I thought I might try to open up a dialogue that I feel could have
profound consequences on the way we view other animals and ourselves.
We know that some animals, take dolphins for instance, can communicate
via sound. Assuming that we could alphabetize, syllabarize, or
logographize the sounds a dolphin makes, could we then develop a
translation to say English?
The problem seems to be in Initial conditions.
Here is a question; given billions of comprehensible excerpts in
English, and then billions of different excerpts in say Spanish, could
we develop a translation between the two?
It seems that we could recognize patterns and that we could find
certain similarities between certain words. Perhaps we (a computer)
could determine that some words are commonly modifiers, or
conjunctions, or objects, and that some words commonly show up
together... but without any initial conditions, wouldn't a computer
possibly confuse [Spanish: caballo (horse)] with [English: cow or
maybe even truck]?
Google translate uses, among other things, user submitted translation
as its initial conditions, but are there perhaps other types of
initial conditions that we might use, better suited to this problem?
Maybe our worlds are just too different. Your thoughts on the matter
are appreciated.
