Yes, that’s how it seems to me as well. And that’s why sometimes, when patterns match in an unfortunate way, it can produce nonsense that Google wouldn’t even dream of (but still smooth-sounding 
)
1 Like
That’s a good idea!
An Anna Blume, Kurt Schwitters,
Nächte
Gedicht 7
Summary
Innige Nächte
Gluten Qual
Zittert Glut Wonne
Schmerzhaft umeint
Siedend nächtigt Brunst
Peitscht Feuer Blitz
Zuckende Schwüle,
O wenn ich das Fischlein baden könnte!
Zagt ein Innen
Zittert enteint
Giert schwül
Herb
Du -
Duft der Braut
Rosen gleißen im Garten,
Schlank stachelt Fisch in der Peitschluft,
Wunden Knie
Wogen Brandung Wonne
O diese Qual, daß ich nicht fliegen kann!
Wonne umtaucht
Geistert Pfiff,
Ich sehne deine Wogen
Du!
Meine Glut fiebert Tod!
Ich umwoge
Result (translation from German to Japanese to English from Deepl):
Night
Poem 7
Heartfelt Nights
The Agony of Gluten
A quivering spark of joy
Painful Unity
The heat of the boiling night
Flames striking lightning
A twitching slurp.
Oh, if I could bathe the little fish
Inner dreads.
Trembling deink.
Yaws sultry
Tarts.
You-.
The bride’s scent
Roses shine in the garden.
Slender projections swim in the whipped air.
Pain in the knees
Bliss in the waves.
This agony of not being able to fly.
Diving bliss
A ghost whistles.
I long for your waves.
You!
My burning fever death!
Ai wu.
Conclusion:
Very legit translation of the original 
And I can see a pattern now, the end is always the best.
1 Like
i mean, it’s sort of in the name:
4 Likes
Der Zipferlake (german adaptation of lewis carroll's jabberwocky)
Verdaustig war’s, und glaße Wieben
rotterten gorkicht im Gemank.
Gar elump war der Pluckerwank,
und die gabben Schweisel frieben.
»Hab acht vorm Zipferlak, mein Kind!
Sein Maul ist beiß, sein Griff ist bohr.
Vorm Fliegelflagel sieh dich vor,
dem mampfen Schnatterrind.«
Er zückt’ sein scharfgebifftes Schwert,
den Feind zu futzen ohne Saum,
und lehnt’ sich an den Dudelbaum
und stand da lang in sich gekehrt.
In sich gekeimt, so stand er hier,
da kam verschnoff der Zipferlak
mit Flammenlefze angewackt
und gurgt’ in seiner Gier.
Mit Eins! und Zwei! und bis auf’s Bein!
Die biffe Klinge ritscheropf!
Trennt’ er vom Hals den toten Kopf,
und wichernd sprengt’ er heim.
»Vom Zipferlak hast uns befreit?
Komm an mein Herz, aromer Sohn!
Oh, blumer Tag! Oh, schlusse Fron!«
So kröpfte er vor Freud’.
Verdaustig war’s, und glaße Wieben
rotterten gorkicht im Gemank.
Gar elump war der Pluckerwank,
und die gabben Schweisel frieben.
DeepL trying desperately to make sense of it
It was so messy, and the glazed cradles
were rumbling in the mud.
The Pluckerwank was miserable,
and the yellow sweat was drifting.
“Beware of Zipferlak, my child!
His mouth is biting, his grip is drilling.
Beware of the flying flail,
The chattering cattle.”
He drew his sword with a sharp hilt…,
to sweep the foe clean,
And leaned against the bagpipe tree
and stood there for a long time turned in on himself.
Germinated, so he stood here,
Then came the Zipferlak, all snuffed up
with a flaming elephant
and gurgled in his greed.
With one! and two! and down to the leg!
The sharp blade, rip-roar-roar!
He severed the dead head from the neck,
And with a leap he went home.
“Have you freed us from the Zipferlak?
Come to my heart, aromatic son!
Oh, blumer Tag! Oh, close the front!”
So he gloated with joy.
It was a mess, and the glazed cradles
Rumbled gorkicht in the mingling.
The pluckerwank was miserable,
And the sweat of the fat was drifting.
google translate also takes surprising amounts of liberties and even drops an f-bomb
It was digestible, and smooth Wieben
rotten gorkicht in the mind.
Elump was the plucking rumble,
and they gave sweat fry.
“Watch out for the Zipferlak, my child!
Its mouth is biting, its grip is boring.
Watch out for the wing flag,
the munching Schnatterrind. "
He draws his sharp-edged sword,
to fuck the enemy without a seam,
and lean on the doodle tree
and stood there withdrawn.
Sprouted in itself, so he stood here,
then the Zipferlak came cranky
cracked with flames
and gurgles in his greed.
With one! and two! and up to the leg!
The biffe blade ritscheropf!
He separates the dead head from the neck,
and wiggling he bursts home.
'You freed us from the Zipferlak?
Come to my heart, aroma son!
Oh, bloomy day! Oh, shut up! "
So he crooked with joy.
It was digestible, and smooth Wieben
rotten gorkicht in the mind.
Elump was the plucking rumble,
and they gave sweat fry.
3 Likes
Sounds interesting but I did not understand it very well, so I had to translate that back to German.
Google Translate:
Summary
Es war bekömmlich und glatt Wieben
fauler Gorkicht im Kopf.
Elump war das zupfende Grollen,
und sie gaben Schweißbraten.
„Pass auf den Zipferlak auf, mein Kind!
Sein Maul beißt, sein Griff ist langweilig.
Achten Sie auf die Flügelflagge,
das mampfende Schnatterrind. "
Er zieht sein scharfkantiges Schwert,
den Feind ohne Naht zu ficken,
und lehne dich an den Doodle-Baum
und stand zurückgezogen da.
In sich gekeimt, also stand er hier,
dann kam der Zipferlak launisch
von Flammen geknackt
und gurgelt in seiner Gier.
Mit einer! und zwei! und bis zum Bein!
Der Biffe-Ritscheropf!
Er trennt den toten Kopf vom Hals,
und wackelnd platzt er nach Hause.
„Du hast uns vom Zipferlak befreit?
Komm in mein Herz, Duftsohn!
Oh, blühender Tag! Ach halt den Mund! "
Also krümmte er sich vor Freude.
Es war bekömmlich und glatt Wieben
fauler Gorkicht im Kopf.
Elump war das zupfende Grollen,
und sie gaben Schweißbraten.
DeepL
Summary
Es war so unordentlich, und die gläsernen Wiegen
rumpelten im Schlamm.
Der Pluckerwank war unglücklich,
und der gelbe Schweiß floss in Strömen.
“Hüte dich vor Zipferlak, mein Kind!
Sein Mund ist bissig, sein Griff bohrend.
Hüte dich vor dem fliegenden Dreschflegel,
Das klappernde Vieh.”
Er zog sein Schwert mit einem scharfen Griff…,
um den Feind wegzufegen,
Und lehnte sich gegen den Dudelsackbaum
und stand dort lange, in sich gekehrt.
Gekeimt, so stand er hier,
Dann kam der Zipferlak, ganz verschnupft
mit einem flammenden Elefanten
und gurgelte in seiner Gier.
Mit eins! und zwei! und hinunter zum Bein!
Die scharfe Klinge, rip-roar-roar!
Er trennte den toten Kopf vom Hals,
Und mit einem Sprung ging er nach Hause.
“Hast du uns von dem Zipferlak befreit?
Komm in mein Herz, duftender Sohn!
Oh, blumer Tag! Oh, mach die Front zu!”
So jubelte er vor Freude.
Es war ein Durcheinander, und die gläsernen Wiegen
rumpelten gorkicht in der Vermischung.
Der Zupfwank war jämmerlich,
Und der Schweiß des Fettes trieb.
Fantastic!
3 Likes
I find Deepl does a pretty good job of taking my English and turning it into passable Japanese. I need to go back and do some clean-up for exact terminology or politeness levels, but I showed a translation to a Japanese co-worker and he approved of it.
Going the other way, from English to Japanese, I still like the translations provided by DeepL better, but Google will give me romaji and let me translate from photos so I still use it.
3 Likes
YT comment : かつて聴いた曲の中で一番鳥肌立って感動した。 恋してるレベルでこの曲に夢中
Google Translate : I was most impressed with the goose bumps of the songs I had ever heard.
She is crazy about this song at the level she is in love with
DeepL : It gave me goosebumps and moved me more than any other song I’ve heard in the past.
I’m in love with this song
10 Likes
That’s not always true, I think. If you write nonsense or misleading Japanese, DeepL will spit out garbage. At least that’s what happened in my case when I had a brainfart and misused the ことがある structure. I got a weird sentence from DeepL which looked “great again” after I fixed my grammar.
I don’t trust the English to Japanese translation as much, though.
2 Likes
The best part about DeepL Japanese → English is that you can click each word in the English translation and select the synonym that you want, and the entire sentence will then re-translate to match the context of the synonym that you selected.
This lets you have a huge amount of control over the final translation.
If you want a technical, literal translation, you can get that. And if your sentence can be interpreted in like six different ways, you can simply change the synonyms to specify the context, and you’ll get a translation that matches your use case.
In comparison, Google translate gives you one or two translations, with no way to specify the context behind the sentence other than to literally spell it out in Japanese, and it often has strange word choice, superfluous (and often wrong) pronouns and super unnatural grammar.
Like, just look at this - how can they show that as an alternate translation? Early rice isn’t a thing. And nobody would ever interpret 早く as referring to rice in that sentence - it’s obviously modifying 食べる. It’s like they took all the words, put them in a bag, translated them, and then pulled them out in a random order.
I’m really glad that DeepL exists, because it means that we don’t have to use Google Translate for Japanese anymore.
6 Likes
Yeah, but that’s only if you already know some Japanese, and have an idea what the Japanese sentence actually is about. But if someone who knows zero Japanese uses Deepl to translate some texts to English, they risk getting perfectly valid output that at the same time can be completely wrong.
But how often do you try to translate something with absolutely zero context?
Because zero context is honestly difficult to have, even if you don’t know the language. Especially when it comes to visual mediums, like manga or anime.
Let’s say someone with no Japanese knowledge translates a manga using DeepL.
By looking at the pages, they know who is saying what. They know where the conversation is taking place, which objects are in the scene, what has been shown before, which characters are part of the conversation, what they’re doing, their facial expressions while talking etc. - in a visual medium, there is so much context available to guide you to a fitting translation.
Even in a non-visual medium, like a novel, each translated sentence will give you a better idea of what the passage is about. Every translation you do gives you more context. And when you have translated a passage, or a page, or a chapter, you can easily tell if one sentence doesn’t make fit in with the rest.
Sure, you can give DeepL some broken Japanese and it will try and fail to give you a decent translation.
But how many times have you gotten completely wrong translations on real sentences?
How often have you actually ran into sentences in the wild that were so unusual and difficult to translate that DeepL got absolutely everything wrong, to the point where you couldn’t find something that made sense with the context in the synonyms?
FWIW, a friend of mine is English-Japanese bilingual and works as a translator. Her view of machine translation is essentially “throw it away and start over.” It’s okay if you’re trying to get the gist of something you’re reading for your own information, but it will make dramatic and embarrassing mistakes if you try to use it to communicate with a native speaker.
It’s easy to go the other way and lose context with each machine translation too though.
Like say someone is working through a novel and uses Deepl to check a complicated quote that doesn’t have an obvious “X said” tag, then to structure the sentence in English Deepl needs to make some assumptions about subjects, which the reader isn’t sure enough to contradict, and they inadvertently come away with a slightly off impression of the line, which makes them think the completely wrong person said it, which then colors how they consider the entire scene.
It might be possible to mitigate that some with Deepl’s context tools and careful rigor, but it’s definitely something to be cautious about, as any translation is necessarily going to destroy or alter, to one degree or another, the structure and subtleties of the original. And if the text is literary, or has any level of implication or insinuation in it, that can make a giant difference, even in cases when the literal meaning is still correct.
I think it’s fine for a last resort, take-with-a-grain-of-salt piece of evidence if completely lost (or for practicalities or when there’s time constraints), but generally for reading I think it’s better to puzzle out the original structure rather than risk losing sight of it for its facsimile.
Basically that’s why after doing this experiment I am really motivated to study Japanese.
A friend of mine is teaching English in an university and she and all her colleagues are increasingly worried about the quality of machine translations between English and German.
1 Like
I follow a bunch of translators on Twitter and just saw this, pretty relevant.
I’d also say that DeepL is better than Google Translate for Japanese though, at least right now.
6 Likes
hmmm… well I like to read the furigana for my english to Japanese translations as well as get suggestions based on common phrases using the chosen vocabulary. I also take screenshots of subtitles as I watch shows so I can go back and review later so I’m looking for the option to import images to scan for the target language. I don’t see any of those options with DeepL.
Furthermore I like to save multiple languages into my translation history so I can swap them out depending on what I’m doing. I don’t see an option for this either.
My Japanese is poop and I need to get used to forming real concise sentences utilizing the most appropiate particles and kanji as possible. Why do you favor DeepL to Google Translate?
I know Google Translate is really bad for Arabic and just about any idiom for any language, but other than that…doesn’t Google translate pretty literally?
For someone who has limited understanding of sentence structure, am I formed bad habits with Google Translate? It does pull from google search history as well as its dictionary it seems. Does DeepL do anything like that?
thanks.
Plus some bonus insanity that we just made up sometime because it sounded more fancy - the word “debt”, for example, was “dette” in Middle English, but they just randomly decided to squeeze the B back in because it was there back in the Latin days (“debitum”) and so made you sound more smarter.
Or… write more smarter. Because the B is silent.
3 Likes
I’m quite happy with English being the de facto world language because the grammar is comparatively simple and it’s easy to incorporate foreign words, but the spelling is so awfully inconsistent.
One of my goals in life is an international spelling reform with phonetic consistency for English. 
3 Likes
Oh yeah, I totally miss the days when we used to pronounce the leading “k” in some words. Those were good days sob sob sob.
And you can do so many things with it, right? Except when you can’t and end up with weird phrases which don’t work 
If you’re relying on it to learn how to form sentences, then yes. Google Translate usually doesn’t produce grammatical errors but it often produces nonsense, really bad prose, awkward word choices, incorrect or extraneous context, and inappropriate politeness level. I wouldn’t really recommend machine translation tools for trying to learn grammar. They excel at getting some kind of rough idea of a text’s meaning when you have absolutely no knowledge of the language. If you’re studying the language, it’s not really going to be very helpful IMO.
Edit:
Also “translating literally” is kind of a weird phrase because it’s really hard to define what exactly that means. If you mean that google translate often misunderstands idioms, yes that does happen at times, and it can be difficult to decipher what the original meaning was.
2 Likes