In general, academic MT observers have tended to view postediting as the weak link in the MT value chain. One of their research targets is to remove the need for postediting as an unfortunate human intrusion into a fully-automatic process.
I’ve already expressed my opinion that it is not possible to do without a human being in the translation process. There are two main reasons for this:
- Inherent ambiguities in natural languages: consider the following entertaining sentence, composed by Martin Gardner:
Wouldn’t the sentence ‘I want to put a hyphen between the words Fish and And and And and Chips in my Fish-And-Chips sign’ have been clearer if quotation marks had been placed before Fish, and between Fish and and, and and and And, and And and and, and and and And, and And and and, and and and Chips, as well as after Chips?
Here quoted and unquoted text get interspersed freely, causing havoc in any attempt to parse it mechanically, even though the meaning is pretty clear once the underlying pattern is picked up. No statistical machine translator will ever be able to get this right.
- New words or new meanings of previous words in the source language: up to a few years ago, “to google” meant nothing (my spellchecker still flags it as a mistake). Statistical engines trained up to the point of perfection up to a particular point in time will need to get the new words and their respective translations (in their respective contexts) from somewhere. This first appearance will necessarily arise from a brand new translation done by a translator (or post-editor) that will eventually find its way into the relevant statistical engines.
This is why our company, although betting heavily on the recent developments in this new technology, also invests in recruiting and training linguists for this challenging and exciting task. If you are a trained translator and want to register with us, you can do so in our registration page. If you are interested in translating documents, you can request a free quote at Translation Services.