Let's discuss the idea that there are no such things as properties, just lots of things which are more or less similar

How do two distinct physical objects—such as a ripe tomato and a stop sign—share the exact same feature, such as being "red"? In metaphysics, the standard *realist* response posits the existence of **universals**: abstract, repeatable entities that exist over and above individual things and are instantiated by them. The thesis you have introduced is the core commitment of **Resemblance Nominalism**, an elegant alternative within *nominalism* (the denial of abstract entities). It claims that the fundamental furniture of the world consists solely of concrete particulars; properties do not exist as independent metaphysical items. Instead, an object is "red" simply because it resembles a set of other concrete, red objects. ``` THE METAPHYSICAL DIVIDE REALISM RESEMBLE NOMINALISM (Universals exist) (Only particulars exist) [ Universal ] [ Object A ] / \ \ resembles [ Object A ] [ Object B ] [ Object B ] ``` ### Mechanisms and Logical Architecture To understand how this operates, consider how a resemblance nominalist constructs what we ordinary speakers call "properties": 1. **Primitive Resemblance:** Resemblance is treated as an irreducible, primitive relation. Objects do not resemble one another *because* they share a property; rather, we say they share a property *because* they resemble one another. 2. **Equivalence Classes:** A "property" is reinterpreted as a maximal group of particulars that resemble each other to a requisite degree. 3. **Paretian / Paradigm Models:** Some versions select paradigm instances (e.g., a standard reference meter or a ideal sample of red) and define membership in a property class by an object’s degree of similarity to those paradigms. ### The Core Debates: Russell’s Regress and Coextension Despite its intuitive appeal, Resemblance Nominalism faces two classical objections in analytic metaphysics: * **Russell’s Regress:** Formulated by Bertrand Russell in *The Problems of Philosophy* (1912), this argument asserts that if you eliminate properties by invoking a relation of *resemblance*, you must explain what all individual acts of resemblance have in common. They must resemble each other, requiring a higher-order resemblance relation, triggering an infinite regress. Russell argued this proves that at least *one* universal—the relation of resemblance itself—must exist. Modern nominalists, such as Gonzalo Rodriguez-Pereyra, counter that resemblance is a multigrade relation that applies directly to particulars without constituting a repeatable universal. * **The Problem of Coextensive Properties:** If two distinct properties belong to the exact same set of objects in the actual world (for example, *renate* [having kidneys] and *cordate* [having a heart]), resemblance nominalism risks treating them as identical because the set of resembling objects is identical. Resolving this often forces nominalists to extend their scope to *possible* objects across modal space, importing heavy cosmological commitments to preserve a lean ontology. ### Broad Implications Resemblance Nominalism aligns closely with cognitive science theories of categorization, such as **prototype theory**, where human concepts are structured around similarity to central exemplars rather than strict, necessary-and-sufficient features. Economically, it offers a sparse ontology—reducing the universe to concrete entities and their concrete relationships—though critics argue that it merely sweeps metaphysical complexity under the rug of "primitive similarity". ## Follow-up questions 1. How does resemblance nominalism handle cases where two distinct properties happen to be possessed by the exact same set of objects? 2. Does Bertrand Russell’s regress argument successfully force the resemblance nominalist to accept at least one universal? 3. How do modern prototype theories in cognitive psychology mirror or challenge the philosophical assumptions of resemblance nominalism?

While both Resemblance Nominalism and Ostrich Nominalism reject abstract universals, they diverge sharply on whether the phenomenon of attribute agreement—objects "sharing" a feature—demands a metaphysical explanation. ``` METAPHYSICAL EXPLANATORY STRATEGIES RESEMBLE NOMINALISM OSTRICH NOMINALISM (Constructive / Reductive) (Austere / Refusal) "a is Red" & "b is Red" "a is Red" & "b is Red" │ │ Requires Grounds Brute Semantic Fact │ │ [ Primitive Resemblance ] [ "a is red" is just true ] ``` ### Explanatory Reduction vs. Radical Austere Refusal D.M. Armstrong coined the term **Ostrich Nominalism** (often associated with W.V.O. Quine and defended by Michael Devitt) to accuse philosophers of hiding their heads in the sand when faced with the "One over Many" problem: explaining how distinct particulars can be same-typed. * **Resemblance Nominalism** accepts the problem. It offers an *explanatory reduction*: object $a$ is red because it stands in a primitive similarity relation to other red objects. It builds a constructive architecture to account for property discourse without universals. * **Ostrich Nominalism** rejects the demand for a metaphysical grounding mechanism altogether. It asserts that predicate application is an ultimate, unanalyzable semantic fact. To say $a$ is red and $b$ is red requires no underlying relational "glue" or set construction; the truth of the simple sentence " $a$ is red" exhausts what needs to be said. | Dimension | Resemblance Nominalism | Ostrich Nominalism | | :--- | :--- | :--- | | **Primary Mechanism** | Symmetric, primitive resemblance relations | Asymmetric, primitive predication | | **Metaphysical Commitment** | Relations among particulars | Only the concrete particulars themselves | | **Stance on "One over Many"** | Solves it reductively | Dismisses it as a pseudo-problem | ### Practical Implications and Tensions The dispute hinges on differing standards of metaphysical parsimony: 1. **Theoretical Economy vs. Explanatory Debt:** The Ostrich achieves the ultimate "lean" ontology—containing only concrete particulars—but critics argue it purchases this simplicity by refusing to explain *why* the universe exhibits order and repeatable patterns. Resemblance Nominalists accept additional structural complexity (resemblance relations) to gain explanatory power. 2. **The Nature of Predication:** In computer science or formal semantics, an Ostrich view aligns with primitive type tagging (treating class membership as an irreducible property of an instance variable). Resemblance Nominalism mirrors vector-space similarity algorithms, where an entity's type is dynamically derived from its proximity to other nodes in a high-dimensional space.