Sean Carroll's new book is (apparently at least partly) about model-dependent realism

Salon has an interview with Sean Carroll in which he discusses his new book, The Big Picture, which releases tomorrow (May 10th). Readers of this blog will no doubt be aware of my fondness for Hawking's model-dependent realism or what I consider to be a cognitive-science-based variation, Lakoff's embodied realism. And while I've often referenced Carroll's old blog post about free will as evidence he shares the basic ideas, some of his comments in the Salon interview certainly have a ring about them that should be familiar to anyone versed in those empirically responsible epistemologies:

Naturalists don’t all agree with each other. On the one end of the spectrum you have the most hard-core variety, who claim that only the most deep-down fundamental description of nature can be said to describe something “real.” They might say that consciousness, or morality, or free will, are all just illusions. On the other end of the spectrum you have naturalists who believe in only the natural world, but are willing to ascribe objective reality to various extra properties it might have – moral judgments, for example, or inner states of conscious experience.

Poetic naturalism sits in between. There is only one world, but we have many ways of talking about that world. And if a particular way of talking gives us a useful handle on what the world is and how it behaves, it’s completely appropriate to consider the concepts it evokes as “real.” Air is really made of atoms, but its temperature and pressure are real, even though the individual atoms don’t have temperatures or pressures. Human consciousness and free will are real, even though they’re not present in the individual particles or cells of which we are made.

Whether he ever mentions model-dependent realism by name in the book, I don't know yet (obviously); but he's certainly in the ballpark. Stephen Hawking essentially says that the question of what is 'real' is meaningless; what matters is the utility of a model in its ability to reliably describe and predict phenomena. Lakoff, similarly, suggests that we consider something to be 'real' when it has a theoretical ontology necessary to explain phenomena. I think Carroll is in good company.