Berichten: 132
Taal: English
erinja (Profiel tonen) 14 februari 2011 00:58:23
At any rate, Todd's personal views on grammar have no relation with his 'politics' as a Raumist, I'm sure. Even if Todd *did* want to change the meaning of yes and no (which he didn't), I wouldn't blame it on his being a Raumist.
sudanglo (Profiel tonen) 14 februari 2011 10:19:01
This morning, I put Google translate to the test. I fed it:
I like coffee
Fruit flies like bananas
Time flies like an arrow
It came back with:
J'aime le café
Les mouches des fruits comme les bananes
Le temps file comme une flèche.
So even with today's technology it correctly interpreted 'Fruit flies' and 'Time flies'
Yes it tripped up on 'like' in the second sentence. But it establishes my speculation that such transalators will often get it right through their knowledge of collocations WITHOUT extra-linguistic knowledge.
Genjix (Profiel tonen) 14 februari 2011 11:45:23
http://translate.google.com/translate?js=n&prev=...
http://translate.google.com/translate?hl=en&sl=n...
.etc
machines:1 humans: 0
sudanglo (Profiel tonen) 14 februari 2011 12:18:21
Edit: I rang up my niece who knows some Chinese and she reports that the three test sentences are all correctly translated into Chinese.
Miland (Profiel tonen) 14 februari 2011 13:04:07
I just put une messe de neuf heures in Google translate. Result: "Mass of nine hours".
T0dd (Profiel tonen) 14 februari 2011 13:27:31
sudanglo:Hang on, Todd. Not so long ago it was you(Raŭmisto) who wanted to change the meaning of jes and ne.If that's what you got from my comments, then I completely failed to communicate.
Surely the point is anyway, that with a little ingenuity any Trojan Horse or other objections can be dealt with by repackaging/rebranding Esperanto.I think that's what Raŭmismo actually does, in fact. It attempts to separate Esperanto from its delusional past (Finvenkismo). It says that what's important and interesting about Esperanto is what it is now, rather than what it could be in some imagined future.
The beauty of campaigning for Esperanto in the schools on purely educational grounds is that it sidesteps all the standard arguments against Esperanto.It doesn't sidestep the sexism and Euro-centrism objections, which the ultra politically correct academic world is very quick to pick up on.
Incidentally, my impression of academics, with regard to Esperanto is that many of them have been quite ready to argue against Esperanto, without knowing the facts, and might not feel it at all necessary to put in 10 minutes of research on the Internet.Sure, but these people won't be moved by your 10 minute educational presentation either.
Look, I'm not saying that this can't work here and there on a small scale. Perhaps there are independent schools willing to try it. But if you're talking about state-controlled school systems then you have to face the fact that they will not make a move without academic approval.
In the example you give, I imagine that it might not be such a problem since the translator would have in its databank an translation for fruit flies as a noun phrase, and not have a translation of time flies as an NP. This would force it to prefer the correct parsing.Sure, but it has to know when to use the noun phrase and when not to. Try "Fruit flies when thrown" and you'll get "Les mouches des fruits une fois jeté"--that is, it interprets the whole input as a noun phrase rather than a complete sentence.
Lest I be misunderstood (again), I'm not arguing that good machine translation is impossible. I am arguing that the gap between "fair" and "good" machine translation is harder to cross than the already crossed gap between "terrible" and "fair". And it's harder still for spoken language, where sentences are often dropped in the middle of other sentences, and other sentences are left--well, you get the point.
sudanglo (Profiel tonen) 14 februari 2011 13:41:41
How many of his examples currently produce the wrong translation is an empirical matter. How many of the supposed problems could not be addressed by a future technology is just guesswork.
By the way 'concrete jungles' which the translator got wrong in the published French translation is correctly handled by Google translate - another case of collocation anlaysis?
sudanglo (Profiel tonen) 14 februari 2011 13:54:13
It says that what's important and interesting about Esperanto is what it is now, rather than what it could be in some imagined future.Thank you Todd, I think this highlights perfectly what I feel is wrong with Raŭmismo as a philosphy for the movement.
Of course, any individual, can learn Eperanto for whatever reason they like, but we are talking about a school of thought not an individual's motivation.
Miland (Profiel tonen) 14 februari 2011 14:35:42
sudanglo:Piron's article is very interesting, but of course based on speculation..He was a professional translator, so I would regard his opinion as better than speculation for the present-day situation, and his predictions as more reliable than those of people outside his profession. In practically the first example I tried from his article, at any rate, the machine did not do too well. Readers are free to examine his other examples if they like.
sudanglo (Profiel tonen) 14 februari 2011 21:00:12
However, nobody, can be quite sure of the future technology that may be developed and he would not have been aware of the capabilities which are currently available.
It is quite clear that there are some theoretical problems, but for whether they might have solutions, one needs an expert on technology.
Incidentally, on a question of evidence; whilst the value of learning Esperanto for later studies of another language seems to have been supported by several studies, as far as I am aware there have been no experimental studies on whether learning Esperanto improves a pupil's use of his own language - heightens his awareness of sentences which are ambiguous or may be misinterpreted, for example.
I find this idea not totally implausible and such evidence would be wonderful ammunition in arguing for Esperanto in the schools.