As Cook (1996: 126) pointed out, “most people are substantially less efficient in their L2 than in their L1”. Having learned English as a second language for virtually ten years, I knew the frustration was real; regardless of how much effort I spent, my L2 proficiency could never be compared with that of a native speaker’s. Nevertheless, I desired to investigate the underlying mechanism that differentiates L2 acquisition (SLA) from the acquisition of L1 (FLA). In this paper, I intend to argue that one plausible reason, which accounts for such difference, is that FLA involves a domain-specific mechanism while SLA makes use of both a domain-general cognitive system as well as a language-specific acquisition device.
Before delving into the discussion of domain-generality and SLA, let us first consider a question proposed by Wittgenstein (1953: 209), “Doesn’t our understanding reach beyond all the [verbalized] examples [we have learned]?” With first language acquisition, the answer is surely affirmative; we are capable of understanding more linguistic units than what we have obtained as inputs. The poverty-of-the-stimulus argument explicitly attends to this constructive aspect of children’s linguistic competence. Even with limited inputs, children’s grammatical competence can “automatically incorporate built-in linguistic structures” to yield grammatical outputs (Cook 1996: 86). According to the black box Language Acquisition Device (LAD) model, children are born with a built-in language faculty so-called Universal Grammar. In the process of acquisition, UG parameters are adjusted accordingly to best adapt to the language that the child is learning. UG can not only process linguistic inputs, but also “gap-fill” syntactical structures to generate novel sentences. All of these theories have suggested that the linguistic inputs we received cannot sufficiently justify the level of competence we are capable of achieving with our L1.
Nonetheless, let us return to Wittgenstein’s question and consider what the answer would be with a second language. Do we understand more of the L2 than what we have been explicitly taught? By interviewing a number of unbalanced bilingual friends of mine, the answer turned out to be 100 percent negative. They all expressed that their lexicon capacity and knowledge of syntactic structures are limited to what they have acquired via ostensive learning. With that said, the poverty-of-the-stimulus argument does not account for SLA the same fashion it does to FLA. In addition, as Chomsky (1995: 18) proposed, all verbalized human beings possess the ability to reach a static state of fully-developed L1 competence, and their L1 proficiency is, in principle, comparable across individuals. SLA, on the other hand, does not initiate at S0 nor terminates at Ss. As Cook (1996: 125) explained, a universally standardized static final state for L2 learning does not exist; L2 competence is a variable that differs from individual to individual. If the LAD were to be the sole mechanism that enables SLA, we should expect that with the same linguistic inputs, L2 learners should be able to gain comparable proficiency levels of outputs. However, that is rarely the case in SLA. Even with the same inputs, L2 learners’ second language competence varies from one another.
In short, neither the poverty-of-the-stimulus argument nor the black box LAD model applies to SLA the same fashion they do to FLA. It is not convincing that a biological mechanism is present to support and sustain the acquisition of an L2. This fact should cast some skepticism on whether or not SLA activates the same brain areas that are responsible for the acquisition of L1. In other words, if we were to adopt a Chomskyan view and presume that FLA is domain-specific, in that the process is specialized and leaning is not observed in other domains (Saffran & Thiessen 2007: 69), SLA, at least, should not be a domain-specific process and is highly probable that it is domain-general.
Modern technology such as fMRI scan provides a great gateway to gain insights into the areas of the brain that are activated for linguistic activities. In a study by Liu & Cao (2016: 71), it was found that an additional area of the brain was activated when L2 was in use. That additional area turned out to be the region responsible for the coordination of sensory system, and late bilinguals showcased a greater activation pattern than early bilinguals did in this particular area. In a longitudinal study by Grant et al. (2015: 44-5), multiple fMRI scan results were compared as subjects’ L2 exposure increased. Researchers found that an additional area that corresponded to cognitive control was activated, however, activation pattern decreased as learner’s L2 exposure increased. In other words, cognitive control was required in the initial stage of SLA but gradually diminished as subjects gained more L2 exposure. In a regression study by Paradis (2011: 229), non-verbal IQ and L1-transfer are both strong predictors of subjects’ L2 competence. These studies have presented empirical support to the argument that SLA requires controls under the domain-general cognitive system. Nonetheless, overlapping areas with the activation patterns of L1 were observed. In addition, activation patterns in the domain-general regions are greater for learners at their initial stage of SLA.
In sum, I have provided both a theoretical approach and some empirical evidence to argue that L2 proficiency is rarely comparable to a speaker’s L1 because a distinct set of acquisition mechanism is involved. Unlike FLA, SLA requires a domain-general processor to function in conjunction with a language-specific acquisition device.
References
Chomsky, Noam, 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Cook, Vivian, and Mark Newson. 1996. Chomsky’s universal grammar: An introduction. Cambridge, MA: Blackwell.
Grant, Angela; Shin-yi Fang; and Ping Li. 2015. Second language lexical development and cognitive control: A longitudinal fMRI study. Brain and Language. 144.35-47.
Liu, Hengshuang, and Fan Cao. 2016. L1 and L2 processing in the bilingual brain: A meta-analysis of neuroimaging studies. Brain and Language. 159.60-73.
Paradis, Johanne. 2011. Individual differences in child English second language acquisition: Comparing child-internal and child-external factors. Linguistic approaches to bilingualism 1.213-237.
Saffran, Jenny, and Erik Thiessen. 2007. Domain-General Learning Capacities. Handbook of language development. 68-86.
Wittgenstein, Ludwig, 1953. Philosophy Investigations. Chichester: Basil Blackwell.
Hi Stephanie Long,
Your argument is interesting. While I was reading, I notice that your assumption that SLA acquisition involves both domain-specific and language-specific device can be used to account for the fast speed of learning in SLA. Even if SLA of L2 speaker is not as competent as L1 speaker; however, they generally acquire it to a good level in a short time. Maybe, learning language by generating explicit rules is more similar to the process of L2 acquisition and learning language by loose cognitive patterns is more similar to L1 acquisition.
Hang Jiang
Hi Hang,
I agree with your last point, in that L2 is similar to generating explicit rules and L1 does not require that significant amount of cognitive inputs. But I think you misunderstood my main point. I was trying to argue that SLA acquisition involves a language-specific device with aids from a domain-general processor. Not everyone can learn a second language in a short span of time. I think the point I was trying to make was that even with the same ostensive teaching inputs, people acquire L2 at a different pace. I was wondering if the reasoning behind such phenomenon was because SLA requires inputs from the domain-general processer. For older people and people with lower intelligent, L2 acquisition can be difficult.
I agree that L2 language acquisition is domain-general, compared to L1 (if babies’ L1 learning requires domain-general system ). Liu & Cao’s study about brain activation is interesting in that it shows that cognitive skill decreases as L2 learner’s exposure to L2 increases. I think this phenomenon is similar to how we learn and solve math problems. When we learn a math equation for the first time, we need some time to get used to it and solve a problem. However, as we get used to it, or “exposed to” the similar question and equation, we find ourselves solving the question without a lot of cognitive control. The equation and question come out “naturally” just like L2 learners are able to express some phrases “naturally.”
Liu & Cao’s study also shows that they found similar activation pattern of L1 in L2 learners’ brain. I think it corresponds to language transfer where L2 learner brings their knowledge from L1. I believe this is where UG comes into play. L2 learners utilize their UG that they activated through L1 to learn L2.
Hi Jungha,
Thank you for bringing the analogy between L2 and a math problem. I think that truly reflects my process of learning a second language. At first, I had to translate everything back to my native language and had to guess and infer from the contexts sometimes. Then as I became more adaptive to living in an English speaking environment, I could respond faster and comprehend a conversation without having to translate. This process was also what brought me to writing this paper and looking at the differences in L1 and L2 acquisition.
I also agree with you in that L2 requires UG inputs. However, I also wonder if that process is aided by the general cognitive processor? When learning your L2 at first, you would not have learned to turn on and off the SVO/SOV/OSV word order “switch”. Nonetheless, when you learn more of the L2 and started to observe overlapping syntactic patterns between your L1 and L2, you have learned to utilized UG to activate through L1 to learn L2. What is your thought on this？