Machine Translation is Good enough to get the gist of the content.

All about language programs, courses, websites and other learning resources
User avatar
mentecuerpo
Blue Belt
Posts: 572
Joined: Sun Jun 23, 2019 6:15 am
Location: El Salvador, Centroamerica, but lives in Phoenix, Arizona.
Languages: Spanish (N) English (B2) Italian (A2) German (A1)
Language Log: https://forum.language-learners.org/vie ... 18#p155218
x 840

Machine Translation is Good enough to get the gist of the content.

Postby mentecuerpo » Fri Oct 09, 2020 4:07 pm

I believe that machine translation at this current time is an essential tool for the foreign language self-learner.

Here I comper DeepL vs. Human, and I think that DeepL gives me a good idea of what is going on; it can be a place to get started quickly, then slow down analyzing the phrases and words.

The language learner can practically read any text from European Languages and use the AI machine translation to learn the content or understand the content, as he/she is exposed to the target language.

See example.
https://youtu.be/4vSu9kNh5aY

LUKE! Die Woche und ich - Luke bei der Miss Germany-Wahl

1: GERMAN
2: Translated with www.DeepL.com/Translator (free version)
3: Human Translation

1. Ich bin ja ein absoluter Befürworter des Feminismus
1. I am an absolute proponent of feminism (machine).
1. I am of course an absolute supporter of feminism, (human).

2. und deswegen hab ich mich mal bei der
2. and therefore I once
2. and that's why I had a look at the

3. Miss-Germany-Wahl zweitausendfünfzehn umgeschaut.
3. Miss-Germany-Election two thousand and fifteen looked around.
3. two thousand fifteen Miss Germany pageant.

4. Hast du 'n Freund? -Äh, ja. -Ah, schade.
Four. You got a boyfriend? -Uh, yeah. -Uh, too bad.
4. You got a boyfriend? -Uh, yes. -Oh, too bad.

5. Möchtest du eine Weintraube? -Oh, ja, gerne.
Five. Would you like a grape? -Oh, yes, please.
5. Would you like a grape? -Oh, yeah, I'd love one.

6. Möchtest du eine Weintraube haben?
Six. Would you like a grape?
6. Would you like to have a grape?

7. Alles Gute.
7. Good luck.
7. Good luck.

8. Du wirst gewinnen, du bist die Beste.
8. You're gonna win, you're the best.
8. You're going to win, you're the best.

9. Das weiß ich ganz genau. Ich glaub an dich.
9. I know you are. I believe in you.
9. I'm totally sure of it. I believe in you.

10. Du bist die Beste, du wirst das machen.
10. You're the best, you'll do it.
10. You're the best, you're going to do it.

11. Soll ich 'n bisschen knabbern?
11. Want me to nibble on some?
11. Shall I nibble a bit?

12. Nee, das... das... das... das nicht.
12. No... that... that... Not that.
12. No, that... that... that... not that.

13. Vielleicht einmal knabbern? -Nein.
13. Maybe a nibble? -No.
13. Maybe one nibble? -No.

14. Einmal nibbeln? -Nein. -Einmal... -Nein!
14. nibble once? -No. -One time... -No!
14. Nibble once? -No. -Once... -No!

15. Wenn der kleine Hunger kommt.
15. when the little hunger comes.
15. When you get a little hungry.

16. Ist denn so eine Miss-Germany-Wahl dann
16. Is a Miss Germany contest
16. Is a Miss Germany pageant like this one then

17. eigentlich wichtiger als so 'ne Bundeskanzlerwahl?
17. actually more important than an election for chancellor?
17. actually more important than a Federal Chancellor election?

18. Hm... -Also, anders wichtig. -Ja, ja.
18. Hm ... -So, differently important. -Yes, yes.
18. Hm... -Well, important in a different way. -Yeah, yeah.

19. Das kann man jetzt nicht so vergleichen, finde ich jetzt.
19. You can not now so compare, I think now.
19. I don't think you can really compare it like that.

20. Aber es ist schon eine... -Genau, eine... -ein Level. Ja.
20. But it is already a ... -Exactly, one... -A level. Yes.
20. But it's certainly a... -Exactly, a... -at that level. Yeah.

21. Inwiefern ist es 'ne politische Verantwortung, Miss Germany zu sein?
To what extent is it a political responsibility to be Miss Germany?
21. To what extent is being Miss Germany a political responsibility?

22. Warum ist Alice Schwarzer so 'ne tolle Frau?
22. Why is Alice Schwarzer such a great woman?
22. Why is Alice Schwarzer such a great woman?

23. Wie bitte? -Warum ist Alice Schwarzer so 'ne tolle Frau?
23. -Excuse me? -Why is Alice Schwarzer such a great lady?
23. What? -Why is Alice Schwarzer such a great woman?

24. Wir grüßen auch mit diesem Beitrag Alice Schwarzer ein Stück weit...
24. we greet Alice Schwarzer with this post as well...
24. We're sending greetings to Alice Schwarzer with this segment, to some extent...

25. Warum? -Als Frauenrechtlerin und, äh... -Ja, aber... warum?
25. Why? -As a feminist and, uh... -Yeah, but... Why?
25. Why? -As a women's rights activist and, uh... -Yes, but... why?

26. Bei der Miss-Germany-Wahl wird einfach für alles gesorgt.
26. everything is taken care of in the Miss Germany contest.
26. At the Miss Germany pageant, they've simply taken care of everything.

27. Alle Größen sind dabei.
27. All sizes are entered.
27. All of the sizes are here.

28. Auch hier für... Miss Sachsen.
28. Also here for... Miss Saxony.
28. Also here for... Miss Saxony.

29. In Gold... das.
29. In gold... this.
29. In gold... this one.

30. Das hier ist dann hier 'n bisschen „Shades of Grey“-Variante, aber...
30. This is a little bit of the "Shades of Grey" variety, but...
30. This here is a bit of a “Shades of Grey” version, but...

31. Ich weiß nicht, was... was hiermit... Ach, jetzt weiß ich, was damit passiert.
31. I don't know what... what to do with... Oh, now I know what happens to it.
31. I don't know what... what with this... Oh, now I know what it does.

32. Und nächstes Jahr Dschungelcamp dann?
Thirty-two. What about next year's jungle camp?
32. And then next year, "Dschungelcamp" ["Jungle Camp," a reality TV series]?

33. Nee, äh, ja, genau... aber vorher Playboy und dann Dschungelcamp.
33. No. Yeah. but first Playboy and then jungle camp.
33. Nah, uh, yes, exactly... but first Playboy and then "Dschungelcamp."

34. Was ist wichtiger bei 'ner Miss-Germany-Wahl:
34. What's more important in a Miss Germany Pageant?
34. What is more important in a Miss Germany pageant:

35. obenrum oder untenrum? -Also, eigentlich...
35. The top or the bottom? -Well, actually...
35. the top half or the bottom half? -Well, actually...

36. Es geht auch nach den inneren Werten.
36. It's all about inner values.
36. It is also about the inner qualities.

37. Es muss schon ebenmäßig sein, ja?
37. It has to be smooth, right?
37. It has to be even, right?

38. Also, es kann keine Höckernase, fliehendes Kinn
38. So, there can't be a hump nose, receding chin.
38. Well, there can't be a bumpy nose, a weak chin,

39. oder 'n rundes Gesicht oder zu speckig.
39. or a round face or too greasy.
39. or a round face or too chubby.

40. Was ist denn...
40. What's the matter...
40. What is...

41. Wo kommen wir denn da hin? -Ha, Kinner [Dialekt: Kinder]...
41. Where are we going? -Ha, Kinner [Dialect: Children]...
41. Where do we go from here? -Come on, kids...

42. Auch in einer alten Kirche kann man die Glocken läuten lassen.
42: Even in an old church you can ring the bells.
42. Even in an old church, you can ring the bells [swing the balls].

43. Das ist richtig.
43. that is correct.
43. That's right.

44. Merkt euch das.
44. remember that.
44. Remember that.

45. Was hast du an der Hand gemacht?
45. what did you do on your hand?
45. What did you do to your hand?

46. Hab ich gebrochen.
46. I broke it.
46. I broke it.

47. Wie hast du das gemacht?
47. How did you do that?
47. How did you do that?

48. Ein Mann hat mir blöde Fragen gestellt.
48. A man asked me stupid questions.
48. A man asked me stupid questions.

49. [...], aber pass auf, was du sagst, Junge, ne [Umgangssprache]?
Forty-ninth, but watch your language, son.
49. [...], but watch what you say, boy, OK?

50. Hast du deinen Fackelmann schon ausgepackt?
50. Have you unpacked your torch man yet?
50. Have you whipped your torch-man out yet?

51. Ja, so.
Fifty-one. Yes, like this.
51. Yeah, like this.

52. Völlig zu Recht erhebt sich jetzt das Volk,
52. The people rise up now, and rightly so,
52. The people are now rising up, and for good reason,

53. denn Harald Glööckler hat den Raum betreten.
53. because Harald Glööckler has entered the room.
53. since Harald Glööckler has entered the room.

54. Die einzig wahre Miss Germany.
54. the only true Miss Germany.
54. The only true Miss Germany.

55. Natürlich kann nur eine gewinnen
55. of course only one can win
55. Of course, only one can win,

56. und eine hat gewonnen.
56th and one has won.
56. and one did win.

57. Aber ich als moderner Mann,
57. but I as a modern man,
57. But I, as a modern man

58. der die Emanzipation einer Frau unterstützt,
58. who supports the emancipation of a woman,
58. who supports the emancipation of women —

59. der [sic] macht da keinen Unterschied.
59. the [sic] makes no difference.
59. he [the modern man] doesn't discriminate in this case.

60. Nein, nein... ich würd sie alle vierundzwanzig knallen.
60. No, no... I would bang all twenty-four of them.
60. No, no... I would bang all twenty-four of them.

61. Zu hart? Zu hart?
61. too hard? Too hard?
61. Too much? Too much?

62. Also, einvernehmlichen Geschlechtsverkehr mit allen haben.
62. so, have consensual sex with all of them.
62. That is, have consensual intercourse with all of them.
6 x

Kraut
Black Belt - 2nd Dan
Posts: 2599
Joined: Mon Aug 07, 2017 10:37 pm
Languages: German (N)
French (C)
English (C)
Spanish (A2)
Lithuanian
x 3204

Re: Machine Translation is Good enough to get the gist of the content.

Postby Kraut » Fri Oct 09, 2020 5:54 pm

DeepL has shortcomings: you have to open the alternatives if you want to find more than getting the gist of it.
24. Wir grüßen auch mit diesem Beitrag Alice Schwarzer ein Stück weit...
24. we greet Alice Schwarzer with this post as well...
24. We're sending greetings to Alice Schwarzer with this segment, to some extent...


DeepL overlooks "ein Stück weit": the human translator has it, also DeepL's alternatives have it.


24. we also greet Alice Schwarzer a bit with this contribution

24. we also greet Alice Schwarzer a little bit with this contribution

24. we also greet Alice Schwarzer somewhat with this contribution...

24. we also greet Alice Schwarzer partly with this contribution...

DeepL always translates into English first, when you come from Spanish and want a German translation, which leads to errors. Overall it is excellent if you know how to deal with it.
Last edited by Kraut on Fri Oct 09, 2020 7:50 pm, edited 1 time in total.
3 x

User avatar
tangleweeds
Green Belt
Posts: 449
Joined: Sat Jul 18, 2015 7:09 pm
Location: Portland, Oregon, USA
Languages: English (N)
beginner: Irish
clearing cobwebs: Japanese
on the shelf: French, Latin
wanderlust: Norwegian, Vietnamese
Language Log: viewtopic.php?t=705
x 1226

Re: Machine Translation is Good enough to get the gist of the content.

Postby tangleweeds » Fri Oct 09, 2020 7:16 pm

mentecuerpo wrote:I believe that machine translation at this current time is an essential tool for the foreign language self-learner.
<snip>
The language learner can practically read any text from European Languages and use the AI machine translation to learn the content or understand the content, as he/she is exposed to the target language.
I agree with the first wholeheartedly, machine translation has gotten so much better it's become a total blessing for the language learner. But I'll just point out that I mostly use the DeepL plugin on Japanese, where it's pretty great too!
3 x
Neurological odyssey is going better! Yay!

Ingaræð
Orange Belt
Posts: 170
Joined: Sat Nov 26, 2016 9:34 pm
Languages: English (N), German (heritage)
Learning: Russian, French, German, Mandarin, Arabic, Spanish.
Mostly forgotten: Italian, Welsh.
x 377

Re: Machine Translation is Good enough to get the gist of the content.

Postby Ingaræð » Fri Oct 09, 2020 11:36 pm

Kraut wrote:DeepL always translates into English first, when you come from Spanish and want a German translation, which leads to errors. Overall it is excellent if you know how to deal with it.

This is an interesting point, as I understand Google Translate also uses English as an intermediary. Does anyone know if there are any machine translators that don't use an intermediary, or at least don't use English as one? I'm guessing that Yandex might use Russian.
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: Machine Translation is Good enough to get the gist of the content.

Postby Iversen » Sat Oct 10, 2020 3:39 pm

For me bilingual texts are an essential part of my studying, and I mostly make them with Google Translate. There are of course lots of human-made translations in the world, but since I don't read literature many of those I can find on the internet are irrelevant for me. I do have some non-fiction books and brochures in parallel versions, but it can be a problem to keep them visible at the same time on a notestand. However I mainly use the parallel texts for intensive study, and since that activity only call for short texts I prefer to make bilingual side-by-side printouts, which are easy to put on a notestand.

Mentecuerpo apparently uses Deepl, and I faintly remember having seen a thread where some recommended it as better than Google Translate. I did have a go at it, but some of the languages I studied weren't on the Deepl at the time (I haven't checked whether they are there now), and now I have become acccostumed to GT - but not to the extent that I trust it blindly (see my thread about its weak points and ways to repair it). As for human translations they are often too 'free' to be of any relevance, and even when they are loyal to the original they have to conform to the rules of their destination language and sometimes also do some cultural adjustments which don't help me to understand to original. I have tried both possibiliies, but in practice I don't see any particular advantage to using human made translations.

So I don't trust GT, but still use it a lot - though mainly in the direction from an original in a weak or mediocre language to something I understand better.The point is that a translation doesn't have to be perfect to be useful. When I'm at the rock bottom miserable newbee stage then I have to check a lot of details in any translation with a dictionary because I need all the help I can get to understand the original. But even at this level there are things I can understand without help and others where I just have to be pointed in the right direction - and sometimes the translation looks so suspicious that I doublecheck it using a dictionary and/or a grammar. However as I get better in a language I also get better at extirpating the useful elements from a suspicious translation - like when a negation has been lost, but I can see that there should have been one.

Final point: I have experimented with both interspersed, 'strict' versus 'loose interlinear and and side-by-side layouts. If you really are a newbee OR you read an article about a language you haven't studied yet then it is of course necessary to keep track of the meaning of each and every word in the original, but later on you can make the connections yourself provided that the translation is reasonably literal, and it is MUCH easier just to put the two versions place side by side in columns in a Word document than it is to make any of the two other formats.

A month or so ago I reused an old trilingual text where I used a combination of interspersed and side-by-side. I made the interspersed version by using a quirk in Word: if you imported a GT translated text 'as text' then you got both languages for free. I don't know whether that still functions in newer versions, and I haven't checked it. As for the interlinear versions I made them with some spreadsheet trickery, but not without investing a lot of time and work in the process - and the result was simply not worth the trouble.
4 x

User avatar
mentecuerpo
Blue Belt
Posts: 572
Joined: Sun Jun 23, 2019 6:15 am
Location: El Salvador, Centroamerica, but lives in Phoenix, Arizona.
Languages: Spanish (N) English (B2) Italian (A2) German (A1)
Language Log: https://forum.language-learners.org/vie ... 18#p155218
x 840

Re: Machine Translation is Good enough to get the gist of the content.

Postby mentecuerpo » Sun Oct 11, 2020 4:07 pm

Iversen wrote:For me bilingual texts are an essential part of my study.

Final point: I have experimented with both interspersed, 'strict' versus 'loose interlinear and and side-by-side layouts. If you really are a newbee OR you read an article about a language you haven't studied yet then it is of course necessary to keep track of the meaning of each and every word in the original, but later on you can make the connections yourself provided that the translation is reasonably literal, and it is MUCH easier just to put the two versions place side by side in columns in a Word document than it is to make any of the two other formats.

A month or so ago I reused an old trilingual text where I used a combination of interspersed and side-by-side. I made the interspersed version.


Irversen, thank you for your contriutions to this post.

I just left Phoenix, AZ yesterday and today I am visiting New York city for a week.

Everything is pretty much closed but I will look for some museums to visit and some second hand book stores to hunt for old language books. I will also look for some specialized language books stores, hope they are open.

Having said that, I am using my phone to compose this note so i will make more mistakes than usual. I am not very good writing things on the phone.

I really enjoy the machine translation and I like it way better than human translation for language learning.
The problem I have with the human translation is that the good translator, conveys the idea and the meaning as close as possible to the translated language. If the idea and the meaning is translated, then the literal translation is lost, and the literal translation helps me understand, in away, how the targel language works. A more literal translation makes my brain work in trying to figure out the meaning by context.

Let me give you an example. I have subscription to Glossika, the program gives you one phrase after the other with its human translations. What I have noticed is that the human translations is faithful to the meaning, so what the speaker means gets translated by a human intelligence (HI), in a phrase that has the exact same meaning in English but many times, differs tremendoulsy from a more literal transtations generated by Artificial Intelligence. The AI is more helpful because it uses word by word translations looking for the best match for the translated word in context. For me as a language student, hunting for words and meanings of those words, a machine translation is the way to go. The human translations is excellent when reading a novel or literary work, but for learning a language in my opinion, the machine translation gets the job done. Glossika, with the human translation, is more interested in teaching you complete phrases where the student learns how to say one phrase in your language to a phrase that conveys the same meaning even if the worrds are completely different.

I am glad that other language studens have made a similar observation i have made about the machine translation and that the students use the machilne translation regluarly for language learning purpose.

Greenting to all from the big apple, it is a bit lonely here, with empty streets, and closed shops and restaurants.

Pardon the spelling mistakes and lack of clarity of my written style, using my phone and i am native Spanish speaker. There is no spell checker AI in this phone.
2 x


Return to “Language Programs and Resources”

Who is online

Users browsing this forum: No registered users and 2 guests