Abstract
I consider two attempts to combine realism with pluralism about the units of selection:
Sober and Wilson’s combination of “model” and “unit” pluralism, and Sterelny and
Griffiths’ “local pluralism”. I argue that both of these attempts fail to show that
realism and pluralism are compatible. Sober and Wilson’s pluralism turns out, on
closer inspection, to be a kind of monism in disguise, while Sterelny and Griffiths’
local pluralism involves a combination of realism and anti-realism about interactors,
and the units of selection, that is fundamentally unstable. My conclusion is that one
must choose whether to be a realist or a pluralist in this area: one cannot be both. The
question of which we should choose is a further question that I do not take a stand on.