Different words and expressions
It is obvious that different languages have dissimilar words and expressions that are used to describe the same phenomenon. An machine translation system frequently has problems to recognise in which situations and how these different words and expressions should be applied. It needs to be emphasised that a machine translator is at a disadvantage for the reason that it typically only has pure text and sound as data input, whereas a human translator is able to use their knowledge in order to convey the meaning of the particular vocabulary item. For obvious reasons, all the above additionally assist the human translator in coming up with a more functional translation in a wider scope of contexts.
As described in previous sections, in order to properly translate the particular expression, it is necessary to fully understand the whole context in which it is used in the original text. In the majority of the above cases the machine translation system may lack the resources to generate an equivalent translation. What is more, there are words that are characterised with polysemantic meanings, which makes the process of MT even more challenging. In order to illustrate the above case, the following situation can be considered. When the person notices a sign in a field which reads “This path across the field is free, however the bull can charge!” they are able to effortlessly understand from the context that the bull in question will not be collecting an admission fee, but it can charge towards the trespasser and chase them. The above presented example shows the way in which the level of contextual input that an machine translation system deficiencies, is related to the level of ambiguity, which may be present in the target text. Therefore, it can be stated that in these situations in which machine translations seem to be illogical the above process usually constitutes the reason of this state of affairs.