Our data is released under various Creative Commons licenses.More information
If you love this content, please consider a donation.
Menu
A->B or B->A
There doesn't seem to be an easy way to tell whether a pair of sentences are
A(Japanese) translated to B(English)
or
B(English) translated to A(Japanese)
I would like to make sure this feature is firmly placed in the wish list for future development.
Is there a specific case where you would need this information?
One not too difficult way is to look at the creation date of each sentence (which you can see in the first entry in the logs). If A was created before B, then it must have been A->B.
> Is there a specific case where you would need this information?
Not as such, but I can explain _why_ I want this information to be recorded / displayed.
One common use of the translated example sentences is to explain what the original sentence means. So it is quite normal to have, for example, obscure English translated (explained) into normal Japanese.
Imagine you have this:
http://tatoeba.org/eng/sentences/show/270663
A. Many a mickle makes a muckle. [Proverb]
B. 塵も積もれば山となる。
B is the Japanese equivalent (and 'translation') of A.
Some well meaning person might decide that hardly anybody knows what "Many a mickle makes a muckle." means and 'correct' it into a different English sentence.
A more common example would be the
http://tatoeba.org/eng/sentences/show/71044
A. I'll make you a present of a doll.
B. あなたに人形をお贈りします。
If B. is the original and A. the translation then you could well say that the English in A is a little odd and should be changed. If A is the original and B is the translation then you could say that the A demonstrates a somewhat old-fashioned phrasing and B explains is.
In those cases it would be desirable to have both phrasings, with uncommon ones marked somehow, so users should be encouraged to add alternate translations rather than 'correcting' existing ones when the translation is not erroneous.
I think this could be handled by attaching more metadata to sentences. Properties like masculine/feminine, proverb, quotation, polite, slang, dated, etc. could be tagged onto a single sentence, so you could have e.g. both a polite and a colloquial translation, and mark them as such.
That would also address the problem where sentences from the Tanaka Corpus are currently marked with tags like [F] on the English sentence, even though the property belongs to the Japanese sentence. That doesn't work that well when the sentences aren't restricted to pairs.