I don't yet do much writing, but he seems to make a lot of sense in saying that grammatical corrections have relatively little impact/use for the learner, but that corrections for more natural phrasing and word order do. However though he does not mention learner level, I suspect that useful type of correction is only useful for more higher intermediate and advanced learners. If one still struggles with grammatical accuracy, the meaning of individual words in context, etc., then usage is likely to be i+3 or something and not yet as comprehensible. I have seen various comments though that such corrections for more advanced learners are harder to come by on Lingq or Lang-8, i.e. when one is past grammatical corrections for gender, case, etc. However Fasulye said in another thread here that on i-talki, at least for German which she corrects and the languages she submits corrections for, one can find such help and without waiting too long for such a correction to be made.
I've definitely seen improvements in advanced learners preparing for written exams. The important thing here is that the correction is often fairly focused, and exam essays are pretty formulaic. This means that a lot of the feedback will be directly relevant to the next task. I certainly agree that (semi-)free writing is a pretty dangerous task to put in front of less advanced learners. They very often overreach themselves grammatically, meaning that they really don't have a solid basis on which to understand the corrections. I also feel that correction is better given face to face than as written feedback, as often the student will need further explanation, and in fact students normally ask lots of questions like "but can I say that in situation X?" "Would that be ... in the past tense?" etc.
Agreed! I used to post texts for correction on italki quite often, but I find that I get so much more out of spending 5 minutes of a Skype session with a native speaker correcting my text. Often things aren't wrong, but they're better expressed in a different way. I do think that getting corrections on writing is a very valuable tool. Regardless of whether you're writing at a beginner, intermediate or advanced level, I think that the key to getting good corrections on sites like italki or lang-8 is to keep texts short. Rather than writing a full paragraph, just write a few sentences - you'll get more (and more detailed) corrections.
To expand further on this point, it's been a while since I've written anything for correction, but this has been my experience with my own learning as well at the lower stages. I would write something with a specific grammatical construct in mind to practice, only to have a correction come back that is completely unrelated to what I was after. This happened on both Lang-8 and Busuu. Worse yet, I'd submit something for correction in Turkish, only to have an Azerbaijani correction with Turkish orthography. Granted, I wasn't always clear that I was after specific usage practice, but too often the corrector is every bit to blame as the person submitting the writing. R. ==
I was interested to hear that (according to the video) there are so many studies out there showing corrections aren't helpful. I've heard about this being the case with corrections made during conversation too. If you just hear the correction and move on, there's probably some benefit, but it might be hard to prove. And with writing, if I write something, get it corrected, and just glance at the corrections briefly before going onto the next writing, I suspect it's the same case. But what if you spend time making sure you understand each correction? What if you go review the grammar and vocabulary behind the errors, and write another essay with the intent of using those things correctly? What if you use the reviewed errors to fix problems with your conversation? My point is, corrections to writing can be very helpful, depending on how they are used. To be fair, Steve eludes to this a little. But he seems so eager to say it isn't helpful that he neglects to define what exactly the learner is doing with the corrections. I think the same is true for corrections made during conversation. Casually noting the correction and moving on probably has some minor benefit. But when I write down the words and grammar I don't know, misuse, etc, and study them later in isolation, then I feel the benefit is much bigger. It all depends on how much time and effort I'm willing to put into it. Bottom line, the warning that just getting something corrected isn't a silver bullet is a good one. But making a general statement that writing corrections aren't helpful could be harmful to beginners.
When I first started writing on sites like busuu, lang-8 and italki, the corrections were often good but didn't really help me to progress much. I finally decided that the problem was often with me, not with the people doing the corrections. Now, I try to follow certain rules and find that in general the corrections are very useful. 1 - Keep it short. You get more and better corrections that way. 2 - When possible, write about a topic that the person correcting is likely to be familiar with. That way, they will be in a better position to follow your logic and to give suggestions about phrasing, word usage and expressions. 3 - Don't be too ambitious. Think about the idea of comprehensible input (i + 1) and change it to comprehensible output. I try to write at a level that I am comfortable with, just pushing a bit here and there, trying out things that I am not quite sure about. The correctors usually know exactly what I am trying to say and give good, focused feedback. 4 - Try not to translate. What I mean by this is to really try to think in the language rather than thinking about what I want to say in English and then translating it to the target language. This sometimes limits what I can write, but I end up using more native-sounding language when I manage to do it. 5 - Recycle. I keep a notebook with corrections, especially those that involve word usage, expressions and collocations. I memorize the sentences that they are in and then try to recycle them in later writing. My writing and skill levels aren't great, but when I keep my rules in mind, I feel like writing corrections help quite a bit in my progress.
Corrections made a huge difference for me in French, as far as I can tell. I cannot reconcile my personal experience with all those studies. If I had time (and wanted to spend money on journal articles), I'd pick apart the study methodologies and see if they're actually applicable to my personal experience, or if they're testing something different. Many language-learning studies, for example, are conducted on a small scale or with toy languages, and they don't always have anything to say about learners who are investing, say, 500 hours of determined study. The history of my French studies is kind of weird: I did Assimil, read a 450 page book, and learned about 1,000 vocabulary words using L1<->L2 word cards (which were absolutely agonizing). I would judge that I was a wobbly A2 at this point, perhaps with B1 reading skills. I put my studies on hold for about two years. During this time, I listened to my wife speak French to our kids. This probably added up to over a thousand hours of comprehensible exposure, but using extremely narrow vocabulary and only one speaker. I also read about a paragraph a day. At the end of this period, I was probably a pretty solid A2. You could argue that I had B1 listening comprehension, but only for my wife. I started speaking French full-time at home. This was really hard, and I made tons of mistakes, but I could communicate. Anyway, around this point, my comprehension was much better than my production. My production was full of nasty errors: I had completely forgotten, for example, that French had an indefinite article, and I just dropped it. I was a classic, highly ungrammatical adult learner with deeply fossilized errors. And input wasn't helping me, because I just glossed over little things like gender, indefinite articles, agreement, etc., when I encountered them in input. Ugly, ugly, ugly: I was a semi-fluent A2 with grammar from hell. I did three things to address this problem: I wrote 100 words per day for 30 days on lang-8, and figured out how to get very good corrections. I worked with a very good French tutor who corrected my extensively and pushed me to speak. Using the feedback from (2) and (3), I paid close attention to my input, and I started actually noticing things like, oh, gender, indefinite articles, various details of the passé composé I had been missing, and so on. I find that this is key: I'm using the corrections to discover previously ignored aspects of input. Once I can perceive these details, I can generally acquire them fairly rapidly. Four months after I started speaking—and getting lots of corrections—I comfortably passed a B2 exam, and my French was a lots better, with decent gender agreement, indefinite articles, and even a spontaneous subjunctive. OK, the gender required some conscious effort to get right, but still—my French was massively better. So based on an entirely unscientific sample size of 1, I'm strongly inclined to believe that corrections can rapidly fix all kinds of random fossilized nastiness, at least under certain circumstances.
I was in the University of Hawaii Second Language Acquisition/English as a Second Language graduate program when some of those studies were being done that showed that corrections don't help (on average). They seemed pretty convincing, on average. The fact that statistical significance in a study was reached does not mean that this is true for every single learner. Swain's "output hypothesis" suggests some ways where production, with a focus on form may facilitate acquisition. It makes sense to me. I too find it difficult to believe that being corrected doesn't help
I'd feel more confident in a video by Steve Kauffman critiquing a specific neurosurgical procedure, or Tiger Wood's golf swing, or Cosmopolitan's latest love potion.
I'm thinking about the 2 negatives in Cebuano, wala & dili (I think it's wala & hindi for Tagalog). I've been briefly corrected on this many times. and I still get it wrong. I've gotten to the point now where I still say it wrong, but right after, I know something went wrong. Does this mean corrections don't help? It can take a while to see results.
I am a big fan of writing corrections! I both like to get as many writing exercises as possible corrected and I myself correct so many "notebook entries" as possible on italki in three languages: German, Dutch and Esperanto. I refuse to correct anything on the "Exchange" on the LingQ - website, because people there are heavily campaigning (see Steve's video!) for corrections by native speakers only. On italki I want to see high quality corrections, it doesn't matter to me, if a teacher overlooks a little mistake, but I favour a systematical marking of the mistakes, so that I can see at once, what is correct and what is wrong. And I have seen several good non - native speakers correcting on italki, because I also closely read the corrections made by my collegues. On italki it's quite common that you get 2 -3 corrections at the same time for one writing exercise, so then you have the possibility to compare the quality of several correctors. I often experience this when correcting Dutch texts that also another corrector corrects the same text. This is transparent, because the writer of the text can decide whether he /she gives the point to only one of the correctors or even two or three points to both or all three correctors (if there are three corrections of his / her text).
