Problem of machine translation is a special case or new universal language problem (look Creation of new universal language). Although it could be solved without universal language. One of approaches to solving is the translation through language-mediator, which role is suitable for new universal language.
At present time in systems of automated translation is used only linguistic model of choosing the translation of starting text. This model practically doesn’t cope with translation of homonyms (grammatical and especially lexical), special terms and multi-meaning words. It conditioned by the fact, that linguistic model based on analysis of the words, different notions could be determined by different words (synonyms) or different notions could be determined be equal words (homonyms, polysems).
Such restriction of linguistic model has already now essentially decreased the efficiency if translation systems and close opportunities for further increasing of translation quality.
On the other side, if you have reflection of different languages into notion space, you will be able to construct direct comparison between notions, described by one natural language, area in notion space, correlated with given word, and word, meaning the same notion in the second (the third, the fifth) language.
The advantage of such approach is that co-ordinate mechanism of the CSNT has insert protection from mistakes, connected with words close in meanings, homonyms and polysems.
Because the reflection of the notion into notion space happens with whole context, it is obviously in what subject area of space notions of start text are located. The calculation of exact location make with help of notion algebra’s mechanism, which helps to chose reaching of the notion to allowed or restricted area (look Allowed and restricted areas and actions) and solve problem of polysems and homonyms before translation.
If we know in what semantic space area the notion, determined by the word to be translated, is located, we will be able to chose right translation and construct opposite correlation of notion and words of target language.
It helps to get intelligent translation from one language to another.