|
 |
My Thoughts On Sentence Mining and others | ||
Tags: Goals | Learning Technique/Method | ||
Share with:
| ||
 Language Learning Forum : Learning Techniques, Methods & Strategies |  | |
| Bao Diglot Senior Member Germany tinyurl.com/pe4kqe5 Joined 5758 days ago 2256 posts - 4046 votes  Speaks: German*, English Studies: French, Spanish, Japanese, Mandarin | Message 41 of 66 25 August 2010 at 4:07am | IP Logged | ||||||||||||||
Sentence mining, or rather any method that is focussed on repetition and input can have the effect that you learn the items in sequence and end up with the sequence stronger in mind than the actual meaning of every single item. That's a problem that has been around ever since somebody invented the word list, and it won't go away by switching to single vocabulary items instead of sentences - if you are the type of person who will easily remember that kind of extra information you need to make sure that the way the information is presented makes use of it when that's possible and avoids it when it would be harmful. I personally find the positive effect of getting collocation, particles and sentence patterns right outweigh the risk of memorizing whole sentences instead of chunks; and that is easily avoided by using two or more example sentences, and maybe doing extra repetitions for single words that prove especially difficult to remember. Oh, and I think we are talking about a lot of different things at the same time, like input/output focus independent/guided comprehension (as in: working out rules and patterns by yourself and then learning the rules vs. learning the rules and then looking at how they're used) repetition of a small number of items/getting a lot of exposure to new material concentrating on single skills/keeping different skills balanced Edited by Bao on 25 August 2010 at 4:27am 3 persons have voted this message useful | |||||||||||||||
| leosmith Senior Member United States Joined 6542 days ago 2365 posts - 3804 votes Speaks: English* Studies: Tagalog | Message 42 of 66 25 August 2010 at 5:23am | IP Logged | ||||||||||||||
I wouldn't say this is true for any focussed repetition, but this is true for sentences and words. However, there is less of this sequential data in words. I try to minimize the amount of flashcards and the amount of data on my flashcards. I wish I could get rid of all of them. So I'm not saying words are good and sentences are bad, I'm saying words are better than sentences. And as another poster mentioned, the problem goes deeper than that. There is a finite number of cards that you keep seeing over and over. So maybe you have only one sentence that starts with "dog", or one compound word that starts with the character for dog. So you see dog, and you give the answer without seeing the rest of the card. Then you go to actually read something, see dog, and assume it ends the same way the card you have memorized. But it might not. This sometimes happens in words, but is much more common with sentences in my experience. My solution is not to try to create more cards, and prolong the out of context phase, but to reduce the number of cards, reduce the data on the cards, and delete the cards after a few months. I use the time I save to actually read, which fixes all the problems.
I prefer getting said positive effects from reading.
I don't think adding more sentences solves the sequence problem, if that's what you're talking about
If I'm off topic, sorry about that. Or did you mean something else? 1 person has voted this message useful | |||||||||||||||
| Lucky Charms Diglot Senior Member Japan lapacifica.net Joined 6941 days ago 752 posts - 1711 votes Speaks: English*, Japanese Studies: German, Spanish | Message 43 of 66 25 August 2010 at 6:56am | IP Logged | ||||||||||||||
|
Thanks for the awesome explanations, leosmith, Cainntear, and Bao. I think I understand now. So it sounds like you dislike sentence mining for the same reason as atamagaii - because you'd rather pick things up from a wide variety of native sentences instead of looking at the same sentence over and over again, right? And you feel that memorizing a sentence as a single vocabulary item won't help you recognize the constituent parts in other contexts. That makes a lot of sense.
It seems to me that using as many example sentences as possible with the same word, as Bao said, would at least partially help avoid this. I usually put about 3 sentences on a single flashcard (although I suppose you could do it on separate flashcard to further avoid that 'sequence' problem) that I find from a combination of my original source, Google, my electronic dictionary, example sentence dictionaries (like ALC for Japanese), and self-made sentences that I confirm with a native speaker. I put the sentence(s) on the front and make the card cloze-deletion style. I sometimes also include the target-language definition and/or a picture to make the word as easy to recall as possible. I think the biggest flaw with this method is that, similar to what Bao said about wordlists, I sometimes can recall the word based on extraneous factors like the first two words of the example sentence, the overall shape of the text on the card, some smeared ink or a crinkled corner if it's a paper card, etc. which I imagine would make it hard to recognize the words in other contexts without these cues. However, I'm not sure how important that is: maybe the point of SRS is to drag the word into the forefront of my consciousness at certain intervals no matter by what means, so maybe it doesn't matter so much whether I 'get the answer' because of the sentence or the definition or the picture, or even the appearance of the card...? In any case, I have a personal rule that I have to read all the example sentences aloud in whatever order I feel like after showing the answer, to insure I get ('re-get'?) that input regardless of what aspect of the card triggered my recall.
I agree that the ideal situation would be to learn everything in context and memorize words because you encounter them in a wide variety of texts. But to rely solely on this without flashcards would take a LOT of reading, preferably in a short amount of time (which is why the demands of L-R are too much for many people). Sentence mining is a good compromise for those who either 1) don't enjoy reading for its own sake 2) can't dedicate a lot of time to reading 3) don't have access to a large quantity and variety of native texts 4) have a poor memory :) The stronger these factors, the more SRS is necessary as an artificial means to ensure regular exposure. Edited by Lucky Charms on 25 August 2010 at 7:16am 1 person has voted this message useful | |||||||||||||||
| Volte Tetraglot Senior Member Switzerland Joined 6431 days ago 4474 posts - 6726 votes Speaks: English*, Esperanto, German, Italian Studies: French, Finnish, Mandarin, Japanese | Message 44 of 66 25 August 2010 at 10:13am | IP Logged | ||||||||||||||
There isn't much to be done about 1, or its counterpart of not enjoying flashcards for their own sake - one can work on doing either and trying to enjoy it more, but ... 3 should be stone-dead, given the internet, for the vast majority of language learners. For 2, I find I learn words a _lot_ faster through L-R than SRS. SRS gives me active but broken use, where I need to make up most of what I'm doing without having seen it in a wide enough range of contexts to have any idea of how things are usually said, faster, but my passive vocabulary grows much more slowly - and as Iversen has written, as one gets stronger at a language, more of the passive words become active. I prefer an SRS over extensive reading/L-R for vocabulary acquisition if I don't have enough time to do a good job of it, and only then. Outside of that, SRS is sometimes a useful supplement, but it seems slower to me on the whole. With L-R, I've been able to map most of the words in, say, "Crime and Punishment" to the English translations after reading it once and investing something like 25 hours in Russian. This isn't even passive vocabulary - it requires the prompting of the English text for most words - but it's a good step towards it. I definitely cannot rival this using an SRS. As for 4, I find I remember things a lot better in context (I'm much more likely to have doubts about a letter or two of word I learned out of context than one that I can quote from song lyrics, for instance, much less usage doubts). Using an SRS (or other rote learning) for vocabulary always makes me feel like I have a poor memory, while other learning methods don't. I can't answer this point definitively, but I suspect it's not definitively cut-and-dried in favor of SRS. Intensive L-R style study helps, but it's not the only way to learn from native texts; read Kato Lomb's book "Polyglot: How I Learn Languages". Learning a decent amount of anything takes time, whether you spend 15 minutes a day on SRS or Assimil, or go straight to native texts. People without a 'lot' of time (however it happens to be distributed) cannot learn languages to a meaningful level, no matter what materials they use. A few phrases or words can be learned quickly; being conversational takes a fair number of hours, and reading literature comfortably takes longer yet, regardless of your method. Personally, I consider the demands of SRS to be far too high to use it as anything but a supplement. 1 person has voted this message useful | |||||||||||||||
| Cainntear Pentaglot Senior Member Scotland linguafrankly.blogsp Joined 6003 days ago 4399 posts - 7687 votes Speaks: Lowland Scots, English*, French, Spanish, Scottish Gaelic Studies: Catalan, Italian, German, Irish, Welsh | Message 45 of 66 25 August 2010 at 11:34am | IP Logged | ||||||||||||||
As leosmith says, learning-in-sequence isn't a problem with SRS as it presents items in a random order.
Avoided, yes; easily, no. Two or more, no -- two hundred or more, perhaps. You're aiming to learn collocations -- that's great, but there's a very real chance of learning collocations that don't actually exist. There is an infinite variety in language, and one thing our brains do is to spot the things that co-occur most often and tag them as collocations/compound terms. By rereading a small number of sentences, you're biasing the statistics in your brain. Your brain sees those words together and starts to think of that sentence as a fixed phrase, a series of collocations. The only way round this is to put an awful lot of similar phrases into your SRS so that you are exposing yourself to the variety of sentence form. Unfortunately the way SRS works, you're going to have to spend a lot of time in there to see all these sentences enough to fully get the variation. That's why I prefer to read books and websites, watch DVDs and online video, and simply talk to people. I surround myself with variety of usage and the collocations emerge naturally where there is a genuine frequency of co-ocurrence. When I'm speaking to people, they aren't aware of this collocations, and often I use them without even noticing it, but I generally start off by knowing the meaning of words to start off with and then let them learn to stick together of their own accord. 2 persons have voted this message useful | |||||||||||||||
| The Real CZ Senior Member United States Joined 5641 days ago 1069 posts - 1495 votes Speaks: English* Studies: Japanese, Korean | Message 46 of 66 25 August 2010 at 1:29pm | IP Logged | ||||||||||||||
|
The SRS isn't meant to replace reading. It's there to help you remember words/grammar that you need/want to remember. Yeah, I may see the same sentence 10 times over 3 months or so, but I don't complain when I hear the same sentence in TV shows all the time. Just because a few people whore out the SRS and don't read, watch TV, talk/chat with others, doesn't mean everyone substitutes the SRS for that stuff. Edited by The Real CZ on 25 August 2010 at 1:31pm 4 persons have voted this message useful | |||||||||||||||
| feanarosurion Senior Member Canada Joined 5273 days ago 217 posts - 316 votes Speaks: English* Studies: Finnish, Norwegian | Message 47 of 66 25 August 2010 at 2:51pm | IP Logged | ||||||||||||||
|
I agree with The Real CZ entirely. On the AJATT website, the focus is always on doing what is fun, and from taking from a variety of sources. For him, that means movies, comics, books, websites, whatever. For any other learner, the specifics will be different, but the focus should be, and usually/often is on the sources, not the sentences.
2 persons have voted this message useful | |||||||||||||||
| Andy E Triglot Senior Member United Kingdom Joined 7095 days ago 1651 posts - 1939 votes Speaks: English*, Spanish, French | Message 48 of 66 25 August 2010 at 4:05pm | IP Logged | ||||||||||||||
|
To go off on a slight tangent.....
It would seem to me that if one of the issues with SRS, is that you're being presented with items that you already "know", then the problem is actually with the "spaced" bit of your Spaced Repetition. Either you absolutely "know" them unequivocally and never-to-be-forgotten, in which case what are they still doing in your SRS? Hit the "don't ever bother me with this again" button or delete them if there isn't one because they're just taking up time and space. Or in fact, they're just being presented too early and too often. I use a Leitner-based system of my own devising (although I'm tinkering with Anki at the moment). New items I'd be happy to see multiple times over 3 months but if I'm still seeing the same material I first looked at several years ago 10 times in 3 months, then the scheduling is wrong and unsurprisingly it's going to bore me to tears. back on track... As far "Sentence Mining" in general goes, I don't see it as replacement for anything - it's just a tool in the toolbox. I only do L1 to L2 as I figure I might as well learn something actively (and get the passive for free) as do passive L2 to L1 and hope it becomes active at some point. 1 person has voted this message useful |
|  | 
Printable version
|
| Copyright 2024 FX Micheloud - All rights reserved No part of this website may be copied by any means without my written authorization. |
