Grammatical habits in written English reveal linguistic features of non-native speakers’ languages

Discussion in 'General Discussion' started by biTsar, Aug 19, 2014.

  1. biTsar

    biTsar Active Member VIP member

    Joined:
    Apr 22, 2014
    Messages:
    237
    “The striking thing about this tree is that our system inferred it without having seen a single word in any of these languages,” Berzak says. “We essentially get the similarity structure for free. Now we can take it one step further and use this tree to predict typological features of a language for which we have no linguistic knowledge.”

    MIT News: Essays in English yield information about other languages
  2. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    Very cool. The main problem I see with the idea of using it to discover stuff about lesser-studied languages is that a lot of those languages will be languages that exist in multilingual areas, so your data will be polluted by second and possibly third languages.

    But more importantly it provides genuine data that Krashen's "universal order of acquisition" is nonsense, and that there is indeed such a thing as L1 interference.
    biTsar likes this.
  3. Peregrinus

    Peregrinus Active Member

    Joined:
    May 27, 2014
    Messages:
    613
    Native Language:
    English
    Intermediate Languages:
    German
    Basic Languages:
    Spanish
    I am not particularly invested in all of Krashen's theories, including the natural order hypothesis. However I am not sure you are correctly drawing conclusions from the study above re the natural order hypothesis.

    The essential aspects of the study linked are 1) probabilistic predictions of a learner's L1 from his/her L2 writings, and 2) quantitative measurements of the degree of similarity between languages through machine algorithms. The latter is the main focus of the article above. Krashen's theory posits that generally some grammatical structures are acquired sooner than others across all languages. But he did not say perfectly acquired, and the article does not speak to the order in which various grammatical structures are learned. The only use of the concept of "order" it mentions is regarding sentence order, like SOV etc., and it does not discuss the chronological sequence of learning them.

    So it is difficult for me to see how this article applies to the natural order hypothesis at all, since it doesn't purport to measure the order in which L2 learners learned various grammatical structures.
  4. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    OK, as usual I'm just looking for another opportunity to call Krashen an idiot (he is). To me, the main outcomes of the study seemed rather obvious. I've generally been fairly confident in identifying texts written by Spanish and French speakers, and that through language interference.
  5. Bob

    Bob Active Member VIP member

    Joined:
    Apr 18, 2014
    Messages:
    339
    Native Language:
    English
    Intermediate Languages:
    Cebuano, Greek_ancient, Spanish
    Basic Languages:
    Filipino, French, Hebrew_clasical, Mandarin, Portuguese, Russian
    no way. They is be making that up.
    biTsar likes this.
  6. Peregrinus

    Peregrinus Active Member

    Joined:
    May 27, 2014
    Messages:
    613
    Native Language:
    English
    Intermediate Languages:
    German
    Basic Languages:
    Spanish
    We all have our favorite axes to grind. I envision your flat having a row of small punching bags with labels like "Benny", "Krashen", etc. If you had their faces screen printed on the bags it would be even more awesome :).


    Yeah determining a writer's L1 through L2 writings doesn't seem particularly noteworthy. But determining similarities between languages through computer analysis does seem to be an important result in that it appears to be at least as accurate as the comparisons made by linguists manually. It would especially be interesting to use their program to compare language isolates quickly to a database of other languages to see if there are any connections trained linguists were unable to detect to any degree of certainty.
  7. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    Unfortunately I had to bin them when I moved away from Edinburgh three years ago. I was always meaning to learn Flash programming to replace them with a little game, but never got round to it.
    Big_Dog likes this.

Share This Page