In the episode, we talked about what happens people learn a third language, or L3, and particularly about whether learners transfer material over from their first language, their second language, or both. But we didn’t get a chance to talk about the Scalpel Model of L3 acquisition, which is too bad, because beyond being an amazing name, it’s also a really cool set of ideas for how learning an L3 works.
So we touched on a few studies in the episode that show that when we learn an L3, it looks like there’s evidence in some cases for transfer from the L1, and in other cases, for transfer from the L2. The Scalpel Model aims to explain how you get transfer from both: it hypothesizes that transfer for the L3 can selectively draw on either existing language, and that it works property by property. Each of the grammars you already have in your head can be broken down into sub-grammars and tagged for what language they came from. And then, you can slice out relevant grammatical properties, and extract them for the new grammar you’re building. Under this model, neither the L1 nor any subsequent language is given a privileged position: pulling from any of them can work, as long as it’s the property you want.
But that transfer isn’t always helpful. If you identify a language as being similar, that doesn’t mean it is, for all the properties you’re concerned about. If you have English and Spanish, and you’re learning French, typologically, you’d figure it to be closer to its fellow Romance language. But Spanish allows you to drop the subject if it’s clear who you’re talking about, where French doesn’t. And you can find L3 French speakers dropping subjects like Spanish, even if they have English around already.
The thing is, scalpels may be good for precise work, but they have limits: they can’t cut through bone, for example. And so the model doesn’t expect everything to be reducible to transfer. How clear and unambiguous or prevalent the L3 input is, or how frequent or complex a construction is, those factors matter, too. For example, a pair of studies looked at speakers of Basque and Spanish (in both orders) learning English. In one, they looked at the ability to leave out objects: you can do it in Basque, you can do it in limited situations in Spanish, and you can’t do it in English. But because the evidence in English is so enormous in favour of needing objects, even when it’s super clear what the meaning would be without it, neither L2 group has problems working this out.
On the other hand, both groups did fairly badly with how English topicalization works. So topicalization is when you front and stress an argument to point up a contrast: take an exchange like this:
- Did Toby fight the sea witch?
- The sea witch, she didn’t fight. She did fight the queen.
So there, we can see “the sea witch” moves up to show the mismatch with the question. And Basque behaves similarly to English on this front. Spanish, on the other hand, has a similar-looking construction (left dislocation) that functions differently in terms of the syntax. If we just got to pick out the stuff that matches, this should be easy to learn for our L3 groups: they could just cut out the Basque construction and adopt it for English. But the Spanish look-alike is about 1000 times more common in usage, so its frequency swamps out the more closely matched option. And so, the L3 learners don’t do great with identifying good English topicalizations. The model covers similar findings regarding language activation and use, as well.
The thrust of the Scalpel Model is that while transfer can get quite sophisticated for what we can take over to form the foundation of our new languages, the realities of language frequency, use, or inaccurate typology comparisons can lead to less native-like outcomes for more language learners. It’s a hypothesis that tries to capture why having more languages really can be helpful, but also can cause problems, depending on the specific linguistic mix. As L3 acquisition keeps getting more research, I think this model’s going to have a lot to say!
So how about it? What do you all think? Let us know below, and we’ll be happy to talk with you about what happens when we learn a bigger pile of languages. There’s a lot of interesting stuff to say, and we want to hear what interests you!