Slashdot Mirror
Assyrian, Babylonian, Sumerian Translator Created
DrJackson writes "A new online translator that can translate Assyrian, Babylonian, Sumerian and Egyptian hieroglyphics (1 of the 3 types anyway) has been developed. This is the first time I ever saw a translator for cuneiform. Something like this would be great for translating interesting historical records like the Amarna Letters."
Machine translation sucks. Among other things, idioms, set phrases, wordplay, and most importantly the fact that there is rarely a one-to-one mapping between languages (often resulting in either a loss of information or requiring missing information to be added, which often requires knowledge of the culture of the language's people) all present challenges that make it unlikely that anything short of human-like AI (or very close) will be able to do good translations. Or to put it more briefly, "Nothing to see here. Please move along."
The whole page is total crap:
The Only thing the "translator" does is taking an English word and match it with lemmata in a lexicon then it takes the first hit and then it goes on. Try typing "I have seen you" you'll get "[I] [have] [see]n [you]" it simply cuts of the "n" of seen and leaves it there because it can only find uninflected forms. This is less than nothing.
And by the way the statement "For best results, use simple words as language has developed a lot since the time of this ancient language." under translation is one of the most stupid things I have read on an academic page language dedicated to some aspect of language. They should just take a Sanskrit dictionary (or whatever ... Maya ... Classical Chinese). Language then and now is pretty much the same, but apparently in some places technology hasn't developed that much, grumblegrumblegrumble...
"Hannibal's plans never work right. They just work." Amy/A-Team
As a person who studied Latin at the high school and collegiate level, I know that much of what is 'worth' translating academically has already been translated by other academics. Sure, a scholar might be able to come up with his own unique translation, but that is not something that can be done by a machine.
A dear friend of mine is an Egyptologist, and I know his struggles in translating writings from different regions of the empire, let alone differences dynasty to dynasty.
Since even the best computer translators (and I mean the corporately deployed ones, not just freebie Web stuff like BabelFish) mangle all but the simplest Spanish, French and German (I can't say anything about Asian languages, as I can't speak or read any) phrases, how can we expect any level of reliability in translating languages that even leading scholars struggle with?
Besides, the most difficult part of translating anything stems from the fact that any person seldom speaks or writes as he should. The rules of language are bent, twisted and altered into regional dialects and strings of ethnic and cultural phraseology. In the Spanish language, a word may take on one meaning in Mexico, and entirely another in Spain. Nevermind the fact that, at least in my experience, Spanish Spanish is significantly different from Mexican Spanish. And those are two languages that diverged only a matter of hundreds of years ago, as opposed to the thousands often seen in dead languages.
This is very interesting to me, but until we have widely-available computers that can understand the subtle nuances of tone, inflection, humor and colloquialisms, the computer translation will never best, or even come close to a careful academic translation, or a translation done by a human fluent in both languages, if not academically trained in both languages.
Message contains 1 attachment: spam.gif