The Polyglot Brain

[rdearman]

https://www.science.org/content/article ... -languages

[Le Baron]

“When you become a specialist at something, you use fewer resources,” Malik-Moraleda says. The study suggests reaching peak cognitive efficiency may be more likely when it’s learned at a young age.

Yes. Though 'specialist' and learning something from a young age might not necessarily be the same thing. I'm thinking more along the lines of 'initial tool for communication', and that this 'first tool' just gets sharpened by long and varied use. It makes sense that habits don't need much work, but that's not really a 'specialist'.

[Iversen]

So in essence your mother tongue is characterized by requiring less brain activity than later acquisitions, but all are dealt with in the same brain area. The difference between the mother languages (in the plural) of bilingual native speakers might also be worth investigating - I once commented on an Italian study that showed that you have to learn a language before the age of five to achieve the lowest amplitudes, so that part of the new results was as I expected. But you can still learn to function at an excellent level even though your brain has to work slightly harder, and it should by now be a proven fact that the window for achieving a reasonably high level doesn't close entirely at the onset of puberty. Maybe pronunciation is harder to learn than other things like vocabulary and grammar, but the reasons for that should also be investigated rather then assumed to be god's will.

To see the influence of time and exposure you would need even more participants in the study and then make a statistical analysis. OK, It must be hard to arrange for double blind studies when we discuss polyglotism, but it could be done for specific study methods, including those you can do without a teacher (even the trade union of language teachers might not like the idea). But even being a desceriptive project rather than an experimental one, this study is already a fine contribution to a sorely neglected topic in science. Long ago (in the HTLAL days) I looked through a complete list over the final dissertations written by students at the 'pedagogical highschool of Denmark' (arguably an institution at a university level), and while there were loads of studies of the problems of bilingual school children (i.e. those from immigrant families) I didn't find one single one about polyglots and/or home learners - which suggests a massive level of herd mentality and possibly censorshop at that particular school. Later on I have read about a few research projects abroad (and I also bought the Babel book), but as it is I wouldn't trust the real expertise concerning polyglottery and/or the efficiency of home study from mainstram pedagogical researchers. How can they know anything if they don't do research in a subject?

[Irena]

Long ago (in the HTLAL days) I looked through a complete list over the final dissertations written by students at the 'pedagogical highschool of Denmark' (arguably an institution at a university level), and while there were loads of studies of the problems of bilingual school children (i.e. those from immigrant families) I didn't find one single one about polyglots and/or home learners - which suggests a massive level of herd mentality and possibly censorshop at that particular school.

Not necessarily. The problem is that studying polyglots is really difficult. You first need to define your terms. Who exactly counts as a polyglot? The article that rdearman linked to says that a polyglot is someone who speaks more than five languages. This is not a universally agreed-upon definition, but they had to start somewhere. Okay, and how well do you need to speak those "more than five" languages in order to count as a polyglot, and who is going to test this and how? And what is a language and what is a dialect? The other problem is that in Western countries you simply aren't going to be able to find very many people who speak more than five languages to any reasonable level. You're more likely to find such people in India or Africa. Meanwhile, researchers who might study these topics are mostly in the West. So, finding research subjects is difficult. Plus, your research subjects may be highly heterogeneous, making any sort of statistical analysis impossible. And that's why people hesitate to do research on the topic...

[Iversen]

OK, it may be difficult to do statistics on such a heterogeneous population as 'polyglots' (whether you set the limit at Hudson's 5 or at 11, like Erard did), but even descriptive studies like the one discussed here would be valuable. As for the results of home study it should be easier to do some statistics, but since you can't hide from the participants which group they are placed in it may be impossible to do the classical double blind experiments. But even statistics on the results of non-blind tests might be interesting. The annoying thing is that as I remember the Danish dissertations there was not much that even hinted at the remote possibility of studying languages at home without a teacher, and that's where I see some kind of bias (though after all the years that have passed since I looked through the list I may also have become more biased ... in favour of homestudy).

[Irena]

OK, it may be difficult to do statistics on such a heterogeneous population as 'polyglots' (whether you set the limit at Hudson's 5 or at 11, like Erard did), but even descriptive studies like the one discussed here would be valuable.

Descriptive studies are valuable - up to a point. Sure, you can do brain scanning, which you may then be able to publish. (Still - tiny sample size.) But anything else is going to be "anecdotal," meaning that other academics will have little trouble tearing it apart, and that's bad for an academic's career. Journalists can write books about polyglots, for which they then get tons of criticism, but that may be fine with the journalist, as long as the book sells well. Now put yourself in an academic's shoes: why wouldn't you leave polyglots to journalists?

As for the results of home study it should be easier to do some statistics, but since you can't hide from the participants which group they are placed in it may be impossible to do the classical double blind experiments. But even statistics on the results of non-blind tests might be interesting.

Yeah... But if you want to do any sort of statistical analysis on language learning, then you have to control for all sorts of things, starting with the study participants' IQ. This isn't impossible to do, but again, are you going to be able to find enough participants in order to get meaningful results?

[Iversen]

Descriptive studies are valuable - up to a point. Sure, you can do brain scanning, which you may then be able to publish. (Still - tiny sample size.) But anything else is going to be "anecdotal," meaning that other academics will have little trouble tearing it apart, and that's bad for an academic's career.

If that's all that is possible then let's by all means get that anecdotal research - and let it be published with the caveat this this may not be hard science (liek physics or chemistry), but that's what can be done. I have a past in the academic humanities, and I would say that there is next to nothing there that would qualify as hard science (except maybe certain parts of grammar and language history, and well, maybe not even those things) - it's all descriptive and anecdotal. And psychology is divided into two parts: anecdotal (and mostly biased) 'research' and hardcore studies of those peripheral details that can be studied with 'hard science' methods (like brain scans), but don't answer the questions that ordinary people expect answers to. Sociology (with its more or less fashionable subdivisions) is structured in the same way.

But if you accept that humanitites and the various subdivisions of psychology and history and antropology and many other things shouldn't be thrown on the midden of history then it should also be possible to find space (and forgiveness) for a few studies of the curious fact that some manage to learn more than language and some even learn them without the assistance of a teacher.

As for criticism from peers ... well, everything that is published within the areas I just mentioned can be attacked, but if the attackers are even less scientific themselves then so be it. Just as a curiosum: when I studied French I once wrote a rather extreme structural analysis of El Desdichado by Gerard Nerval. The responsible teacher (whose bias was more in the direction of psychoanalysis) rejected it with one silly advice - that I should try to imitate a certain analysis done by a more established name. My answer to that was to publish my own analysis unchanged in the institute magazine and write a new essay for another teacher. But after that I had to warn some of my co-students not to deliver a structuralistic analysis of any kind to teacher no. 1.

So if you can't make all happy then at least be brazen about it.

[Irena]

Iversen, here's a very simple question: when it comes to the study of polyglottery, what exactly can an academic do better than a journalist could? Neuroimaging is the obvious answer. Okay, other than neuroimaging? I'm not sure there is much of anything else, to be honest.

[Iversen]

Study study methods?

I have done some testing on myself, and because I'm in a double role it definitely doesn't qualify as science. But in 2014 I wanted to know whether my triple-column wordlist layout functioned, so I went through a Serbian dictionary with some 15.000 words from А to Ш and included most, but not all words in my lists. Then I took sample pages before and after the point I had reached and subdivided the words into known and not known (in later wordcounts I have introduced a middle group called 'la-la'). The basic result was that I would know (or be able to guess) about a third of the words before the use of a wordlist, but two thirds after. Which would mean that I had learned 5000 words in two months.

There are all possible sorts of things wrong with this approach, first and foremost that I tested myself, and that there wasn't a clear criterion for how well I should know a 'known' word - and also that I didn't check whether I also remember the words a half year later. It is also a clear weakness that I asked myself about the words after I had found them in the dictionary - but the methods I used to check my short-term memory (with 2-3 rounds and the last one 2 weeks after the wordlist date) did to some extent suggest that my judgments weren't too impaired by this. And if so, then I can't see what else I could have done - except maybe involve a native bilingual speaker as my judge.

As I said I did do a primitive test on the short-term effect. In my wordlists I have columns for target language base language (here English) and target language again, and I checked my memory by covering the columns with translations in them and running through the words in the first column after one day, and by and large i remembered (or could guess) at least 80% of the words without reading the translations - and not much less after 2 weeks, but the words I remembered then might not be exactly the same as those I had remembered after one day. It is nevertheless interesting that the level didn't fall like a stone in water during those two weeks.

However the whole thing hinges at the problem: how well do I have to remember the meaning of a word - must I come up with exactly the same word (or words) as in the dictionary, or could a synonym or a Danish word with more or less the same meaning be accepted?

Well, anyway this wasn't science, and I have never pretended that it was, but the setup could be transmogryphed into something that wasn't too much less scientific than most of the things pedagogical researchers have produced. And the lack of 'blindness' could to some extent be remedied by letting two groups try first wordlists and Anki and then Anki and wordlists.

Counter question to Irena: can you point to anything done by language researchers that by your standards is both useful AND scientific? And would you blindly trust anything written by a researcher that hadn't studied a certain topic and maybe had a preconceived meaning about it? In that case it wouldn't be much worse to rely on a journalist - I have seen so much rotten unscientific babble from serious employees of universities that I'm no longer impressed just by an academic title (and that includes my own).

[Irena]

Study study methods?

How? Without a sufficiently large sample size, it's all going to be anecdotal. And if it's anecdotal, then why not leave it to journalists? They're probably going to be able to write a more interesting story anyway.