Do plants compute? The blunt answer is “yes”. Plants compute insofar as they manipulate representational states. The sine qua non of representation‑based competency is off‑line adaptive behavior. Reactive behavior differs from truly cognitive one because it fails to meet the principle of dissociation (the states of a reactive system covary continuously with external states). Off‑line competencies thus mark the borderline between reactive, noncognitive, cases of covariation and the cognitive case of intentional systems. Nocturnal reorientation in Lavatera cretica leaves is not to be interpreted in reactive terms, since such a competency is not explained by means of online forms of covariation. (Garzon, 2007, pp. 210-1.)As I noted earlier, Paco is a representationalist embodied cognitionist (that's a pretty awful name). Here, however, he is a computationalist, representationalist, embodied cognitionist. This, again, makes it harder to be definitive about what advocates of embodied cognition believe.
Maybe someone should put forth the argument that because the advocates of embodied cognition can't agree on what they mean, they should give up that whole enterprise. It's hopeless. But, that someone wouldn't be me.