Consider the color of a ruby and the fact that it sits inside a velvet case. One of these features seems to belong to the ruby itself; the other depends on something beyond it. This intuitive contrast between intrinsic and relational properties is among the most fundamental distinctions in metaphysics—and among the hardest to make precise.

The distinction matters far beyond taxonomic tidiness. Whether a property is intrinsic or relational shapes debates about duplication, supervenience, causation, and the individuation of natural kinds. If mass turns out to be relational rather than intrinsic, our picture of physical objects changes dramatically. If shape is implicitly relational, geometry's metaphysical status shifts.

Yet the boundary between intrinsic and relational is not as clean as introductory treatments suggest. Borderline cases proliferate, proposed tests face counterexamples, and the very notion of "depending on nothing external" requires unpacking. What follows is a systematic examination of where the distinction holds firm, where it bends, and what the pressure points reveal about the nature of properties themselves.

The most accessible entry into the relational/intrinsic divide distinguishes purely relational properties from impure relational ones. A purely relational property like being taller than something involves a relation to an existentially quantified entity—some object or other satisfies the relational slot. An impure relational property like being taller than Socrates fixes a specific individual in that slot. Both are relational, but they differ in logical structure and in what they demand of the world.

This distinction matters for intrinsicality assessments because impure relations anchor a property to a particular external entity. If an object has the property of being three miles from the Eiffel Tower, that property cannot be characterized without reference to a specific landmark. Purely relational properties are more abstract—being three miles from something—yet they still require the existence of an external relatum. Both types fail the intuitive test for intrinsicality: neither belongs to the object considered entirely on its own.

However, the picture complicates when we consider properties that appear intrinsic but harbor relational structure under analysis. Take being a sibling. On the surface it looks like a classificatory label attached to an individual. But it is logically equivalent to having a parent in common with someone distinct from oneself—a purely relational property in disguise. Recognizing this structural ambiguity is essential: surface grammar does not reliably indicate whether a property is intrinsic or relational.

The pure/impure distinction also raises a subtler question. Are impure relational properties genuinely distinct properties, or merely instances of pure relational properties with a parameter filled in? If being taller than Socrates is just the pure relation being taller than applied to a specific case, then perhaps impure relational properties are not fundamental features of reality at all—they are derivative. This has consequences for property ontology: it suggests that the most metaphysically basic relational properties are the pure ones, and that individuals enter the picture only at the level of instantiation, not at the level of property identity.

Takeaway

Not every property wears its relational structure on its sleeve. Determining whether a property is intrinsic requires looking past surface grammar to the logical anatomy underneath—specifically, whether the property's characterization demands reference to anything beyond the object itself.

The hard cases are where metaphysics earns its keep. Consider shape. Intuitively, being spherical seems as intrinsic as any property could be—it characterizes an object's geometry without reference to anything external. But shape properties are defined in terms of spatial relations among an object's parts. A sphere is spherical because every point on its surface stands in a specific distance relation to its center. If shape reduces to internal spatial relations, is it really intrinsic, or is it relational in a way that merely stays within the object's boundaries? This forces a decision: do we count intra-object relations as consistent with intrinsicality, or does any relational structure—even internal—disqualify a property?

Next, consider velocity. In classical mechanics, velocity might seem intrinsic to a moving body. But special relativity reveals velocity as frame-dependent: an object's velocity is always relative to a reference frame. No object has a velocity simpliciter. If velocity is irreducibly relational in this way, then a property once considered paradigmatically intrinsic turns out to depend on external structure. This is not a merely verbal point. It means that the metaphysical status of properties can shift as our best physical theories develop—a sobering reminder that armchair analysis and empirical science must remain in dialogue.

Mass presents a different puzzle. In Newtonian physics, mass appears intrinsic—an object's resistance to acceleration, a feature it carries regardless of context. But on a dispositional reading, mass is the disposition to respond in certain ways to applied forces. Dispositions are often analyzed relationally: to have a disposition is to be such that if certain conditions obtained, certain outcomes would follow. If mass is fundamentally dispositional, and dispositions are relational, then mass is relational too. The categoricalist who insists mass is a brute intrinsic quality and the dispositionalist who reads it as relational reach incompatible conclusions about the same physical magnitude.

These borderline cases collectively reveal that the intrinsic/relational distinction is not a simple binary but a spectrum of dependence. Some properties depend on external objects, some on internal parts, some on reference frames, and some on counterfactual scenarios. Each type of dependence exerts different pressure on intrinsicality. A satisfactory account of the distinction must either explain why some forms of dependence are compatible with intrinsicality and others are not, or accept that the boundary is inherently vague—a conclusion many metaphysicians resist but few can entirely avoid.

Takeaway

Shape, velocity, and mass each challenge the intrinsic/relational boundary in a distinct way—through internal relations, frame-dependence, and dispositionality respectively. The distinction is less a clean line than a family of questions about different kinds of dependence.

David Lewis proposed an influential criterion: a property is intrinsic if and only if it is shared by every possible duplicate, including a lonely duplicate—an exact copy existing entirely alone in an otherwise empty possible world. The test is elegant. It captures the core intuition that intrinsic properties are those an object would retain even if everything else were stripped away. Being spherical passes: a lonely duplicate of a sphere is still spherical. Being three miles from a tower fails: in an empty world, there is no tower to be three miles from.

The lonely test handles a wide range of cases cleanly. It correctly classifies paradigmatic intrinsic properties—having a certain mass, charge, or chemical composition—and paradigmatic relational ones—being someone's neighbor, being the tallest in a room. It also elegantly handles disjunctive properties like being either round or within a mile of a cube: a lonely round object satisfies this property, but only via the intrinsic disjunct, so the test helps isolate the relational component without being tricked by logical packaging.

Yet the test faces pressure from essential relational properties. Consider a view on which certain objects have their origins essentially—Saul Kripke argued that a person could not have originated from a different sperm and egg than the ones that actually produced them. If origin is essential, then being born of these particular gametes is a property the object has in every possible world in which it exists, including lonely ones. But it is plainly relational—it references specific external entities. The lonely duplicate would have this property necessarily, yet it is not intrinsic by any intuitive reckoning. Lewis's test, which ties intrinsicality to duplication across worlds, struggles here because the relational property travels with the object across all worlds by metaphysical necessity.

This challenge has generated several responses. Some philosophers refine the notion of duplication to exclude essential relational properties by fiat. Others argue that the test should be supplemented with an independence condition: a property is intrinsic only if its instantiation is logically independent of the existence or nonexistence of any wholly distinct entity. This patches the problem but at the cost of the test's original parsimony. The deeper lesson is that no single criterion may capture intrinsicality perfectly—the concept may be analytically primitive, resistant to reductive definition, and best illuminated by a family of tests and paradigm cases rather than a single necessary-and-sufficient condition.

Takeaway

Lewis's lonely duplicate test is the most powerful single tool for diagnosing intrinsicality, but essential relational properties reveal its limits. The concept of intrinsicality may ultimately resist full reductive analysis, serving instead as a foundational notion we triangulate rather than define.

The intrinsic/relational distinction is not a tidy partition but a site of genuine metaphysical difficulty. The pure/impure relational contrast provides initial orientation. Borderline cases involving shape, velocity, and mass reveal that different kinds of dependence—on parts, frames, and counterfactuals—stress the distinction in different ways.

Lewis's lonely test offers the best available criterion but cannot accommodate every case, particularly properties that are both relational and essential. This suggests intrinsicality may be a concept best understood through convergent indicators rather than a single reductive formula.

What emerges is a richer picture of properties themselves. How a property relates to its bearer, to other objects, and to modal structure tells us something deep about the architecture of reality—and about the limits of our conceptual tools for mapping it.