As Iina stated in a previous post, “context is machine translators’ weak point”. Likewise, Jonathan Downie wrote a guest post, which is a reply to this post of mine, where he states that context is a fundamental factor within the communication process. He sustains that context changes the perception of the message, and therefore it must be included as an element that “rearranges” the traditional sender-message-medium-receiver diagram. In general terms, I cannot disagree with the remark. Nevertheless, context can and should be thought of not only as a barrier to communication (when contexts are different), but also as something that allows communication (when it is shared).
Context is an essential concept with which any translator must count, although it is quite difficult to grasp what exactly “context” is, in its nature. For example, when reading online translation or dictionary forums, very often the person attempting to answer a doubt asks for the context around the doubt, which is typically a word, phrase or expression. The context, thus, is always “around” and never “is” the point we are trying to understand. Similarly, context seems to always pre-exist the linguistic doubt, and also verbal communication. In fact, it pre-exists any kind of communication. This is one reason why context is conceptually difficult to grasp: it is a pre-condition of any communication involving two or more parties. Therefore, any attempt to communicate what “context” is, like this one, is already presupposing it, making it arguably impossible to regress into its specific and distinguishable nature.
There are two important and interesting philosophical enquiries that can follow from this idea. The first is perhaps angst-inducing and hits the common assumption that when we talk, we are talking about things that exist in the world and we are expressing them as they are. However, if indeed we can only communicate through shared context, what we are achieving is communication between two agents. The communication between them may hold absolutely no reference to the world outside their communicational process, although they most likely believe otherwise. If this premise is correct, what follows is that we are not able to assert definitively that the world is intelligible through our communication. In other words, reality may remain forever a foreign dimension, even though we believe we are stating an undisputable fact when we say that “the chicken is crossing the road”.
The second possible implication concerns AI and translation. As Wittgenstein stated, the construction of meanings, in language, can arise only from the Lebenswelt of the communicational agents. This shared world of life can be thought of as context itself. It is built not only on the formal linguistic architecture, like vocabulary and grammar, but, to a similar or even greater extent, on petit-perceptions, everyday stimuli and non-philosophical lifestyles. Only by imagining a computer that acquires higher and different capacities and skills than the ones programmed in it can we speculate about perception of context in a machine. This is still science fiction domain. For now, the hope to achieve optimized communication and translation must rely on humans that are biologically programmed for life and for “shared life”, capable of sharing contexts and, perhaps, incapable of knowing what it is and what lies beyond it.
Software localization is much easier with a visual editor. You’ll see if the translation looks good and fits into the place.
Try Multilizer localization tool to see it yourself.