Introduction
Multilingualism is now seen as the norm rather than the exception in an age of migration and supranational entities, and where minority language rights and the consequent educational policies have become more common. The field of applied linguistics reflects that transition: second language acquisition (sla) research is slowly being replaced by research on multilingualism, which includes third language (L3) acquisition. Indeed, there is a growing list of studies that are ‘normalizing’ third language acquisition by studying bilinguals learning a new language but not considering bilingualism a variable (e.g. Stafford, Sanz & Bowden Reference Stafford, Sanz and Bowden2012; Lado et al. Reference Lado, Bowden, Stafford and Sanz2014; Pérez-Vidal Reference Pérez Vidal2014; Cox & Sanz Reference Cox and Sanz2015). In this modern global context, researchers have produced empirical research on L3 acquisition that can be divided into three main categories depending on its focus: (a) classroom studies conducted in bilingual communities and schools with students learning a third language as part of the school curriculum (Cenoz Reference Cenoz2013); (b) research on cross-linguistic influence investigating sources of transfer from L1 or L2 into the L3 (e.g. Sanz, Park & Lado Reference Sanz, Park and Lado2014) and other possible directions for transfer (González Alonso et al. Reference González Alonso, Rothman, Berndt, Castro and Westergaard2016); and (c) laboratory research – that is, studies outside of classroom or immersion contexts, in which dependent and independent variables can be tightly controlled by the researcher – conducted within a cognitive approach, the focus of this timeline. Despite its potential diminished ecological validity, this last strand is characterized by the robustness of its design and its improved overall validity, and by the manipulation of external conditions and the measurement of internal variables related to cognition.
Laboratory research on L3 development started with comparisons of monolinguals’ and bilinguals’ processing of a new language. Early work was conducted by cognitive psychologists interested in understanding the role of experience in the acquisition of new knowledge and looked at bilinguals as experienced (bilingual) learners compared to novices (monolinguals). They were also interested in the interaction between experience and explicit and implicit conditions, with results suggesting that experienced learners (1) habitually exert more effort when processing novel verbal stimuli; (2) use cognitive processing strategies that facilitate the construction of formal rules; and (3) are better able to shift strategies to restructure their language systems.
For linguists, however, this literature was limited in two important ways, both having to do with research methodology. First, these studies operationalized ‘language learning’ as ability to predict the grammaticality of stimuli consisting of strings of letters lacking any meaning (artificial grammars) while linguists take a functionalist approach that looks at language development as making form–meaning connections and view language as a vehicle of communication. From this perspective, a definition of grammar as rules that determine letter string-combinations appears as lacking. Also, sampling methods in terms of language background were not as strict as desired.
Researchers tried to solve these issues by carefully selecting their samples of monolingual and bilingual language learners and by including mini-grammars of natural languages in their designs; they also broadened the target structures to include morphosyntactic, phonological and lexical development. Importantly, these researchers manipulated learning conditions in ways that approximated real pedagogical conditions, i.e. explicit conditions included provision of grammar rules prior to or during practice, and implicit conditions presented opportunities to practice without explicit grammar instruction. Finally, and following contemporary trends in SLA research, research on L3 acquisition began to look at bilingualism as an individual difference itself related to other individual differences, mostly socio-cognitive in nature, such as working memory capacity, learning strategies and motivation. This move placed laboratory research on L3 acquisition squarely within interactionist approaches that try to capture the multiple relationships between internal and external variables. Naturally, variables intrinsically related to L3 development, such as age and level of L2 acquisition/bilingualism, were placed front and center.
Laboratory research on L3 acquisition began in the domain of phonology, with questions of bilingual advantages in phonetic production and perception. As in other domains, one chief debate has been whether any bilingual advantages are dependent on typological similarities to first language (L1) and/or second language (L2) or stem from a broader cognitive advantage that applies to entirely novel segments (Sanz et al. Reference Sanz, Park and Lado2014). From the 1970s to the current day, this question continues to be of interest, as designs become more sophisticated to include variables such as age of acquisition of L2. To explain contradictory results, researchers have identified subclassifications of novel segments, such as those that are ‘easy’ or ‘difficult’ to learn. In addition, while early studies investigated learning without instruction, more recent studies have considered the role of instruction in learning. Since 2010, studies have also examined the question of segmenting speech stream into words.
Research on L3 lexical development was a natural offshoot of the study of phonetic perception and cognitive capacity, such as phonological short-term memory (PSTM). Some studies have found that a bilingual advantage in PSTM explains a bilingual advantage in word learning, while others do not support this conclusion. As in phonology and morphosyntax, studies have considered the role of type of instruction, and modality of instruction (aural, visual, or combined). Another common thread across domains has been whether bilingual advantages are limited to typological similarities with their L1 and/or L2, or are also present when learning entirely novel items. Research questions particular to lexical development include the type of stimulus in terms of concreteness and familiarity of the referent. Recent research has also expanded the timeframe of study to not include only immediate learning but also retention.
As for L3 morphosyntax, with the intention of finding explanations for the advantages in L3 morphosyntactic acquisition identified for bilinguals over monolinguals (Klein Reference Klein1995), and to further probe into differences in deployment of attention as a potential explanation, Sanz developed, with Bowden and Stafford, a research program that investigates the interaction between language experience and input varying in degrees of explicitness, and that includes cognitive variables such as aptitude as possible moderating variables. The target is the use of word order, case, and number morphology in the assignment of semantic functions (or who does what to whom) in L3 Latin by native speakers of different L1s (English, Spanish, Chinese) and L2s (English, Spanish, Japanese, Arabic). Departing from previous studies that compared monolinguals with bi/multilinguals, The Latin Project delves into the role of prior experience on additional language learning by including bilinguals with different degrees of proficiency in their non-primary languages. The Latin Project has also incorporated an often-underrepresented population in multilingual acquisition research: older adults.
Although laboratory research on L3 acquisition started in the 1960s, the majority of the studies date from the 2000s, so the methodology is innovative and highly eclectic, with designs borrowed both from linguistics and psychology by way of SLA research. Cognitive L3 language acquisition research is complex and requires sophisticated designs with sampling, constructs, measurements, and analyses that present important challenges. Proficiency/degree of bilingualism is especially problematic because it includes a multiplicity of elements (e.g. productive and receptive skills) and is often measured with tests designed for monolinguals and not sufficiently fine-grained to evaluate highly skilled bilinguals; also, responses to self-rating questionnaires may be influenced by different factors (e.g. attitude towards the bilingual's minority language, for example). Regarding analyses, L3 acquisition and use are multifactorial phenomena that demand an interactive approach, which itself demands large samples, yet another challenge. The most popular quantitative procedures include analyses of variance (ANOVAs), correlations and regressions, but the next years will see an increase in publications that include linear growth analysis to investigate systematic change and inter-individual variability in this change, as well as the inclusion of new technologies such as eye-tracking and event-related potentials (Grey 2013: see timeline) as tools to understand the way previous language experience influences additional language learning. Latency, with accuracy, has been the dependent variable of choice in this strand of research. For further details, we refer the reader to Lado & Sanz's chapter on methods in multilingualism research in volume 10 of the Encyclopedia of language and education.
In conclusion, this timeline represents laboratory research on L3 acquisition following a cognitive approach. In accordance with the historical developments in SLA, bilingualism, and linguistics over the past decades, we have identified the following major trends:
-
A. Linguistic target
-
A1. Isolated phonemes or syllables
-
A2. Artificial grammar
-
A3. Isolated but meaningful words
-
A4. Semi-artificial language
-
A5. Miniature language
-
-
B. Type of instruction
-
B1. Inspired by cognitive psychology, where implicit is memorization and explicit is rule-search
-
B2. SLA/Processing Instruction-inspired, where explicit is grammar instruction and implicit is no grammar instruction
-
-
C. Role of individual differences (IDs)
-
C1. IDs related to bilingualism: age of acquisition, level of proficiency, number of languages known, typology of languages known
-
C2. Sociocognitive IDs: strategy, learning style, metalinguistic awareness
-
C3. Cognitive IDs: cognitive control, working memory, phonological memory
-
Cristina Sanz is Professor of Spanish & Linguistics at Georgetown University, and director of summer, intensive, and School of Foreign Service Spanish programs. Her research looks into the interactions between learning contexts and individual differences in multilinguals. She is currently writing a Handbook for Teachers of Spanish with Wiley and editing a handbook on study abroad for Routledge. Her volume Mind and context in adult SLA received the 2006 MLA's Mildenberger Prize.
Jessica G. Cox is Assistant Professor of Spanish and Linguistics at Franklin and Marshall College in Lancaster, Pennsylvania. Her research interests lie at the intersection of bilingualism and individual differences in SLA, including L3 acquisition, aging, and type of instruction. She has also conducted research on language learning during study abroad. Her publications can be found in journals such as The Modern Language Journal and Studies in Second Language Acquisition.
