Spanish-English Articulation and Phonology of 4- and 5YearOld Preschool ChildrenAn Initial Investigation

1

Spanish-English Articulation and Phonology of 4- and 5-Year-Old Preschool Children

An Initial Investigation

Alejandro E. BriceUniversity of South Florida, St. Petersburg

Cecyle K. CarsonJennifer Dennis O’BrienUniversity of Central Florida

In this study, it was postulated that typically developing (i.e., normally developing without incidence of a speech or language delay or disorder) Spanish/Englishspeaking children ages 4 to 5 years old would show different articulation productions and phonological patterns in both languages. Sixteen participants from Florida were tested with Spanish and English articulation and phonology tests. For articulation, two manner or articulation comparisons were found to be significant (i.e., plosives and liquids/glides). In addition, two phonological patterns (i.e., stopping and velar fronting) were significantly different. Normative articulation and phonological Spanish and English data were obtained and should be useful for today’s public school speech-language pathologists. Further research should include normative data for bilingual children with articulation and/or phonological disorders to develop more appropriate treatments. In addition, it is recommended that other languages be investigated as the nation is also experiencing growth in languages beyond Spanish.

Keywords: Spanish-English articulation; Spanish-English phonology; bilingual phonology development; bilingual lan-guage development; preschool language

R ecent statistics indicate that 22% of the U.S. popu-lation under 5 years of age is Hispanic (Facts for

Features, 2008). Hispanics in the United States have shown a higher preschooler population than any other culturally and linguistically diverse group (U.S. Census Bureau, 2005). In Florida, more than 3.6 million resi-dents in 2008 were Hispanic (U.S. Census Bureau State and County Quickfacts, 2008). The rise of bilingual chil-dren, in particular, Spanish/English-speaking children (U.S. Census Bureau, 2005), supports the need for rele-vant bilingual developmental information.

Bilingual (i.e., Spanish-English) articulation and phono-logical development has apparent differences when com-pared to monolingual English (Goldstein & Cintron, 2001; Goldstein & Washington, 2001). Goldstein and Washington (2001) noted that

almost all data that exist on phonological development in preschool children come from monolingual speakers, in particular, monolingual English speakers. . . . Little

information exists concerning phonological development in bilingual speakers. It may be the case that phonological acquisition in bilingual speakers is different from that of monolingual speakers. (p. 154)

It also has been noted that there is an inclination to misdiagnose children as articulation and/or phonologi-cally disordered when comparing bilingual, in particu-lar, Spanish-English speakers, with monolingual English speakers (Goldstein, Fabiano, & Washington, 2005). When acquiring a second phonemic and phono-logical system (e.g., English), errors due to articulation

Communication Disorders Quarterly

Volume XX Number XMonth XXXX xx-xx

© 2009 Hammill Institute on Disabilities

10.1177/1525740108327447http://cdq.sagepub.com

hosted athttp://online.sagepub.com

Authors’ Note: This research was funded by an American Speech-Language-Hearing Association Office of Multicultural Affairs grant. In addition, the authors would like to recognize the contributions of Dr. Linda Louko who was initially involved with the project and the early data collection. Please address correspondence to Alejandro E. Brice, College of Education, University of South Florida, 140 Seventh Avenue South COQ 235, St. Petersburg, FL 33701-5016; e-mail: [email protected].

Communication Disorders Quarterly OnlineFirst, published on March 6, 2009 as doi:10.1177/1525740108327447

or phonological interference may occur (Goldstein et al., 2005). Consequently, the purpose of this study was to provide speech-language pathologists information on how Spanish-English articulation and phonology pro-ductions differ to assist in more accurate assessment and diagnoses.

Bilingual language acquisition is influenced by the rules of both languages, that is, at the articulation and phonological levels. Bilingual children may possess flu-ency in more than one language, for example, showing varying proficiency levels in one of their languages to complete fluency in both languages (Gutierrez-Clellen, 1999). McLaughlin (1995) defines bilingual and bilin-gualism as the ability to speak, listen, read, and/or write in more than one language with varying degrees of profi-ciency. The rules of the first language can positively influence the second language and result in articulatory and phonological transference. However, negative inter-actions may also occur, which can lead to articulation and/or phonological interference. Therefore, it is essential for speech-language pathologists to be knowledgeable of articulation (phonetic) and phonological (phonemic) production patterns of young Spanish-English, bilingual-speaking children and the influence that each language may have on the other.

Spanish phoneme acquisition and phonological pattern use have been researched to a moderate extent (Goldstein et al., 2005). In Spanish, development is generally the same across the different vernaculars; however, some variations differentiate the dialects within Spanish. Therefore, dialects should be considered changes from more dominant forms from which specific variations may be measured (Goldstein & Iglesias, 2001).

Cotton and Sharp (1988) stated that six major dialects of Spanish are spoken in the Americas: (a) Mexican; (b) Central American; (c) Caribbean; (d) Chilean, Paraguayan, Uruguayan; (e) Highlandian; and (f) Argentinean. Of these, the two most prevalent in the United States are the Mexican American dialect and the Caribbean dialect (e.g., Cuban, Puerto Rican; Goldstein & Iglesias, 1996a). Spanish dialects are characterized mainly through differ-ences in consonants involving fricatives, liquids, glides, and nasals (Goldstein & Iglesias, 1996b).

Phonetic Development in Spanish

English has 24 consonants, 17 syllabic nuclei, and 2 semivowels (i.e., , glides); Spanish has 19 consonants, 10 syllabic nuclei, and 2 semivowels. Consonant blends

involving /s/ in Spanish do not appear in initial syllable or word positions but often occur in English as seen in street (Goldstein et al., 2005). In addition, Spanish words end only in vowels or the following consonants /n, r, s, 1, d/. For further information regarding cross-linguistic Spanish and English phonetic comparisons, see Canfield (1981), Hammond (1989), Lafford and Salaberry (2003), Stockwell and Bowen (1965), and Whitley (2002).

Studies on phonetic development have been primarily limited to a few dialects, that is, studies examining the speech of Mexican or Mexican American children. For example, results from an earlier developmental study by Jimenez (1987) have been used to depict general Spanish development. The Jimenez study examined 120 Mexican American participants between the ages of 3 years 0 month and 5 years 7 months from south Texas who spoke Spanish as their first language. Results revealed that the children had acquired 50% of Spanish consonants by age 4, with the exception of: /d/, /g/, /s/, /ñ/, /x/, /s/, trilled /R/, and tap /r/. By age 5, only the /s/ and trilled /R/ were not mastered (i.e., produced accurately by 90% of the participants).

Spanish-English speakers produce speech differently from monolingual Spanish speakers (Grosjean, 1989). Goldstein and Washington (2001) stated, “There is some evidence that the phonological system of bilingual speakers develops somewhat differently from that of monolingual speakers of either language” (p. 154). It is unavoidable that Spanish-speaking children in the United States will come into contact with English. The likeli-hood of finding only monolingual Spanish speakers who have not had any English contact or influence from English is rare and unlikely. Therefore, it is well docu-mented English affects Spanish articulation and Spanish affects English articulation.

Spanish-English Phonology

There is very quick phonological development between ages 2 and 5 (Hodson & Edwards, 1997; Hodson & Paden, 1991). Universal phonological patterns in Spanish demonstrated by age 3 include: (a) cluster reduction, (b) liquid gliding, and (c) deletion of the stop while keeping the liquid. By age 4, most Spanish speakers have mas-tered most consonants and vowels (Goldstein & Iglesias, 1996b).

Martinez (1986) found the following phonological patterns in Mexican (Spanish-English speaking) 4-year-old children: (a) tap/trill /r/ deficiencies, (b) consonant

2 Communication Disorders Quarterly

sequence reductions, (c) deaffrication, (d) stopping, (e) affrication, (f) fronting, (g) assimilations, (h) palataliza-tion, (i) metathesis, (j) migration, (k) vowel deviations, and (1) sibilant distortions. In monolingual children Becker (1982) found the following after age 4 to be evi-dent among children who spoke Mexican Spanish: (a) consonant sequence reduction, (b) deaffrication, (c) non-phonemic deviations, and (e) tap/trill deficiencies.

Goldstein and Iglesias (1996b) stated that cluster reduction, especially involving deletion of the /1/, was the sole phonological pattern that had an occurrence rate greater than 10% (i.e., 15.2%) in the 3-year-old partici-pant group. The most frequently occurring pattern in the 4-year-old participant group was final consonant dele-tion (8.6%). Participants in the Goldstein and Iglesias study consisted of 54 Spanish-speaking children of Puerto Rican descent who were enrolled in a bilingual Head Start program (Spanish to English) and were between 3 (n = 24) and 4 (n = 30) years of age. Production data were gathered from the Assessment of Phonological Disabilities (Iglesias, 1978), an instrument designed spe-cifically for Spanish-speaking children.

Yavas and Goldstein (1998) found that syllable reduc-tion and/or deletion are frequently occurring processes among Spanish/English-speaking children. Spanish structure does not put emphasis on final-syllable conso-nants. Postvocalic omissions and syllable reduction are also heavily influenced by the Spanish dialect and should not be viewed as errors (Goldstein et al., 2005; Yavas & Goldstein, 1998).

Goldstein and Washington (2001) examined the pho-nological patterns of 12 typically developing (no appar-ent speech, language, or hearing delays or disorders) Spanish/English-speaking 4-year-old children. Their participants were recruited from the Camden, New Jersey area. The participants spoke primarily Puerto Rican Spanish (n = 19), 2 spoke Dominican Spanish, and 1 spoke Nicaraguan Spanish. A discrepancy between Spanish and English fricatives and affricates were found (i.e., greater than a 10% difference). The Spanish tap/flap and trill consonants were produced with less than 80% accuracy. Phonological processes that differed (i.e., greater than 5 percentage points) between Spanish and English included cluster reduction, stopping, and spiran-tization. Goldstein and Washington also found that instances of Spanish influenced English and English influenced Spanish (which we refer to as instances of interference in this article). Of particular interest was the influence of Spanish on English as these productions are what English-speaking speech-language pathologists

may interpret to be as articulation errors. Final consonant deletion, Spanish tap to English liquid substitutions, and deaffrication were influence errors noted. In addition, Goldstein and Washington found that their participants substituted /v/ for /b/. It should be noted that phonetically a /v/ does not exist in Spanish, but rather the bilabial fricative /β/ is produced for the letter v and that a substi-tution of the biliabial fricative /β/ for the Spanish plosive /b/ is an acceptable substitution without change in word meaning (Canfield, 1981; Cotton & Sharp, 1988; Hammond, 1989; Whitley, 2002). Therefore, it should be stated that the biliabial fricative /β/ is commonly substituted by the /b/ plosive and may be perceived by the non-Spanish speaker as constituting a /v ≥ b/ substitution.

Statement of the Problem

Limited information exists for Spanish-English artic-ulation and phonology development in young bilingual children living in the United States. Some studies have documented Spanish phonetic and phonological devel-opment in young children who are either monolingual or bilingual speakers of the Southwestern Mexican dialect (e.g., Acevedo, 1989; Becker, 1982; Gildersleeve-Neuman, Kester, Davis, & Peña, 2008; Martinez, 1986). The study by Gildersleeve et al. (2008) appears limited because they examined phonological development of 3-year-old children. Phonological development is still incomplete at that age. Future studies need to examine phonological development of 4-year-old children. In addition, Gildersleeve et al. examined bilingual children’s phonol-ogy in English only. Goldstein and Washington (2001) stated, “Thus, there is a need to describe both the English and the Spanish phonological skills of bilingual chil-dren” (p. 154). In addition, most studies have focused on a single dialect of Spanish, that is, the Mexican or Mexican American Spanish dialect. The Caribbean dia-lect, that is, Puerto Rican (Goldstein & Iglesias, 1996b) or Cuban dialects, or other dialects of Spanish have not been extensively studied. The need for typical develop-mental articulation and phonological Spanish informa-tion with regard to the various dialects of Spanish spoken in the United States is vital for appropriate assessment and subsequent diagnosis of young bilingual, Spanish/English-speaking children.

Hypotheses

It was postulated that typically developing (i.e., normally developing without incidence of a speech or language delay or disorder), bilingual, Spanish/English-speaking children of various Spanish dialects, ages 4 to 5, would show

Brice et al. / Spanish-English Articulation and Phonology 3

different articulation productions and/or phonological patterns in English and Spanish. These children were measured with English and Spanish articulation/phonol-ogy tests: the Goldman–Fristoe Test of Articulation 2 (GFTA-2; Goldman & Fristoe, 2000), the Khan–Lewis Phonological Analysis–Second Edition (KLPA-2; Khan & Lewis, 2002), and a Spanish articulation/phonology research instrument, the Comprehensive Assessment of Spanish Articulation–Phonology (CASA-P) (Brice, Carson, & O’Brien, 2008).

The hypotheses for this study stated that differences would occur for the 16 bilingual children (i.e., of various Spanish dialects) for manner of English and Spanish articulation (i.e., plosives, fricatives/affricates, stridents, liquids/glides, and nasals). Articulation was assessed by percentage of consonant correct (PCC). Phonological patterns or processes in English versus Spanish for the 16 bilingual children (i.e., of various Spanish dialects) were assessed by percentage of occurrence (POC) errors.

The dependent variables consisted of: (a) the PCC manner productions for Spanish and English (i.e., plosives, fricatives/affricates, stridents, liquids/glides, and nasals) and (b) the phonological pattern POC errors in Spanish and English. The independent variables con-sisted of the two languages (Spanish and English) for intragroup comparisons.

Method

Participants

A total of 16 bilingual (Spanish/English-speaking) 4- to 5-year olds (eight 4-year-olds and eight 5-year-olds) par-ticipated in this study. The children were enrolled in either a Head Start or preschool program. Two children were removed from the data collection and study because they failed the speech and/or language screening. Possible articulation and phonological difficulties may have been present in the excluded participants. The children in the study consisted of 10 males and 6 females. The children attended a preschool in central Florida, a Head Start pro-gram also in central Florida, and a preschool in the south Florida area. The participants’ and/or their parents’ country of origin and language consisted of the following: (a) Puerto Rico (n = 9), (b) Nicaragua (n = 2), (c) Cuba (n = 1), (d) Bolivia (n = 1), (e) Argentina (n = 1), (f) Columbia (n = 1), and (g) Dominican Republic (n = 1). In our review of the literature we are unaware of other studies investigating as diverse a population as this. Goldstein and Washington (2001) primarily used Spanish speakers of a Puerto Rican dialect (9 of their 12 participants). They noted this to be a

limitation of their study. Therefore, it can be stated that a diverse and varied Spanish dialect was spoken by the par-ticipants, which should be viewed as a strength of this study.

Participants were selected based on the following criteria: (a) the children spoke Spanish and English fluently; (b) the children did not possess any speech, language, or hearing difficulties such as delays or disor-ders; and (c) the children were between 4 and 5 years of age. All children included in this study met these criteria (n = 16). Participants were also selected randomly from all children who met the criteria at these sites.

Information regarding language use was obtained via parent report. All the children in the study spoke Spanish and English either at home or the school and had spoken both languages since birth (i.e., simultaneous bilinguals). Parent report indicated that all children were, however, predominantly exposed to Spanish (i.e., greater than 50% or a majority of the time). All the children enrolled in the study were typically developing with no previous history of a speech or language delay or a hearing disor-der. A speech and language screening was also adminis-tered to decrease the possibility of the children in the sample having a speech, language, or hearing disorder.

Fluency in both Spanish and English was a prerequisite for the participants and was determined from an oral language speech sample gathered before each data-collec-tion session. The children were rated by the first researcher (a high-level fluent speaker of Spanish and English) and determined to have a score of 4 or higher on a Likert scale using the International Second Language Proficiency Ratings (ISLPR; Wylie & Ingram, 1999). The ISLPR consists of a Likert-type scale from 0 to 5 (low to high abilities) with plus interval ratings (i.e., 0, 1, 1+, 2, 2+, 3, 3+, 4, 4+, 5). A Level 4 score is considered to indicate fluent language abilities (Wylie & Ingram, 1999). All 16 participants scored at least a 4 in both Spanish and English. The participants were all simulta-neous learners of Spanish and English. According to Haynes (2008), these participants would be classified as Stage 4 bilinguals or showing intermediate fluency.

The ISLPR is a well-established proficiency scale involving the macro-language skills of listening, speak-ing, reading, and writing. Only Spanish/English-speaking skills were evaluated for this study. There are two models of the ISLPR: the general proficiency model and the specified purpose model (e.g., English for court of law purposes). Because the participants were rated for over-all speaking proficiency in both Spanish and English, the general model was chosen. Only the overall speaking scores were obtained for this study.

English speech and language development was infor-mally assessed by the preschool teachers and the

4 Communication Disorders Quarterly

researchers in conversing with each child and was deter-mined to be within the normal range. In addition, the parents all indicated that their child’s speech and language devel-opment in both languages was developmentally age appro-priate. Because Spanish was not spoken in the preschool environments, the Structured Photographic Expressive Language Test–Preschool, Spanish (Spanish SPELT-P; Werner & Kresheck, 1989) was administered as a formal Spanish language screening. The Spanish SPELT-P is a measure of Spanish syntax including age-appropriate vocabulary. All the children passed the Spanish SPELT-P screening items (i.e., a majority or half of the items). The researchers informally judged a conversational speech and language sample and informally and formally scored (i.e., use of the ISLPR and Spanish SPELT-P) each participant’s articulation, phonological, semantic, and syntactic abili-ties. No problematic behaviors in the classroom, at home, and/or through the screenings were noted.

All the participants passed a hearing screening con-sisting of pure tones at 500, 1,000, 2,000, 4,000, and 6,000 at 25 dB. In addition, an oral-peripheral exam was given to rule out any oral structure or function abnor-malities that could have interfered with articulation and/or phonological productions. All the children passed the English and Spanish speech and language screenings, hearing screening, and oral-peripheral exam.

Procedures

A Spanish articulation and phonology test, the CASA-P, was given to the group of 4- and 5-year-old Spanish/English-speaking participants. Please see the appendix for a list of the words in the CASA-P. The chil-dren were assessed in a quiet room. Instructions were provided in both Spanish and English by the first author. English articulation was assessed by administration of the GFTA-2 (Goldman & Fristoe, 2000) and subsequent phonological analysis with the KLPA-2 (Khan & Lewis, 2002). All sessions were audio and video recorded with a Panasonic Super VHS 456 Pro-Line video camera.

Development of the CASA-P

The CASA-P is a research tool created by two univer-sity researchers that assesses Spanish articulation and phonology. The first researcher is proficient in Spanish and knowledgeable about Spanish-English articulation and phonology. The second researcher is knowledgeable about English articulation and phonology.

The CASA-P assesses all Spanish initial and final consonants. It was created to be an age-appropriate

preschool assessment tool (i.e., 3 to 5 years of age). Words were chosen to elicit all initial and final sounds in Spanish so that 10 or more occurrences for each phono-logical process were elicited during the articulation/ phonological assessment. Words were also chosen to be age appropriate. The CASA-P was created to allow com-parisons between Spanish and English articulation and phonology. Establishing content validity for the CASA-P was achieved by the following means.

1. A comprehensive and systematic review of the literature of Spanish articulation and phonology was conducted. The literature (Acevedo, 1989; Brice, 1996; Eblen, 1982; Goldstein et al., 2005; Goldstein & Iglesias, 1996a, 1996b; Jimenez, 1987; Mann & Hodson, 1994; Martinez, 1986; Paulson, 1989; Rivera-Umpiere, 1988; Villanueva, 1990) revealed that the following 10 phonologi-cal patterns were highly occurring in Spanish phonology: (a) final consonant deletion, (b) palatal fronting, (c) velar front-ing, (d) stopping, (e) liquid simplification, (f) assimilation, (g) syllable reduction, (h) cluster simplification, (i) prevo-calic singleton omission, (j) stridency, and (k) tap/trill defi-ciency. Seven processes are common to both Spanish and English (Acevedo, 1989; Brice, 1996; Eblen, 1982; Goldstein & Iglesias, 1996b; Jimenez, 1987; Martinez, 1986; Paulson, 1989; Rivera-Umpiere, 1988; Villanueva, 1990; Yavas & Goldstein, 1998). As a result, 7 concurrent patterns were analyzed between Spanish and English: (a) final consonant deletion, (b) palatal fronting, (c) velar fronting, (d) stopping, (e) syllable reduction, (f) liquid simplification, and (g) cluster simplification. See Table 1 for a summary of the phonological patterns assessed by the CASA-P.

2. Five native speakers of Spanish reviewed the items for standard Spanish pronunciation and age appropriate-ness. The reviewers consisted of two Cuban Spanish speakers, one Puerto Rican Spanish speaker, one Costa Rican Spanish speaker, and one Peruvian Spanish speaker. All bilingual speakers transcribed the stimuli using narrow and broad transcription over six meetings (approximately 1-hr session). One hundred percent con-sensus or agreement was achieved during these sessions regarding the transcriptions.

Two CASA-P versions were created so that order effect would not affect the presentation of the stimuli. Both versions contained the same stimuli, differing only in the presentation order. Form A was given to 11 par-ticipants; Form B was given to 5 participants. The authors had expected to administer Form B to more par-ticipants, hence, the slight discrepancy in the number given to each group. Because these were equivalent forms it was not expected that this minor discrepancy would affect the overall results.

Brice et al. / Spanish-English Articulation and Phonology 5

The CASA-P was administered to all the participants by a fluent Spanish-English speaker (i.e., either by a Spanish/English-speaking graduate student or by the bilingual, Spanish/English-speaking researcher). Each participant was shown the stimulus picture and was asked, “What is this?” in Spanish (i.e.,“Que es esto?”). The children were to name each picture. If the participant did not answer, a delayed imitation response was elicited. Delayed imitation was elic-ited by naming the item, showing the next item, and return-ing to the unnamed item for another elicitation. Direct imitation was accepted only if all other attempts were unsuccessful. The researcher transcribed all errors on indi-vidual response forms.

Reliability: Intrarater and Interrater

Item coding and transcription training exercises were conducted between the student researcher and the bilingual faculty researcher. A 100% training agreement was achieved by transcribing 20% of the entire data sample (i.e., all items for all participants). This took three sessions lasting approximately 3 hrs. Therefore, consensus was achieved for both transcription and coding of the data.

Intrarater reliability was established by reanalyzing a different 10% of the data. A high 94% intrarater reliability

agreement was achieved by the student researcher. In addition, the student and faculty researchers analyzed a different 10% of the data to establish interrater reliability. Each researcher separately scored the same participants, achieving an interrater reliability consensus of 84.5%.

Equipment

A Panasonic Super VHS 456 Pro-Line video camera was used to obtain the sample (i.e., audio and verbal responses). The camera microphone has a frequency response of 50–20,000 Hz with a 47dB signal-to-noise ratio. Camera placement was approximately 3 to 6 ft from the children.

Data Analysis

Means and standard deviations were calculated for articulation manner production and for phonological pat-terns in both Spanish and English. Only initial and final phonemes found in both Spanish and English were ana-lyzed according to manner of production. Comparisons were analyzed for within-group differences (i.e., Spanish vs. English phonetic and phonemic productions for the one group of bilingual children); consequently, separate dependent t tests were performed for the Spanish versus

6 Communication Disorders Quarterly

Table 1 Phonological Patterns Assessed by the Comprehensive Assessment of Spanish Articulation–Phonology

Goldstein & Mann & Rivera- Yavas & Phonological Acevedo Brice Eblen Iglesias Jimenez Hodson Martinez Paulson Umpiere Villanueva Walters Goldstein Patterns (1989) (1996) (1982) (1996b) (1987) (1994) (1986) (1989) (1988) (1990) (2000) (1998)

Final X X X X X X X consonant deletionFronting X X X XStopping X X X XLiquid simplification X X X X XAssimilation X X XCluster simplification X X X X X X X X X XSyllable reduction X X X X XPrevocalic singleton omission X X X X X Tap/trill deficiency X X X X X X X Stridency deficiencies X X X X X X X X X

English articulation manner and phonological pattern comparisons. Alpha level was set at p < .05. The Bonferroni correction was not used as the data were not analyzed simultaneously. Therefore, no alpha-level adjustment was necessary (Weisstein, 2008).

Results

Independent and Dependent Variables

The independent variables consisted of the GFTA-2 (Goldman & Fristoe, 2000) and the CASA-P. The depen-dent variables consisted of results from: (a) the GFTA-2, (b) the KLPA-2 (Khan & Lewis, 2002), and (c) the CASA-P. PCC was analyzed for all articulation out-comes in Spanish and English; POC errors were ana-lyzed for all phonological patterns in both languages.

Paired T Tests

Prior expectation that significant results would occur between the two languages justified the use of a one-tailed test (alpha set at p < .05). Comparisons among Spanish and English manner consonant production (i.e., plosives, fricatives/affricates, stridents, liquids/glides, and nasals) indicated that only two manner productions (plosives and liquids/glides) yielded significant differences: plosives, t(15) = 1.861, p = .0415, and liquids/glides, t(15) = 2.076, p = .0275. Effect size (Cohen’s d) for plosives resulted in an effect size of r2 = .312, demonstrating a shared variance of 31.2%. Cohen’s (1988) criteria (i.e., small = .10–.29, medium = .30–.49, large >.50) indicated this effect size to be medium. Effect size for liquids/glides resulted in r2 = .338, indicating a shared variance of 33.8%, or a medium effect (see Table 2).

Seven phonological patterns found in both languages were analyzed: (a) final consonant deletion, (b) stopping, (c) velar fronting, (d) palatal fronting, (e) liquid simplifi-cation, (f) cluster simplification, and (g) syllable reduction. Two phonological processes between the two languages were statistically significant: stopping, t(15) = –6.526,

p = .000, and velar fronting, t(15) = –2.355, p = .033]. Effect size for stopping resulted in an r2 of 0.57 with a shared variance of 57%. This effect size was large (Cohen, 1988). Effect size for velar fronting indicated an r2 of .147 and a shared variance of 14.7% (i.e., a small effect size). All of the other phonological pattern com-parisons were found to be nonsignificant (see Table 3).

Descriptive Analyses: Spanish and English

Evaluating the means for the individual sounds in Spanish and English revealed that differences in means occurred only for the Spanish and English initial sounds of /b, k, j, w/.

The English /b/ was produced with 100% accuracy compared with 94% accuracy in Spanish, and English /k/ was said with 94% accuracy with 88% accuracy in Spanish. Descriptive data indicated that among the liq-uids and glides, the /j/ phoneme was articulated with 94% accuracy in Spanish and 69% accuracy in English. The production of the phoneme /w/ was articulated 100% correct in English and was produced correctly with 94% accuracy in Spanish. Three of the four differ-ences indicated more accurate English articulation pro-ductions than Spanish productions.

Final consonants that were analyzed for both Spanish and English included /d, 1, s, n/. Analysis revealed that the final Spanish /d/ resulted in a 63% PCC when compared with the English final /d/, which resulted in 100% PCC. The production of the /1/ phoneme was 100% correct in Spanish and 81% correct in English. The final /s/ was 63% correct in Spanish and only 88% correct in English. Final /n/ was less accurate in English (81%) than Spanish (100%). Production accuracy seemed to be equal in Spanish and English for common final sounds (see Table 4).

Descriptive Analyses: English Only

Some sounds occur only in English. Therefore, descriptive analyses of English document how Spanish has influenced English productions. Twenty English initial

Brice et al. / Spanish-English Articulation and Phonology 7

Table 2 Spanish and English Manner Consonant Percentage of Occurrence Comparisons

Percentage of Occurrence Manner Mean SD T (one-tailed paired samples t test) df p

Plosives .3750 .8062 1.861 15 .0415Liquids/glides .3125 .6020 2.076 15 .0275

and final sounds were also descriptively analyzed. They consisted of: (a) initial sounds of /d/, /J/, /θ/, /ð/ and (b) final sounds of /p, b, t, k, g, f, v, z, m, t∫, ∫, θ, ð, ∂, d, J, η/. Analysis revealed that the initial voiceless and voiced /θ, ð/ were articulated with less than 75% accuracy. The final consonant English sounds of /b, v, t∫, ∫/ were said with less than 75% accuracy.

Interference Errors

Interference errors consist of production deviations from what is expected based on the structure of the first language imposed onto the structure of the second language, i.e., in articulation, phonology in this study (McLaughlin, 1985). These errors were determined through careful cross-linguis-tic analysis, that is, investigating how Spanish phonetic and phonemic productions would be produced in English given the natural differences between the two languages (Brice & Brice, 2009). Interference errors are still judgments made by the researcher/clinician/teacher. Currently, there are no

means for establishing reliability measures for these types of errors. Only the first author has knowledge of Spanish articulation and phonology. Therefore, an interrater reliabil-ity agreement of interference errors with the other authors was not possible. The authors are aware of this as a limita-tion of the study.

Five of the 16 children exhibited Spanish interference errors (i.e., errors ranged from 2% to 8%). Nine children showed deaffrication errors (i.e., ranging 33% to 67%), for example, /t∫/ ≥ /∫/ or /dJ/ ≥ /j/. Three children exhib-ited initial devoicing errors (i.e., ranging from 0% to 4%), for example, /θ/ ≥ /t/. Eight children exhibited final devoicing errors (i.e., ranging from 3% to 9%), for example, /t∫/ ≥ /t/.

Discussion

The purpose of this study was to compare Spanish and English articulation and phonology in a group of normally

8 Communication Disorders Quarterly

Table 3 Means, Standard Deviations, and T-Test Comparisons for Spanish and English

Phonological Process Comparisons

Percentage Consonant Correct Mean SD Phonological Process (Spanish, English) (Spanish, English) t df p

Final consonant deletion 1.750 1.9493 –0.466 15 .648 2.250 4.9598 Velar fronting 0.0000 0.0000 –2.355 15 .033* 2.312 3.928 Palatal fronting 0.0000 0.0000 –1.000 15 .333 0.6875 2.750 Stopping 0.0000 0.0000 –6.526 15 .000* 8.312 5.095 Liquid simplification 1.750 3.255 0.000 15 1.000 1.750 4.123 Cluster simplification 1.750 3.255 0.000 15 1.000 1.750 4.123 Syllable reduction 1.750 5.9721 0.975 15 .345 0.3750 1.0878

*p < .05.

Table 4 Spanish and English Individual Phoneme–Sound Mean Comparisons

p b t d k g 1 (initial) j w 1 (final)

Spanish 1.000 .9375 1.000 1.000 .875 1.000 1.000 .938 .9375 1.000English 1.000 1.000 1.000 1.000 .937 1.000 1.000 .688 1.000 .8125

developing simultaneous bilingual children. The children were between 4 and 5 years of age. It was expected that there would be manner of articulation differences for the bilingual children in Spanish versus English. Two manner consonant productions were found to be significantly differ-ent: plosives and liquids/glides. English sounds were gener-ally more accurate. It is likely that the children in this study were exposed to more English than was initially expected. Another possibility may have been that the children had developed greater accuracy in English because English was acquired at a later age, being developmentally more accurate in English. These results merit further investigation.

It was expected that there would be phonological pat-tern differences for the Spanish/English-speaking group of children as measured by PCC errors in Spanish versus English. Two phonological patterns were found to be significantly different: stopping and velar fronting. It seemed that stopping was more complicated in English (9% occurrence) than Spanish (0% occurrence) as more fricatives are found in English that can subsequently be stopped. In Spanish only one fricative can be stopped. The fronting pattern occurred 2% in English and 0% in Spanish. Spanish is generally more fronted than English (i.e., corresponding sounds produced more anteriorly; Brice, 2002), which could have affected these results.

Descriptively, the /b/ phonemes showed different mean productions between Spanish and English. It is known that Spanish and English differ with regard to when these phonemes are voiced, that is, voice onset time (Brice, Castellon-Perez, & Ryalls, 2004; Flege, Munro, & MacKay, 1995; Zampini, 1998). Therefore, it is possible that voicing interference occurred when say-ing this plosive sound in Spanish and English. Further investigation is warranted.

The descriptive data indicated that the initial /j/ pho-neme was produced with 94% accuracy in Spanish and with only 69% accuracy in English. It appeared that the children were not transferring from Spanish to English. One possibility for this could have been that initial Spanish consonants are usually tense, resulting in phone-mic interference. For example, some children may have made the articulation of this phoneme more complicated (i.e., more tense) as seen in the example of substituting the glide in yellow for the affricate Jello.

The /w/ phoneme was produced with slightly more accurate productions in English than in Spanish. According to Nash (1977), the English /w/ has a higher frequency of occurrence (4%) rate than the Spanish /w/ (2%). This slightly higher use of the /w/ sound in English may have contributed to slightly better articulation.

The final Spanish /d/ phoneme was produced with 63% accuracy whereas English production was 100%

correct. The children seemed to be stressing the /d/ final consonant in English while in some situations deleting it in Spanish. Deletion of final consonants is typical in some Spanish dialects. The word mitad said as mita serves as an example (Goldstein & Iglesias, 1996b).

The production of the final phoneme /1/ was more accurate in Spanish and less accurate in English. It appears that the English /1/ is developmentally difficult for the Spanish-speaking children to produce. Quite a few of the children vowelized the phoneme. For example, shovel was articulated as shovo. The Spanish-speaking children did not appear to generalize correct production to English. The final /1/ in Spanish is more frequently occurring than the English /1/ (Canfield, 1981); there-fore, the children in this study may have not yet gen-eralized from a frequently occurring sound in one language to the same less occurring sound in the other language.

Final /s/ production was 63% accurate in Spanish and 88% accurate in English. Production of the /s/ phoneme in both Spanish and English was aspirated and/or inter-dentalized. This could have been due to the complex nature of the /s/ phoneme. This phoneme is a later occur-ring sound and is not developmentally age appropriate to produce in either language. In addition, it is known that in Caribbean dialects of Spanish that final consonant dele-tion, and particularly /s/ deletion, is a common dialectal feature (Goldstein & Iglesias, 1996b). Therefore, Spanish dialectal interference from Spanish to English may have also occurred.

Final /n/ was less accurate in English, that is, 81% accuracy compared to 100% accurate in Spanish. Imperfect auditory phoneme perception or simplification could have occurred. Some participants changed the /n/ to an /m/, that is, gun was articulated as gum.

Initial English voiceless /θ/ and voiced /ð/ were artic-ulated with less than 75% accuracy, indicating difficulty producing these sounds as they are not in the Spanish language consonant inventory. Targeting the voiceless /θ/ and voiced /ð/ phonemes resulted in either substitutions, deletions, and/or distortions.

Final consonant English /b, v, t∫, ∫/ were articulated with less than 75% accuracy. The /b, t/ are common to both languages; however, the English /b/ and Spanish /p/ display common and overlapping voice onset time fea-tures (Flege et al., 1995; Zampini, 1998). The Spanish /t/ also tends to be more anteriorly produced than the English /t/ (Brice & Brice, 2009). Therefore, these chil-dren could have experienced voicing and placement interference errors between Spanish and English. The English sounds of /∫, v/ also do not occur in Spanish. Less accurate production could have been been affected

Brice et al. / Spanish-English Articulation and Phonology 9

by having to learn a new set of sounds; hence, these errors can be attributed to Spanish language interference.

Five participants demonstrated second language phone-mic interference errors, 9 showed deaffrication errors, 3 demonstrated initial devoicing errors, and 8 showed final devoicing errors. These patterns are all examples of interfer-ence errors between Spanish and English.

This study is similar to the study by Goldstein and Washington (2001) in terms of: (a) selecting a sample size (16 vs. 12 participants), (b) gathering data from bilingual 4-year-olds, and (c) using articulation and pho-nology data. Some of the results from this study support those of Goldstein and Washington. For example, both studies found that the bilingual students had difficulties with the tap/trill sounds (noted as liquids/glides in this study). Stopping was a process that the 4-year-old chil-dren still exhibited. With regard to the Spanish influ-ences on English (described here as interference errors) it should be noted that children in both studies deaffri-cated the /tJ/ sound, resulting in /∫/. Some of the children in this study tended to delete final consonant sounds as did children in the Goldstein and Washington study. However, there were also differences between the two studies. Children in this study had difficulty with plo-sives, whereas those in the Goldstein and Washington study had more difficulty with fricatives and affricates. Velar fronting was noted in this study, whereas cluster reduction and spirantization were noted in the Goldstein and Washington study. Differences in geographic sam-ples, use of different phonological instruments, and dif-ferent dialects spoken by the children all may have contributed to the different results. These findings illus-trate that further research is needed to better establish a more robust profile of Spanish-English articulation and phonology among 4- to 5-year-old children.

The articulation and phonological production described here have illustrated that Spanish/English-speaking chil-dren demonstrate different types of articulation and pho-nology errors than do monolingual children. It is imperative that speech-language pathologists obtain information from bilingual children speaking a particular dialect of Spanish that corresponds to the child they are assessing. Some articulation sounds transfer between languages, and some phonological rules transfer between Spanish and English. However, it should be apparent that not all aspects transfer. By examining typically develop-ing children, readers are exposed to what is expected in bilingual children.

One of the limitations of this study is that a small sample of children was gathered. The data were also limited in that only single-word productions were elic-ited; therefore, conversational intelligibility conclusions

cannot be drawn from the data. The study was limited by the geographic location of the participants (i.e., one state), yet two large metropolitan cities served as loca-tions for data gathering. The study was also limited by not having a more standard measure of articulatory/pho-nological interference. Although, several dialects were obtained in the data sample, it is insufficient to make generalizations regarding each dialect. Also, although the parents reported more Spanish than English being spoken in the home, the influence of English in the pre-school and day care environments is unknown. These factors should be investigated in future studies. With all of these limitations, clinicians are warned not to make too sweeping of generalizations regarding the results and conclusions drawn from this study. Including more chil-dren speaking different dialects of Spanish and including samples from different geographic regions of the United States need to be completed.

In conclusion, normative articulation and phonological information in Spanish is useful for today’s public school speech-language pathologists. The speech-language pathol-ogist should be able to distinguish language differences versus articulation and phonological disordered productions in bilingual, Spanish/English-speaking children. It is antici-pated that this study has provided some developmental Spanish-English articulation and phonological information that will be useful for clinicians in making diagnoses and providing appropriate intervention.

It is recommended that further research include Spanish/English-speaking participants with articulation and/or phonological disorders because the Hispanic population is dramatically increasing in the nation. In addition, future studies should incorporate larger sample sizes to obtain more accurate representations of the Spanish/English-speaking population. It is also recom-mended that other languages be investigated as the nation is seeing growth in the use of other languages.

10 Communication Disorders Quarterly

Appendix Comprehensive Assessment of Spanish Articulation–Phonology List of Words

Words Sound Articulation English Meaning

Planta /pl/ PlantPuerta /p/ and /a/ DoorNadar /n/ SwimCuchara /k/ SpoonElla /e/ SheTigre /t/ TigerManí /i/ PeanutChicle /t∫/ GumClub /kl/ Club

(continued)

References

Acevedo, M. A. (1989, November) Typical Spanish misarticulations of Mexican-American preschoolers. Paper presented at the annual meeting of the American Speech-Language-Hearing Association, St. Louis, MO.

Becker, M. C. (1982). Phonological analysis of speech samples of monolingual Mexican 4-year-olds. Unpublished master’s thesis, San Diego State University, San Diego, CA.

Brice, A. (1996). Spanish phonology: A review of the literature. Florida Journal of Communication Disorders, 16, 14–17.

Brice, A., Carson, C., & O’Brien, J. D. (2008). Comprehensive Assessment of Spanish Articulation-Phonology (CASA-P). Unpublished assessment tool, University of South Florida St. Petersburg, St. Petersburg, FL.

Brice, A. (2002). The Spanish language. In A. E. Brice (Ed.), The Hispanic child: Speech, language, culture and education (pp. 21–29). Boston: Allyn & Bacon.

Brice, A., & Brice, R. (Eds.). (2009). Language development: Monolingual and bilingual acquisition. Upper Saddle River, NJ: Prentice Hall.

Brice, A., Castellon-Perez, Y., & Ryalls, J. (2004). Speech recogni-tion of code switched words by proficient Spanish-English bilin-guals. Journal of Distinguished Language Studies, 2, 13–22.

Canfield, D. L. (1981). Spanish pronunciation in the Americas. Chicago: University of Chicago Press.

Cohen, J. (1988). Statistical power analysis for the behavioral sci-ences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

Cotton, F., & Sharp, J. (1988). Spanish in the Americas. Washington, DC: Georgetown University Press.

Eblen, R. E. (1982). Some observations on the phonological assess-ment of Hispanic-American children. Journal of the National Student Speech, Language, Hearing Association, 10, 44–54.

Facts for Features. (2008). Hispanic heritage month: September 15–October 15, 2006. Retrieved September 3, 2008, from http:// www.census.gov/Press-Release/www/releases/archives/facts_ for_features_special_editions/007173.html

Flege, J., Munro, M., & MacKay, R. (1995). Factors affecting strength of perceived foreign accent in a second language. Journal of the Acoustical Society of America, 97, 3125–3134.

Gildersleeve-Neuman, C. E., Kester, E. S., Davis, B. L., & Peña, E. (2008). English speech sound development in preschool-aged children from bilingual English-Spanish environments. Language, Speech, and Hearing Services in Schools, 39, 314–328.

Goldman, R., & Fristoe, M. (2000). Goldman–Fristoe Test of Articulation 2. Circle Pines, MN: American Guidance Service.

Goldstein, B., & Cintron, P. (2001). An investigation of phonological skills in Puerto Rican Spanish-speaking 2-year-olds. Clinical Linguistics and Phonetics, 15, 343–361.

Goldstein, B., Fabiano, L., & Washington, P. (2005). Phonological skills in predominantly English, predominantly Spanish, and Spanish-English bilingual children. Language, Speech, and Hearing Services in Schools, 36, 201–218.

Goldstein, B., & Iglesias, A. (1996a). Phonological patterns in Puerto Rican Spanish-speaking children with phonological disorders. Journal of Communication Disorders, 29, 367–387.

Goldstein, B., & Iglesias, A. (1996b). Phonological patterns in normally developing Spanish-speaking 3- and 4-year-olds of Puerto Rican descent. Language, Speech, and Hearing Services in Schools, 27, 82–89.

Goldstein, B., & Iglesias, A. (2001). The effect of dialect on the analysis of phonological patterns in Spanish-speaking chil-dren. American Journal of Speech-Language Pathology, 10, 394–406.

Goldstein, B., & Washington, P. S. (2001). An initial investigation of phonological patterns in typically developing 4-year-old Spanish-English bilingual children. Language, Speech, and Hearing Services in Schools, 32, 153–164.

Grosjean, F. (1989). Neurolinguists, beware! The bilingual is not two monolinguals in one person. Brain and Language, 36, 30–15.

Gutierez-Clellen, V. (1999). Language choice in intervention with bilingual children. American Journal of Speech-Language Pathology, 8, 291–302.

Hammond, R. M. (1989). American Spanish dialectology and phonology from current theoretical perspectives. In P. C. Bjarkman & R. H. Hammond (Eds.), American Spanish pronunciation (pp. 137–169). Washington, DC: Georgetown University Press.

Haynes, J. (2008). Stages of second language acquisition. Retrieved August 5, 2008, from http://www.everythingesl.net/inservices/ language_stages.php

Hodson, B., & Edwards, M. (1997). Perspectives in applied phonology. Gaithersburg, MD: Aspen.

Brice et al. / Spanish-English Articulation and Phonology 11

Appendix (continued)Words Sound Articulation English Meaning

Ventana /v/ WindowFlecha /fl/ ArrowIglesia /i/ ChurchBloque /bl/ BlockMuneca /nj/ DollLlave /j/ KeyBoca /b/ MouthOlla /o/ PanLago /1/ LakeHueso /h/ BoneMariposa /m/ ButterflyFlor /R/ FlowerFresa /fr/ and /s/ StrawberryGlobo /gl/ GlobeMitad /d/ HalfDinero /o/ MoneySobre /s/ Over, envelopeBrazo /br/ ArmFeliz /f/ HappyQuince /e/ FifteenGallo /g/ RoosterLapiz /s/ PencilBicicleta /a/ BicycleSofa /s/ and /f/ SofaHuevo /w/ EggTrompeta /tr/ TrumpetHierba /j/ GrassAndar /a/ Walk, goGigante /h/ GiantAprisa /h/ FastDendtadura /d/ TeethSopa /s/ SoupVaso /β/ GlassCruzar /kr/ To cross

Hodson, B., & Paden, E. (1991). Targeting intelligible speech. A phono-logical approach to remediation (2nd ed.). Austin, TX: Pro-Ed.

Iglesias, A. (1978). Assessment of phonological disabilities. Unpublished assessment tool, Ohio State University, Columbus, OH.

Jimenez, B. C. (1987). Acquisition of Spanish consonants in children aged 3–5 years, 7 months. Language, Speech, and Hearing Services in Schools, 18, 357–363.

Khan, L., & Lewis, N. (2002). The Khan–Lewis Phonological Analysis (2nd ed.). Circle Pines, MN: American Guidance Service.

Lafford, B. A., & Salaberry, R. (2003). Spanish second language acquisition: State of the science. Washington, DC: Georgetown University Press.

Mann, D., & Hodson, B. (1994). Spanish-speaking children’s pho-nologies: Assessment and remediation of disorders. Seminars in Speech and Language, 15(2), 137–147.

Martinez, R. (1986). Phonological analysis of Spanish utterances of normally developing Mexican-American Spanish speaking 3-year olds. Unpublished master’s thesis, San Diego State University, San Diego, CA.

McLaughlin, B. (1985). Second language acquisition in childhood. Hillsdale, NJ: Lawrence Erlbaum.

McLaughlin, B. (1995). Fostering second language development in young children (Report No. EDO-FL-96–02). Washington, DC: Office of Educational Research and Improvement. (ERIC Document Reproduction Service No. ED 386 960).

Nash, R. (1977). Comparing English and Spanish: Patterns in pho-nology and orthography. New York: Regents Publishing.

Paulson, D. M. (1989). Phonological systems of Spanish-speaking Texas preschoolers. Unpublished master’s thesis, Texas Christian University, Fort Worth, TX.

Rivera-Umpiere, E. (1988). Phonological analysis of 3- and 4-year-old monolingual Puerto-Rican Spanish-speaking children’s utter-ances. Unpublished master’s thesis, San Diego State University, San Diego, CA.

Stockwell, R. P., & Bowen, J. D. (1965). The sounds of English and Spanish. Chicago: University of Chicago Press.

U.S. Census Bureau. (2005). Homepage. Retrieved August 4, 2005, from http://www.census.gov

U.S. Census Bureau State and County Quickfacts. (2008). Florida. Retrieved February 29, 2008, from http://quickfacts.census .gov/qfd/states/12000.html

Villanueva, R. (1990). Phonological analysis of Spanish utterances of Puerto Rican children with speech disorders. Unpublished mas-ter’s thesis. San Diego, CA: San Diego State University.

Walters, S. (2000, November). Phonological development in three two-year old simultaneous bilingual children. Paper presented at the American Speech-Language-Hearing Association Annual Convention, Washington, DC.

Weisstein, E. W. (2008). Bonferroni correction. Retrieved February 29, 2008, from Math World Web site: http://mathworld.wolfram.com/BonferroniCorrection.html

Werner, E. O., & Kresheck, J. D. (1989). Spanish Structured Photographic Expressive Language Test–Preschool. Sandwich, IL: Janelle Publications.

Whitley, S. M. (2002). A course in Spanish linguistics: Spanish/English contrasts (2nd ed.). Washington, DC: Georgetown University Press.

Wylie, E., & Ingram, D. E. (1999). International Second Language Proficiency Ratings. Brisbane, Australia: Griffith University.

Yavas, M., & Goldstein, B. (1998). Phonological assessment and treatment of bilingual speakers. American Journal of Speech-Language Pathology, 7, 49–60.

Zampini, M. L. (1998). The relationship between the production on perception of L2 Spanish stops. Texas Papers in Foreign Language Education, 33, 85–99.

Alejandro E. Brice, PhD, CCC-SLP, is an associate professor of secondary/ESOL at the University of South Florida, St. Petersburg.

Cecyle K. Carson, PhD, CCC-SLP, is an associate professor of communication sciences and disorders at the University of Central Florida.

Jennifer Dennis O’Brien, MA, CCC-SLP, is a practicing speech-language pathologist in Jacksonville, Florida.

12 Communication Disorders Quarterly