(1) For all x, if x is introduced to explain the overt physical behavior of paradigmatically physical objects, then x is a physical object.
(2) The special ontological commitments of cognitive science (mental representations, information processing mechanisms, etc.) are introduced to explain the overt physical behavior** of animals' bodies.
Therefore, (3) The special ontological commitments of cognitive science are physical objects.
I think the argument is also sound if we replace both occurrences of "special ontological commitments" with "novel linguistic forms (e.g. relations to mental representations)" and "physical objects" with something like "physicalistic linguistic form."
I want to call this an identity theory (maybe a bit cheekily) because (3) is, as far as I can tell, just what all identity theories have in common. It would be a new version because it is an identity theory no matter whether or how the posits and theorems of cognitive science are identical to various paradigmatic objects and theorems in and about the body.
Perhaps some won't like (3) because the posits of cognitive science lack spatial location, and aren't physical for that reason. I myself am willing to give up the intuition that something has to have a spatial location in order to be a physical object. If I weren't so willing, I would note that, as far as I can tell, nothing is lost by stipulating that these posits are located diffusely throughout the body.
Seriously, is this totally undefensible for reasons I can't see?
* - I tend to think phenomenal consciousness is fundamentally different. Let's restrict "mentalistic talk" and "mentalistic ontology" to the theory and ontology of cognitive science, narrowly construed.
** - Where "overt physical behavior" is taken to include some number of physiological states.