Brian Weatherson (home page, blog) has authored an SEP article on David Lewis. In it he ably takes on the difficult task of summarizing in a few dozen pages the work of one of the most prominent and productive philosophers of the latter part of the twentieth century.
One section addressed something I had idly wondered about Lewis. As Weatherson discusses in his section 5, entitled “Humean Supervenience”, much of Lewis work involved reduction: whether the subject was the mind, language, laws of nature, or causation (modality was a rather spectacular exception), he presented arguments that all the truths about a world ultimately supervene on a set of (perfectly natural) properties and relations in that world. He then argued that these, in turn, are intrinsic properties of point-sized objects and spatiotemporal relations. (The “Humean” aspect of this is that he follows the spirit of Hume’s “dictum” that there are no necessary relations between distinct entities.)
What I had wondered was: did Lewis ever acknowledge that his picture of locally distributed point-sized objects in a space-time container was a classical view, inconsistent with modern physics?
Those who challenged this aspect of Lewis’ work over the years argued that his supervenience base is incapable of supporting some of the truths about the world, and so his attempts at reduction fail. However, most critics haven’t quibbled with the assumptions about physics, rather they argue that some additional higher-level properties and/or relations need to be added to explain the world. Even many materialist philosophers think additional brute metaphysical structure beyond the basic physical entities is required (this debate was the subject of this old post); some philosophers think necessary connections are needed to explain causation; dualists will see a need for fundamental mental properties and perhaps psycho-physical laws.
But, Weatherson notes that a second way to argue against Lewis’ project is indeed to point out that modern physics is inconsistent with Lewis’ notions. At a minimum, he says, quantum physics seems to need non-spatiotemporal relations to explain Bell’s theorem.
Weatherson explains that Lewis was indeed aware that his picture was inconsistent with quantum physics, but still thought it was extremely valuable to defend the thesis regardless of that fact. Weatherson explains the idea as follows: to the extent his picture of physics is wrong, it is because physics has more content in it (e.g. entanglement relations) rather than less. So, if Lewis can successfully defend reduction of various folk theories on his terms, it will remain a valuable accomplishment.
That makes sense, but I still wonder if explicitly reckoning with quantum physics and its interpretation couldn’t helpfully reframe these philosophical debates. There is more to the shift from classical physics to quantum physics than just adding some non-local connections. There is the quantum measurement event itself, which I see as an ineliminable new addition to the basic constituents of nature. I think that mind and causation might be the targets of successful reduction given an improved supervenience base which featured a network of measurement events in its basic ontology. (Also, one neat thing about taking this approach is that we could hold on to Hume’s dictum. He may not have known about QM, but I think the indeterminism involved in measurement events is consistent with his insight.)