Since the early 1920s, much research has been done on bilingualism, with the earliest investigations finding mainly negative cognitive effects attributed to bilingualism (Pintner & Keller, 1922; Saer, 1923). However, ever since Peal & Lambert (1962) counter-claimed that bilinguals actually do perform better in certain cognitive tasks (e.g. symbolic manipulation types of non-verbal tasks), more recent studies have solidified the current view that bilingualism enhances ones cognitive flexibility and metalinguistic awareness. Yet, most research on language acquisition only focuses on one target language and neglects other languages already acquired or are being acquired by the learner. Research on third language acquisition (TLA) or the acquisition of additional languages (AAL), which attempts to fill this gap by bringing together the two traditionally detached fields of bilingualism and language acquisition, has only begun to accelerate during the late 1990s (Cenoz, 2008; Falk & Bardel 2010).
The effect of bilingualism on TLA is one of the main areas of interest in research concerning third language (L3) studies. In order to sufficiently answer this question, we would have to define what we mean by "bilingualism" and "third language acquisition". Different linguists define bilingualism differently over a broad spectrum, from the maximalist view of 'equal native-like competency in two languages' to the minimalist interpretation of a 'minimal competency in two languages'. For the purpose of this paper, we will define "bilingualism" as the ability to communicate effectively in two languages. Similarly, henceforth, "third language acquisition" will be defined as the process of learning and acquiring of a non-native language in a secondary context (i.e. language is acquired in a structured setting) by a learner who have already acquired two other languages previously.
General consensus today prescribes to the notion that bilingualism brings about various cognitive benefits. It can be postulated that third language learners have a distinct advantage over second language learners due to their enhanced cognitive capabilities brought about by their bilingualism. Hakuta & Bialystok (1994) wrote that "the knowledge of two languages is greater than the sum of its parts". Most studies tend to indicate advantages in bilinguals over monolinguals in language acquisition, especially when the learner's bilingualism is additive rather than subtractive (Cenoz, 2003). While there are numerous other factors affecting TLA including language similarity, L1/L2 influences, L1/L2 proficiency, recency of use, language status, role of Universal Grammar, cross-linguistic influences, early multilingualism and age of TLA (De Angelis 2007; Cenoz, 2008), we will be focusing solely on the effects of bilingualism on TLA. In particular, we will be discussing the effects of meta-linguistic and meta-procedural gains on TLA due to bilingualism.
Bilinguals have shown to display cognitive flexibility and enhanced metalinguistic awareness over monolinguals, and this in turn enhances bilinguals' TLA (McLaughlin & Nayak, 1989; Cenoz, 2003). Studies of bilingual children have shown bilinguals to have higher scores in tests targeting creative or divergent thinking. Research has also shown bilinguals to possess better abilities to control and utilise linguistic knowledge and to have an overall predisposition in word awareness tasks (Cummins, 1991; Bialystok, 2001). Furthermore, in Cummins (1991), the interdependence hypothesis was proposed which suggested that skill transfers take place from a bilingual's L1 into L2. One can only assume that these enhanced cognitive and metalinguistic abilities would have a positive effect in a bilingual's acquisition of a L3. It can also be logically deduced that the skill transfers from L1 into L2 based on the interdependence hypothesis can also occur from a L3 learner's L1 and L2 into the L3. In an effort to further this hypothesis, De Angelis (2007) wrote that TLA differ significantly from second language acquisition (SLA) due to the fact that L3 learners can transfer linguistic elements from their prior knowledge in both L1 and L2, and this gives L3 learners an advantage over L2 learners. Studies have also shown that L3 learners often leverage upon their access to two different linguistic systems during the process of TLA (Herdina & Jessner, 2002). All these suggested that the enhanced cognitive and metalinguistic abilities a bilingual possesses positively affects the acquisition of a L3.
Enhanced cognitive abilities are not the only skills transferable during TLA. Syntactic transfers have also been shown to occur in L3 leaners. Different languages have different syntactic rules, and most bilinguals would have had access to the different syntactic rules governing the different languages they know (assuming that the two languages are syntactically dissimilar). This knowledge of two different sets of syntactic rules would assist a bilingual in the learning of a L3. Flynn, Foley & Vinnitskaya (2004) proposed that language learning is cumulative in the Cumulative-Enhancement Model, suggesting that each language learned previously influences and enhances subsequent language learning, adding to a cumulative advantage for third and additional language learners. However, a more recent study by Bardel & Falk (2007) on Swedish and Dutch L3 learners found only positive transfers of syntactic properties from L2 but not L1 in the learning of a L3. It was additionally proposed that "in L3 acquisition, the L2 acts like a filter, making the L1 inaccessible".
If the above researches holds true, bilingualism could have either a positive or negative effect on third language acquisition depending on the learners' L2. If the L2 is syntactically similar to the L3, the L3 learner would experience positive transfers of the syntactic properties of L2 into L3, aiding the learners' acquisition of L3. Conversely, if the L1 but not the L2 is syntactically similar to the L3, the L3 learner would not be able to access the benefits of his knowledge in his L1 that he would otherwise have been able to reap had there not existed an L2 to act as a filter. Nevertheless, both studies while disagreeing on the existence of syntactic transfer from a L3 learner's L1, agreed that they both found no negative (only positive or neutral) transfers regardless of the similarity or dissimilarity of the L3 learner's L2. Therefore, irrespective of the fact that syntactic transfer from L1 occurs or not, bilinguals would have an advantage (or at least no disadvantage) in language learning as compared to monolinguals.
Similarly, studies have also proved that transfers in the bilingual's knowledge of different orthographic systems give the bilingual an advantage in TLA. While numerous languages in the world are similar orthographically, many written languages have vastly different orthographic systems. In many cases, a learner of an additional language would need to learn a new orthographic system. Abu-Rabia & Sanitsky (2010) compared students with knowledge in two orthographies (Hebrew and Russian) against students with knowledge in only one (Hebrew) in the acquisition of English as an additional language. While the results found that transfers of orthographic skills occurred in both groups of students, the learners who were familiar with two orthographic types outperformed those with knowledge in only one type of orthography. It was also noted that "the rich orthographic experience in different orthographies is an advantage for trilingual speakers". However, numerous studies have found limited orthographic transfers when two vastly different orthographic systems were involved (e.g. Wang, Perfetti & Liu, 2005 on Chinese-English; Wang, Park & Lee, 2006 on Korean-English). In particular, Bialystok, Luk & Kwan (2005) compared Spanish-English, Hebrew-English and Chinese-English bilinguals with monolinguals, and found that all three groups of bilinguals obtained higher levels of literacy than the monolinguals. They also found that Spanish-English and Hebrew-English bilinguals had a greater advantage than Chinese-English bilinguals (Spanish and Hebrew are both written alphabetically albeit in different scripts; Chinese is written logographically).
While evidence shows that bilinguals would have an advantage in TLA due to orthographic transfers, the magnitude of such gains would be dependent on the similarities (if any) of the orthographies involved. It can be assumed that as the benefits of orthographic transfers exist due to the bilinguals' enhanced awareness of different orthographic and script systems, bilinguals who use the same orthography and script in both L1 and L2 would find little or no advantage over monolinguals in the acquisition of an additional language which uses a different orthographic system. The same bilingual would however experience benefits in learning a L3 which uses the same or similar orthography and script as both the L1 and L2. However, what remains to be researched is the amount of orthographic transfer when the L3 is orthographically similar to either but not both the L1 and L2. It would also be interesting to find out if the 'filtering effect' as proposed earlier by Bardel & Falk (2007) on blocking of syntactic transfers from the L1 by the L2 likewise holds true for orthographic transfers.
TLA is similar to SLA in many ways, but yet it has been reported that language learners benefit from not only meta-linguistic gains but also from meta-procedural gains from prior language learning experiences. A study by McLaughin & Nayak (1989) speculated that "expert learners use different information-processing strategies and techniques than do more novice learners", and noted that multilinguals use a wider range of strategies and are more flexible in language learning than monolinguals. Similarly, Kemp (2007) showed that multilinguals' experience in language learning helps them formulate better learning strategies which "speeds acquisition through freeing up working memory". Kemp further noted that the number and frequency of strategies used in language learning is positively related to the number of languages the language learner already knows. From this, we can conclude that the more languages a person knows, the easier it will be for him to acquire an additional language. We can thus deduce that bilingualism gives the bilingual an advantage in L3 acquisition brought about by the prior language learning experience.
However, most studies on L3 learners' meta-procedural gains have compared L3 learners with L2 learners, neglecting the differences between L3 learners who are simultaneous or sequential bilinguals. For L3 learners to reap the benefits from meta-procedural gains, logically, the learner must have had prior 'textbook learning' of a language. Simultaneous bilinguals who grew up learning two languages simultaneously in a naturalist setting would not have had prior language learning experiences to tap upon. Further studies on TLA comparing simultaneous and sequential bilinguals would fill this gap and offer additional evidence of such meta-procedural gains if results prove that sequential bilinguals perform better in TLA over simultaneous bilinguals.
During the course of this paper, we have discussed how, more often than not, bilingualism has shown to exhibit positive effects on the acquisition of a L3 be it meta-linguistically or meta-procedurally. It has to be noted that while most studies conducted on the effects of bilingualism on TLA tend to claim advantages for bilinguals, not all reports on the effects of bilingualism on third language acquisition is positive. Some studies have instead shown no significant advantages in TLA by bilinguals (Soler, 2008). Cenoz (2003) further pointed out that "even if bilingualism has an effect on third language acquisition, it does not have to affect all aspects of third language proficiency in the same way, and different conclusions can be drawn depending on the dimension of language proficiency taken into consideration". Moreover, in Del Puerto (2007), it was pointed out that balanced bilinguals tend to perform better than less balanced bilinguals in third language learning. In other words, the degree of proficiency in L1 and L2 must also be taken into account when evaluating the effects of bilingualism on third language acquisition. Logical reasoning would direct that a bilingual who is less proficient in or is suffering from language attrition in either of his languages would perform less impressively than a balanced bilingual.
As Del Puerto (2007) aptly sums it, third language acquisition is an "exceedingly complex process and a multicausal phenomenon than can be affected by multifarious factors". We have but discussed some of the effects of bilingualism on third language acquisition, but the multitude of causal factors involved in third language acquisition means that much more can be discussed on this subject.