Why do online translators not work?

Answer 1

Online translators indeed become confused by English words that have more than one meaning (which is a STINKIN LOT of words). For example, the word 'you' has a different grammatical meaning in the following two sentences.

I give the flower to you.
You give me the flower.

(In the German language, the direct object you is "dich" and the subject you is "du"). Machine translations have no way of distinguishing.

Other languages also have words with double meanings - for example, "sie" in German means both "you" and "they" - and they have the same verb conjugation as well.

Finally, as some other answerers have mentioned, culture. English idioms such as "it's raining cats and dogs" do NOT translate into their actual meaning. The Germans say "es tut mir leid" to apologize, but this literally means "it gives me sorrow." There is no international, inter-language standard as to grammatical and cultural consistency.

Hope this helped :)

Answer 2

There are several reasons for that but I'll give you a simple one. You know that online dictionaries (dictionaries which just translate single words and expressions) do work quite ok. But you also know that if you translate a single word to another language, you get several possible translations for it. This is because most words have more than one meaning, I think this is true for any language.
Let's say that the average number of different translations for one single word to another language is three. (In fact I think it's much more).
Now, imagine a whole sentence of ten words. And you try to translate this sentence. If every word in the sentence has three different meanings and therefore three different translations, you get approximately sixty thousand different possible translations for the whole sentence! Software translators return only one of these possible translations, which they might deem the most likely one from their algorithms, but in most cases they are dead wrong.
Translating does not mean a mechanic replacement of words by their translations. Translating means full understanding of a text in the original language first (which often requires additional information apart from the text itself) and then finding equivalent ways in the other language for expressing the same meaning. This is not a mechanical process. It suits the human mind much more than a mechanic processor, however complex. Essentially it is the same problem as making a machine understand natural language at all - which also is far from solved yet.

Answer 3

Because language is a tool of communication. Just think of the ways in which we communicate! Words are only part of the total process. We use smiles, gestures, inflection of tone, etc. Where we have to rely on the written word, we introduce those elements in subtle ways by our use of language. The machine can never pick up those things. It can only render the words literally, with no concept of the subtleties behind their use. Asking a machine to communicate for us is like asking an actor to say his lines in a monotone without any movements or requiring a dancer to perform steps with perfect technical accuracy, but no feeling.

Answer 4

Because of the colloquialisms in every language and their inability to distinguish how a word is being used in a sentance. (Do you have a watch or do you watch it?)