This sabbatical project aimed at
(1) formulating syntheses of our more than 20 year research program on age of acquisition and advanced second language learning,
(2) reporting our recently completed studies on the relative impact of early/late age of acquisition vs. mono-/bilingualism on near-nativelike language learning, and
(3) identify new directions for research on maturational constraints and critical/sensitive periods for language learning.

The focus has been on how the study of near-native language learning – in childhood as well as adulthood – can provide important clues to how brain maturation affects learning ability, and how the phenomenon of nativelike language learning/use at different ages of acquisition can be used in social-psychological research on how expectations and stereotypes affect our perception, memory and assessment of second language speakers’ proficiency. The project aimed at four internationally published articles, but resulted in six, three of which have been published, while three are still being reviewed by international journals. All articles/chapters will be made open access via the last revised manuscript version in Diva; some may also be freely available on the journals’ web pages. Below is a summary of the project’s scientific contribution, with reference to the six articles and chapters.

In a chapter entitled “Age effects on language acquisition, retention and loss: key hypotheses and findings” [1], we report the empirical results of our now 20+ years research program on age, maturational constraints, and the the classic hypothesis of a critical period for language acquisition (The Critical Period Hypothesis, CPH). Here we discuss systematically – on the basis of our empirical data on early and late second language learning, first language attrition, language learning and ‘first language remnants’ in international adoptees, language learning aptitude, implicit and explicit learning, and the relative impact of monolingualism and bilingualism – how our published results relate to the criteria for the falsification of CPH as set out in the international literature. We find that age of acquisition (AoA) is the most crucial factor for ultimate attainment in a second language (compared to, e.g., length of residence/exposure and language use); that the relationship between AoA and ultimate attainment is discontinuous rather than linear, and that AoAs 6–7 and 12–13 y/o, respectively, represent typical breakpoints (depending on the linguistic phenomenon studied); that those very unusual adult second language learners (but also average young learners) who in everyday language usage are perceived as native speakers of their second language, in fact prove to be non-nativelike when their language competence and behavior is analyzed in detail; that learners’ delayed exposure to the second language, rather than their bilingualism, is the main reason for non-nativelikeness; and that adults’ learning and processing of the second language, to a greater degree than in children, is compensated via the explicit/declarative memory system rather than handled by the implicit/prodeural system, which is supported by the fact that highly advanced/near-native adults always turn out to possess an innate language aptitude above average, and that their development of different linguistic phenomena (e.g., phonetic perception and grammatical intuition) seems more sporadic and non-parallel (unlike children, who develop all aspects of the language in parallel). Even though we refrain from claiming that our studies constitute definitive evidence of maturational constraints and a critical period for language learning, we dare to claim that our results cannot be used to reject the CPH.

A clear trend in the literature over the past decade has been to interpret the non-nativeness of advanced second language learning (in both childhood and adulthood) as an effect of the bilingualism of these individuals – not of their late AoA. One position often expressed is that monolingual-like mastery is not to be expected from bilingual individuals, since monolingualism and bilingualism must be seen as separate competences, and that the current practice of comparing second language speakers with monolingual control groups constitutes a ‘monolingual bias’. One frequently proposed (but hitherto never implemented) methodological feature is to engage simultaneous bilingual individuals – that is, people who, from birth, were raised bilingually in the home – as control participants. In the same spirit, it has been suggested that individuals who are adopted into a new language environment in childhood can learn the new language effortlessly, quickly, and to completion only because their first language is lost within a few months after adoption (as opposed to childhood learners who maintain competence and use of their first language). The hypothesis is that the neural network subserving language can be “reset” by the abrupt and total language loss, which, in turn, allows for monolingual learning. In our article entitled “Age of acquisition – not bilingualism – is the primary determinant of less than nativelike L2 ultimate attainment” [2], we argue against these positions with results from a large-scale study of 80 participants who’s production and perception of Swedish was investigated with 7 different linguistic (phonetic, morphosyntactic, lexical) instruments. With a 2x2 factorial design, the ultimate attainment of monolingual first language speakers of Swedish, simultaneously bilingual first language speakers of Swedish and Spanish, sequentially monolingual second language speakers of Swedish (adopted at ages 3–8 y/o from Spanish-speaking countries to Sweden), and sequentially bilingual second language speakers of Swedish (with Spanish as first language, who immigrated to Sweden in childhood at ages 3–8 y/o). The results clearly indicate that AoA is the decisive factor, while bilingualism has only limited or no effect (depending on the linguistic phenomena being studied). In an article entitled “Revisiting the lexical deficit in bilingualism: The impact of age of acquisition” [3], we report on data from two additional lexical tests from the above study, and the results point in the same direction. An adjacent discussion is found in the article “But first, let's think again! (Commentary on Mayberry & Kluender)” [4]. In the keynote article which I comment on, the authors argue in a similar theoretical direction as described above, and suggest that AoA effects and a critical period are relevant for first language acquisition only, but considered irrelevant for second language learning. My commentary article focuses on the fact that studies of apparently native-language second language speakers have generally used test instruments and analyses that were insensitive to the small differences that indeed exist between native and near-nativelike language proficiency.

In addition to syntheses of our previous and ongoing studies, this sabbatical also aimed at identifying and initiating new directions and promising pathways for research on age, maturational constraints, and critical/senitive periods. In an article entitled “Early vs. late-acquired near-nativeness as a unique ground for exploring age effects in SLA: A research agenda” [5], I make explicit proposals of concrete research topics and designs within three inter-related sub-inquiries: (1) whether AoA or bilingualism is the cause of these individuals near-nativeness, and whether their relative impact varies among linguistic phenomena; (2) whether younger and older second language learners’ near-nativeness rests on the same or fundamentally different neurocognitive systems; and (3) how native listeners reactons to near-native speech, and how stereotypes and expectations about age and language acquisition influence their perception and memory processes. As regards this last inquiry, the last article, entitled “Native listeners’ perception and reconstructive memory of near-nativelike speech: effects of expectancy violations and stereotype incongruites” [6], reports on an empirical “verbal guise” study of how 90 native speakers of Swedish evaluate, judge, and remember one and the same audio clip of an adult second language speaker who speaks with a very subtle foreign accent, depending on whether the pre-information to the clip indicated that the speaker immigrated to Sweden in childhood or in adulthood. The results showed that the listeners’ evaluation and recall of the speech sample was affected by the information, but not their actual, objective assessment of the speech.

In addition to the publications, the project has resulted in one Master’s thesis, one new PhD project, the formation of a small research group at the Center for Research on Bilingualism with a focus on perception of accent, and a systematized speech material that will be available for students doing Bachelor’s and Master’s projects. A new specific research question that was raised in the project is whether a so-called ‘lexical deficit’ characterizes bilingual individuals (as currently suggested in the literature), or if this is an expression of publishing bias and an overlook of the age factor. This question has resulted in a project application (which will be submitted in early 2019) together with a research group from Åbo Akademi University, Finland. Finally, during my one month stay in Canada (Simon Fraser University, Vancouver, and University of Alberta, Edmonton), I gave three guest lectures on the reported studies and articles, and made valuable contacts with a number of linguists, phoneticians and psychologists working on second language acquisition and/or bilingualism.