Bidirectional Translation

Ask specific questions about your target languages. Beginner questions welcome!
DaveBee
Blue Belt
Posts: 952
Joined: Wed Nov 02, 2016 8:49 pm
Location: UK
Languages: English (native). French (studying).
Language Log: https://forum.language-learners.org/vie ... =15&t=7466
x 1386

Re: Bidirectional Translation

Postby DaveBee » Fri Dec 08, 2017 1:38 pm

tastyonions wrote:Putting a number of days between translations so that your memory of the original text has faded is best. This removes the possibility that you will simply be relying on memory rather than actually translating.
Isn't recalling things held in short term memory a good way to move them to long term memory?
1 x

User avatar
tastyonions
Black Belt - 1st Dan
Posts: 1576
Joined: Sat Jul 18, 2015 5:39 pm
Location: Dallas, TX
Languages: EN (N), FR, ES, DE, IT, PT, NL, EL
x 3866

Re: Bidirectional Translation

Postby tastyonions » Fri Dec 08, 2017 7:53 pm

DaveBee wrote:
tastyonions wrote:Putting a number of days between translations so that your memory of the original text has faded is best. This removes the possibility that you will simply be relying on memory rather than actually translating.
Isn't recalling things held in short term memory a good way to move them to long term memory?
Sure. How you use it depends on whether you want to train translation or memorization.
0 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Bidirectional Translation

Postby Iversen » Fri Dec 08, 2017 9:29 pm

Since I generally don't have problems translating, but need to memorize things to absorb a new language, my preference would be to do the retranslation while the original still is present somewhere in my mind, but close to slipping away. That's where I need something that can grab and nail it so that it doesn't get lost in the mists forever. And immediate retranslation does that, where as postponed retranslation is more like trying to hit the jackpot in a lottery - with brain circuits that just as well could have been used for free language production.
1 x

Online
Cainntear
Black Belt - 3rd Dan
Posts: 3469
Joined: Thu Jul 30, 2015 11:04 am
Location: Scotland
Languages: English(N)
Advanced: French,Spanish, Scottish Gaelic
Intermediate: Italian, Catalan, Corsican
Basic: Welsh
Dabbling: Polish, Russian etc
x 8662
Contact:

Re: Bidirectional Translation

Postby Cainntear » Sun Dec 10, 2017 10:58 am

Now we're getting into territory that's beyond any of us, because we're essentially talking about the limits of neuropsychology.

For my personal anecdote, during high school, I came to the realisation that my brain was creating French sentences two ways simultaneously -- one through a slightly laborious mechanical think-it-through process, and one... just doing it. I had been kind of ignoring the latter for a while, even though it was essentially screaming "I KNOW THE ANSWER", and my conscious brain was saying "just give me another moment to check...." I can remember walking out of the French class (..actually, having said that... it must have been the Italian class because of what part of the building I associate the memory with) and having an image in my head of two paths: one of multiple arrows in a big long line, and the other a single arrow going round the outside as a short-cut. I resolved to just trust my brain from then on. My conclusion at the time was pretty much that it was a "fake it till you make it" situation.

A Krashen-esque explanation of that was that I had "learned" the conscious stuff and was just cheating until I had time to acquire it.

However, I studied a bit of AI at university and one of the concepts presented was the "computational neural network", which was based on a simplified theoretical model of biological brains. Neural networks can be trained with or without corrective feedback, but corrective feedback is the quicker route. (In fact, at the time, computer learning without feedback was largely theoretical as very few people had the computing power to attempt it.) We could produce our own tiny networks that could learn how to produce the right answers only if they were trained with enough examples, and I figured that was how I'd learned French and Italian at high school -- that my conscious declarative knowledge was the correct/incorrect feedback that the intuitive procedural system needed in order to learn.

But the reason I learned was because that intuitive procedural system was trying to make the sentence anyway, hence making an error that could be corrected. For this to work, you need to be using both systems simultaneously.

The weakness in many language classrooms is that the exercises give immediate reward to successful use of declarative rules, and never give reward use of procedural processing... then two years later the teacher basically berates the student for continuing to work as they always have and not magically doing things a different way from what they were told. Except that worldwide, language teach has improved in many ways, and I don't want people to think I'm having a dig at my own profession!!



Anyway, after that long and poorly though-out ramble, I should get to the point, shouldn't I?

Translating both ways is a very complex task type. You're looking to use the original as corrective feedback to try to encourage yourself to reproduce the sentence from its components. If you still remember the original, that can become corrective feedback inside your brain, as long as your brain isn't discouraged from working because it already knows the answer. The brain aims for efficiency, and if your brain decides that memorisation is the quickest route, it's bloody hard to convince it to do otherwise. Iversen's clearly developed a habit that stops his brain taking the easy way out, but there's no guarantee that anyone new to the technique will be able to do that.

This is probably going to be very much about you as an individual.

What you have to start to think about is whether you're "feeling" the meaning of the words as you're saying them. If you can look at the first few words of a sentence and know exactly what the translation is without thinking about what the words mean, your brain is just memorising and you need to change technique.
2 x

User avatar
tommus
Blue Belt
Posts: 957
Joined: Sat Jul 04, 2015 3:59 pm
Location: Kingston, ON, Canada
Languages: English (N), French (B2), Dutch (B2)
x 1937

Re: Bidirectional Translation

Postby tommus » Sun Dec 10, 2017 12:20 pm

Cainntear wrote:Anyway, after that long and poorly though-out ramble, I should get to the point, shouldn't I?

Your post is very intriguing. I sense that there is a very important technique lurking in that explanation. I know you said it may well be very individualised. But I wonder if you could offer a bit of a step-by-step process (sort of a cookbook) that you think would generally describe how a person might try to implement this technique?
2 x
Dutch: 01 September -> 31 December 2020
Watch 1000 Dutch TV Series Videos : 40 / 1000

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Bidirectional Translation

Postby Iversen » Sun Dec 10, 2017 8:18 pm

Cainntear's answer was quite interesting, but it is hard to see exactly where it takes you. I do understand that there is a difference between deliberately constructing sentences and having sentences flow from some hidden well in your mind, but for me these situations are just extremes of the same process. Even when I speak Danish I'll sometimes get a metalinguistic thought wheezing through my brain while I speak - mostly on the stylistical level, but sometimes also concerning grammar or vocabulary.

With weak languages such thoughts need to be more numerous (and more conscious) since weak languages function less automatically than strong ones, but without an automatic component I'm just constructing sentences or even worse: repeating fixed phrases. Actually: if there isn't some level of automatic sentence production involved I wouldn't even pretend to speak the language. I do however avoid such situations in conversations because know that my brain isn't fast and clever enough to concoct messages in languages where everything has to be looked up and checked with a grammar first. When writing you have time to do a certain amount of conscious construction work, but even here it will be slow, laborious and frustrating to do so without some level of help from your automatic language systems.

What then about the corrections? Well, they are necessary to ensure that your utterances don't veer out into the territory of green colorless ideas and utter ungrammaticality, but while you get them they don't further your ability to use your existing resources in an active way. Or in other words, corrections do hamper your intuitive procedural system, as mentioned by Cainntear. On the other hand: when I do immediate retranslation (or something that amounts to more or less the same - another kind of extensive cuing could do the same job) the best sign that things are going well is that I feel the sentences have become my own, and that they come forward 'because they already are there' and just have to be let loose. It may only be a travesty of truly free speech production, but can be done long before you are able to converse fluently.

By the way, according to Chomsky (in at least some of his avatars) .. what happens if I want to write this very sentence? In his system the first things that comes is the famous S = NP + VP. In my brain it seems that some member of the word family that includes "write" comes first, then maybe either "sentence" or "phrase" (skipping "Chomsky") and then something in my brain finds a suitable way to embed those words into a grammatical structure with all the appropriate bells and whistles. And no, I didn't rule out that some kind of abstract meaning came first, but then on a level where those vague and fluffy semithoughts led directly to the choice of a number of specific content words. And only then the syntax machine started up and more or less automatically concocted a full sentence, which I then proceeded to write down in the edit window.
3 x

Online
Cainntear
Black Belt - 3rd Dan
Posts: 3469
Joined: Thu Jul 30, 2015 11:04 am
Location: Scotland
Languages: English(N)
Advanced: French,Spanish, Scottish Gaelic
Intermediate: Italian, Catalan, Corsican
Basic: Welsh
Dabbling: Polish, Russian etc
x 8662
Contact:

Re: Bidirectional Translation

Postby Cainntear » Mon Dec 11, 2017 6:47 am

Iversen wrote:By the way, according to Chomsky (in at least some of his avatars) .. what happens if I want to write this very sentence? In his system the first things that comes is the famous S = NP + VP.

Which just goes to show how little Chomsky understood about language. Can you break a question in English into NP+VP? No. Pro-drop languages and VSO languages don't work that way, or verb-second Germanic languages either. Even his stuff about grammaticality only works where people have been acculturated to think of grammar in a mechanical, abstract way. Repeat the test in a language setting where such a culture does not exist, and meaningless sentences are rejected as not grammatical.
4 x

User avatar
Iversen
Black Belt - 4th Dan
Posts: 4768
Joined: Sun Jul 19, 2015 7:36 pm
Location: Denmark
Languages: Monolingual travels in Danish, English, German, Dutch, Swedish, French, Portuguese, Spanish, Catalan, Italian, Romanian and (part time) Esperanto
Ahem, not yet: Norwegian, Afrikaans, Platt, Scots, Russian, Serbian, Bulgarian, Albanian, Greek, Latin, Irish, Indonesian and a few more...
Language Log: viewtopic.php?f=15&t=1027
x 14962

Re: Bidirectional Translation

Postby Iversen » Mon Dec 11, 2017 7:36 am

This leads me to think about defective verbs in Latin and verbs for weather etc. which allegedly only can exist in the 3. person singular ("it rains"). But what happens if you are Zeus, who according one legend once clandestinely entered the room of a fair lady in the shape of a golden rain? He could definitely have used the verb in the first person singular, so whatever the official grammar say the absence of a first person form could be seen as an adaption to conditions in the real world rather as a morphological restriction.

As for the infamous "colorless green ideas" sentence the only weird thing about it is that it has some inbuilt semantical contradictions, but none that break the normal syntactical rules of the English language.

Grammatical (including morphological) rules are not platonian entities with an autonomous existence - they are tentative rationalizations based on actual speech patterns, and if those speech patterns have fuzzy edges then the same must be the case for the rules we formulate to remember them. If something doesn't exist or is explained away as 'meaningless' or 'ungrammatical' then the reason simply could be that there aren't any cases where we would use a certain formulation - although it could also be the result of a choice made long ago to use another form. This explains why it isn't OK to say "goed" in English - we have an implied convention that it should be "went" or "gone" according to the context. But if a sufficiently large number of native speakers AND foreigners decide that "goed" is OK then it has become part and parcel of the language .. at least for those speakers.The irony is that it might be the advanced foreign speakers who remain most adamant about sticking to the traditional rule which they have learnt from a book.

Which just goes to show how shaky foundations the naive belief in an unambiguous separation line between grammatical and ungrammatical utterances is. Chomsky had a good reason to accept judgements from native speakers about invented formulations alongside examples culled from a corpus in order to elucidate the workings of the English language through experiments, but it is simply too dangerous to expect the world to be nicely divided into yes yes and no no. Sometimes the correct evalution will be "well, it may be OK, but.." or "some might accept it, but I don't" or "I use it, but I know it is wrong".
1 x

Online
Cainntear
Black Belt - 3rd Dan
Posts: 3469
Joined: Thu Jul 30, 2015 11:04 am
Location: Scotland
Languages: English(N)
Advanced: French,Spanish, Scottish Gaelic
Intermediate: Italian, Catalan, Corsican
Basic: Welsh
Dabbling: Polish, Russian etc
x 8662
Contact:

Re: Bidirectional Translation

Postby Cainntear » Fri Dec 15, 2017 1:07 pm

tommus wrote:
Cainntear wrote:Anyway, after that long and poorly though-out ramble, I should get to the point, shouldn't I?

Your post is very intriguing. I sense that there is a very important technique lurking in that explanation. I know you said it may well be very individualised. But I wonder if you could offer a bit of a step-by-step process (sort of a cookbook) that you think would generally describe how a person might try to implement this technique?

The only thing I can say is that if you are doing an exercise and you are not thinking about the meaning of the sentence at all, it's not language, it's just word juggling. I do not believe you can learn language by juggling words. This goes for both bilingual/translation exercises and monolingual exercises.
0 x

Online
Cainntear
Black Belt - 3rd Dan
Posts: 3469
Joined: Thu Jul 30, 2015 11:04 am
Location: Scotland
Languages: English(N)
Advanced: French,Spanish, Scottish Gaelic
Intermediate: Italian, Catalan, Corsican
Basic: Welsh
Dabbling: Polish, Russian etc
x 8662
Contact:

Re: Bidirectional Translation

Postby Cainntear » Fri Dec 15, 2017 1:26 pm

Iversen wrote:This leads me to think about defective verbs in Latin and verbs for weather etc. which allegedly only can exist in the 3. person singular ("it rains"). But what happens if you are Zeus, who according one legend once clandestinely entered the room of a fair lady in the shape of a golden rain? He could definitely have used the verb in the first person singular, so whatever the official grammar say the absence of a first person form could be seen as an adaption to conditions in the real world rather as a morphological restriction.

I wouldn't use the verb "rain" there, though. I just couldn't -- it doesn't feel right.
(Is this the earliest reference to a golden shower in history??)

As for the infamous "colorless green ideas" sentence the only weird thing about it is that it has some inbuilt semantical contradictions, but none that break the normal syntactical rules of the English language.

Only if you define syntactic rules in such a way as to permit this phrase.

When we make a part-of-speech distinction between countable and uncountable, then "*a box of iPad" and "*a litre of waters" break the rules of syntax. But if your parts of speech only go as deep as "noun", then both the sentences are perfectly "grammatical", but that's purely arbitrary.

It is equally arbitrary to say that there is no distinction between non-physical nouns and physical ones; but if we do make such a parts-of-speech distinction, then we have the syntactic rule that non-physical nouns cannot be classified by adjectives denoting purely physical properties, including colours.

Lucien Tesnière, a French linguist working contemporaneously with Chomsky, came up with the term "valency grammars" (an idea more generally expressed as "dependency grammars" in modern literature) to describe a different kind of parse tree to Chomsky's. Tesnière's model puts the verb as the first node of the tree, and everything that comes below a node depends on what things the semantic element at the node can combine with.

Sleep, for example, cannot take "furiously", and it has to have a concrete physical noun as logical subject.

Chomsky had a good reason to accept judgements from native speakers about invented formulations alongside examples culled from a corpus in order to elucidate the workings of the English language through experiments, but it is simply too dangerous to expect the world to be nicely divided into yes yes and no no.

He had good reason, but his argument was roundly demolished in the following years. What irritates me is how everything that came after gets lost in the thunder of what-Chomsky-said.
0 x


Return to “Practical Questions and Advice”

Who is online

Users browsing this forum: Cainntear, jackb and 2 guests