|
 |
Is comprehension measurable? | ||
Share with:
| ||
 Language Learning Forum : General discussion |  | |
| montmorency Diglot Senior Member United Kingdom Joined 4834 days ago 2371 posts - 3676 votes  Speaks: English*, German Studies: Danish, Welsh | Message 193 of 211 27 August 2014 at 6:45pm | IP Logged | ||||||||||||||||||||
... You seem to have jumped from comprehension to production. You have moved the goalposts, to use an English expression. (But do you understand the expression as an English native-speaker would...? :-)) 3 persons have voted this message useful | |||||||||||||||||||||
 emk Diglot Moderator United States Joined 5538 days ago 2615 posts - 8806 votes Speaks: English*, FrenchB2 Studies: Spanish, Ancient Egyptian Personal Language Map | Message 194 of 211 27 August 2014 at 7:51pm | IP Logged | ||||||||||||||||||||
Personally, I appreciate corrections, and I believe that they help me. Many foreign language students and teachers would agree. However, there's a fair bit of Second Language Acquisition research that suggests that the benefits of correction are largely illusory—and some other research which says correction can help. For a good summary and links to further studies, see this article for language teachers:
There are several hypotheses about Second Language Acquisition, including: 1. The input hypothesis. 2. The noticing hypothesis. 3. Comprehensible output hypothesis. Above, I suggested that learners ought to "internalize" the rules of French gender agreement to the point that *la-F ____ vert-M sounded weird and wrong. This is technically a version of (2), the Noticing Hypothesis, suggesting that learners should train themselves to notice gender markers in their input. I can easily imagine exercises to help build this awareness. The problem is that a lot of SLA researchers believe in the Input Hypothesis, and pretty much everybody agrees that the Noticing Hypothesis is very difficult to state in any kind of testable form. Do you have to be able to explain the actual grammar rules that govern gender? Or is it sufficient to be aware of which word forms are actually used in any given sentence?
I'm not sure that a learner who regularly made mistake mistakes like this could be described as a "relatively advanced speaker." Generally speaking, students taking the DELF B2 are expected to make occasional errors, but to make very few consistent errors. And that sentence has consistently awful gender. The errors are "unnatural", by which I mean that a student with any mastery whatsoever of gender would be unlikely to make them. Here are some mistakes that would be more "forgivable" at a B2 level, assuming they happened relatively rarely:
Any of these errors, in isolation, would be perfectly normal for a B2 student, as would a certain amount of guessing the gender of rare words. A student whose typical errors fall into one of these categories has at least partially acquired the rules of French gender, even if they have trouble getting everything correct in fast speech. But a student who constantly says things like *mon voiture or *cet beau café a été detruite would have to have remarkably good skills elsewhere to do well on production sections a C1 exam. Personally, I could easily believe that a student who consistently makes these blatant errors has excellent comprehension of French, because you could erase the gender from a French text completely with little loss of meaning. But this is one of several reasons that I don't believe in a strict version of the input hypothesis.
This is an excellent point. I would go so far as to say that a student who has thoroughly mastered the gender of everything except nouns is actually in pretty good shape—at the very least, they should be able to learn the gender of individual nouns quickly.
Honestly, I don't always bother with native proofreading. My rules of thumb are: 1. Sales copy on my website should be written by a native speaker with copywriting experience. 2. Résumés and cover letters should be proofread carefully by a native speaker. 3. Technical blog posts should be proofread when possible. 4. Routine professional emails get proofread once or twice by me and then fed through Bon Patron to catch stupid mistakes, but that's about it. 5. Personal emails are simply written and sent, although I might occasionally look up a gender or feed an idiomatic expression through Linguee. Error correction and feedback are nice, but they're simply not practical for routine emails, especially past a B2 level. I mean, I'm perfectly well aware that I still make errors, but the majority of my errors are subtle enough that I need to call in favors (or pay a tutor) to get good corrections. And it doesn't make sense to do that for routine emails.
In northern New England, we call that "tourists from New Joisey." :-) Or more seriously, as a speaker of a pluricentric language with hundreds of common accents, some of my least favorite accents will inevitably be native ones. I'm sure the good folks from New Jersey feel the same way about me. More seriously, if you did want to invent a numerical scale for speaking, I think you pretty much need to imitate the CEFR exams: 1. Make a list of things you care about: accent, vocabulary sophistication, grammatical correctness and richness, etc. 2. Assign a number of points to each category. 3. Write out standards saying, "a 2-point score in grammatical correctness looks like this or this, but a 3-point score looks like that or that." 4. Train your examiners well. 5. Choose speaking tasks (at least above B1) with a goal of defeating Boris Shekhtman's tricks. Done well, this will produce results that are semi-repeatable, within a known margin of error. And that's really all you can ask for. Edited by emk on 27 August 2014 at 7:52pm 2 persons have voted this message useful | |||||||||||||||||||||
| tarvos Super Polyglot Winner TAC 2012 Senior Member China likeapolyglot.wordpr Joined 4713 days ago 5310 posts - 9399 votes Speaks: Dutch*, English, Swedish, French, Russian, German, Italian, Norwegian, Mandarin, Romanian, Afrikaans Studies: Greek, Modern Hebrew, Spanish, Portuguese, Czech, Korean, Esperanto, Finnish | Message 195 of 211 27 August 2014 at 7:57pm | IP Logged | ||||||||||||||||||||
Why is it relevant whether the irregularity is in the stem or the ending? Eventually you pronounce a different thing, so push comes to shove the actual words you emit are different. I don't quite see the relevance of this comment.
You're missing a big issue here. French ORTHOGRAPHY was not updated. Remove that from the equation and things suddenly start making a lot more sense. What is wrong with speakers choosing to speak a way that you personally don't like? Do languages have to be efficient? I didn't see that being a prerequisite when language was invented. In fact, for comprehension, it's better, because you add in a layer of redundancy - even if you miss the nous in speech, you might still hear the -ons, which indicates it must be "we" who are talking. Etc. Redundancy is a common feature of engineering systems and a very useful one.
Does it matter that it does?
I thought healthy referred to there being a significant number of speakers. Edited by tarvos on 27 August 2014 at 7:57pm 3 persons have voted this message useful | |||||||||||||||||||||
| Stolan Senior Member United States Joined 4038 days ago 274 posts - 368 votes Speaks: English* Studies: Thai, Lowland Scots Studies: Arabic (classical), Cantonese | Message 196 of 211 27 August 2014 at 8:40pm | IP Logged | ||||||||||||||||||||
|
[QUOTE=tarvos]
-Why is it relevant whether the irregularity is in the stem or the ending? Eventually you pronounce a different thing, so push comes to shove the actual words you emit are different. I don't quite see the relevance of this comment. Principal parts! -You're missing a big issue here. French ORTHOGRAPHY was not updated. Remove that from the equation and things suddenly start making a lot more sense. How? -What is wrong ...... Then why is redundancy completely absent in the rest of the world's languages? -Does it matter that it does? Yes, because it confirms what I said is true. And the idea these features are normal is the very reason why there are fewer excellent "Iversen" "Emk" and "Medulin" types in the world and far more "S_Allard" "Fnord" types. -I thought healthy referred to there being a significant number of speakers. What languages in the world must inflect adjectives to agree with nouns? Some don't have adjectives like Arabic where nouns are used and Japanese which use verbs and nouns. What languages have so many decaying inconsistent paradigms, so many competing methods, so many random extensions, useless agreements, so many archaism that refuse to die, so much returns to bad habits such as aspect pairs and more that cause nearby languages to do so as well. Proto Finnish/Estonian did not inflect adjectives, but now they do so an extra 3 parts to memorize for Estonian because IE! Chechen too but both are the smallest offenders, they only need remove Indo European influence and they would not be so misshapen. Russian for example could not still be Russian with the same treatment. You do know how irregular Latin adjectives are? those irregularities are not from sound changes, they are clearly the very definition of the problem all Indo European language possess or used to possess. Edited by Stolan on 27 August 2014 at 8:44pm 2 persons have voted this message useful | |||||||||||||||||||||
| tarvos Super Polyglot Winner TAC 2012 Senior Member China likeapolyglot.wordpr Joined 4713 days ago 5310 posts - 9399 votes Speaks: Dutch*, English, Swedish, French, Russian, German, Italian, Norwegian, Mandarin, Romanian, Afrikaans Studies: Greek, Modern Hebrew, Spanish, Portuguese, Czech, Korean, Esperanto, Finnish | Message 197 of 211 27 August 2014 at 10:23pm | IP Logged | ||||||||||||||||||||
Don't know what that is, don't care. The point is that people pronounce in chunks. When I speak, I don't think in morphemes and analytics. It's too complex a process to operate concomitantly while also trying to speak. So I switch off the analysis for the most part and simply speak automatically, and the chunks that come out are not "je... vais... manger" but simply j'vais manger (chunk), tu viens manger (chunk)avec moi (chunk), I might even say "tu viens avec?" (it's Belgian). I analyse language in this type of phrase and what I utter is analysed in the context of that phrase. Not in terms of its constituent parts. What matters is the chunk that I utter in order to get my message across.
French not dropping pronouns is forced by the fact that in speech, most verb endings are the same phonetically speaking (as discussed above), but written differently. If French orthography reflected its phonology better that would be a lot clearer and the discrepancy would be much less noticeable. But you already know all of that!
It's not. Why does Korean have classifiers? You could simply say "two book" and it'd be good enough. Why do they have to have a different classifier for book? Or even one at all? Redundancy. All good systems build in buffer zones to specify more clearly in case the context is not obvious enough. Language is not a theoretical framework that you can simply force on people and strip of all regularities, because it's a human construct that is subject to human habits and priorities. Even Esperanto is subject to this, and it's a conlang. Otherwise, we should all simply speak Lojban. But we don't.
This is just such a non-sequitur I don't even know where to start.
Indo-European languages, because that is what they do and because IE languages clearly decided not to adhere to arbitrary Stolan standards of correctness. For the record, I enjoy aspect pairs - they provide me with the tools to make certain distinctions that otherwise would be left to context. In fact, that's how I see all these features; I use them to make explicit certain assumptions that would be left implicit in languages that are more context-dependent like Mandarin. I'm sorry, but if you're going to make bad blanket statements about grammar tendencies (aspects are bad habits? What the feck?) you have to look at the other side of the coin.
Well, I guess you won't study Chechen then. Poor life the Chechens must have. 5 persons have voted this message useful | |||||||||||||||||||||
| s_allard Triglot Senior Member Canada Joined 5436 days ago 2704 posts - 5425 votes Speaks: French*, English, Spanish Studies: Polish | Message 198 of 211 28 August 2014 at 12:21am | IP Logged | ||||||||||||||||||||
|
I confess that I can't follow the discussion on French verb morphology. I'll restrict my comments to some points
more relevant to the thread.
There may be some debate about the effectiveness about forms of error correction in the classroom environment but I believe that there is absolutely no doubt that at the individual level, i.e. one-on-one instruction, coaching or some form of individual editing, error correction or critical feedback is not only effective, it is downright essential. I would think that nearly all performance artists, athletes, authors and professionals in many fields work with coaches, editors, mentors, critics or advisors --whose role is to provide feedback and advice. Every article in a good newspaper is read by at least three different people. Ditto for any serious magazine. Every major author works with an editor. What professional golfer does not have a swing coach? How does an actor work on acquiring a dialect for a role? If an adult language learner wants to acquire a native-like accent, this will most likely require the services of an accent reduction specialist and lots of daily work. Anybody who has seen a masterclass given by a great music teacher or musician will know that individual constructive criticism and error correction are essential for improvement. Similarly, it is not important to have everything you write proofread by a native speaker. On the other hand, I'm sure for certain things, as emk has rightly pointed out, you either have a native speaker write the original or have your writing checked. The same could be said of preparing for the writing part of language tests such as the CEFR. Since you know that you will be asked to write, let's say, a 400-word essay on a topic, I think that you would prepare for this test by actually writing some practice essays and have them corrected by a tutor. This does not preclude tons of input of all kinds, but this will not replace actually producing writing that will be corrected in collaboration with a good tutor or coach. As for developing a system of using percentages for productive skills like writing and speaking, I agree that it probably isn't worth the effort when the current systems of assessment work quite well. As a matter of fact, using percentages for estimating comprehension based on feeling in my opinion is not necessarily a bad thing. It makes people feel more comfortable and they can use them for personal benchmarking. But this is not the same as measuring. The problem is that there is no form of external assessment. When someone says they understand a text 100%, there is no way to know what and how that person actually understands. If that person says they understand six sentences out of ten, can we know what the person really understands and does not understand? As has been discussed endlessly here, a common approach to assessing understanding, outside the world of professional language testing, is to count words. For reasons that I won't repeat here, I think this approach is at best illusory. It confuses vocabulary coverage with understanding and, above all, it ignores very fundamental issues of grammaticality. 1 person has voted this message useful | |||||||||||||||||||||
| Stolan Senior Member United States Joined 4038 days ago 274 posts - 368 votes Speaks: English* Studies: Thai, Lowland Scots Studies: Arabic (classical), Cantonese | Message 199 of 211 28 August 2014 at 2:21am | IP Logged | ||||||||||||||||||||
Can you name a single non-Indo European or IE influenced language that inflects adjectives inconsistently for case number and gender, has derivational morphology in place of inflectional, and is filled with broken syntactical redundancy? French conjugation only adds irregularity, it just sits there as a feature to brag about. 2 persons have voted this message useful | |||||||||||||||||||||
| s_allard Triglot Senior Member Canada Joined 5436 days ago 2704 posts - 5425 votes Speaks: French*, English, Spanish Studies: Polish | Message 200 of 211 28 August 2014 at 3:26am | IP Logged | ||||||||||||||||||||
|
When we write or speak, our mistakes are visible or audible. Often we can correct them. I write very quickly and often make mistakes here and end up going having to correct them, often with some embarrassment. But at least I and others can see my mistakes. When I say I understand something how do I know that I'm not making a mistake? I can't know. After all, I'm convinced that I understand. There's no way for somebody to point out my mistaken understanding of something. This is not a problem when I'm reading something with really technical or exotic terminology where I have no clue. Then I can safely say that I don't understand. But it's more of a problem when the words are quite common, as in a piece of literature. For example, I have mentioned in another thread that I recently read the novel Mike and Psmith by the great British writher P.G. Wodehouse. This book, written in 1907, revolves around life in a British boarding school and the game of cricket. Here is a paragraph talking about a game: "Psmith had always disclaimed any pretensions to batting skill, but he was undoubtedly the right man for a crisis like this. He had an enormous reach, and he used it. Three consecutive balls from Bruce he turned into full- tosses and swept to the leg-boundary, and, assisted by Barnes, who had been sitting on the splice in the usual manner, he raised the total to seventy-one before being yorked, with his score at thirty-five. Two minutes later the innings was over, with Barnes not out sixteen, for seventy-nine." There are 88 words and four sentences in this paragraph. The only word that I had never seen before in my life was "yorked". Here is the Oxford dictionary definition: (Of a bowler) bowl out (a batsman) with a ball that pitches under the bat. If I try to estimate my understanding of this passage in percentages, I don't really know where to begin. The very first sentence seemed clear enough. The others were problematic because of what was evidently cricket terminology. I knew all the words, but I really didn't know what the author was talking about. And more importantly, even when I found the terminology in the dictionary I really didn't know the significance in the context. For example, I never really understood what "Two minutes later, the innings was over, with Barnes not out sixteen, for seventy-nine." meant. Was that good or bad? Did I understand 25% - one sentence out of four? If I counted the terminology words, maybe I could claim I understood about 80% of this text. Of course, I would call the whole thing Some understanding and simply skim over the whole passage. Normally, something like this is not that important in the grand scheme of things, but this is the last chapter of the book and the cricket game plays a pivotal role in the ending. I eventually understand that the good guys win but I still don't understand how. My problem of course was not knowing the game of cricket. Knowing the words without knowing the game was basically useless. How much did I understand the last chapter? Not much I fear, but I did enjoy the ending. 2 persons have 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. |
