16 / 68
14
S
p eech
P
athology
A
ustralia
MULTICULTURALISM AND DYSPHAGIA
grounds with SLI. Certainly, further research in the area with
children of diverse language backgrounds may provide
clearer recommendations about the appropriateness of certain
tests for certain populations.
Another issue that concerns assessment of children learning
the language of instruction is that of nation-wide or district
assessments. Assessing the academic potential of students
who are still in the midst of learning English is likely to result
in misleading impressions of their potential as well as their
academic progress (Cummins, 2000). Cummins highlighted
that students who have been learning English for about three
years in school perform at about one standard deviation
below age-equivalent norms in English academic skills. If
such data are not incorporated into the interpretation of test
scores, many students learning English as a second language
in academic programs can be deemed as language delayed or
language impaired.
Intervention
There are a number of implications for intervention from the
material discussed so far. A summary is provided in table 3.
Ideally, speech pathologists working with bilingual clients
should themselves be bilingual. However, this is not a
practical solution – we cannot be competent in all the
languages our clients may present with. However, it is our
argument that all clinicians have an obligation to work
bilingually. They need to use the language skills and
knowledge they have, in addition to the skills of interpreters,
teachers, parents, language teachers and others. Achieving
this may require the acquisition of new skills on the part of
the clinician and changes in the way speech pathology
students are educated.
normal communicative behaviours in bilinguals of all levels
of language proficiency (Brice, 2000; Brice & Anderson, 1999;
Mahon et al., 2003; Oller & Pearson, 2002); however, it can
also appear as a consequence of confusion or inadequacy in
one language. There are few guidelines in the literature to
assist the professional in deciding whether the child is code
switching normally or is confused. Brice and Anderson (1999)
found that a hierarchy of syntactic elements is common in
code mixing, for example nouns in subject and object position
were most common, followed by verbs, then verb phrases (p.
21). This hierarchy may be used as a guideline for assessing
the appropriateness of code mixing in bilingual children.
However, further research is needed to ascertain the appro
priateness and efficacy of assessing code mixing in this
manner.
The fact that the bilingual person is likely to use pragmatic,
syntactic and morphological elements of another language
when communicating in English should also be part of the
assessment. An analysis of a language sample in English
needs to apply information on language characteristics of the
speaker’s first language (such as that found in Swan & Smith,
2001). For example, since Chinese speakers do not formally
distinguish parts of speech, we may expect English–Chinese
bilingual speakers to use different English word classes such
as
difficult
and
difficulty
interchangeably, and that this is a
feature of bilingualism not a language disorder.
Standardised tests for which the normative population is
not bilingual should not be used, or at the very least no
standard scores should be reported. Paul (2007, p. 182) made
a number of suggestions for how standardised tests might be
used for this population, and all of them involve treating
them as criterion-referenced tests and not using the standard
scores as they are meaningless. Unfortunately, a normative
sample of bilingual children from the same language
background would be difficult to obtain because of the
diversity of languages and varied levels of language exposure.
There are some psycholinguistic markers (rapid automatic
naming and non-word repetition skills) that have come to
prominence in recent literature which seem to be relatively
language-free and which identify children with language
impairments. For example, Wiig, Zureich and Chang (2000)
found that monolingual English-speaking children with
language disorders had a statistically significant reduction in
rapid automated naming for tasks across semantic groups
such as colour and shape. Estes, Evans and Else-Quest (2007)
conducted a meta-analysis of the differences in non-word
repetition skills between children with and without SLI and
found children with SLI to exhibit significantly poorer skills
in non-word repetition skills compared with children without
SLI. However, the apparent promise of such findings may not
carry over to other languages as we might have wished.
Stokes, Wong, Fletcher and Leonard (2006) found that there
was no significant difference in repeating Cantonese non-
words between children with SLI and typically developing
children of the same age. On the other hand, Kohnert and
Windsor (2004) found that monolingual English-speaking
children with SLI performed significantly poorer than
typically developing bilingual Spanish and English-speaking
children on non-linguistic processing tasks such as choice
visual detection. This involved determining how accurate the
children were in associating a particular colour presented
with an associated response button. However, the bilingual
and language-impaired monolingual groups of children
performed comparably on identifying real and non-words as
well as on response time for picture naming tasks (Windsor &
Kohnert, 2004). It seems that the nature of psycholinguistic
tasks as well as the languages spoken by the children are
contributing factors in determining whether these tests can be
used reliably to identify children from multilingual back
Table 3. Some possible “new” intervention skills, goals
or methods
n
A monolingual SLP working in a bilingual “team”
– Primary interactant can be an interpreter, family
member, or ESL teacher
– Conduct of session can be primarily in a non-
English language
– Conduct of session can be 50% in one language,
50% in another
n
Teach code switching
n
Use one language in the teaching of another;
explanations in the “better” language
n
Specifically teach enhanced metalinguistic skills:
point out differences and similarities
n
Teach the clinician the other language, with the
client/team as expert.
– Uncover those things that are unique to that
language – no equivalents in English
– Work with those activities/functions only used in
the non-English language – cultural activities,
family interactions, talking on the phone to
community members, etc.
An example is the issue of code switching discussed earlier.
As we reported, skilled bilinguals exhibit considerable code
switching, which enhances their interactions with other
bilinguals (Brice & Anderson, 1999; Grosjean, 1989). Logically
therefore, competent code switching should be a goal in
bilingual intervention. However, it is doubtful that this goal is
being used in any speech pathology education program at
present and there are no clear protocols for achieving it.