ghoti

blog

Understanding translation through corpus

Do we really need to know grammar to understand language? What if we can have billions and billions of words of data to work with and put it through a computer?

Well, that is pretty much what corpus linguistics is.

Here is a nice article about how Google Translate works. It may not be full proof but it is better than anything else we have. A human being will have heard an uncountable number of words in his or her lifetime and we use that input to understand what others are trying to say. So why not a computer with lots of etext in it?

That is I guess the attraction I have to corpus linguistics.

Published by

Warren Tang
I write about linguistics and philosophy. I also write poetry and take photographs.

3 responses to “Understanding translation through corpus”

  1. signature103

    James,
    I absolutely agree.

    From a translators point of view (this translator) machine translation will never replace human translation as long as machines fail to understand the meaning of the utterance. And any meaning of an utterance is complicated by context and by what the speaker/writer wants to to convey. The meaning is complicated a second time by his/her interpretation of the context or his/her view of the context, and then by what the listener/reader believes to be the meaning (especially if one is a translator, say, of a literary text).

    So for a machine to understand an utterance it needs to be embodied as humans are and active within the system the speaker/writer is in to have a chance to understand the meaning.

    Clients in need of translation services will generally have no understanding of the complexity of this if they are completely unaware of what language actually entails, especially if the client is from a monolingual background who has had little or no exposure to the second language experience. So, yes, you do need to educate them about this.

    The academic side of me is interested in how far artificial intelligence can occur, and this is what I have blogged about here.

    Like

  2. James

    When it comes to Google Translate, translators all agree on the fact that cannot be absolutely considered as a reliable tools to translate. However, many people tend to still use it a lot. The most important issue today is to let people know how bad it can be. It is central for clients to understand that they need to rely on a professional translation agency for the job they set out. The hard thing to do is to convince them about it! Certainly these blogs can help in this sense, so that awareness of quality can be raised among clients.

    Google translate not only gives a general wrong view of the foreign text, but the worst thing is that it is believed to be working in a very good and efficient way! This is only because people do not understand the foreign language in question and rely on it for everything.

    Like

  3. Muhammad Shakir Aziz

    I work as a translator and agree with this article. Mostly we encounter similar sentences specially in specialized domains e.g. Localisation, Software Translation. Thus the role of translation memory tools becomes important.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

Gravatar
WordPress.com Logo

You are commenting using your WordPress.com account. ( Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. ( Log Out /  Change )

Cancel

Connecting to %s

%d bloggers like this: