Cross-linguistic and multilingual aspects of speech sound disorders in children

Environmental and personal factors relevant to the SLP  SLPs will never be able to speak all of the languages that are spoken by their clients.  During assessment, all SLPs should “develop sensitivity to their own social interactive styles and cultural biases relative to language sampling variables” (Stockman, 1996, p. 365)

Best practice for assessment  Evidence-based practice has primarily focussed on intervention for monolingual-English children  Assessment practices with multilingual children now are receiving greater attention  Assessment of children from non-dominant cultural backgrounds remains a challenge  However, as SLPs from around the world continue to share their knowledge, our practices are enhanced, and multilingual children’s assessment (and intervention) is more informed McLeod, S. & Goldstein, B. A. (Eds). (2012). Multilingual aspects of speech sound disorders in children. Bristol, UK: Multilingual Matters. Williams, A. L., McLeod, S. & McCauley, R. J. (Eds.) (2010). Interventions for speech sound disorders in children. Baltimore, MD: Paul H. Brookes Publishing. McLeod, S. (Ed). (2007). The international guide to speech acquisition. Clifton Park, NY: Cengage.

Seyhun Topbaş - Anadolu University , Turkey

Linguistic & cultural variety  There are many speech communities in Turkey whose first language is not Turkish.  Kurdish, Arabic, Persian, and other European languages such as Greek, Bulgarian, and Armenian have also been spoken since the time of the Seljuk and Ottoman Empires.  International immigration and mobility have also been changing the structure of the society.

Why concern for SLP?  Turkish is the official language and the medium of instruction in education is in Turkish as L1.  A second language, mainly English, is taught as a foreign language in primary schools begining from grade 4 ( about 9 years of age).  Thus, a child whose L1 is not Turkish, is exposed to Turkish as L1 in primary schools and then to a foreign language.

Why concern for SLP?  Currently, there is a growing interest for bilingual population mainly because:  Those children whose mother tongue is not Turkish do not receive necessary formal (both nursery and primary) additional support in either language.  This contributes negatively to the level of linguistic development of children in both languages.  There are vast regional disparities, where the lowest enrolment rates in pre-primary and primary education are observed in South-eastern and Eastern Anatolian provinces.

 The largest bilingual population of these minority languages is Kurdish-Turkish. Kurdish (Kurmanji and Zaza dialects) is the first language of about 12.98% inhabitants around South-eastern and Eastern Turkey.  Kurdish children generally learn Kurdish at home and begin learning Turkish as their second language (L2) via the media at home and outside and/or formally at 7 when they go to school (Derince, 2010).  Most Kurdish-Turkish speaking children do not fully fit the definition of either simultaneous or sequential bilinguals.

 Because Kurdish is not allowed as the medium of instruction in the primary schools, a mismatch occurs between the language spoken at home and the medium of instruction at school.  In the school context, children have to learn literacy skills and instructional content of science and mathematics in Turkish which may lead to some reading and writing problems, and academic learning difficulties in L2 (i.e., Turkish).

 The danger is that this situation may lead to disadvantage, and stigmatization of being language impaired.  Due to the reduced amount of both languages the language characteristics of these children may superficially resemble children with specific language impairments and it is often difficult to differentiate language impairments at the early phases of sequential bilingualism.

 It is important to identify these difficulties in languages in order to understand, whether the difficulty is stemming from a lack of exposure to either language or specific to impairment and if so, what intervention or remedial approaches are needed.  Phonology allows one of the promising assessment domains.

ASSESSMENT  Currently, there are no tools to assess speech and language performance of Kurdish children.  An SLP is competent enough to collect data, analyse the differences and similarities from the speech samples in the two languages and can demonstrate whether the differences are due to impairment or not.

Case Example Age: 3;6 Target Turkish Child /yatak/ [yatax] (bed) (ear) /kulak/ [qulax] (newspaper) (girl) /gazete/ [xestə] (spoon) (armchair) /kız/ [qıs] /kaşık/ [qaîx] /koltuk/ [qotıx]

IMPLICATIONS  A person who is not an SLP may interpret such a difference as a disorder; however, either the place or manner of articulation was preserved in most of the words, resembling a dialectal difference.  Identification of these difficulties in languages is important to show whether the difficulty is stemming from a lack of exposure to either language or specific to impairment and in either case, what intervention or remedial approaches are needed.  Thus, phonology allows one of the promising assessment domains for disentangling the two and to observe how speech impairment may be manifested.

Raúl F. Prezas - Texas Christian University, Fort Worth, TX Raúl Rojas - University of Texas at Dallas, TX

Bilingual Spanish-English children in USA  1990-2005  Bilingual children in school (“ELLs”) more than doubled  75% of ELLs are native Spanish-speaking children  Explosive growth of bilingual children  Growing demand for bilingual speech-language services

Assessing bilingual children  The Federal Government (IDEA, 2004) and ASHA (2010) concur  Non-biased and accurate assessment of bilingual children  Assess the native and second language, “unless it is clearly not feasible to do so” – IDEA, Section 612(a)(6)(b)

Speech sound disorders (SSDs)  Majority of clinical caseloads in US schools?  Children with SSDs! (ASHA, 2011)  Developmental milestones and assessment protocols available for diagnosing SSDs in English monolinguals  Similar resources scarce for bilingual children  Contextual Probes of Articulation Competence-Spanish (CPAC-S; Goldstein & Iglesias, 2006)  Hodson-Prezas Assessment of Phonological Patterns (Hodson & Prezas, 2010)

Bilingual assessment of SSDs  Recommended practices  Collecting background information  Family concerns  Dialect(s) spoken  Program (English immersion, dual-language)  Use of support personnel, if needed  Interpreters  Collaboration with bilingual speech-language pathologists (SLPs)  Testing phonological skills in both languages  Determination – scores low in one language or both?

 Assessment practices  Single word + connected speech samples  Phonetic inventory – strengths/weaknesses both languages  Estimates of intelligibility in known/unknown contexts  Relational Analysis  Accuracy of shared (e.g., /p/ in both languages) Cs  Accuracy of unshared (e.g., Spanish trill /r/; English /v/) Cs  Error Analysis  Rule out cross-linguistic effects (e.g., substituting American liquid /ɹ/ for Spanish flap /ɾ/ and trill /r/)  Phonological Pattern Analysis  Common patterns (e.g., cluster reduction)  Uncommon patterns (e.g., initial consonant deletion) Fabiano (2007)

Dialectal considerations  Highest prevalence in US – Mexican & Puerto Rican Spanish (consonant differences)  Children from Mexican descent with adult-like speech vs. other dialects:  Sequences with /s/  Escuela = /eskwela/ vs. /ehkwela/ or /e_kwela/  Pescado = /peskaðo/ vs. /pehkaðo/ or /pe_kaðo/  Final Consonants (e.g., /s/)  Dos = /dos/ vs. /doh/ or /do_/  Guantes = /wantes/ vs. /wanteh/  Liquids (i.e., /l/ and /r/)  Verde = /beɾðe/ /vs. /belðe/ Goldstein (2004); Prezas (2008)

Severity continuum Deviation Types and Phonological Repertoires Severity Omissions Substitutions Repertoires Profound Extensive Frequent Negligible Severe Moderate Frequent Extensive Limited Mild Occasional Occasional Emerging Rare Limited Near typical Adapted from Hodson (2007)

Metaphonological assessment  Rhyming & alliteration/onset  Oddity & matching  Segmenting & blending  Word, Syllable, Onset & Rime (e.g., sh+eep); Body & Coda (shee+p); Phonemes (sh+ee+p)  Manipulation  Substitutions, Deletions  Invented/developmental spelling  Naming (production tasks)  Multisyllabic Words, Pseudowords, Rapid Naming

Collaboration with personnel  Classroom teacher  Communication and collaboration  Training teacher to serve as second listener  Using teacher “expertise” (e.g., Spanish)  Spanish-speaking co-worker/colleague  Other teacher on campus  Secretary/aid  Bilingual SLP staff  District diagnostician  Collaboration with monolingual SLPs

David Ingram - Arizona State University, USA

Two Aspects of Assessment  1. Dimensions: Single dimension vs. Multiple dimensions, e.g. Processes vs. Multiple levels  2. Similarities (Differences): across children (inter- child variation, and within children (intra-child variation), e.g. inter-child: monolingual vs. bilingual acquisition intra-child: change over time

Dimensions 1. Basic Analysis (Ingram & Ingram, 2001): 4 levels whole-words (complexity & proximity); word shapes ; phonetic inventories (onsets & codas); matches and substitutions (onsets and codas) 2. MAPS (Phonological Assessment of Phonological Similarity, Ingram & Dubasik 2011): 9 quantitative measures across the 4 levels of the basic analysis; used for both inter- and intra-child analyses

Similarity  How to measure similarity?  Measure similarity across measures, and quantify similarity into a percentage score from 100% to 0% for each measure  First developed to compare phonological acquisition in Spanish-English bilingual twins (Ingram, Dubasik, Liceras & Fuentes Fernandez 2010)

Some Results  1. Inter- and Intra-child analyses of Spanish- English bilingual twins  2. Comparison of English-speaking twins vs non- twins  3. Comparison of Spanish-speaking typically developing three-year-olds vsl age-matched children suspected of having a speech sound disorder

Bilingual Twins  Measures Leo Simon ----------------------------------------------------------------------------------------------------------  pMLU Targets 100% 80%  pMLU Child 60% (Spanish) 70% (Spanish)  Proximity 30% (Spanish) 80%  Prop. Monosyllables 50% (English) 40% (English)  Preferred Syllables 80% 90%  Articulation Scores Onsets 100% 70% (Spanish)  Articulation Scores Codas 80% 60% (English)  Number of Matches Onsets 60% (Spanish) 60% (Spanish)  Number of Matches Codas 90% 60% (English)  ------------------- ------------------  Overall Similarity 73% 68% Children showed language separation

Bilingual Twins  Measures Leo vs. Simon (English) Leo vs. Simon (Spanish)  ---------------------------------------------------------------------------------------------------------  pMLU Targets 100% 90%  pMLU Child 100% 100%  Proximity 90% 70% (Leo)  Prop. Monosyllables 100% 100%  Preferred Syllables 100% 100%  Articulation Scores Onsets 80% 100%  Articulation Scores Codas 90% 90%  Number of Matches Onsets 90% 100%  Number of Matches Codas 50% (Simon) 100%  ------------------- ------------------  Overall Similarity 88% 94%  Children were highly similar within languages

Twins vs Non-Twins Measures Leo Jane Rachel Jennika Simon Lucy Samuel Daniel ----------------------------------------------------------------------------------------- pMLU Targets 100% 80% 60% 100% pMLU Child 100% 90% 70% 90% Proximity 90% 50% 90% 80% Monosyllables 100% 100% 50% 100% IA: Onsets 80% 60% 80% 0% IA: Codas 90% 90% 90% 60% RA: Onsets 90% 80% 80% 0% RA: Codas 50% 100% 100% 100% ------------------------------------------------ Overall Similarity 88% 85% 81% 68% Identical twins were more alike than non-identical twins; children closer in age were more alike than children further apart in age

Spanish Speaking 3-Year-Olds Comparison of Phonological Systems at Time 1 vs Time 2 (approx. 9 months later) Measures Child 1 Child2 Child3 ----------------------------------------------------------------------------------------- pMLU Child 100% 70% 80% Proximity 100% 60% 70% IA: Onsets 90% 90% 70% IA: Codas 100% 90% 90% RA: Onsets 80% 80% 80% RA: Codas 80% 80% 20% -------------------------------------------------------- Overall Similarity 92% 75% 68% Child 1 showed some improvement in final consonant matches; child 2 improve across levels, especially in whole word measures; child 3 showed most gains, especially in final consonant matches

Conclusion  Multidimensional phonological analyses combined with a measure of phonological similarity provides insights into:  Monolingual vs. bilingual acquisition;  Twins vs. non-twins;  Age differences between siblings  Intra-child acquisition over time

Carol K. S. To - The University of Hong Kong, SAR China

Number of Chinese-Speakers in USA  Population 5 years or above: 280,950,438 (US census Bureau, 2007)  % spoke a language other than English at home: 19% Spoken a language other than English at home: 100% • Spanish or Spanish Creole 62.3% • Other Indo-European languages 18.6% • Asian and Pacific Island languages 15% • Other languages 4.1%

 “Asian and Pacific Island languages include Chinese; Korean; Japanese; Vietnamese; Hmong; Khmer; Lao; Thai; Tagalog or Pilipino; the Dravidian languages of India, such as Telugu, Tamil, and Malayalam; and other languages of Asia and the Pacific, including the Philippine, Polynesian, and Micronesian languages.” (US Census Bureau, 2007, p. 2 )

Chinese  Chinese is a language family consisting of 7 varieties .  The varieties shared the same writing system (Standard Chinese) but phonologically we are mutually unintelligible.  Mandarin (Putonghua)  Wu (includes Shanghainese)  Yue (includes Cantonese & Taishanese)  Min (includes Hokkien, Taiwanese & Teochew)  Xiang  Hakka  Gan

Hong Kong 88

Official Languages in Hong Kong 1842 1974 1997 2010 (Local) Cantonese (1842 - ) English (1842 - ) Chinese (Modern Standard Chinese, MSC) (1974- ) “Biliterate & Trilingual Policy” Putonghua “Compulsory Chinese medium 89 of instruction policy”

 Hong Kong is a multilingual city.  HKC has undergone drastic sound changes over the past 160 years.  Speech-language pathologists (SLPs) need an up-to- date information about acceptable variants of standard pronunciations of HKC.

Standard Cantonese in 1940 Tones (10) Phonemes High-Level (55), H-Fall (52), H-Stop (5), H-Rise (35), Mid-Level (33), M-Stop (33) Midlow-Level (22), ML-fall (21), ML-Rise (23), ML-Stop (22) Initial /p-, ph-, t-, th-, k-, kh-, kw-, kwh-, w-, l-, j-/ consonants /ts- (tʃ-), tsh- (tʃ h-), f-, s- (ʃ -), h-, m-, n-, ŋ-/ /-p, -t, -k, -m, -n, -ŋ/ Final consonants /i, y, ɛ, œ, a, ɔ, u, ɐ/ Vowels Diphthongs /ai, ɐi, аu, ɐu, ei, ɵy, ɔi, ui, iu, ou/

Hong Kong Cantonese in 21st century Tones ((190)) Phonemes High-Level (55), H-Fall (52), H-Stop (5), H-Rise (35), Mid-Level (33), M-Stop (3) Midlow-Level (22), ML-fall (21), ML-Rise (23), ML-Stop (2) Initial /p-, ph-, t-, th-, k-, kh-, kw-, kwh-, w-, j-, m-, nn--, l~- l-// consonants /ts-, (tʃ-), tsh-, (tʃ h-), f-, s-, (ʃ-) , h-, ŋ- / Final /-p,,--tt, ~-k-k, ,-m, --nn, ~-ŋ /-ŋ/ consonants Vowels /i, y, ɛ, œ, a, ɔ, u, ɐ/ Diphthongs /ai, ɐi, аu, ɐu, ei, ɵy, ɔi, ui, iu, ou ɛu //

Conclusions  Sound changes in HKC is taking place at a very rapid rate.  Cantonese is an oral language. The Chinese writing system belong to the logographic language which carries limited cues for pronunciation.

Clinical Implications  Sound changes relating to syllable-initial and syllabic consonants can be treated as acceptable variants in speech sound assessments.  The four syllable-final sound changes are taking placing quickly. They can be accepted as variants during speech sound assessment.

Helen Grech - University of Malta, Malta

THE MALTESE ISLANDS  Archipelago in Mediterranean Sea  Malta – main island; Gozo - sister island  93 km south of Sicily; 288 km north of Tunisia  1/3 million population; 1 million tourists annually  Independent language since 1090 AD; Semitic in origin with influences of Romance languages & English

2 National Languages Maltese English Maltese acquired as L1 by >90% of population (Borg, Mifsud & Sciriha, 1992). Parental report: 56.9% 4.7% 38.6% Home language (Grech & Dodd, 2008) Maltese English Maltese + only only English

Language structures  Maltese & English are distinctively different in phonology, morphology and syntax.  Maltese has complex phonotactics.  Maltese is highly inf lective (sentence could be composed of 1 lexeme with 3-4 bound morphemes typically signifying gender, plurality and tense).

The National Minimum Curriculum Bilingualism is the basis of the educational system. Most children are taught a 3rd language in schools, normally Italian. > 80% of children in church or independent schools and 10% in state schools are taught a 4th language (Sciriha, 1999).

Language Mixing Language mixing is typical in the Maltese culture with English carrying a higher social status. Children increasingly engage in language mixing, particularly at an early age, (Gatt, 2010). By 3-5 yrs most children are bilingual to varying degrees. 52% of children (2;0-6;0 yrs old) spontaneously used translation equivalents (Grech & Dodd, 2008).