Response selectivity to speech in the left hemisphere

There's a figure in Hickok & Poeppel 2007 showing activations to speech sounds (CVs, etc.) across several recent functional imaging studies, most of which contrast these speech sounds with various acoustic control stimuli. The figure shows bilateral activation consistent with our claim that both hemispheres are capable of processing speech sounds, and somewhat at odds with the common view that "phonemic" processing mechanisms are left dominant. What is not evident from that figure (but is noted in the text of the paper), is that many of the studies show more extensive left hemisphere activation to speech sounds when contrasted with non-speech controls, and a couple show only left hemisphere activation. This asymmetry seems to be the driving force behind the view that phonemic processing is left dominant. But here is something to keep in mind: speech activates left and right superior temporal regions rather symmetrically. What is asymmetric (at least in some studies) is the response to non-speech control stimuli. So the question is, how do we interpret computationally, the difference between a region that responds to speech as well as similar non-speech sounds vs. another region that also responds to speech but not as well to non-speech sounds?