Apparently, it is hardly possible to quantify sufficient linguistic input in both simultaneous and successive bilingualism:
We do not know what the ‘normal’ amount of exposure to a language is. The number of child-directed utterances or, more specifically, of verbal interactions varies enormously across families, monolinguals as well as bilinguals. Consequently, we have an only approximate idea of what the minimum amount of input is for a child to be able to develop a native competence. It should definitely not drop below 20 per cent of the total amount of verbal interactions a child receives. More reduced amounts of exposure to a language may suffice to trigger linguistic development, but this will happen at a significant slower rate and it is unlikely to result in native competence (Meisel: 177 f).
Among the benefits and advantages of child bilingualism, as could be seen, was the opening up to another culture and the ability to, at least in part, express the same ideas in two languages. In order to facilitate this, although not measurable in definite numbers, a minimum threshold level of language exposure does exist and children would profit enormously from verbally rich environments. The result of such an acquisition process—especially from early on—“will almost certainly be vastly superior to what is normally achieved in foreign language classrooms” (ibid.: 185).
1.4 The Gap Between Knowing and Doing
However, studies in language acquisition also need to look at the gap between knowing and doing, in other words the difference between competence and performance. Whereas the most important components of a speaker’s linguistic knowledge are their lexicon and grammar, the ability to activate their linguistic competence in communicative interaction is equally decisive. As an illustration the well-known “tip-of-the tongue” phenomenon (TOT) is often quoted, a situation where speakers fail to recall a word or term, although they feel retrieval is imminent, “on the tip of their tongue”. It occurs at all ages and happens frequently enough to interfere with learning or in everyday situations:
Speakers experiencing it are momentarily not able to retrieve a word that they have used many times before and that they can use in spontaneous speech shortly afterwards. Thus, they obviously do know it. In fact, when offered a list of words, they reject the wrong ones. Even more interestingly, they may be able to identify certain properties of the word, like its initial sound, its number of syllables, or its meaning; or they can cite words that are similar in sound and meaning. This shows that the irretrievable element is not missing; yet it is only partially, not fully, accessible. It furthermore suggests that items in mental lexica can be retrieved in a number of different ways, not merely by their initial sounds (ibid.: 16).
There is no one explanation for this phenomenon. Either a speaker’s memory strength may not be strong enough to recall an item (direct-access view) or incorrect, but similar responses to a query come to mind; in this explanation these answers are termed blockers because they block the possibility to recall the correct item, once they are overcome retrieval will be possible again (blocking hypothesis; cf.: ibid.). If there are, then, different ways of activation, a successful retrieval depends on appropriate cues and lexical access, at the same time, occurs in stages. “Language learners must therefore develop both a grammatical competence and the proficiency to put it to use in a wide array of communicative contexts” (ibid.: 16). Thus, the question arises: What about bilinguals? Do they experience similar phenomena or is there a double load in their mental lexicon because of their increased linguistic load?
Experiments conducted by the American Psychological Association (APA 2020: n.p.) indicate that bilinguals, compared to their monolingual peers, are involved in a “doubled processing load” which can lead to difficulties in specific situations but otherwise be advantageous. Given the descriptions of visual objects and people’s names, Spanish-English bilinguals showed relatively better results in retrieving proper names, being more difficult to reproduce, than monolinguals:
… for a harder task, bilinguals showed relatively better performance. Bilingual disadvantages may be limited to representing multiple forms for individual meanings; proper names improved naming because they have essentially the same form across languages (ibid.).
1.5 The Dual System Hypothesis
Whether this can also be seen as a confirmation of the Dual System Hypothesis (DSH; cf. Meisel: 48), also called Dual Language System Hypothesis (cf. Genesee 1996) is open to debate, but at least it hints at further differences in linguistic knowledge and performance between monolinguals and bilinguals—that are useful to know when it comes to study CLIL performance in more detail. The DSH proposes the
idea of early differentiation of linguistic systems, characteristic of the developmental pattern through which simultaneous bilinguals typically proceed … a hypothesis widely accepted today on the basis of findings obtained by a steadily increasing number of case studies carried out over the past 30 years (Meisel: 48, 72).
According to the proponents of the DSH, bilingual children acquire two different language systems from the beginning of language acquisition, which as a process results in two distinct mental grammars (cf. Meisel: 52); their rules do not cross over from one language to the other (cf. Bradley: 2020 n.p.) without denying a linguistic interaction between the two systems. Interaction would occur “primarily in settings where one language is dominant and the other one weak, either in the social context or in the mind of speakers” (Meisel: 53). Since bilingual programs like CLIL inevitably have to deal with a stronger (L1) and a weaker language (L2), this would have a number of consequences for teaching strategies and learning processes, which are discussed when looking at CLIL tools and skills (cf. chapter 8). But despite this assumption based on observation and common sense the terms “strong” and “weak” need some qualification:
Empirical justifications for qualifying a language as ‘weak’ or as ‘non-dominant’ commonly refer to one or more of the following observations: a) The language in question is rarely or not at all used actively; b) the other language is strongly preferred; c) the development of the allegedly weak language is less advanced than that of the other language (ibid.: 130).
The strong point pertinent to the DSH is, then, that bilingual children do not move through stages to be eventually able to distinguish between two languages. On the contrary, they are able to have two language systems and are aware which system is related to which language (cf. Bradley: 2020 n.p.). There are recent studies that ascertain bilinguals’ capacity to use the correct system for each language without “mixing” them, and an individual’s capability to differentiate which language is spoken—Ava’s example comes to mind—and a test showing that bilingual children responded with the same language spoken to them (cf. ibid. and Genesee: 1996).
This shows us that bilingual children under the age of two could differentiate the language being heard and respond with the same language, which supports the idea that children have two different language systems (Bradley: 2020 n.p.).
How code-switching and the concept of interlanguage can be accounted for is part of the challenges still connected with CLIL and dealt with in a final look at desiderata of this approach (chapter 10).
Overall and also according to the MHH-Review, studies provided a mixed and complex picture in terms of the relationship between mother tongue (L1) and the foreign language (L2) acquired. Established was some transfer from L1 to L2 and vice versa and hence the existence of two language systems, not independent but inter-connected. The question of a unitary phonological system was regularly abandoned in favor of the evidence of two separate systems by a majority of the studies reviewed. The observation that phonemes are shared in both languages was substantiated by the experience that they are acquired in one language before another. This interconnectedness does not only account for the fact that researchers found robust evidence for transfer from the dominant language to the language acquired later (L1 to L2), but was also mirrored by the existence of only two older studies (from the 1970s and 1980s) arguing for a unitary language system. This is to say that empirical data help to determine the validity of language acquisition theories of past