Try Googling for "Chrysippus' dog". That, at least, brings up hits like http://my.fit.edu/~aberdein/DogLogic.pdf. Just reading that paper made me realize that there were at least two dimensions in which the statement
"different species being able to deal with different abstractions (eg, certain dogs being able to recognize conditionals but not predicates)"
could be be misinterpreted. Assuming you were looking (as was I) for experimental evidence supporting that specific statement, I'm guessing we'll both be disappointed. But maybe not!