1 Introduction
It is well known that to obtain a fair representation of language skills in multilingual people, all their languages must be assessed (Bedore and Peña 2008). However, due to a lack of comparable tools across languages, this is not always possible. In many cases, assessment tools are only available in well-studied languages, often those with high social status or the languages of schooling. For pre-school children, being assessed only in the majority language may render misleading results if that language is not spoken in the child’s home and the child has had little or no exposure to it. Regardless of whether appropriate assessment tools are available, language assessment may be important and necessary for many purposes – amongst others to allow identification of a possible language impairment or delay. In the case of multilingual children, multilingualism needs to be disentangled from language impairment, as the language of children who are multilingual may share characteristics with that of children with language impairment. For instance, similarly to children with language impairment, multilingual children might lag behind their monolingual peers – when only one of their languages is taken into account. Whereas one does not want to diagnose typically developing multilingual children as language impaired, under-diagnosing language impairment in multilingual populations, and thereby denying them access to whatever support may be available in their contexts, is detrimental to multilingual and monolingual children alike.
The lack of assessment tools for multilingual children was specifically addressed by a 2010–2013 European Union–funded network of child language researchers (COST Action IS0804, Bi-SLI, “Language Impairment in a Multilingual Society: Linguistic Patterns and the Road to Assessment,” www.bi-sli.org), and their activities still continue. In this network, a collective, international effort was made to develop assessment tools for different language domains across a wide range of languages, resulting in a battery of tools under the umbrella name LITMUS (Language Impairment Testing in Multilingual Settings; see, e.g. Armon-Lotem, de Jong, and Meir 2015). Here, we will focus on vocabulary, and take as our point of departure our experience with two different tools for assessing vocabulary in multilingual children, namely the MacArthur-Bates Communicative Development Inventory (CDI) and the LITMUS Cross-Linguistic Lexical Tasks (CLT). These tools are aimed at different age groups, and they were developed in very different ways, but they both exist in linguistically and culturally comparable versions across many different languages. As such, a basic assumption is that those versions can be used for assessing the different languages of a multilingual child, and that they will indicate strengths and weaknesses in the child’s languages in an accurate manner. However, as we will show, our work across different populations of multilingual children indicates that however comparable our child language tools are meant to be, they need to be used with care, and their results need to be interpreted reflectively.
One of the reasons for treating the language assessment results of multilingual children with caution is that there is no such thing as The Multilingual Child. Multilingualism, also childhood multilingualism, can take many forms (see, e.g. Butler 2013; Lanza 2007; Wei 2013). A child could, for instance, be acquiring two or more languages simultaneously (from birth) or consecutively/successively (where exposure to languages other than the first language (L1) takes place after the age of 2 years); early (before the age of 6 or 7 years) or late; and additively or subtractively, where the latter refers to the child learning the second language (L2) to the detriment of the first, as is often the case when the L1 is a minority language. The relative exposure to each language may vary, as well as where and how each language is learnt. The child could comprehend and speak the L2 or could be a passive multilingual who is able to understand the L2 but does not speak it. All languages could be acquired in the child’s home context, or some could be acquired through exposure at the childcare institution or in the community at large. There could be a lot of community support for the child’s languages or little to no such support for one or more of the languages. Such heterogeneity amongst multilingual children calls for caution when interpreting research findings and assessment results across populations and contexts.
In Norway, albeit generally considered a monolingual country, most children grow up multilingually in the sense that they are exposed to Norwegian at home, in childcare, at school, and in all spheres of society; have access to books, television shows, and other educational and entertainment artefacts in Norwegian, but start acquiring English early on through activities on the internet such as games, YouTube videos, and pop music. Some children are introduced to English in kindergarten, and all are exposed to it at school, from the age of 6 years. More than one-fifth, and maybe up to one-third, of Norwegian children are exposed to other languages in the home, having one or more parents speaking one of the traditional languages in Norway, such as Sámi or Kven, or more recent immigrant languages such as Polish, Somali, Swedish, or Turkish (Statistics Norway 2020a). Some of the languages have higher status in Norway than others, and the Norwegian community typically supports the acquisition of, for example, English and Swedish over Polish, Somali, and Turkish. Swedish, Danish, and Norwegian are also (near) mutually intelligible, providing a linguistic advantage to children acquiring a combination of these languages. Overall, more than 92% of children aged 1 to 5 years attend kindergarten (Statistics Norway 2020b). The age of kindergarten entry varies to some extent between cultural groups, but increasingly children start kindergarten already at the age of 1 year. The access to educational and entertainment artefacts in the various languages differ; children speaking Somali, for example, will find very little written or visual artefacts of their home language.
In South Africa, multilingualism is the norm, and childhood multilingualism takes many forms, which differ from the kinds of multilingual contexts in which Norwegian children grow up. Unlike Norway, South Africa does not have a country-wide dominant language in terms of the number of speakers. Of the country’s 11 official languages, isiZulu (at 25%) has the largest percentage of home language speakers and isiNdebele (at 1.6%) the smallest. Although English is the lingua franca in South Africa, it is spoken as a L1 by only 8% of the population (Statistics South Africa 2018b). The majority of South African children do not receive sufficient exposure to English in their daily lives to allow them to enter school proficient in the language. Yet, English is the predominant and preferred language of education. Note that almost half of the children under the age of 6 years (47%) do not attend childcare institutions before school entry and thus receive the vast majority of their language input at home. Home childcare is, however, not always of high quality in terms of language interaction; for instance, just over half of South African children aged 0 to 6 years are frequently read to or told stories by members of their household, and about a third of this age group are not entered into conversation with by members of their household (Statistics South Africa 2018a), thereby reducing the quantity of input received in any of their languages. The majority of ...