Wednesday December 28, 2011

I was thinking and doing som reserch about one of my kurent projects: Making an English languaj that is syntakialy limited such that it makes posible automatik translation into lojikal formalism. I stumbled akros som prety interesting artikles listed below: http://en.wikipedia.org/wiki/Controlled_language One of them has a website wher one kan find an introduktion to the system. It is aktualy very good and worth reeding. It is a 80 paj powerpoint presentation turned into PDF. http://en.wikipedia.org/wiki/Basic_English The 1944 paper kritikal of Basic English: How Basic Is Basic English? http://en.wikipedia.org/wiki/Simplified_English http://en.wikipedia.org/wiki/Plain_English http://en.wikipedia.org/wiki/Plain_language The benefits of using plain languaj ar rather obivus and konkreet. Using non-plain languaj makes komunikation take longer and proseed les optimal. This is mostly just waste of time but somtimes it is a mater of life and deth. My (it is shared) projekt has som on-going diskusion in my forum. However, the languaj i hav in mind is mor similar to formalism than ACE is (the one linked to erlyr). I think that it is too problematik to handle nested konditionals with quantifyrs like:

F1. (∀y)(∀xFxy→Gxy)→Fy

in sylogistik languaj, i.e., as in sentenses like:

S1. “All men are human.”

Rather, one needs sentenses that ar harder to understand and les like ordinary English but beter for formalization like:

S2. “For any X, if X is a man, then X is a human.”

In simple kases, such as the example sentenses with the form:

F2. ∀xMx→Hx

ther is no need for mor advansed sentense syntax, but in the kase of the formalization F1 ther is need for such sentenses.