Jared Oliphint, "Using a two-dimensional model from social ontology to explain the puzzling metaphysical features of words"
Synthese, 2022
When a student or interested observer has asked me what the field of metaphysics is about, I have found it helpful to say something like, “Well, part of what metaphysics does is closely examine the things we talk about and refer to.” What kind of thing is the Eiffel Tower, or the number 5, or your mother, or Spiderman? What are these things like? Tell me more about these things you refer to, and how they relate to or resemble other things in the world we could mention.
At some point when I was a graduate student at Texas A&M, it dawned on me that I hadn’t yet seen any philosophers ask those metaphysical questions about the very things that are doing the referencing, i.e. the objects of language. So I began the scholarly search for philosophical work that focused on the metaphysics of language, while trying to come up with answers on my own. But I was failing to come up with anything satisfying: I soon found Kaplan’s groundbreaking 1991 paper that eventually (but very slowly) prompted work on the topic (Cappelen, Epstein, Wetzel, Miller, and others), but I wasn’t satisfied with the answers I was seeing to the metaphysical questions about words. How do we individuate words? How do they retain their identity? What kind(s) of things are they? Part of my problem, and why it continued to be a source of continual confusion for me, was that I was using classical metaphysical categories and questions to think through what a metaphysical account of words might look like, as if words are simply objects to be found “out there” that bear resemblance relations to other objects of the same kind, and objective identity conditions. The word “tree”, for example, can not only show up in other languages (if we make the right individuation and identity assumptions) and look very different from the English language instances, but the word can be expressed by other objects besides inky and pixely characters: it can be expressed through voice (sound waves), signs (hand gestures), Braille (raised bumps), and so forth. Try finding relevant resemblance properties between sound waves and raised bumps that accounts for those objects belonging to the same kind or class.
It wasn’t until I took a closer look at Brian Epstein’s work in particular that I had the “Aha!” moment: there is an entire field within metaphysics called “social ontology” that focuses on objects and kinds that depend at least in some way on social agents, like communities or individuals. (If you’re thinking to yourself, “Well, of course. Seems like it took you longer than it should to find such an obvious sub-field”, you’re probably right.) Objects used in language seem like paradigmatic social objects that depend in many ways on social communities to be whatever they are; they are not objects we discover, but objects we create. After struggling to figure out how particular words counted as tokens of a type, and being dissatisfied with attempted explanations that used types and tokens to explain the metaphysical features of words, I turned to Epstein’s work as an alternative. Without going into all the details, (roughly) he distinguishes between two kinds of metaphysical explanations: the familiar relation of grounding between an object and some kind (or type) in which the object is a member (or token), and his own innovative relation he calls anchoring between an object and the membership conditions for that kind (or type). I realized that it isn’t that the type-token relation is inherently defective, it is simply incomplete as an explanation of the metaphysical features of words, because a full explanation needs to include the various membership conditions that social agents give for a particular word like “tree”, for example, and all its non-resembling objects as instances.
So that’s the story of how the paper came to be. I introduce and motivate the problem, give a few points of data that a metaphysics of words should consider, show how the type-token model is inadequate for a full explanation of the metaphysical features of words, and finally show how the full two-dimensional model that includes the anchoring relation does a better job of explaining those features. A version of the paper ended up in my dissertation, and now the metaphysical features of words are not as mystifying as they were for me. This niche topic has been a fun entry point into other issues within metaphysics that I now see have more to do with social features and conditions. So I hope to continue working on the issue more broadly, assessing what features of objects might depend on social agents in some way, and in the philosophy of religion trying to understand how the theistic origins of language also play some role in a more robust metaphysics of language.