The question at the center is an old one. Theseus kept his ship in the harbor as a monument. Over the years, planks rotted and were replaced one by one. Eventually nothing of the original remained. Is it still the same ship? The puzzle has lasted centuries without resolution. What follows argues that the failure to resolve it is not an accident — it is what we should expect.
I. A Problem About Naming
I think the puzzle is not about identity in some deep metaphysical sense. It is about naming — about when a name still applies to a changing collection of matter.
We name things by resemblance. A ship gets a name because it resembles something we have tracked before, or something we care about tracking. Different people track different things. A sailor cares whether the vessel is seaworthy. A museum curator cares whether the object has historical provenance. Both are responding to real features of the ship. Neither, in my view, is more correct than the other. They are simply operating with different criteria for when the name applies.
The mistake, I think, is to assume there is one criterion hidden in the nature of things, and that the puzzle arises from not having found it yet. There is no such criterion. The world does not come pre-labeled. What is conventional is where we draw the line — and lines can be drawn in more than one place without either placement being wrong.
II. The Trouble with Resemblance
Resemblance is not transitive. The ship after one replaced plank resembles the original closely enough to keep the name. After the second plank, it still resembles the previous stage closely enough. Step by step, nothing seems to change. But at the end, when every plank is gone, the conclusion is that something with almost nothing in common with the original is still “the same ship.”
This is the Sorites paradox in nautical form — the same structure that makes “heap” and “bald” impossible to pin down. One replaced plank does not make a new ship. But enough planks do — and the convention does not say when.
The right response, I think, is not to fix the convention. It is to recognize that any naming practice based on local resemblance will produce exactly these cases. The puzzle is a structural consequence of the practice, not a defect in it.
There is also the question of what the resemblance is assessed against. Compare each stage to the original, and the resemblance degrades gradually until the name must eventually be revoked — though the convention will not say exactly when. Compare each stage only to the one before it, and the resemblance never degrades enough to revoke the name at all. My view is that neither reading is privileged. The reference point is part of the convention, not a fact about the world.
When two people argue about the ship and cannot agree, I think the productive move is for each to make their criteria explicit. They might discover they care about the same thing and had not noticed — or that they care about genuinely different things, in which case the disagreement dissolves rather than gets resolved. Either outcome is better than assuming there is a hidden answer neither of them has found.
III. Zoom In and the Problem Moves
One response is to name things more precisely. Specify the exact arrangement of atoms. Then any replacement of a plank produces a different object by definition. No vagueness.
But consider two physicists examining an atom whose constituents have been partially rearranged. Is it still the same atom? The puzzle reappears one level down. It will reappear at every level. Whatever resolution you get from a finer description, the finer description faces the same problem.
The Ship of Theseus is not, in my view, a local quirk in how we talk about ships. It is what happens whenever discrete names are applied to things that change. It shows up at every scale, from the macroscopic down to whatever physics bottoms out at — if it bottoms out at all.
Whether there is a fundamental level where identity is finally sharp is an open question in physics. My intuition leans toward thinking there probably is not — every time we have thought we found the bottom, there has been something smaller inside. But I hold this lightly; it is a bet on how physics will go, not something I can establish from the armchair. If a bedrock level exists, higher-level conventions borrow precision from it. If not, every resolution is provisional all the way down. Either way, the point stands: resolutions are local, borrowed from a finer level, and that finer level faces the same problem.
IV. Nothing Persists
Take the fine-grained view seriously and a stark conclusion follows. If identity just is strict material constitution, then replacing a single plank already produces a different ship. Nothing persists through any change. As Heraclitus put it, you cannot step in the same river twice.
I think this is correct, strictly speaking. Ordinary claims like “same ship” or “same person” are not claims about strict identity. They are claims of sufficient resemblance for the purpose at hand. When the harbormaster says “same ship,” he means: same enough to dock, same enough to insure, same enough to track. That, in my view, is all the claim ever was — and it is enough.
Ordinary object-talk is technically false but practically adequate. This should not be alarming. We do not expect the word “chair” to track some deep metaphysical category. We expect it to reliably pick out things to sit on. It does that job well. What more could we want from it?
The same analysis applies, I think, to persons. The name “I” is attached to a continuously changing material configuration — cells replaced, atoms exchanged, neural states shifting from moment to moment. The name persists not because there is an unchanging self beneath it but because the trajectory is continuous enough, and similar enough across time, to make the name useful. Strictly speaking, there is no persisting self. There is a causal trajectory through matter, with a name attached.
V. The Language Problem
There is a difficulty that runs deeper than the puzzle itself. This analysis was conducted in natural language — and natural language is built around persisting subjects. Every sentence with “I will” presupposes exactly the continuity this view denies. As Nietzsche observed, grammar seduces us into positing subjects because every verb demands one. “It lightnings” became “lightning strikes,” and suddenly there is something doing the striking — conjured entirely by syntactic necessity. Language evolved for creatures who needed to track stable objects across time. It was not built for a view on which nothing persists.
A more honest formulation replaces “I will have this belief” with something like: at this moment, what we approximately call this person acts on the basis of the neural configuration that currently exists. No persisting self required. But even this is imperfect — the phrase “what we call” still implies a subject doing the calling.
The tool is imperfect. It can still do real work. But it is worth knowing that the medium resists the message.
VI. Can Philosophy Track Anything?
If language is an evolutionary instrument — shaped by selection pressure, not by a drive toward truth — then what exactly is philosophical argument doing? I think of it the way I think of eyes. Vision evolved for navigation and threat detection, not for perceiving beauty. But it can still perceive beauty, because the capacity generalizes beyond the pressure that produced it. Language evolved for coordination and signaling. My view is that it can still do philosophy — or something that functions like it.
But there is a disanalogy worth taking seriously. When vision goes wrong, there is an external correction available — you walk into the wall, your prediction fails. For philosophical argument, the correction mechanism is weaker: social pressure, predictive success, the sense of coherence when a view holds under pressure. None of those are as clean as the wall.
The feeling of having resolved a problem is not reliable evidence that the problem has been resolved. Consider a community of whales that develops a call signaling good food. The call reliably produces a feeling of satisfaction in the community — and yet the fish they are being directed to are venomous, gradually poisoning them with each meal. They would have no idea until the effects became catastrophic. I think we could be doing something structurally similar when we philosophize. The phenomenology of insight is an output of the same evolved system whose reliability is in question.
Likewise, people were once flat earthers, or held geocentric models of the solar system, and felt the deep satisfaction of a coherent worldview. The satisfaction was real. The worldview was wrong. There is no guarantee that what feels like resolution is resolution.
There is, as far as I can tell, no vantage point outside the system from which to check. This is not a special problem for this view — it is the situation of every philosophical position. My response is to hold these conclusions openly: as the best available position given current argument, open to revision, without stronger warrant than that.
I find this clarifying rather than troubling. It means only that we have been asking for something the world was never going to provide — a deep fact of identity underneath the conventions. Once we stop looking for it, what remains is enough: a world of matter, conventions calibrated to practical purposes, and the inquiry itself, held without illusions about its foundations. There is a certain lightness to that — what one might call, without embarrassment, just vibing.
Polished by Claude.
Discover more from Investigations in Forms
Subscribe to get the latest posts sent to your email.
