5;1–5;5, mean = 5;3), and 5 bilingual

Table 1. Sociolinguistic variables for bilingual, predominantly Spanish, and predominantly English children.

Years of exposure Age Bilingual B1 B2 B3 B4 B5 Average SD

Spanish

5;0 5;5 5;0 5;0 5;0 5;1 .2

English

Proficiency Spanish

Input

Output

English

Spanish

English

Spanish

English

33.85 58.52 37.50 55.05 74.39 51.86 16.5

66.15 41.48 62.50 44.95 25.61 48.13 16.5

36.98 69.89 48.86 55.05 74.39 57.03 15.3

63.02 30.11 51.14 44.95 25.61 42.96 15.3

5 5 5 5 5 5.0 0

5 5 5 1 2 3.6 1.9

1 4 2 3 4 2.8 1.3

4 3 4 2 2 3 1

Predominantly Spanish S1 5;1 S2 5;5 S3 5;3 S4 5;3 S5 5;3 Average 5;3 SD .1

5 5 5 5 5 5.0 0

1 1 1 2 0 1 .7

3 4 4 4 4 3.8 .44

0 2 1 3 1 1.4 1.1

100.00 91.86 100.00 89.58 100.00 92.28 5.1

0 8.14 0 10.42 0 3.71 5.1

100.00 91.86 93.75 89.58 89.58 92.95 4.3

0 8.14 6.25 10.42 10.42 7 4.3

Predominantly English E1 5;0 E2 5;2 E3 5;5 E4 5;3 E5 5;3 Average 5;3 SD .2

5 5 5 0 0 3.0 2.7

5 5 5 5 5 5 0

1 0 2 0 0 0.6 .9

3 3 4 4 4 3.6 .5

13.44 0 17.76 0 18.75 9.99 9.3

86.56 100.00 82.24 100.00 81.25 90.01 9.3

0 0 15.13 0 0 3.02 6.8

100.00 100.00 84.87 100.00 100.00 96.97 6.8

Note. Age is in years;months.

204

L ANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

SCHOOLS • Vol. 36 • 201–218 • July 2005

was no significant difference in language output for English and Spanish in the bilingual group, t(1, 4) = 1.02, p = .366, although there was a large effect size (d = .91), indicating somewhat greater output in Spanish for this group of bilingual children. There was a significant difference, however, in language output between English and Spanish for the PS group, t(1, 4) = 22.29, p = .000, with a very large effect size (d = 19.94), indicating, as expected, much greater Spanish output for the PS children. There was also a significant difference in language output between English and Spanish for the PE group, t(1, 4) = 15.52, p = .000, with a very large effect size (d = 13.88), indicating, as expected, much greater English output for the PE children. On the basis of parent report of language history, the bilingual children all were exposed to Spanish from birth and to English for an average of 3.6 years (SD = 1.9; range = 1–5). Finally, the English dialect of the children was one that is common to the northeastern United States, and the Spanish dialect of the children consisted of Puerto Rican, Dominican, or Mexican Spanish.

Procedures A single-word phonological assessment for monolingual and bilingual (Spanish–English) Latino children was used to gauge children’s speech sound productions (Peña, Gutiérrez-Clellen, Iglesias, Goldstein, & Bedore, in development). The assessment contains 31 separate target items for English and 29 separate target items for Spanish and has been used previously to assess the phonological skills of Latino children (e.g., Goldstein & Washington, 2001, where the measure is described in detail). The Spanish version of the assessment was administered to the children in the PS group, and the English version of the assessment was administered to the children in the PE group. Both the Spanish and English versions were administered to the children in the bilingual group. To elicit the target words, represented by photographs of the items, the examiner prompted a response by asking, Qué es esto? (“What is this?”). If the child did not respond with the target word, the examiner either described the function of the stimulus (“It is used for...”) or used a fill-in-the-blank sentence in order to elicit the target word (e.g., for the target word “doctor,” the sentence is, “When I was sick, I saw the...”). If the child still did not name the picture, the examiner elicited a production by delayed imitation. The examiner phonetically transcribed the children’s productions at the time the children produced the words. All samples were recorded on a Marantz Model PMD222 portable cassette recorder. The examiners were bilingual graduate students in speech-language pathology who had previous coursework in phonetics and phonology and who were trained in transcribing the speech of English- and Spanishspeaking children. Both inter- and intrajudge reliability on transcription of the English and the Spanish samples was completed. Using audiotapes of the children’s productions, interjudge reliability was completed between the first and second authors on all of the productions of all of the children in the study and was 99% for both English and Spanish. Intrajudge

reliability was completed by the second author on all of the productions of all of the children in both English and Spanish after a 4-week interval and was found to be 99% in each language.

Analyses In order to measure the phonological skills of the children in the three groups, six analyses were completed. First, overall percentage of consonants correct2 (PCC–R; Shriberg, Austin, Lewis, McSweeney, & Wilson, 1997) and PCC–R by manner class were measured. For Spanish, manner classes included stops, nasals, fricatives, affricates, liquids, glides, flap, trill, and spirants. It should be noted that although spirants are not phonemic in Spanish, their accuracy is being calculated on the basis of previous findings that accuracy on spirants is higher for monolingual children than for bilingual children (Goldstein & Washington, 2001). For English, manner classes included stops, nasals, fricatives, affricates, liquids, and glides. Second, PCC–R for early-, middle-, and late-developing sounds was calculated. Data for early-, middle-, and late-developing sounds were taken from Shriberg (1993) for English and were derived from Acevedo (1993) for Spanish. Given the number of phonemes in each language, English and Spanish are divided into three categories: early 8 (/m, b, j, n, w, d, p, h/), middle 8 (/t, N, k, g, f, v, tS, dZ/), and late 8 (/S, T, s, z, D, l, ®, Z/) for English and early 6 (/p, d, n, t, j, w/), middle 6 (/k, g, x, m, f, ≠/), and late 6 (/tS, b, l, r, |, s/) for Spanish. The developmental validity of these subdivisions has been established previously for English (Shriberg, 1993) and for Spanish (Fabiano, Goldstein, & Washington, 2003). Third, an error analysis noting consonant and vowel substitution patterns in each of the three groups was completed. This analysis was completed in order to determine atypical error patterns and crosslinguistic effects. Fourth, type and rate of cross-linguistic effects were determined. Cross-linguistic effects were defined as segments not occurring in one language being used as substitutes in the other language. For example, the substitute of the English prevocalic [®] for a target trill /r/ in a Spanish production would be considered as a crosslinguistic effect because the substitute [®] does not occur in Spanish. The substitute of [tS] for /s/ in a Spanish production, however, would not be scored as an example of a cross-linguistic effect because the substitute [tS] does occur in Spanish. Modifications of syllable structure also would be considered cross-linguistic effects. For example, an English production of [esto„] for /sto„/ “store” would be scored as a cross-linguistic effect because the speaker is modifying the English production on the basis of the syllable structure of Spanish. Types of cross-linguistic effects were taken from those that have been documented in the literature (e.g., Kayser, 1993; MacDonald, 1989; Perez, 1994; Roseberry-McKibbin, 1995; Wolfram, 1971). 2 Although PCC–R was designed for and validated on productions from connected speech samples (Shriberg & Kwiatkowski, 1982), it has been applied to single words (e.g., Bernhardt & Stemberger, 2002) and is reported to correlate significantly with single-word productions (Hodson, 2004).

Goldstein et al.: Phonological Skills in Bilingual and Predominantly Spanish and English Speakers

205

Fifth, percentage of occurrence of phonological patterns was determined. Eleven phonological patterns were measured: unstressed syllable deletion, cluster reduction, final consonant deletion, stopping, backing, fronting (palatal and nasal), final devoicing, final voicing, assimilation, and spirantization. These patterns were chosen because they are exhibited routinely in the speech of Spanish- and Englishspeaking and bilingual children (e.g., Bichotte, Dunn, Gonzalez, Orpi, & Nye, 1993; Gildersleeve et al., 1996; Shriberg & Kwiatkowski, 1980). Finally, accuracy of syllable types was measured. The syllable types V, VC, CV, CVC, CCV, and CCVC were examined in Spanish; there were at least three opportunities for each of those syllable types to be produced, with the exception of VC syllables (one opportunity). The syllable types VC, VCC, CV, CVC, CVCC, CVCCC, and CCV were examined in English; there were at least three opportunities for each of those syllable types to be produced, with the exception of VCC, CVCCC, and VC syllables (one opportunity each). All analyses were completed taking into account the children’s dialect features. That is, features known to be characteristic of the children’s dialect were not scored as errors (e.g., final /s/ deletion in Puerto Rican and Dominican speakers) (e.g., Goldstein & Iglesias, 1996, 2001). To examine within-subjects effects, a Pearson product– moment correlation, using r2 for effect size, was computed to determine the relationship between frequency of output (i.e., the percentage of time that each child used English and/or Spanish) and overall PCC–R; consonant accuracy by manner classes; accuracy of early-, middle-, and latedeveloping sounds; phonological patterns; and accuracy of syllable types. Effect sizes were interpreted as follows: .1– .29 = small, .3–.59 = medium, and .6 and greater = large (Cohen, 1988). To examine between-subjects effects, comparisons were made between (a) the Spanish of the PS children and the Spanish of the bilingual children and (b) the English of the PE children and the English of the bilingual children using a series of univariate ANOVAs. Effect size was calculated using partial eta squared (ηp2) with interpretation using the following guidelines: 00–.10 = negligible, .10–.25 = small, .25–.50 = moderate, .50–.80 = large, and .80–1.00 = very large (Fiestas & Peña, 2004). For both within- and between-subjects effects, Bonferroni corrections were made to the preset alpha level of .05 for multiple comparisons; adjusted p values are indicated in the Results section.

RESULTS Consonant Accuracy PCC–R for all segments, by manner class (adjusted p value of .005 for Spanish and .007 for English) and by early-, middle-, and late-developing sounds (adjusted p value of .017 for Spanish and English), are represented in Table 2 for Spanish productions and in Table 3 for English productions. Spanish productions. Results of a Pearson product– moment correlation, with r2 as effect size, did not show

206

L ANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

any significant correlations between amount of output and PCC-R. Effect sizes were all small, with the exception of stops (a moderate effect size of .53), indicating a modest relationship between amount of output and accuracy on stops. Overall PCC-R across all groups averaged more than 90%, indicating mastery of the sound system as a whole. Although not significantly higher, F(1, 9) = 3.00, p = .121, ηp2 = .273, the results did indicate that overall PCC-R was slightly higher for the Spanish of the bilingual children (95.2%) than for the Spanish of the PS children (91.4%). Of the 5 children in the bilingual group and the 5 children in the PS group, only 1 child, S2, exhibited an overall PCC-R lower than 90% (86.4%). A series of univariate ANOVAs did not reveal one significant between-group difference for PCC-R and accuracy of sound classes (adjusted alpha of .005), although accuracy for stops approached significance, F(1, 9) = 14.42, p = .005, with a large effect size, ηp2 = .64 (Table 2). Overall, PCC-R for individual sound classes was relatively high. In fact, there were only three sound classes for which PCC-R was less than 85%. PCC-R for the flap, trill, and spirants was 76.2%, 76%, and 71.4%, respectively, for the PS children. Although Goldstein and Washington (2001) found higher accuracy on spirants for monolingual than for bilingual children, the opposite result was found for the children in this study. Accuracy for spirants was higher for bilingual children (85.7%) than for PS children (71.4%), perhaps as a result of their having overall greater phonological accuracy (as measured by overall PCC-R) than their PS peers (95.2% and 91.4%, respectively). Although accuracy for the groups was relatively high, some individual children exhibited relatively low accuracy for some sound classes. Accuracy on fricatives for S3, on the flap for B1 and S5, and on the trill for B3, S2, and S5 was at least 1 SD below the mean as compared to the group average. In order to examine consonant accuracy further, percentage correct for early-, middle-, and late-acquired sounds was calculated. There was no significant correlation between amount of output and accuracy on early-, middle-, and late-acquired sounds. In addition, there was no significant difference for early, middle, and late accuracy between the Spanish of the bilingual children and that of the PS children, although there was a large effect size for middle-6 sounds, ηp2 = .59 (Table 2). For both groups, percentage correct for early-, middle-, and late-acquired sounds showed a developmental trend. Accuracy was highest for earlydeveloping sounds and lowest for late-developing sounds. It should be noted, however, that percentage accuracy for middle- (89.03%) and late-acquired (89.94%) sounds for the PS children was practically identical. English productions. Results of a Pearson product– moment correlation, with r2 as effect size, showed no significant correlation between amount of output and consonant accuracy and small effect sizes, indicating little relationship between the variables. Overall PCC-R was not significantly higher for PE children (96.5%) than for bilingual children (94.8%). This trend held for all sounds classes. It should be noted, however, that effect sizes were large for stops (η p2 = .89) and affricates (ηp2 = 1.0).

SCHOOLS • Vol. 36 • 201–218 • July 2005

Table 2. Within- and between-subjects comparisons for consonant accuracy in Spanish productions.

PCC-R

Stops 26

Nasals

Fricatives

10

12

Opportunities

84

B1 B2 B3 B4 B5 Average SD

91.67 98.81 92.86 95.24 97.56 95.22 3.0

100.00 100.00 100.00 96.15 100.00 99.23 1.7

100.00 100.00 100.00 100.00 100.00 100.00 0

S1 S2 S3 S4 S5 Average SD

94.05 86.42 90.24 96.39 90.00 91.42 3.9

92.31 92.31 96.15 96.15 96.15 94.61 2.1

100.00 90.00 100.00 100.00 100.00 98.00 4.5

Affricates

Liquids

1

8

5

84.21 94.74 89.47 94.74 100.00 92.62 6.0

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 100.00 87.50 100.00 97.50 5.6

100.00 100.00 100.00 80.00 100.00 96.00 8.9

100.00 81.25 76.47 88.89 89.47 87.22 9.0

100.00 100.00 100.00 100.00 100.00 100.00 0

87.50 100.00 87.50 100.00 100.00 95.00 6.9

100.00 100.00 100.00 100.00 100.00 100.00 0

Within subjects ra p r2

–.212 .557 .045

–.729 .017 .531

–.274 .443 .075

–.047 .898 .002

b

–.255 .477 .065

Between subjects F p η p2

3.00 .12 .27

14.42 .005 .64

1.00 .38 .11

1.26 .30 .14

b

.40 .55 .05

9

6

7

B1 B2 B3 B4 B5 Average SD

55.56 100.00 100.00 100.00 75.00 86.11 20.2

100.00 100.00 33.33 100.00 100.00 86.66 29.8

85.71 85.71 71.43 85.71 100.00 85.71 10.1

100.00 100.00 100.00 95.45 100.00 99.09 2.0

94.73 100.00 100.00 94.73 100.00 97.89 2.9

86.11 100.00 88.88 97.22 93.54 93.15 5.7

S1 S2 S3 S4 S5 Average SD

77.78 71.43 88.89 100.00 42.86 76.19 21.6

100.00 50.00 80.00 100.00 50.00 76.00 25.1

100.00 71.43 42.86 71.43 71.43 71.43 20.2

95.65 91.30 100.00 100.00 100.00 97.39 3.9

95.45 81.82 90.00 90.91 86.96 89.03 5.1

92.11 85.71 88.57 100.00 83.33 89.94 6.5

Between subjects F p η p2

24

Middle-6

1.0 .35 .11

Trill

Within subjects r p r2

Early-6

.321 .366 .103

Flap Opportunities

Spirants

Glides

22

Late-6 38

–.099 .786 .010

–.043 .906 .002

–.227 .529 .051

–.278 .437 .077

–.532 .113 .283

–.022 .951 .0005

.56 .48 .07

.37 .56 .05

1.99 .20 .20

.75 .31 .09

11.62 .009 .59

.68 .43 .08

Note. B = bilingual (i.e., the Spanish productions of the bilingual children); S = Spanish (i.e., the Spanish productions of the predominantly Spanish-speaking children). a Correlations between output level and consonant accuracy, bnot computed because all values = 100%.

Goldstein et al.: Phonological Skills in Bilingual and Predominantly Spanish and English Speakers

207

Table 3. Within- and between-subjects comparisons for consonant accuracy in English productions.

PCC-R Opportunities

Stops

73

Nasals

Fricatives

11

14

30

Affricates 3

Liquids

Glides

13

2

B1 B2 B3 B4 B5 Average SD

98.63 100.00 94.52 90.41 90.53 94.81 4.5

100.00 100.00 90.00 96.67 91.67 95.66 4.7

100.00 100.00 100.00 100.00 100.00 100.00 0

92.86 100.00 100.00 85.71 83.33 92.38 7.8

100.00 100.00 66.67 100.00 66.67 86.66 18.3

100.00 100.00 100.00 69.23 94.44 92.73 13.4

100.00 100.00 100.00 100.00 100.00 100.00 0

E1 E2 E3 E4 E5 Average SD

97.26 100.00 100.00 89.58 95.88 96.54 4.3

100.00 100.00 100.00 86.67 93.33 99.46 1.2

100.00 100.00 100.00 86.67 93.33 96.00 6.0

92.86 100.00 100.00 77.78 94.44 93.02 9.1

100.00 100.00 100.00 100.00 100.00 100.00 0

92.31 100.00 100.00 88.89 94.44 95.13 4.9

100.00 100.00 100.00 100.00 100.00 100.00 0

Within subjects ra p r2

.219 .544 .048

Between subjects F p η p2

.39 .55 .27

.539 .108 .291 3.11 .12 .89

Early-8 Opportunities

-.485 .155 .235

.045 .903 .002

2.25 .17 .14

.01 .91 .19

.545 .104 .297 2.67 .14 1.0

Middle-8

24

26

.108 .766 .017

b

.14 .72 .02

b

Late-8 23

B1 B2 B3 B4 B5 Average SD

100.00 100.00 91.66 100.00 100.00 98.33 3.7

100.00 100.00 92.30 92.30 86.66 94.25 5.7

95.65 100.00 100.00 78.26 86.66 92.11 9.5

E1 E2 E3 E4 E5 Average SD

100.00 100.00 100.00 93.75 96.88 98.13 2.8

100.00 100.00 100.00 93.33 93.55 97.38 3.6

91.30 100.00 100.00 83.33 93.33 93.59 6.9

Within subjects r p r2

–.145 .689 .021

Between subjects F p η p2

.405 .245 .164

.21 .66 .14

1.07 .33 .65

.094 .796 .009 .08 .79 .20

Note. B = bilingual (i.e., the English productions of the bilingual children); E = English (i.e., the English productions of the predominantly English-speaking children). a Correlations between output level and consonant accuracy, bnot computed as all values = 100%.

208

L ANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

SCHOOLS • Vol. 36 • 201–218 • July 2005

Despite the relatively high group averages for consonant accuracy, there were some individual children who showed relatively low accuracy levels for some sound classes in comparison to the group averages. For example, accuracy on fricatives for E4 was almost 2 SD below the mean in comparison to the group average. Accuracy on affricates for B3 and B5 was 1 SD below the mean compared to the bilingual group, and accuracy on liquids for B4 was 2 SD below the mean in comparison to the bilingual group average. As was the case for Spanish productions, there was no significant correlation between amount of output and early-, middle- and late-developing sounds. In addition, there was no significant difference for early, middle, and late accuracy between the English of the bilingual children and that of the PE children, although there was a large effect size for middle-8 sounds, ηp2 = .65 (Table 3). The trend in accuracy for early-, middle-, and late-developing sounds was developmental for PE children and bilingual children. Except for 2 children, individual variation on early-, middle- and late-developing sounds was within 1 SD of the group mean. Accuracy for late-8 sounds for B3 and E4 was more than 1 SD below the mean.

Substitution Patterns The type and frequency of substitution patterns for the Spanish productions are outlined in Table 4 and in Table 5 for the English productions.

Table 4. Number and types of substitution patterns in Spanish.

Percentage useda

Pattern

Tokens

Predominantly Spanish

[D] ➝ [d] [V] ➝ [g] /|/ ➝ [l] /r/ ➝ [|] [V] ➝ [g] /r/ ➝ [h] [B] ➝ [b] /g/ ➝ [t] /g/ ➝ [Ø] /x/ ➝ [j] /e/ ➝ [i]

4 3 3 3 2 2 1 1 1 1 1

18.2 13.6 13.6 13.6 9.1 9.1 4.5 4.5 4.5 4.5 4.5

Bilingual

/r/ ➝ [|] /|/ ➝ [l] /f/ ➝ [k] /f/ ➝ [v] /s/ ➝ [z]b /|/ ➝ [n] [D] ➝ [d] [D] ➝ [j] [B] ➝ [|] /r/ ➝ [j] [V] ➝ [v]b /k/ ➝ [tS]

3 3 1 1 1 1 1 1 1 1 1 1

18.75 18.75 6.25 6.25 6.25 6.25 6.25 6.25 6.25 6.25 6.25 6.25

a

Based on total number of substitutions; bexample of transfer.

Table 5. Number and types of substitution patterns in English.

Pattern

Tokens

Percentage useda

Predominantly English

/v/ ➝ [b] /|/ ➝ [t] [D] ➝ [d] /T/ ➝ [s] /m/ ➝ [n] /|/ ➝ [n] /m/ ➝ [|]b /l/ ➝ [|]b /ø/ ➝ [œ]

1 1 1 1 1 1 1 1 1

11.1 11.1 11.1 11.1 11.1 11.1 11.1 11.1 11.1

Bilingual

/l/ ➝ [j] /tS/ ➝ [S] /s/ ➝ [t] /t/ ➝ [s] /t/ ➝ [S] /z/ ➝ [s]

3 2 1 1 1 1

33.3 22.2 11.1 11.1 11.1 11.1

a

Based on total number of substitutions; b(possible) example of transfer.

Spanish productions. Overall, there were relatively few substitution errors made by the children: 22 for the PS children and 16 for the bilingual children. In general, the children showed the largest number of errors on the spirants and the later developing flap and trill. For the PS children, 18 of 22 substitution errors were on spirants (10 tokens), flap (3), and trill (5). For the bilingual children, 12 of 16 substitutions occurred on spirants (4 tokens), flap (4), and trill (4). Other targets in error included /f/, /s/, and /k/ for the bilingual children and /g/, /x/, and /e/ for the PS children. English productions. For both groups, the distribution of substitution patterns was more equivalent than for the Spanish productions. For PE children, the nine error tokens represented nine different types. For the bilingual children, the nine tokens represented six types; only substitutions for /l/ and /tS/ occurred more than once. For the bilingual children, the types of substitutions were different in their Spanish (Table 4) and English (Table 5) productions. That is, there was no overlap in the types of substitutions exhibited in each language. In the bilingual children’s Spanish productions, flap for trill and [l] for flap were the most commonly occurring substitutions as compared to [j] for /l/ in their English productions.

Cross-Linguistic Effects The substitution analysis also was used to measure the frequency and types of cross-linguistic effects. Of the 1,790 consonant segments produced by the children, there were only three examples of cross-linguistic effects (0.17%). In the bilingual children, there was one example of [z] substituting for /s/ and one example of [v] substituting for [V]. Both of these examples occurred in the bilingual children’s Spanish productions. There were no examples of cross-linguistic effects in the bilingual children’s English productions.

Goldstein et al.: Phonological Skills in Bilingual and Predominantly Spanish and English Speakers

209

Phonological Patterns

In the PS and PE children’s productions, there were two (questionable) examples of cross-linguistic effects. There was one occurrence each of [|] substituting for /l/ and of [|] substituting for /m/, both in the production of 1 PE child (E3). Those examples are questionable because the flap is an allophonic variation of /t/ and /d/ in English and is in the repertoire of English speakers, even though it is not phonemic. Thus, these specific examples probably should not be considered examples of cross-linguistic effects.

In order to gauge general patterns of phonological production, percentage of occurrence for phonological patterns that are commonly measured in children of this age was calculated for Spanish productions (Table 6) and for English productions (Table 7). Spanish productions. On the basis of an adjusted alpha of .005, there was no significant correlation between

Table 6. Percentage of occurrence and correlationsa for phonological patterns in Spanish productions.

Unstressed syllable deletion Opportunities

Cluster reduction

68

Final consonant deletion

Liquid simplification

16

4

23

Stopping 20

Backing

Fronting

78

78

B1 B2 B3 B4 B5 Average SD

1.47 0 2.94 1.47 0 1.17 1.2

12.50 0 0 6.67 0 3.83 5.6

0 0 0 0 0 0 0

0 0 0 12.50 0 2.50 5.6

5.00 5.00 0 0 0 2.00 2.7

1.19 1.19 2.38 0 0 0.95 .9

1.19 0 0 2.38 0 0.71 1.0

S1 S2 S3 S4 S5 Average SD

2.94 0 4.41 0 0 1.47 2.1

12.50 0 25.00 0 18.75 12.60 11.2

0 0 0 0 0 0 0

12.50 0 12.50 0 0 5.00 6.8

0 11.76 16.67 0 10 6.58 7.4

0 2.47 1.25 1.20 1.28 1.38 .9

0 3.70 0 1.20 0 0.76 1.6

ra p r2

–.014 .969 .028

–.070 .847 .005

–.047 .897 .002

b

.252 .482 .063

.255 .477 .065

Final devoicing

Final voicing

Opportunities

4

3

B1 B2 B3 B4 B5 Average SD

0 0 0 0 0 0 0

S1 S2 S3 S4 S5 Average SD

0 0 0 0 0 0 0

r p r2

b

25.00 0 0 0 0 5.00 11.2

.372 .290 .138 Assimilation 70

Spirantization 60

0 0 1.30 0 1.40 0.53 1.2

0 0 0 1.56 0 .32 .7

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 1.56 1.56 3.33 1.27 1.4

–.613 .060 .376

–.310 .383 .096

.312 .381 .097

Note. B = bilingual (i.e., the Spanish productions of the bilingual children); S = Spanish (i.e., the Spanish productions of the predominantly Spanish-speaking children). a Correlations between output level and consonant accuracy; bnot computed as all values = 100%.

210

L ANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

SCHOOLS • Vol. 36 • 201–218 • July 2005

Table 7. Percentage of occurrence and correlationsa for phonological patterns in English productions.

Unstressed syllable deletion Opportunities

Cluster reduction

53

Final consonant deletion

Liquid simplification

25

18

13

0 0 11.11 0 5.56 3.33 4.9

B1 B2 B3 B4 B5 Average SD

0 0 0 0 0 0 0

0 0 9.10 18.18 9.1 7.27 7.6

E1 E2 E3 E4 E5 Average SD

0 0 0 1.89 0 1.62 .8

5.56 0 0 12.00 0 12.71 5.3

ra p r2

.345 .329 .119

-.281 .432 .079

Final devoicing

0 0 0 0 0 0 0

Final voicing

Opportunities

4

10

B1 B2 B3 B4 B5 Average SD

0 0 0 0 0 0 0

25.00 0 0 0 0 5.00 11.8

E1 E2 E3 E4 E5 Average SD

0 0 0 0 0 0 0

r p r2

b

21

Backing

Fronting

73

61

0 0 0 0 0 0 0

5.88 0 0 11.76 4.76 4.48 4.8

1.39 0 0 1.39 2.15 .98 1.0

0 0 0 2.78 0 .55 1.2

0 0 0 0 0 0 0

5.88 0 0 4.76 4.76 6.41 2.8

1.39 0 0 2.13 2.11 2.90 1.1

0 0 0 2.13 0 1.70 1.0

b

-.442 .828 .195

Stopping

–.079 .828 .006 Assimilation 68

.111 .760 .012

–.019 .958 .0004 Spirantization 56

0 0 1.32 0 1.35 0.53 1.2

0 0 0 1.56 0 .32 .7

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

–.080 .826 .006

–.547 .102 .299

–.288 .421 .083

Note. B = bilingual (i.e., the English productions of the bilingual children); E = English (i.e., the English productions of the predominantly English-speaking children). a Correlations between output level and consonant accuracy, bnot computed as all values = 100%.

amount of output and percentage of occurrence for the 11 phonological patterns and there were small effect sizes, with the exception of final voicing (a moderate effect size of .38), indicating little relationship between the variables. Given the high consonant accuracy exhibited by the children, the percentages of occurrence for the phonological patterns were relatively low. Across all 11 phonological patterns, only 1 pattern was exhibited with a frequency of occurrence greater than 10%, cluster reduction in the PS

children (12.6%). Four patterns were not exhibited by any of the children: final consonant deletion and final devoicing by both groups and final voicing and assimilation by the PS children. There were other patterns in which only 1 or 2 children evidenced tokens; for example, liquid simplification, final voicing, and spirantization. Between-subjects effects for the Spanish of the bilingual children and the Spanish of the PS children were calculated on the four phonological patterns with percentages of

Goldstein et al.: Phonological Skills in Bilingual and Predominantly Spanish and English Speakers

211

occurrence greater than or equal to 5%. On the basis of an adjusted alpha level of .01, univariate ANOVAs indicated no significant difference for any of the four patterns and small effect sizes: cluster reduction, F(1, 9) = 1.75, p = .22, ηp2 = .18; liquid simplification, F(1, 9) = .40, p = .55, ηp2 =.05; stopping, F(1, 9) = 2.58, p = .15, ηp2 = .24; and final voicing, F(1 ,9) = 1.00, p = .35, ηp2 = .11]. English productions. There were no significant correlations (adjusted alpha of .005) between amount of output and percentage of occurrence for the 11 phonological patterns and there were small effect sizes, demonstrating the lack of a robust relationship between the variables. As was the case for the Spanish productions, the frequency of occurrence for these patterns was relatively low; again, cluster reduction was the only pattern to be exhibited with a frequency greater than 10% (12.7% by the PE children). In English, six patterns were not exhibited by any children in one of the two groups: liquid simplification and final devoicing for both groups; unstressed syllable deletion by the bilingual children; and final voicing, assimilation, and spirantization by the PE children. Between-subjects effects for the English of the bilingual children and the English of the PE children were calculated on the three phonological patterns with percentages of occurrence greater than or equal to 5%. On the basis of an adjusted alpha level of .02, a series of univariate ANOVAs indicated no significant difference for any of the patterns and small effect sizes: cluster reduction, F(1, 9) = .82, p = .39, ηp2 = .22; stopping, F(1, 9) = .31, p = .35, ηp2 = .17; and final voicing, F(1, 9) = 1.00, p = .35, ηp2 (not calculated given a 0% occurrence for children in the PE group).

Percentages of occurrence for some phonological patterns were somewhat different in bilingual children for their Spanish and English productions. Unstressed syllable deletion and liquid simplification showed somewhat higher percentages of occurrence in Spanish, and cluster reduction, final consonant deletion, and stopping showed somewhat higher percentages of occurrence in English. Percentages of occurrence for the other patterns were comparable across the two languages.

Syllable Types To judge the children’s phonotactic skill, syllable accuracy was measured by calculating percentage correct for syllable types in Spanish (Table 8) and in English (Table 9). Spanish productions. There was no significant correlation (adjusted alpha of .007) between amount of output and syllable accuracy and there were small effect sizes. Overall syllable accuracy was higher in the bilingual children (96.2%) than in the PS children (89.8%). A large portion of that discrepancy can be accounted for by the low accuracy of S3 on V, VC, and CCV syllable types (50%, 0%, and 62.5%, respectively). Interestingly, this was not the child with the lowest overall PCC–R in the PS group. Between-subjects effects for the Spanish of the bilingual children and the Spanish of the PS children were calculated on overall syllable accuracy. Results of a univariate ANOVA indicated no significant difference between the groups, F(1, 9) = 1.18, p = .31, ηp2 = .31. Four syllable types showed identical (or almost identical, ≤5% difference)

Table 8. Percentage correct and correlationsa for syllable types in Spanish.

Overall Opportunities

V

73

VC

CV

CVC

CCV

CCVC

7

7

44

6

8

3

B1 B2 B3 B4 B5 Average SD

90.39 100.00 96.67 93.70 100.00 96.15 4.2

75.00 100.00 100.00 100.00 100.00 95.00 11.1

92.31 100.00 80.00 90.00 100.00 92.46 8.2

100.00 100.00 100.00 97.22 100.00 99.44 1.2

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 100.00 75.00 100.00 95.00 11.1

75.00 100.00 100.00 100.00 100.00 95.00 11.1

S1 S2 S3 S4 S5 Average SD

90.81 91.25 68.75 99.59 98.50 89.78 12.4

100.00 100.00 50.00 100.00 100.00 89.58 22.4

100.00 100.00 0 100.00 100.00 87.5 44.7

94.87 97.62 100.00 97.56 100.00 98.16 2.1

100.00 100.00 100.00 100.00 100.00 100.00 0

75.00 87.50 62.50 100.00 75.00 78.13 14.3

75.00 100.00 100.00 100.00 50.00 84.38 22.4

ra p r2

.452 .190 .204

.002 .997 .000

–.144 .691 .021

.423 .223 .179

b

–.502 .139 .252

–.121 .739 .015

Note. B = bilingual (i.e., the Spanish productions of the bilingual children); S = Spanish (i.e., the Spanish productions of the predominantly Spanish-speaking children). a Correlations between output level and consonant accuracy; bnot computed as all values = 100%.

212

L ANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

SCHOOLS • Vol. 36 • 201–218 • July 2005

Table 9. Percentage correct and correlationsa for syllable types in English.

Overall Opportunities

VC

VCC

CV

CVC

CVCC

CVCCC

CCV

2

1

18

12

3

1

4

B1 B2 B3 B4 B5 Average SD

100.00 100.00 98.61 96.88 76.44 94.39 10.1

100.00 100.00 n.a n.a 100.00 100.00 0

100.00 100.00 100.00 100.00 0 80.00 44.7

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 100.00 100.00 100.00 100.00 0

E1 E2 E3 E4 E5 Average SD

98.57 100.00 73.01 84.25 100.00 91.17 12.1

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 41.18 76.47 100.00 85.34 25.8

100.00 100.00 90.90 81.82 100.00 88.1 8.1

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 100.00 100.00 100.00 100.00 0

100.00 100.00 25.00 100.00 100.00 76.92 33.5

b

b

ra p r2

.079 .829 .006

b

.510 .132 .260

–.299 .402 .089

–.404 .636 .163

CCVC

CCCV

9

1

B1 B2 B3 B4 B5 Average SD

n.a. n.a. 88.88 75.00 88.88 84.25 8.0

100.00 100.00 100.00 100.00 0 80.00 44.7

E1 E2 E3 E4 E5 Average SD

90.00 100.00 100.00 100.00 100.00 78.75 4.8

n.a. 100.00 0 0 100.00 28.57 57.7

Opportunities

ra p r2

.712 .047 .507

–.171 .636 .029

–.087 .824 .008

Note. B = bilingual (i.e., the English productions of the bilingual children); E = English (i.e., the English productions of the predominantly English-speaking children). a Correlations between output level and consonant accuracy, bnot computed as all values = 100%.

accuracy between the Spanish of the bilingual children and the Spanish of the PS children: V, VC, CV, and CVC. Two syllable types showed higher accuracy in Spanish for the bilingual children than for the PS children: CCV and CCVC. There were no syllable types that showed higher accuracy in Spanish for the PS children than for the bilingual children. English productions. There was no significant correlation (adjusted alpha of .005) between amount of output and syllable accuracy and there were small effect sizes, with

the exception of CCVC syllables (a moderate effect size of .51). Between-subjects effects for the English of the bilingual children and the English of the PE children were calculated on overall syllable accuracy. Results of a univariate ANOVA indicated no significant difference between the groups, F(1, 9) = .21, p = .66, ηp2 = .20. As was the case for Spanish, overall syllable accuracy in English was higher for bilingual children (94.4%) than for PE children (91.2%). Three syllable types showed identical (or almost identical, ≤5% difference) accuracy between the

Goldstein et al.: Phonological Skills in Bilingual and Predominantly Spanish and English Speakers

213

English of the bilingual children and the English of the PE children: VC, CVCC, and CVCCC. Four syllable types showed higher accuracy in English for the bilingual children than for the PE children: CV, CCV, CCVC, and CCCV. There were two syllable types, VCC and CVC, which showed higher accuracy in English for the PE children than for the bilingual children. For the bilingual children, accuracy for syllable types was largely commensurate across their Spanish and English productions. There were two exceptions: Accuracy for VC syllables was higher in English (100%) than in Spanish (92.5%), and accuracy for CCVC syllables was higher in Spanish (95%) than in English (84.3%).

DISCUSSION The purpose of this study was to examine the relationship between amount of output in Spanish and English and phonological skills and to determine if and how phonological skills in Spanish-English bilingual children compared to those of PE and PS children. Recall that language output here was based on parent estimates of language use (cf. Peña et al., 2003). The results indicated that there were no significant correlations between amount of output in each language and phonological skill (segmental and syllabic accuracy and percentage of occurrence of phonological patterns). Although there were some comparisons that showed at least a moderate effect size, the majority of effect sizes were negligible to small. These results indicate a relatively limited relationship between level of output (at least as measured by parent report) and phonological skill. This lack of a relationship between output and phonological skill indicates that overall volubility in each language cannot be used to specify relative phonological skill. Given the distinction between groups on parent report of output, a link between output and phonological performance within the three groups would have been expected. Such a link has been found for other areas of language (e.g., semantics; Peña, Bedore, & Rappazzo, 2003). In this study, such a distinction largely was not found. In fact, there were moderate-to-large effect sizes between level of output and phonological skill only for stops and final voicing in Spanish and for stops and affricates in English. Moreover, there were some negative correlations between output and phonological performance (e.g., almost all of the correlations between sound class accuracy and output for Spanish productions). These results indicate that parent report of relative output, for example, may have limited validity for establishing a relationship between parent estimates of language and measures of phonological skills, if only as it relates to the limited number of children in this study. In the future, independent measures of language output, other than those of parent report, should be gathered to determine if they more adequately link with phonological skills. The results of the current study also demonstrated that there was no significant difference in segmental accuracy, syllabic accuracy, or percentage of occurrence of phonological patterns between either the Spanish skills of PS and

214

L ANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

Spanish–English bilingual speakers or the English skills of PE and Spanish–English bilingual speakers. It is interesting to note, however, the large effect size for middle-developing sounds between bilingual and predominantly monolingual children. Future studies will be needed to determine if this finding extends to children with different linguistic profiles and children who are older and younger than those in this study. The lack of a distinction between groups contrasted somewhat with previous studies. For example, Gildersleeve et al. (1996) noted more advanced phonological skill in monolingual children than in bilingual children. There are two main factors that likely contributed to this discrepancy. First, the age of the children under investigation differed between the studies. Gildersleeve et al. (1996) and Goldstein and Washington (2001) primarily examined 4-year-olds whereas the current study examined all 5-yearolds. It is possible that the phonological skills of bilingual children become similar to those of monolingual children as they get older. Gildersleeve-Neumann and Davis (1998) found that as age increased for bilingual children, their accuracy became commensurate with that of monolingual children. Second, differences between monolingual and bilingual children may have been an unintended consequence of differences in methodology (e.g., number of participants, data collection procedures, and analyses) rather than distinctions in phonological skill. Goldstein and Washington noted the difficulty in trying to compare data from 4-year-old Spanish–English bilingual children to previous studies of monolingual English- and Spanishspeaking children. Goldstein and Washington’s retrospective comparison resulted in some disparate findings between groups. In the current study, phonological skills of bilingual children were compared prospectively with those of agematched PS and PE peers. That prospective comparison obviated differences in participant selection, data collection, and analyses. Just as Vihman and McLaughlin (1982) hypothesized, prospectively comparing the phonological skills of bilingual and PS and PE children largely yielded commensurate performance between groups. These results are limited, of course, by the relatively few participants in each group that might yield enough power to find differences between groups. Including additional children in all three groups might result in enough power to identify between-group differences should they exist. In addition, even though single-word samples are routinely used for phonological analysis, measures of connected speech might be gathered to increase the overall phonotactic complexity of the samples. Although the phonological skills of Spanish–English bilingual children were proportionate with those of PS and PE children, they were not identical. There were group and individual differences in performance across groups for sound class accuracy; accuracy for early-, middle-, and late-developing phonemes; and phonological patterns. For example, affricates in the English of the PE children were more accurate than affricates in the English of bilingual children. On the other hand, fricatives were slightly more accurate in the Spanish of bilingual children than in the Spanish of PS children. Fabiano et al. (2003) found a similar result in a prospective study of Spanish–English

SCHOOLS • Vol. 36 • 201–218 • July 2005

bilingual children and monolingual Spanish and English speakers aged 4;5 to 5;6. In their study, accuracy in English for most sound classes (e.g., fricatives and liquids) was higher for bilingual children (as was their overall PCC–R) than for monolingual children, although the monolingual children exhibited greater accuracy on some sound classes (e.g., stops). In Spanish, however, accuracy across sound classes was relatively equal between bilingual and monolingual children with three exceptions: Accuracy on the flap, the trill, and spirants was higher for the bilingual children than for the monolingual children, just as it was for the children in the current study. For the children in the current study, the slight advantage in overall accuracy between the Spanish of bilingual children (95.2%) and the Spanish of PS children (91.4%) may have contributed to this discrepancy. Accuracy for early-, middle-, and late-acquired sounds followed the trend of sound class accuracy. Although not a specific focus of this study, data from early-, middle-, and late-acquired sounds continue to validate this metric for Spanish-speaking children. Data from this study and those from Goldstein et al. (2003) indicate that Shriberg’s (1993) system for differentiating developmental groups of sounds in English can be applied validly to Spanish. The one difference between English and Spanish is the number of phonemes in each group—eight for each group in English and six for each group in Spanish. Of course, this schema will need to be applied to Spanish-speaking children with phonological disorders as has been done with Englishspeaking children (Shriberg, 1993). There were also some differences in percentage of occurrence for phonological patterns across languages. For example, two phonological patterns, cluster reduction and final consonant deletion, showed higher percentages of occurrence in English than in Spanish. Unstressed syllable deletion, however, was exhibited with a higher percentage of occurrence in Spanish than in English. These differences are likely due to the phonotactic constraints of each language. Word-initial clusters are somewhat more complex and word-final consonants are more replete in English than in Spanish (Hammond, 2001). Thus, there are more opportunities for children to exhibit cluster reduction and final consonant deletion in English than in Spanish. Goldstein and Iglesias (1996) found a similar result for monolingual Spanish-speaking children in comparison to monolingual English-speaking children. On the other hand, unstressed syllable deletion was higher in the children’s Spanish than in their English. This result is expected given that, on average, words in Spanish are longer than those in English (Vihman, 1996); thus, there are more opportunities for unstressed syllables to be deleted. Despite these rather isolated differences across languages, percentages of occurrence (and consonant and syllable accuracy) were comparable in the Spanish and English productions of bilingual children. Again, age may be the determining factor. As bilingual children mature, their skills across languages become less variable (Gathercole, 2002; Gildersleeve-Neumann & Davis, 1998), although there was individual variation, as expected (e.g., Romaine, 1995). Individual variation was common throughout all of the

results (segmental and syllabic accuracy and percentage of occurrence for phonological patterns). For example, even though accuracy for the flap and trill for the bilingual and PS children was 86% and 76%, respectively, some children showed quite low and/or variable accuracy on these sounds. B1’s accuracy on the flap was only 56%, but was 100% on the trill; S5’s accuracy was low on both sounds (43% for flap and 50% for trill). This variability extended to other aspects of phonological skill as well. For example, the percentage of occurrence for final voicing in English was only 5%, although it was 25% for 1 child (B1). In fact, B1 was the only child who exhibited that pattern at all. Thus, the combination of group and individual results provides a baseline to speech-language pathologists (SLPs) who use a single-word elicitation task to differentiate TD bilingual children from those with phonological disorders. This type of task is routinely used by SLPs for this purpose with children this age, thus giving it some ecological validity. Two other issues regarding bilingual phonology center on the types of errors that bilingual children exhibit: crosslinguistic effects and the typicality of substitution errors. One of the attested characteristics of bilingual children’s phonology is cross-linguistic effects (e.g., Keshavarz & Ingram, 2002). There have been questions, however, as to whether these effects are the result of the influence of one language on the other (e.g., Perez, 1994) or whether they are relatively common substitutions exhibited by monolingual speakers as well (e.g., Vihman & McLaughlin, 1982). In the current study, the frequency of cross-linguistic effects was extremely low (.17% of all the consonants produced). Two other studies examining the frequency and types of these effects also found a similar result. Goldstein and Washington (2001) found the frequency for crosslinguistic effects to be .3% of all sounds produced. This finding is consistent with previous studies (cf. Goldstein et al., 2003; Goldstein & Washington, 2001). Some researchers have claimed that bilingual children exhibit atypical error patterns in comparison to their monolingual peers (e.g., Dodd et al., 1996). In the current study, bilingual children largely produced substitutions that also have been attested in monolingual children. Thus, bilingual children did not exhibit a large proportion of atypical error patterns. In Spanish, the bilingual children produced only four substitution types (/f/ ➝ [k]; /f/ ➝ [v]; /V/ ➝ [B]; /k/ ➝ [tS]) that were not attested in monolingual Spanish-speaking children (Goldstein, Belen, & Ballard, 2002). In English, the bilingual children produced only two substitution types (/t/ ➝ [s], /t/ ➝ [S]) that also were not exhibited by monolingual English-speaking children (Smit, 1993). Thus, although, bilingual children may produce some atypical errors (i.e., those not usually exhibited by monolingual children), those types of errors are relatively uncommon, at least in this relatively small group of 5-year-olds.

CLINICAL IMPLICATIONS According to the results of this study, there is no clear link between bilingual children’s output in each language, as

Goldstein et al.: Phonological Skills in Bilingual and Predominantly Spanish and English Speakers

215

measured by parent report, and phonological skill. Thus, in gauging the phonological skills of bilingual children, SLPs may need to use measures other than parent report to determine that relationship. For example, teacher report, classroom observation, or direct independent measures of language skill during connected discourse, as well as discrete criterion-referenced tasks, might be other alternatives. The results from this study also suggest that the phonological skills of bilingual and PS and PE children (and the Spanish and English skills of bilingual children) are more similar than different (remembering, of course, that individual variation was common throughout all the results). Despite the similarities between 5-year-old PS and PE and bilingual children, the evidence from this study indicates that the phonological skills of the groups are not identical. For example, accuracy for the flap, trill, and spirants was higher in the Spanish of the bilingual children than in the Spanish of the PS children. Also, the percentage of occurrence for unstressed syllable deletion was higher in Spanish than in English. This study highlights the need to include multiple analyses to examine the phonological skills of bilingual children. Completing a limited number of analyses (e.g., examining percentage of occurrence for phonological patterns) is insufficient for measuring the complete phonological system of bilingual children. To adequately assess phonology, SLPs should complete comprehensive independent and relational analyses (e.g., Stoel-Gammon & Dunn, 1985). These results also reinforce the concept that bilinguals are not simply the amalgamated product of their two languages, but that each language operates largely independently with constraints from the specific ambient language being acquired (e.g., Grosjean, 1989). Thus, SLPs should examine phonological structures that are common to both languages (e.g., stops, CV syllables) and also examine features that are unique to each language (e.g., trill in Spanish and /®/ in English) (Goldstein, 2001b). Results from the current study indicate mastery of most, but not all, sounds and phonotactic patterns in Spanish or English by TD Latino children entering kindergarten at age 5. This information indicates that, in the absence of timely identification and appropriate intervention, Latino children who are less skilled on phonological production in both of their languages may be at risk for delays.

ACKNOWLEDGMENTS We are grateful to the families who and schools that participated in this project. Thanks to Ferenc Bunta for providing comments on an earlier draft of this manuscript. This study was funded in part by NIH Contract N01-DC-8-2100.

Amayreh, M. M., & Dyson, A. T. (1998). The acquisition of Arabic consonants. Journal of Speech, Language, and Hearing Research, 41, 642–653. Anderson, R., & Smith, B. (1987). Phonological development of two-year-old monolingual Puerto Rican Spanish-speaking children. Journal of Child Language, 14, 57–78. Austin, D., & Shriberg, L. (1997). Lifespan reference data for ten measures of articulation competence using the Speech Disorders Classification System (Tech. Rep. No. 3). Madison, WI: University of Wisconsin–Madison, Phonology Project. Bernhardt, B., & Stemberger, J. (2002). Intervocalic consonants in the speech of English-speaking Canadian children with phonological disorders. Clinical Linguistics and Phonetics, 16, 199–214. Bichotte, M., Dunn, B., Gonzalez, L., Orpi, J., & Nye, C. (1993, November). Assessing phonological performance of bilingual school-age Puerto Rican children. Paper presented at the annual convention of the American Speech-Language-Hearing Association, Anaheim, CA. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Earlbaum. De la Fuente, M. T. (1985). The order of acquisition of Spanish consonant phonemes by monolingual Spanish speaking children between the ages of 2.0 and 6.5. Unpublished doctoral dissertation, Georgetown University, Washington, DC. Dodd, B., So., L., & Li, W. (1996). Symptoms of disorder without impairment: The written and spoken errors of bilinguals. In B. Dodd, R. Campbell, & L. Worrall (Eds.), Evaluating theories of language: Evidence from disorder (pp. 119–136). London: Whurr. Eblen, R. (1982). A study of the acquisition of fricatives by threeyear-old children learning Mexican Spanish. Language and Speech, 25, 201–220. Fabiano, L., Goldstein, B., & Washington, P. (2003, July). Phonological development in monolingual English, monolingual Spanish, and bilingual (Spanish–English) Latino children. Poster session presented at the 2003 Child Phonology Conference, Vancouver, Canada. Fantini, A. (1985). Language acquisition of a bilingual child: A sociolinguistic perspective (to age 10). San Diego, CA: College Hill Press. Fiestas, C., & Peña, E. (2004). Narrative discourse in bilingual children: Language and task effects. Language, Speech, and Hearing Services in Schools, 35, 155–168. Gathercole, V. C. M. (2002). Monolingual and bilingual acquisition: Learning different treatments of that-trace phenomena in English and Spanish. In D. K. Oller & R. Eilers (Eds.), Language and literacy in bilingual children (pp. 220–254). Clevedon, U.K.: Multilingual Matters. Gildersleeve, C., Davis, B., & Stubbe, E. (1996, November). When monolingual rules don’t apply: Speech development in a bilingual environment. Paper presented at the annual convention of the American Speech-Language-Hearing Association, Seattle, WA.

REFERENCES

Gildersleeve-Neumann, C., & Davis, B. (1998, November). Learning English in a bilingual preschool environment: Change over time. Paper presented at the annual convention of the American Speech-Language-Hearing Association, San Antonio, TX.

Acevedo, M. A. (1993). Development of Spanish consonants in preschool children. Journal of Childhood Communication Disorders, 15, 9–15.

Goldstein, B. (1995). Spanish phonological development. In H. Kayser (Ed.), Bilingual speech-language pathology: An Hispanic focus (pp. 17–38). San Diego, CA: Singular.

216

L ANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

SCHOOLS • Vol. 36 • 201–218 • July 2005

Goldstein, B. (2001a). Assessing phonological skills in Hispanic/ Latino children. Seminars in Speech and Language, 22, 39–49. Goldstein, B. (2001b). Transcription of Spanish and Spanishinfluenced English. Communication Disorders Quarterly, 23, 54–60. Goldstein, B., Belen, L., & Ballard, D. (2002, November). Substitution patterns in the phonology of Spanish-speaking children. Poster presented at the annual convention of the American Speech-Language-Hearing Association, Atlanta, GA. Goldstein, B., Fabiano, L., & Iglesias, A. (2003, May). Phonological development in monolingual English, monolingual Spanish, and bilingual (Spanish–English) Latino children. Poster session presented at the 4th annual International Symposium on Bilingualism, Tempe, AZ. Goldstein, B., & Iglesias, A. (1996). Phonological patterns in normally developing Spanish-speaking 3- and 4-year-olds of Puerto Rican descent. Language, Speech, and Hearing Services in Schools, 27, 82–90. Goldstein, B., & Iglesias, A. (2001). The effect of dialect on phonological analysis: Evidence from Spanish-speaking children. American Journal of Speech-Language Pathology, 10, 394–406. Goldstein, B., & Washington, P. (2001). An initial investigation of phonological patterns in 4-year-old typically developing Spanish–English bilingual children. Language, Speech, and Hearing Services in Schools, 32, 153–164. Gonzalez, A. (1981). A descriptive study of phonological development in normal speaking Puerto Rican preschoolers. Unpublished doctoral dissertation, The Pennsylvania State University, State College. Grosjean, F. (1989). Neurolinguists, beware! The bilingual is not two monolinguals in one person. Brain and Language, 36, 3–15. Grunwell, P. (1997). Natural phonology. In M. J. Ball & R. Kent (Eds.), The new phonologies: Developments in clinical linguistics (pp. 35–75). San Diego, CA: Singular. Gutierrez-Clellen, V., & Kreiter, J. (2003). Understanding child bilingual acquisition using parent and teacher reports. Applied Psycholinguistics, 24, 267–288. Haelsig, P., & Madison, C. (1986). A study of phonological processes exhibited by 3-, 4-, and 5-year-old children. Language, Speech, and Hearing Services in Schools, 17, 107–114. Hammond, R. (2001). The sounds of Spanish: Analysis and application (with special reference to American English). Somerville, MA: Cascadilla Press. Hodson, B. (2004). Hodson Assessment of Phonological Patterns (3rd ed.). Austin, TX: Pro-Ed. Johnson, C., & Lancaster, P. (1998). The development of more than one phonology: A case study of a Norwegian–English bilingual child. International Journal of Bilingualism, 2, 265–300.

Paradis, J. (2001). Do bilingual two-year-olds have separate phonological systems? International Journal of Bilingualism, 5, 19–38. Pearson, B, Fernandez, S., Lewedeg, V., & Oller, D. K. (1997). The relation of input factors to lexical learning by bilingual infants. Applied Psycholinguistics, 18, 41–58. Peña, E., Bedore, L., & Rappazzo, C. (2003). Comparison of Spanish, English, and bilingual children’s performance across semantic tasks. Language, Speech, and Hearing Services in Schools, 34, 5–16. Peña, E., Gutiérrez-Clellen, V., Iglesias, A., Goldstein, B., & Bedore, L. (in development). Bilingual English Spanish assessment. Unpublished assessment tool. Perez, E. (1994). Phonological differences among speakers of Spanish-influenced English. In J. Bernthal & N. Bankson (Eds.), Child phonology: Characteristics, assessment, and intervention with special populations (pp. 245–254). New York: Thieme Medical. Romaine, S. (1995). Bilingualism (2nd ed.). Oxford, U.K.: Blackwell. Roseberry-McKibbin, C. (1995). Multicultural students with special language needs. Oceanside, CA: Academic Communication Associates. Schnitzer, M., & Krasinski, E. (1994). The development of segmental phonological production in a bilingual child. Journal of Child Language, 21, 585–622. Schnitzer, M., & Krasinski, E. (1996). The development of segmental phonological production in a bilingual child: A contrasting second case. Journal of Child Language, 23, 547–571. Shriberg, L. (1993). Four new speech and voice-prosody measures for genetics research and other studies in developmental phonological disorders. Journal of Speech-Language-Hearing Research, 36, 105–140. Shriberg, L., Austin, D., Lewis, B., McSweeney, J., & Wilson, D. (1997). The percentage of consonants correct (PCC) metric: Extensions and reliability data. Journal of Speech, Language, and Hearing Research, 40, 708–722. Shriberg, L., & Kwiatkowski, J. (1980). Natural process analysis. New York: John Wiley and Sons. Shriberg, L., & Kwiatkowski, J. (1982). Phonological disorders III: A procedure for assessing severity of involvement. Journal of Speech and Hearing Disorders, 47, 256–270. Shriberg, L. D., Kwiatkowski, J., Best, S., Hengst, J., & Terselic-Weber, B. (1986). Characteristics of children with phonologic disorders of unknown origin. Journal of Speech and Hearing Disorders, 51, 140–161. Smit, A. (1986). Ages of speech sound acquisition: Comparisons and critiques of several normative studies. Language, Speech, and Hearing Services in Schools, 17, 175–186.

Kayser, H. (1993). Hispanic cultures. In D. Battle (Ed.), Communication disorders in multicultural populations (pp. 114–157). Boston: Andover Medical.

Smit, A. (1993). Phonologic error distributions in the Iowa– Nebraska articulation norms project: Consonant singletons. Journal of Speech and Hearing Research, 36, 533–547.

Keshavarz, M., & Ingram, D. (2002). The early phonological development of a Farsi–English bilingual child. International Journal of Bilingualism, 6, 255–269.

Smit, A., Hand, L., Freilinger, J., Bernthal, J., & Bird, A. (1990). The Iowa articulation norms project and its Nebraska replication. Journal of Speech and Hearing Disorders, 55, 779–798.

MacDonald, M. (1989). The influence of Spanish phonology on the English spoken by United States Hispanics. In P. Bjarkman & R. Hammond (Eds.), American Spanish pronunciation: Theoretical and applied perspectives (pp. 215–236). Washington, DC: Georgetown University Press.

Stoel-Gammon, C., & Dunn, C. (1985). Normal and disordered phonology in children. Austin, TX: Pro-Ed. Vihman, M. (1996). Phonological development: The origins of language in the child. Oxford, U.K.: Blackwell.

Goldstein et al.: Phonological Skills in Bilingual and Predominantly Spanish and English Speakers

217

Vihman, M. (2004). Later phonological development. In J. Bernthal & N. Bankson (Eds.), Articulation and phonological disorders (5th ed., pp. 105–138). Boston: Allyn & Bacon. Vihman, M., & McLaughlin B. (1982). Bilingual and second language acquisition in preschool children. In C. Brainerd & M. Pressley (Eds.), Verbal processes in children: Progress in cognitive development research (pp. 35–58). New York: Springer Verlag. Watson, I. (1991). Phonological processing in two languages. In E. Bialystock (Ed.), Language processing in bilingual children (pp. 25–48). Cambridge, U.K.: Cambridge University Press. Wolfram, W. (1971). Overlapping influence and linguistic assimilation in second generation Puerto Rican English. Paper

218

L ANGUAGE, SPEECH,

AND

HEARING SERVICES

IN

presented at the meeting of the American Anthropological Association, New York. (ERIC Document Reproduction Service No. 057665)

Received January 6, 2004 Revision received April 26, 2004 Accepted September 8, 2004 DOI: 10.1044/0161-1461(2005/021) Contact author: Brian Goldstein, Department of Communication Sciences, 110 Weiss Hall, Temple University, Philadelphia, PA 19122. E-mail: [email protected]

SCHOOLS • Vol. 36 • 201–218 • July 2005