why do computer-based translators have so much trouble going between Chinese and English? Can’t someone have come up with an efficient algorithm by now for making sense between the two?

808 views

why do computer-based translators have so much trouble going between Chinese and English? Can’t someone have come up with an efficient algorithm by now for making sense between the two?

In: Technology

5 Answers

Anonymous 0 Comments

Languages are very complex, and English and Chinese are like polar opposites.

Imagine a language where “blue”, “fish”, “sky” and “rock” are all represented by the word “frog”. And there is no verb conjugation.

Translating a sentence like, “I see frog”, becomes slightly more difficult… You don’t know if the person is talking about seeing a fish, the sky, a rock, and you don’t know *when* they saw it – are they seeing it now, did they see it in the past? Did they see it one time or many times (is it an habitual action?)

edit: The same can be said of languages that are even more similar. Time tenses are expressed differently in different languages and different tenses can imply different nuances (like the difference between “I will go there” or “I plan on going there” or “I am going there”). Sometimes a language has one word that requires a many-word explanation in the other. The Swedish verb “orka” is a great example – it means “to have the energy or force to do something”. As you can see, in English, that takes an entire sentence on its own to explain, and there are even many other ways that you will find it translated based on context.

Anonymous 0 Comments

Do you speak both Chinese and English? My understanding would say there is not a basic algorithm to link the two, and a very complex logic system would have to be developed

Anonymous 0 Comments

Chinese language is more than just the words. It’s cultural meaning. There are a number of classic xiangsheng (Chinese comedy skit) about the Chinese language. One that stuck out was “chi le ma?”. Literally translated the words means “have [you] eaten?” But the phrase is used as a casual greeting from cultural usage. There are ALOT of these in common everyday use. Just like “what’s up?” doesn’t mean asking about the ceiling or the sky, there are numerous phrases in Chinese where the phrase meaning has little to do with the words in the phrase.

Some examples

http://carlgene.com/blog/2012/02/45-mandarin-sentences-with-chinese-characteristics/

My fav is “bie ge wo bu san bu si, would ge ni yan se kan kan”

Literal word translation “don’t give me no 3 no 4, I’ll give you color see see”

Meaning “don’t tell me any BS or I’ll put the hurt on you”

Anonymous 0 Comments

A short answer would be that the two languages are just too different to be translated into each other mechanically. Some might even say fundamentally different. As a native speaker of mandarin who is also fluent in English I can assure you translating between the two is really hard, when I switch to English mode I construct sentences and use expressions almost completely differently. And as someone already mentioned, you need to have a really deep understanding of the culture behind a language to successfully understand all the subtle things in it in order to translate.

Anonymous 0 Comments

Translation is hard. You can’t directly translate each word, you must first understand the meaning of a sentence to translate. So in order to translate from Language A to Language B you must first be able to understand A (that is, extract meaning from the words).

This is very difficult to do – even in English. Think of the following sentence:

> “I shot an elephant in my pajamas.”

Can you answer who is wearing the pajamas? There are 2 possible answers: me or the elephant.

It turns out that natural languages are very ambiguous (you can read sentences multiple ways), which makes ‘calculating’ their meaning very hard, which in turn makes translation very hard. And ambiguity is just one of the many difficulties of natural language processing (the part of computer science dedicated to understanding languages).