Social Ontology
A twenty-pound note is a piece of polymer with no unusual physical properties. Its being money is not a physical fact about it, yet it is a fact — one can be wrong about it, and counterfeits are counterfeits whatever anyone thinks in the moment. Borders, debts, marriages, universities, and parliaments are all like this: real enough to constrain behaviour and to be discovered, while depending for their existence on human attitudes.
Social ontology asks what such entities are and how they fit an inventory. It is treated here as an ontological question — what exists and in what category — and not as a political one.
Institutional facts
Searle's account remains the standard framework. Its central device is the constitutive rule:
counts as in context .
This piece of paper counts as a twenty-pound note in the United Kingdom; that person counts as chairman in this committee; this line counts as the border in international law. Constitutive rules differ from regulative ones (drive on the left) in that they do not govern antecedently existing activity but make it possible: there was no chess to regulate before the rules of chess.
Two further notions do the work:
Status functions. The term assigns a function the object cannot perform in virtue of its physics alone. A screwdriver's function is physically grounded; money's is not, and a note works as money only because it is accepted as such.
Collective intentionality. The acceptance must be collective — a matter of we-intentions rather than an aggregate of individual ones. Searle takes this as primitive and irreducible to individual intentions plus mutual belief, which is a substantive ontological claim and much contested.
Institutional reality is built by iterating the formula: the of one rule becomes the of the next, so that pieces of paper become money, money becomes capital, capital becomes a corporation's assets. The resulting structures can be enormously deep while resting on physical facts at the base.
Do groups exist?
A committee decides; a company is liable; a nation declares war. Are there group entities over and above their members?
Eliminativism paraphrases: talk of the committee is shorthand for talk about its members and their relations. Attractively economical, and it faces the same difficulty as mereological nihilism — the paraphrases get long, and it is not obvious they preserve inferential role. It also struggles with groups that survive complete turnover of membership: the university is the same university though no one from 1900 remains, which the members-and-relations account must capture some other way.
Realism about groups posits group entities. List and Pettit argue for group agents: where a collective aggregates its members' judgements by a procedure, the group can hold attitudes that no member holds. The discursive dilemma shows this formally — majority voting on premises can yield a collective conclusion contrary to the majority verdict on the conclusion itself. If a group's attitudes are not a function of its members' attitudes, the group is not obviously reducible to them.
The ontological question is whether this establishes a new entity or only a new pattern — which is the same fork encountered over ordinary objects, and it admits the same real-patterns answer.
Social construction
The claim that something is "socially constructed" is made about a wide range of things and means different things in each case. Hacking's taxonomy is the standard disambiguation, and it is worth the trouble because most disputes here are disputes about which claim is being made.
- Construction of ideas. The concept of a kind is a human product, shaped by history and interest. This is usually uncontroversial and is compatible with the kind being fully natural — the concept of an electron is a human product.
- Construction of objects. The things themselves are brought into existence by social practice. Money and borders are clear cases.
- Interactive kinds. Hacking's own contribution. Classifications of people can change the people classified, who know the classification and respond to it — a "looping effect" absent for natural kinds. Electrons do not know they are electrons; patients given a diagnosis often behave differently in consequence.
Applied to race and gender, the substantive ontological questions are: which sense of construction is intended; whether the kind is nonetheless real (constructed kinds can be perfectly real — money is); and whether it supports induction and explanation, which is the test the natural kinds page applies. Haslanger's ameliorative approach adds a further methodological move: rather than analyse the concept we have, ask what concept would best serve legitimate purposes, and adopt that. This is not a claim about what the kind is, which is why keeping it apart from the descriptive question matters.
The dependence structure
Social entities are the clearest illustration of the dependence apparatus at work.
| Entity | Depends on | Kind of dependence |
|---|---|---|
| A twenty-pound note | this piece of polymer | rigid, on the term |
| Money as such | collective acceptance | generic, ongoing |
| The University of X | its founding acts | rigid, historical |
| A debt | the parties and the agreement | rigid |
| A border | treaty and recognition | generic, ongoing |
The important distinction is between dependence on an origin and dependence on continuing attitudes. A university depends historically on its founding but would not vanish if everyone briefly forgot it; money would cease to be money if acceptance ceased entirely. The two kinds of dependence make for very different modal profiles, and conflating them produces bad arguments about whether social entities are "really" real.
Where social ontology sits
This belongs to the inventory and category questions, and its contribution to the section is a class of entities that are uncontroversially real and uncontroversially mind-dependent. That combination is instructive, because much of the section's argument has treated mind-dependence as a mark against reality — the objection to predicate nominalism was that it made the world depend on our language, and the objection to artifacts is that they depend on intentions.
Money shows the inference is too quick. What is objectionable in predicate nominalism is not dependence on us but dependence of the wrong things on us: the redness of the tomato does not wait on our vocabulary, whereas the monetary status of the note does wait on our practices, and there is nothing suspect about that. Sorting which is which is exactly the work an ontology of social entities has to do.