The ship of Theseus provides a classic example of the problem associated with holding on to the notion of an identity that persists through time. The ship gradually undergoes replacement of its constituent planks until by the end none of the original planks remain. Does it then retain the same ‘identity’ as the original ship? To further complicate the picture, all the replaced planks are reassembled into a ship ‘identical’ to the original ship, with the same design. The problem then is the stark contradiction in asserting that the replaced ship and the newly reassembled ship are the same. An Aristotelian resolution of this problem would be to assert that the replaced ship is the ‘same’ ship as the original ship owned by Theseus. The argument rests upon the preservation of the essence (design) despite the accidental changes (replacement) of the constituent planks. However according to the same argument the reassembled ship should be identical to the original ship implying that it should be the same as the replaced ship, which clearly cannot be the case. Therefore the contradiction seems to persist. I argue against the Aristotelian resolution of the problem and claim that a Humean analysis of the problem not only suggests the problematic nature of holding onto a persistent identity but also provides a possible resolution by rejecting the idea of a persistent identity.
Before delving into the heart of the contradiction, it is important to realize the need to clearly define certain important terminologies that appear throughout such a discussion. These terms, identical, same and identity are noted in the above paragraph in quotation marks. However I will not define these terms just yet. Rather than providing a general definition, I will derive them from the respective philosophical positions in question and at the end provide an explanation as to why the definitions upheld by Aristotelian position eventually lead to the contradiction. Nevertheless I will make the reasonable assumption that identity is transitive:
1) A and B have the same identity
2) B and C have the same identity
=> A and C have the same identity
For the purposes of this discussion the ship of Theseus sets sail every year and on returning it has a plank replaced by a new one. Year after year, eventually all the planks are replaced and the removed planks are then reassembled into a new ship with the original design. The annual replacement of a plank ensures that the process is gradual, which I argue is an important criterion for such a notion of identity to persist. To make this obvious, imagine replacing all the planks of the ship within a day, it is definitely more difficult to retain this notion of a persistent identity and to call it the same ship.
Now, from an Aristotelian standpoint things can undergo either essential or accidental changes. Essential changes alter the ‘essence’ or in our case the design of the ship. An accident change does not alter the design, in our case replacing the planks that constitute the ship. So in the Aristotelian sense to say that despite the accidental changes, the ship remains the same suggests that the definition of identity coincides with the definition of essence. As essence persists through these changes (accidental) so does the identity of the ship. Therefore at the end even though all the planks have been replaced, the ship still retains its identity or the essence or the design (shape, size). So to say two things are same in this sense is not to assert that they are indistinguishable but that they have the same identity (essence). Identical on the other hand refers to the case where they not only share identity (essence) but also the accidental properties (the planks) and are indistinguishable. This then suggests that the reassembled ship is identical to the original ship, as it not only has the same planks but also the same shape and design and therefore the essential properties. And by the transitivity of identity, the replaced and the reassembled ship would have to be the same (share identities) leading to the contradiction that it sought to avoid. However a possible means of avoiding this contradiction would be to reject the premise that the reassembled ship has the same essential properties (therefore the identity) as the original ship. This would then avoid the conclusion that it has the same essential properties as the replaced ship. However this would imply that there is more to essential properties than the design, the shape, and the size of the ship. Whatever the reassembled ship lacks, it does not seem to be sensible to the senses otherwise it could be fixed and made identical. It has to be a property inaccessible to the senses that the original ship had and retained through the replacement of its planks. However this seems too easy a resolution. I argue that positing a property inaccessible to the senses doesn’t add anything to the ship. By the virtue of its inaccessibility to our senses, we have no means of determining either its existence or non-existence. What then allows us to deny the existence of this property in the reassembled ship? So the best the Aristotelian standpoint could do is to assert that if that property (essential) is absent in the reassembled ship, it wouldn’t be identical to the original ship. But we have no means of determining whether this is true and therefore it does not provide a conclusive resolution to the contradiction.
This is exactly what David Hume would have rejected. In his A Treatise of Human Nature, Hume argues against the idea of an identity (or as he calls it the substance) that inheres the properties since we have no means of discovering it. According to him, things are no more than the collection of their properties. And therefore there is no persistent identity that survives through the changes. I argue that the Humean notion of identity avoids the contradiction mentioned above and provides a possible resolution. According to the Humean standpoint, the replaced and the reassembled ships are clearly not the same as they are made up of different planks. This clearly eliminates the contradiction. However this seems to lead to a different contradiction. Since there is no identity that persists through the changes, the replaced ship cannot be the same as the original ship; share the same identity.
Before tackling this counterargument it is important to redefine the notions of identity, same and identical. Unlike in the Aristotelian framework we can no longer talk about a persistent essence as the source of identity. In the Humean sense there is no persistent identity. So then what is an identity? I argue that the way to include the notion of an identity into such a position is to propose a dynamic identity. Identity in this sense is nothing but the mental label given to a collection of properties. It is dynamic in the sense that it evolves with the changes to the properties. Therefore an object that undergoes the change would no longer be the same (share identities) with the object before the change. This then implies that when I replace a plank of that ship, its identity changes. Therefore identity is no longer understood as independent and inherent of the properties but as a product of their union.
The above argument suggests that it was a wrong question to begin with to ask whether the replaced and the original ships are the same. There is no inherent identity that possesses the properties that persists through the changes. Identity as we understand it evolves and is dependent on the union of the properties that make up a thing. This I argue is the only sustainable notion of identity (if we want to persist with one) that avoids the contradiction.
However there seems to be a problem with this resolution. Despite its ability to not only resolve but also to redefine the problem, it is disturbing to assert that there is no persistent identity on intuitive grounds. It is quite clear that on changing a plank we perceive the ship as the same as the one before (the replacement). But our intuitions are by no means the definite check for accuracy. And the reasons why we have these intuitions are neither significant for the purposes of this paper nor within its scope. However, the fact that these changes are gradual might be important for this notion of a persistent identity. And the fact that the changes are small and often insignificant in comparison to the portion of the object that remains unchanged. In our example changing a plank a year is not only gradual but the change itself is almost negligible compared to the hundreds of planks that remain unchanged. On returning after the second year, our notion of the identity of the ship has now evolved to incorporate the replaced ship. So it is no longer equivalent of replacing two planks on the original ship but replacing one on this ship with the changed notion of identity. By the time all the planks have been replaced the ship shares none of the constituent materials as the original ship but since we have clung on to this idea of a persistent identity that remains the same despite the changes to the ship’s constituents, we find the need to posit the existence of such an identity. In other words, such an identity can be seen as a mechanism of resolving the discrepancy in our perception of the ship. And this is exactly what the Aristotelian explanation of the problem seems to do. It has this idea of essence that persists through the accidental changes, which we call the identity. But that as shown before is not an adequate an explanation to resolve the contradiction.
It is clear that the idea of a persistent identity as defended from the Aristotelian position fails to adequately account for the contradiction posed in the problem of the ship of Theseus. And despite the fact that it runs contrary to our intuitions, the Humean rejection of a persistent identity not only helps resolve the contradiction but also drastically redefines the problem.
- Hume, David, A Treatise of Human Nature, (New York: Dover, 2003 edition)
- Gallois, Andre, “Identity Over Time”, The Stanford Encyclopedia of Philosophy (Summer 2012 Edition), Edward N. Zalta (ed.), URL = <http://plato.stanford.edu/archives/sum2012/entries/identity-time/>.
- Cohen, S. Marc, “Aristotle’s Metaphysics”, The Stanford Encyclopedia of Philosophy (Summer 2012 Edition), Edward N. Zalta (ed.), URL = <http://plato.stanford.edu/archives/sum2012/entries/aristotle-metaphysics/>.
- Image source: http://phnur.blogspot.com/2012/04/improbable-fiction-continuity-and.html, last accessed: Dec, 15, 2012.