Speaking in Tongues

Desert1 writes "Carnegie Mellon's renowned computer science department has developed a system which allows for conversation between two different languages called Tongues. Currently this has been used between Croatian and English, perhaps one day they will be able to develop one that will allow politicians to talk to normal folks and be understood." It's been in development for a while.

Comments

Brute Force

[chill]

I've long wondered why someone doesn't just brute force translation.

Create a human translated database of damn near EVERYTHING in two languages, like English and Spanish. Then, just do fast lookups.

Computing power is such that this would be possible.

Re:Brute Force

[kalidasa]

I've long wondered why someone doesn't just brute force translation. Create a human translated database of damn near EVERYTHING in two languages, like English and Spanish. Then, just do fast lookups. Computing power is such that this would be possible.

Because natural languages don't work like that. Natural language translation is an AI problem, it cannot be solved with brute force. Hell, translation by human agents is nearly intractable. This is a problem which has been worked on since the 1960s (if you think Babelfish is bad, you should understand that it's superb compared to the early efforts).

The syntaxes of natural languages do not map 1:1. For instance, some languages are ergative - they use agents rather than subjects to explain who is doing what. Though it sounds like all you'd have to do is map Tamil::Agent = English::Subject, and restructure the rest of the statement, it doesn't work that way. Vocabularies are the same way: each word has a semantic field which overlaps the semantic fields of other words in the same language, and overlaps the semantic fields in other languages, but almost never is the semantic field of one word in one language the same as that of another word in a second language. Try to translate the word "know" to French or "love" to ancient Greek and you'll understand the problem (French uses a different word depending on whether what you know is a person or an idea, e.g., while in Greek the words eros, philia, and agape all refer to different concepts that English speakers describe as "love").

It can't be translated the way suggested, either, with "bank" being translated as either "river's edge" or "money storage place" - leaving aside that the second definition is culturally contingent, there could be connotations in either phrase which are distinct from those in the source phrase. There are also problems with ambiguity - natural languages are by definition laden with ambiguities and contingencies, which require the ability to reason to find some corresponding structure in a target language.

This is a problem which will continue to be worked on for pretty much all our lives. If it's ever solved, HAL won't be far around the corner.

Esperanto...

[Average]

In the pre-computer days, some folks noticed that a neophyte (basic idea, needs dictionary)translation into Esperanto was much more comprehended at the other end than a neophyte translation to the destination language or a neophyte translation by the recipient.

The reasoning was that the process of translating into a more formal mechanical language clarified and codified ideas.

Once again, it's the dividing line between human and machine that's the problem. Millions of people train themselves to C or the shells. Fewer to assembly. But it takes some wetware work to push the human/computer boundary closer to the computer.

Like most programming has a learning curve, usually less than ASM, leaving language translation completely to the machine will be fraught and ambiguous. Good translation requires some push from normal speech, but maybe not so far as mastering every other possible language...

Re:Esperanto...

[kalidasa]

The problem with this is that Esperanto doesn't have enough resources to map syntactical and vocabulary fields from all possible source languages (and all parts of any given source language) to all possible target languages (or every part of any given target language). And the adding the resources would be a relatively intractable problem. This is an AI issue, not something that can be solved entirely within the machine.

Not quite real translation...

[Theaetetus]

The team built the system's translator using a technique known as example-based machine translation. Essentially, they created a database that holds a massive list of English phrases and their Croatian equivalents, culling data from bilingual Internet sites and university textbooks. When the engine receives a text phrase in one language, it provides the equivalent text in the other.

So, basically, it's a lookup function, translating the incomming speech and then comparing in a database... So, while they could have a huge dictionary that could cover most situations, they aren't really doing a 'translation' per say...
Although, then again, for anyone who has taken language classes, but are not fluent in the second language, isn't that what we do? I know that while I was taking French and Latin, to come up with phrases I would do phrase translations because I was still thinking in English. I wasn't fluent enough to think in those other languages, so I couldn't formulate phrases directly properly.

I suppose, in essence, this will work as a translator, but it is neither a babel-fish type universal translator nor is it any replacement for fluency.

Still cool, though. Now, can they get it to run on a Palm?

-T

Re:Long way to go

[achurch]

People will always do the translation gig better.

Oh, absolutely. I'm bilingual as well (Japanese and English), and particularly with Japanese, the language itself is so ambiguous that even native speakers don't always understand each other--you can imagine how difficult that made it to learn the language. ;)

But I don't think the point of machine translation is necessarily to get a perfect translation out; for that, the machine would have to be able to think like a human, and that would bring up all sorts of difficulties I don't even want to touch. But if the computer can do a good-enough translation, then the humans involved can figure out the rest. For example, another poster suggested the ambiguity of "bank"--a place where you store money vs. the edge of a river--but even if the machine translation got it wrong, the humans involved could figure things out in the end. (You could say "the edge of the river" instead, for example.)

I'm personally looking forward to progress in machine translation. While there will never be any substitute for actually learning and understanding a foreign language, realtime translation could go a long way toward improving intercultural understanding, and could help stem the loss of languages due to the spread of English and other "core" languages.

Re:Take it to a Pentecostal meeting.

[Jogar the Barbarian]

Just a bit of info "from the horse's mouth" as it were... :)

It is VERY RARE that glossolalia (speaking/praying in tongues) is comprehensible to any mortal man. Scripture refers to it as "groanings that cannot be uttered", and that when "[your] spirit prays, [your] mind is unfruitful". I take that to mean you don't understand what you're saying, either. I know I don't when I do it.

However, there are scattered reports of someone delivering a message in tongues, which was followed up by the interpretation (as God commands there to be), but that the original message was comprehensible by one or more strangers who just "happened" to come to that specific church meeting, and heard speech from their foreign, exotic dialect. (a miracle)

Messages in tongues are, IMO, distinct manifestations of the supernatural from merely "praying" in tongues. Praying in tongues I believe is was is described as being used to "edify your spirit", and is what Paul was referring to when he said "I thank God that I speak in tongues more than you all." That means he had a extraordinarily vibrant prayer life, one that was immersed in the supernatural. Messages, in contrast, are brought to edify an entire body of believers (and to "wow" the unbelievers), but only when it is accompanied by the interpretation... otherwise, it's just gibberish.

So, Message + Interpretation functions the same as the spiritual gift of Prophecy, it's just a two-phase form of the same manifestation: a message from God.