It’s a desideratum of any mereological theory that it be able to capture pretty much any scientific theory. By this I mean that it must be possible to formulate that scientific theory within the terms of the relevant mereological theory. If there is no version of mereological nihilism that can capture general relativity, that is an awful result for nihilism.
Let us designate what must be captured by any mereological theory C. You might think that C must contain all sorts of things. You might think it contains folk views. You might think it contains all possible scientific theories. You might even think it contains moral or aesthetic statements. Definitely it contains the true scientfic theories. We will remain agnostic for the moment on what else it might contain.
Let us designate the individual bits within C that roughly correspond to statements (though of course they cannot be statements yet) ca,cb,cc…ci.
Let M be the set of mereologies (m*, m%, m$…mi), that can capture all of C, no matter how difficult the required convolutions.
Let Fmz be a function which maps any cx to an appropriate English sentence sx, where the sentence expresses the relevant content in terms acceptable to the relevant mereological theory (in this case, mz). Really Fmz will not be a one to one function; rather it will map bits of C onto many possible acceptable expressions.
Call the set of all such sentences which each express some bit of content from C (i.e. all those sentences which Fmz maps onto), Lmz.
What I want to suggest is that to say mz is true, or to put mz forward as correct in anyway is to express a preference for speaking Lmz. To deny mz is to express a preference for not talking in mz. It is easy to see how such an account could be, in principle, generalised to other metaphysical theories. It’s also easy to see how you could transform the theory into subjectivism, or even cultural relativism or some other relativism instead of expressivism.
But, you might think, aren’t the nihilist and the universalist happy to speak the same language in an everyday context (largely, the language of the universalist)? Yes, but presumably there will be some formal contexts were they are not happy to speak language. On this account perhaps they prefer the nihilist language in some contexts they take to be more definitive or important.
If the nihilist still insists they are willing to use the same language as the universalist in all contexts, if they insist that they are talking about something beyond language, something that really is, they are simply speaking nonsense.
Further problems to dispense with include
1. What are the boundaries of C?
2. What do the individual particles of C amount to, if they are not statements, but abstract contents of a sort.
3. Accounting for different senses of statements. So for example you might mean “there is no chair in there” in a metaphysical sense (as a thesis about ontology) or as a simple statement of fact. How do we distinguish these grades of meaning? Possibly we will have to adopt a more sophisticated expressivism which allows for both a factive and expressive content (where the factive content will basically just be bits of C).