Concerning the question of the subject:
1. Google Translate is based off of statistical methods. It doesn’t actually know the rules of language, and it has no idea about context. It’s just aping the patterns based on what it’s seen before! It’s a “fill in the blanks” sort of approach, which has its merits and demerits (yes, there are other approaches to machine translation; they suck too, just in their own ways). It definitely doesn’t do any thinking, though. This gets really bad if it has to fill in details or resolve ambiguity to get the job done at all; going English to Spanish, “Dr. Smith” might be correctly “Dr. Smith” or “Dra. Smith”, but if Dr. Smith identifies as, say, a woman... then as far as I know only one of these is proper and respectful. And some words/phrases simply have 15 definitions; it’s just not likely to pick the obscure one, even if it happens to be the only correct interpretation.
2. I hear tell that Google Translate doesn’t have as many English–Japanese text pairs as they’d like. For starters, because Japanese is an official language of neither the UN nor the EU and that’s where they got a metric truckton of their parallel texts.
3. The less similar the languages, the less well statistical methods work. Word order is a particularly obvious problem here; word order in translation is a problem that requires judgment more than it does correlation. And Japanese word order is reversed compared to English in so many
4. When you deal with specialized terminology, statistical methods don’t always have the right material to work from.
You can read about how these methods generally work on Wikipedia
Concerning the passage, this is the official line from the Japanese dub, courtesy of this site
, which has rated this translation 最低 (horrible), and it’s still better than Google Translate:
Already there is an incongruity between 船長のログ (Captain’s Log, “Log” here in English) and 航星日誌 (something like “star-sailor’s diary”, I’m guessing it was chosen to evoke the translated title of Stanisław Lem’s “The Star Diaries”, but I didn’t check that theory too closely), the term used for the TOS and TNG dubs. More poetic. Better.
スター日付 is literally “star + date” with “star” in English and the correct “date”. Mashing them together as if a compound word is not as good as the official convention 宇宙暦 (“space calendar”, more or less idiomatic Japanese just like “stardate” sounds in English). Really, compounding English and Japanese words is done all the time in practice, but it is stylistically to be avoided in general in the same way one should avoid compounding Greek and Latin or English and Latin if you don’t want to sound just awful
This one gets a little tougher; I’ll color-code the parts that roughly correspond:Starship Enterprise on training mission to Gamma Hydra
, Section 14, coordinates 22-87-4.宇宙戦艦エンタープライズ号
is not good as 宇宙戦艦 (space-battleship) in the dub (it is, as emphasized throughout the series, not to be treated as a ship of war). The usual translation for ST’s “starship” is 宇宙艦 (space-ship). As far as I know, there’s nothing strictly wrong with 星船 except that’s never ever ever been the ST term and it’d still be a strange term in Japanese SF.on training mission
- trimmed-down language as appropriate for log. It’s not “Enterprise is
training mission”. 訓練航行中 conveys this tone (though it’s not exact; 航行 means “sailing” or “navigating” more than it does “mission”). 訓練ミッションの is horrid, making it sound like training missions are the only thing the ship does.to Gamma Hydra
- Both translations treat “Gamma Hydra” as a place name, but it’s really meant to be Gamma Hydrae, a star in the constellation Hydra. The site I linked suggests 海へび座ガンマ星系 (the Gamma Hydrae system).Section 14, coordinates 22-87-4.
- Notice how GT splits this up into two, not understanding that these need to go together, and that the ordering of mission then place was orderly format. (Also note that the official translation says “15”, which is a mistranslation, hence the rating of “crap”). Again, GT chooses an English word where there is a more euphonic kanji-based equivalent; some of this has to do with modern trends towards preferring English, but it’s still overdoing it. Then again, it had no idea what “section” here meant, so what better word could it have used? It could have disambiguated, but instead seems to have played it safe.
All systems normal and functioning.
全装置順調に作動。 - Not exactly great according to our critical site, but at least it gets the intended meaning across and I wouldn’t object because I don’t have that fine a grasp of the language.
すべての通常のシステムと機能。 - means “all of the normal systems and functionality”. Not the sentence fragment we were looking for. Garbage.