ESI Professional and the Split Function - Part VI
We have given you a few examples of how ESI is capable of splitting a MWE (multi-word expresión) when it does not fit in the context. We could furnish hundreds more to illustrate this kind of behavior. For instance, take the following sentence:
The verb “have” followed by the particle “to” forms a MWE which has a totally different meaning. It implies an obligation: Tengo que (I must) . ESI recognizes this context and translates accordingly.
However, if we change the sentence slightly thus:
This is the only sheet of paper I have to write your article.
we then see that the MWE “have to” no longer fits in the context. It should now be SPLIT into its two components “have“ + “ to”, with the meaning of “possession”
Now we will see some examples where ESI fails to automatically SPLIT the MWE and you have to do the job instead.
It just so happens that using exactly the same sentence pattern as above, but changing the words “sheet of paper” to “idea”:
“This is the only idea I have to write your article”,
the translation comes out wrong!
It says: Ésta es la única idea que tengo que escribir su artículo.
Why does ESI fail to automatically split the MWE? Shouldn’t it always do its job correctly? Actually, yes! It should do it, but for some reason it fails! Remember that no system can ever be 100% correct. Not even humans are always 100% correct.
In due time we will get to the problem of why ESI is not splitting some MWEs, analyze it, and come up with a solution. This is always the case with each successive update. It is a never-ending job because, as we have explained it in the other articles, the possibilities of incorrect translation are endless. It is an infinite number.
On the other hand, this shows off the real value of the Interactive system available in ESI Professional. With just one click (on the SPLIT button) you can always fix the whole sentence without rewriting or editing it manually.