Picking Up a Second Language Is Predicted by Ability to Learn Patterns

psychologicalscience.org/ind … terns.html

According to this article, how well and fast you can pick up another language depends largely on one’s ability to learn statistical patterns…

[quote]The data revealed a strong association between statistical learning and language learning: Students who were high performers on the shapes task tended to pick up the most Hebrew over the two semesters.

“It’s surprising that a short 15-minute test involving the perception of visual shapes could predict to such a large extent which of the students who came to study Hebrew would finish the year with a better grasp of the language,” says Frost.[/quote]

hmm… now i want to take that test…

[quote=“hansioux”]http://www.psychologicalscience.org/index.php/news/releases/picking-up-a-second-language-is-predicted-by-the-ability-to-learn-statistical-patterns.html

According to this article, how well and fast you can pick up another language depends largely on one’s ability to learn statistical patterns…

[quote]The data revealed a strong association between statistical learning and language learning: Students who were high performers on the shapes task tended to pick up the most Hebrew over the two semesters.

“It’s surprising that a short 15-minute test involving the perception of visual shapes could predict to such a large extent which of the students who came to study Hebrew would finish the year with a better grasp of the language,” says Frost.[/quote]

hmm… now I want to take that test…[/quote]

Students who STUDIED. The same universe of thought that believes a bell curve is a normal outcome for language – a universal human ability. Acquisition does not rely on being a high performer on a statistics test.

People are cognitively biased Bayesian reasoners about almost everything else. That we do the same thing with language to some extent isn’t too amazing.

“Numerosity” is a universal human (and crow) ability, as well, but we can obviously measure whose understandings of mathematics are deeper and more natural to them, and there is a bell curve for mathematical proficiency. Some of it is innate brain structure (that is, young, professional mathematicians usually demonstrate a high acumen for it in their youths), but some of it is learned knowledge, especially since mathematics doesn’t carry intuitively past our fingers or assisting diagrams in most cases (though it does for people like Ramanujan). Part of it is innate, but that doesn’t mean that a substantial part of it isn’t learned. If UG turns out to be a series of protocols for filtering an open array of parameters, and those parameters are set for languages via induction (or selective falsification), then it seems to argue for some Bayesian reasoning affecting language acquisition, as well.

The same is true with lots of partly innate abilities (like pro-social behavior). Some of it may be hardwired. Some of it is obviously learned. (To counter absolute innateness of pro-social behavior, we still don’t answer trolley car problems intuitively and consistently.)

Chomsky, one figurehead of extreme linguistic nativism, more and more, is relying on “just thinking about these things” for getting to the correct understanding of the acquisition v. learning distinction, in part because the neurological research isn’t supporting his exact view anymore, no matter how many times he just states that the research is “empty” (which in Chomskyanese means that he doesn’t agree with it).

No, people are cognitively based reasons when they’re put into a situation where they must overtly analyze – like a language class, or someone trying to teach language based on comparison of a million sentences. Every normal human being has a native language. There is nothing preventing any human being from acquiring a second one, absent studies like this that make people believe up front that they “can’t”.

I think that you’ve misread “biased” as “-based.” We do suffer cognitive biases, but we appear to use some form of Beyesian reasoning with other cognitive interferences.

Immaterial. The point is that people acquire languages without any regard to whether or not they are geniuses at, or suck at, statistics, or math, or any other sort of logic. The most illogical people are perfectly fluent and have just as good an ability to acquire language as people who think in symbols. The study is looking at people STUDYING language, not people acquiring it. This is a measure of how well people succeed in school, not at how well they acquire language.

Do you believe that language acquisition is inductive?

Not in parallel to math or statistics.

That’s immaterial to the article. Bayesian induction isn’t necessarily a conscious procedure, not any more than calculus is when you have to orient yourself in space to catch a ball. There’s a difference between the the knowledge involved in formally representing a procedure, and the knowledge obtained through exposure to and engagement with a procedure; but in all other fields, training in either end usually improves abilities in the other. That is, some level of formal instruction on riding a bike does, along with actual practice riding a bike, improve people’s abilities to ride bikes.

This research could very well bear out that any pattern detection training primes one to understand languages better, even if the procedure has nothing to do with languages.

My point is that the research is NOT testing the acquisition of language – it’s testing a very small sample who are being taught language in a specific, overtly analytical manner. That’s skewing the results, and based on the fact that language is a human universal, the results are counterintuitive.

Acquisition of language is a human universal, as a child. I can throw a 4 year-old who can’t speak a word of Chinese in a kindergarten and they’ll be speaking on the level of the peers in a matter of months. Take an adult and put them in a similar environment and they’ll learn nothing. The neural mechanisms behind acquisition during childhood and adulthood are clearly different, so pretending that seeing as we all learn as children, we can all equally learn as adults, and in the same way, is quite obviously not accurate. Behind anything difficult in life is someone saying that it’s actually really easy (and they’ll show you, for a fee of course), when in reality it just IS difficult.

I’m not really surprised that the ability to recognize and learn patterns is predictive of 2nd language success, but there’s clearly more to it. I know intelligent people that suck at language, and others who are quite decent but can barely tie their own shoelaces.

This is not really entirely true.

The neural mechanisms behind language acquisition are not “clearly different” between adults and children. There’s no proof of that. The issue of accent indicates that the phonological neurology may be somewhat more limited later on, but that too depends on input and quantity and quality of input. Kids get 10,000 hours of the native language by age 3. Adults rarely get that much high-quality targeted input, so really the comparison is not moot because it’s adults vs children, it’s moot because the quality of the input is so different.

If you give an older learner lots of high-quality, 100% comprehensible, targeted input, they will easily acquire the language. We do it all the time. We do NOT just “throw them in” – that’s the point. Immersion doesn’t work for most adults. “Optimized immersion” does. The key is establishing meaning, which immersion typically fails to do.

And just FYI, I have shown people in Taiwan that it really is easy – and I didn’t charge them, either.

[quote=“R_jay”]Acquisition of language is a human universal, as a child. I can throw a 4 year-old who can’t speak a word of Chinese in a kindergarten and they’ll be speaking on the level of the peers in a matter of months. Take an adult and put them in a similar environment and they’ll learn nothing. The neural mechanisms behind acquisition during childhood and adulthood are clearly different, so pretending that seeing as we all learn as children, we can all equally learn as adults, and in the same way, is quite obviously not accurate. Behind anything difficult in life is someone saying that it’s actually really easy (and they’ll show you, for a fee of course), when in reality it just IS difficult.

I’m not really surprised that the ability to recognize and learn patterns is predictive of 2nd language success, but there’s clearly more to it. I know intelligent people that suck at language, and others who are quite decent but can barely tie their own shoelaces.[/quote]
If you can put an adult into a learning environment where they get the same amount of input at their level of interest and ability then they will be great in a matter of months. The problem is that there are no kindergartens for adults and putting an adult in a kindergarten won’t work either. There are ways to try and replicate that, but the problem is that adults have a natural tendency to revert back to what they know to convey meaning whereas kids do not have the sophistication to do that. The level of input in a kindergarten is also remarkably different than what an adult would deal with in immersion. The problem lies in the environment, not in the person. The idea that kids can acquire up to a high level in no time but adults can’t is nothing more than folklore.
A thread on forumosa is hardly a scientific discussion but I agree with the research. The ability to recognize patterns is key to language learning. Those patterns are not necessarily an entire sentence but could be chunks of language which is why the TPRS guys teach phrases or chunks, the TBL and PPP guys are keen on collocation in vocab teaching and have focus on form and there are methods like the Lexical approach. They all try to replicate an element of real life and comprehension, which is what is necessary for learning as much as processing what you receive in input. Adults who can think outside of their own language and notice those patterns learn language easier? It appears very plausible.

Standing upright is a human universal. People with naturally stronger, uncompensated postural muscles tend to stand with better ease, posture, and comfort. We can intervene on that, too, by actually training people to adjust their seated posture first.

Acquiring languages is a human universal. Languages are patterned strings. Therefore, people who can do that patterning more easily can learn languages more easily. Perhaps we can intervene on that, too, even without language. That would be a very basic experiment to run.

Claiming that the results are counterintuitive is counterintuitive.

But this was all that I needed (for the deductive argument):

Humans learn languages by induction.
Induction, for humans, is Bayesian inference.
Therefore, humans learn languages by Bayesian inference.

Or this (which I’ve weakened to an inductive one for accuracy):

Languages are patterned strings.
Smith can detect all previously tested patterned strings more quickly than Jones can.
Therefore, Smith can learn/acquire the next patterned string more quickly than Jones can.

But here’s some enlightening listening on the problem of taking “counterintuitiveness” seriously:

This is a research study that tested a very limited condition (the ability to learn a language in a structured class, most likely grammar-based, probably based on the results of a discrete item paper-based test) and related it to another condition (the ability to recognize patterns in a test situation) and then generalized wildly to conclude that pattern recognition = ability to “pick up” (horribly imprecise phrasing, too) a new language.

That’s irresponsible research in stating such a “conclusion”, but it’s the trend in popular science.

I take this as the purport of the article, and it’s nothing as strong as the conclusion, “Pattern recognition skill equals new language acquisition ease,” so I don’t think that the writers of the study or the article are where the irresponsibility lies (but maybe I’m reading too deeply into your abuse of the symbol ‘=’). It smells suspiciously like straw-manning the article.

Are you griping about the imprecise phrasing of a two-page summary of a psychological study, and are you just guessing about how they conducted the study? Whatevs… I just e-mailed the guy for the actual study.

Also, it’s not irresponsible research when they guard their conclusions with claims that their causal claim is just one more feasible response in light of the research that they conducted.

I am familiar with academic research, thanks.

[quote=“ironlady”]This is a research study that tested a very limited condition (the ability to learn a language in a structured class, most likely grammar-based, probably based on the results of a discrete item paper-based test) and related it to another condition (the ability to recognize patterns in a test situation) and then generalized wildly to conclude that pattern recognition = ability to “pick up” (horribly imprecise phrasing, too) a new language.

That’s irresponsible research in stating such a “conclusion”, but it’s the trend in popular science.[/quote]

you use the words likely and likely when saying what they did, then accused them of being irresponsible for stating a conclusion. But if you don’t know what they did, surely you are the irresponsible one jumping to a conclusion.

A pattern I have seen on here is you guessing ( educated or otherwise ) what others fob and running them down.

I stand by what I said. I know how to read research and the difference between research and press releases aimed at the general public. I also know the difference between “learning” and “acquisition”, which the people at the “Association for Psychological Science” do not seem to feel is germane to an endeavor relating to language. It is extremely germane.

This web write-up is popular science. If you want to deny the fact that this study is linking only results from the formal traditional classroom-based study of language and statistical ability, go for it. But that’s precisely what it’s doing, based on what they are saying. And it fools the majority of people because they aren’t going to get the journal. Next thing it’ll be a Facebook meme or picked up in people’s RSS feeds, with the soundbyte getting smaller and smaller each time.

Ok. I am interested. The only “research” study I can find on TPRS is in IJFLT of March 2012 and the students performed equally well in reading and listening but the TPRS students were better in speaking and writing. Problem is that the procedure required the author, who is clearly pro TPRS, to teach both groups and thus I think he probably didn’t do as a good job with the other class. That negates the results right there and you would know that since you know how to “read research”.
It is very possible to have comprehensible input in a variety of methods, so now I am dying to have you post a link or a title of a research study that clearly shows that TPRS is better. Until then, please stop referring to what you are doing as “true” acquisition and what everybody else is doing as rubbish.