In his “Of Personal Identity,”1 David Hume criticizes the idea of personal identity. “The identity which we ascribe to the mind of man is only a fictitious one” (168). In this paper, I argue for an interpretation of the passage in Hume’s Treatise where he talks about the fiction called personal identity. In the passage that I will examine in this paper (pp. 169-170), Hume characterizes the identity that we ascribe to the perceptions that make up a single person as relations of resemblance or causation. I argue that there are two ways to interpret what Hume is doing in this passage. Either he is giving the necessary and sufficient conditions for the situations in which we judge identity to hold, or he is merely explaining where our idea of identity of persons comes from. In this paper, I will present an account of, and arguments for, the first interpretation and then reject it. I will then argue that, in the passage about resemblance and causation, Hume is presenting an account of how we form the idea of a person. This interpretation puts us into a better position to understand later passages in Hume’s Treatise.
First, as background for these two interpretations, it is important to get clear on what it might mean for perceptions to bear the relations of resemblance and causation. The relation of resemblance seems like it requires no clarification. Hume writes that “an image necessarily resembles its object,” so perceptions of the same object will necessarily resemble each other, and these resembling perceptions placed next to each other will “make the whole seem like continuance of one object” (170). Since the mind only perceives one thing at a time, placing ideas next to each other can only be done by the faculty of memory. Hume defines the memory as the “faculty by which we raise up the images of past perceptions” (169). It seems safe to say that resemblance between perceptions need not be so strict that perceptions of an object from multiple angles do not resemble each other. With such a strict requirement for resemblance hardly any of our perceptions would bear the relation. Hume’s idea of resemblance is not spelled out fully, but it seems to be something like ‘appearing, sounding, feeling, smelling, or tasting nearly the same.’
A causal relation between perceptions is a bit more difficult given Hume’s particular views on causation. Hume believes that in general, and not just with the separate perceptions in the mind, we do not perceive connections between things. Thus, we never actually perceive causal connections between things in the world. When we see something of a certain type always followed by something of another type, we posit a causal connection to account for this regularity. So, according to Hume, the idea of a causal relationship of the form ‘A causes B’ is established by repeatedly observing events of type A followed by events of type B. These ideas give us the general idea of causation which states that every event has a cause.
Thus, for perceptions to bear causal relations to each other, we must observe some regularity in how they succeed each other. Hume writes, “Our impressions give rise to their correspondent idea; and these ideas, in their turn, produce other impressions” (170). Impressions of a certain thing are regularly followed by the idea of that thing. This constant pairing of impressions with corresponding ideas leads us to posit a causal connection between them. Ideas also in turn are followed by other impressions. Hume calls these ‘impressions of reflection.’ For example, we might say that the impression of a spider causes the idea of a spider, and then that the idea of a spider causes the impression of fear. Ideas can also cause other ideas, such as someone who thinks of money every time she hears something about Wall Street. In that case, the idea is causally related to the idea of money because of the constant pairing of the two ideas. In these ways, perceptions can bear causal relationships with the perceptions that come after them. But they can also bear a causal relation to those that come before. New perceptions “destroy, influence, and modify” (170) older ones. To understand this it is helpful to remember that Hume likens the mind to a stage. At any one time there is one perception on the stage. Almost as soon as a perception ‘takes the stage’ so to speak, it is replaced by a new perception. Each new impression causes a new idea which forces the current perception ‘off the stage.’ Regardless of the direction of causation, our idea of identity entails that many of the perceptions of a person, if not all, are related causally.
Now, the first interpretation of resemblance and causation between perceptions is that they are intended as necessary and sufficient conditions for the times when we intuitively ascribe identity over time to a person. It takes Hume’s intent here to be that he wants to give an account of our concept of identity that incorporates all the situations in which our intuitions of personal identity are clear and also give predictions for cases that are unclear or that we had not considered yet. According to this interpretation, Hume will have succeeded if there are no cases where his proposed conditions hold and yet we do not intuitively ascribe personal identity, or vice versa. This kind of account is probably prompted by arguments in which Hume seems to be arguing based on our intuitions. For example, right after the passage with the explanation of the two relations we are considering, Hume argues against the Lockean memory criteria. He says that identity can continue even when memory does not. He argues this by appealing to our intuitions: “Will [anyone] affirm that, because he has entirely forgot the incidents of [past] days, that the present self is not the same person with the self of that time?” (171). This argument against Locke implies that these perceptions stand in the appropriate relations whether we remember them or not, and that he thinks he can account for this intuition because we judge identity based on these relations and not by whether they are accessible to memory. In other words, this interpretation takes ‘identity’ to be a shorthand for some other relation or relations that can hold between perceptions. Hume would then be simply spelling out what those relations are.
To make this interpretation plausible, one must make a few assumptions not explicitly spelled out in the text. Of course, several different persons can have perceptions that resemble each other. Thus, in order for us to attribute identity to a set of perceptions they must be aligned in a chain where each perception is related by resemblance and causation to other perceptions in the chain. According to this interpretation every perception of a particular person bears both causal and resemblance relations to some other perception or perceptions belonging to him. This is not too much of a stretch since the relation of causation already presupposes a resemblance relation. Recall that for Hume, causal relations are established by repeated observations of events of recognizable types. These resemblance relations are brought about by memory. Our memory places a copy of some past perception ‘on the stage’ as a result of our habitual connection between things. We are able to compare perceptions and identify resemblance relations only because we are able to retrieve past perceptions. With a few additions that Hume did not explicitly say, this view seems like a plausible view that someone could in fact hold.
This interpretation accounts for most of our intuitive judgments about which person a particular perception belongs to and the ways in which our judgments about whether someone is the same person differ in special cases. For example, any person living today is judged to be a different person from Socrates because his or her memory cannot produce perceptions that bear resemblance relations to those perceptions Socrates actually had. Also, any person who can bring up an image of the past through memory, must be the same person as the person who originally perceived the content that memory. Similarly, what one person currently perceives has no causal connection with what another person perceives at the same time. Much less is there a correlation constant enough to establish a causal link in Hume’s sense. And finally, if we know that two perceptions directly caused each other, it is probably safe to say that they belong to the same person. These principles seem to guide our judgments of personal identity over time in most cases.
Although these criteria account for the judgments we make about identity in most cases, we can still think of other cases in which these connections do not hold and yet we would say that they are perceived by the same person. For example, consider when a person gets knocked out in a fight and ‘the next thing he knows’ he wakes up in a hospital. This kind of case brings up the difficulty of specifying resemblance as a sufficient condition for establishing identity. We already saw that too strict of a definition of resemblance would disallow many of our own perceptions from bearing that relation. However, if we define resemblance loosely enough to account for this case—such that any perception of a hospital resembles any other perception of a hospital, for instance—we allow the perceptions of many persons to bear the relation. It cannot be a necessary condition for our judgments of identity that all perceptions bear the relation of resemblance to some other perception. And if perceptions do not have to be related by resemblance, they will not need to be related causally either since perceptions need to resemble each other in order for us to posit the causal link. We can also think of cases where the relations do hold between several perceptions and yet we would not say that they were perceived by the same person. For example, we can imagine a mind reader whose perceptions resemble and are caused by the perceptions in another person’s mind. Even if one thinks that such a case is impossible, if she still judges that in this imaginary case the perceptions belong to separate minds, then that is enough to prove that the judgment was not made by appealing to the relations of resemblance and causation. These perceptions would bear the relations in question and yet we would not say that they must be held by the same mind. So these relations cannot be sufficient conditions describing our judgments either.
In light of these criticisms, we might take another look at why Hume talks about resemblance and causation in relation to personal identity even though he has already denied that such an identity exists. Hume thinks that most of our ideas are derived from actual impressions. Yet, it is also possible that we combine impressions or ideas together such that the resulting idea has no corresponding impression. For example we might put together our idea of a horse and a bird to come up with the idea of a pegasus. So even though we never had the impression of a pegasus we are able to form the idea of it. If it is impossible to ever have an impression of such an idea, Hume says that we have made a mistake in combining ideas. So we will do better to understand Hume as simply telling the origin of our idea of personhood, or telling why we are led to make the mistake of forming the idea of a person. In the section on resemblance he says that perceptions that seem to us to resemble each other “convey the imagination more easily from one link to another” (170). If the perceptions resemble each other enough, we will mistake them to be exactly the same perception. Hume thinks we make this move with external objects as well. When an object changes by a small of imperceptible amount, we are very likely to call it the same thing. Certain classes of the perceptions that resemble each other display regularities for which we posit causal relationships. That our chain of perceptions displays these regularities convinces us all the more that there are connections between them. Thus, Hume is saying, not that we judge or should judge identity when perceptions resemble or seem to cause each other, but that when we think these relations hold, we will posit, incorrectly, that they belong to the same person.
The cases mentioned above do not pose counter-examples if we understand Hume to be offering conditions in which the idea of person arises. If this is true, Hume will not need to account for all the nuances of the situations in which we would or would not judge personal identity to hold. In fact, he writes that “we have no just standard by which we can decide any dispute concerning the time when they acquire or lose a title to the name of identity” (171). Nor does the problem of rigorously defining the relation of resemblance come up. We see that some of our own perceptions resemble each other, and thereby posit that both were perceived by the same person, namely ourselves. Other passages around the descriptions of resemblance and causation clarify Hume’s intent. “Our notions of personal identity proceed entirely from the smooth and uninterrupted progress of the thought along a train of connected ideas, according to the principles above explained” (169). He also says that these comments concern “the first origin and uncertainty of our notion of identity, as applied to the human mind” (171).
However, even if we take this to be Hume’s intention, the counter-examples from above still pose problems for Hume. We judge the knocked-out fighter to be the same person even while he is unconscious. We use the idea of person in cases where resemblance and causal relations do not hold, so our idea of personhood must contain more than just those two relations.
Hume noticed these problems and their implications as well. Later in the Treatise, Hume admits that he his account leads him to unsatisfactory results. Hume thinks that strictly speaking, we do not perceive that these relations hold between perceptions. We do not perceive any connection between any perceptions. But, something makes it the case that all of my perceptions are unified with each other while my perceptions are not unified with any of those of someone else. In his own words, Hume believes his account leads to two conflicting conclusions, “that all our perceptions are distinct existences and that the mind never perceives any real connexion between distinct existences” (175-176). These beliefs are only irreconcilable if we also hold that all of the perceptions of what we generally call a person are actually connected somehow. The way Hume phrases his result shows that he know he cannot account for these connections. His use of ‘Our perceptions’ implies that perceptions are owned by someone or something. But how can a perception be owed when all that exists besides it is more perceptions? All of my perceptions are connected and all of your perceptions are connected but my perceptions are not connected to yours. To even state this, one must recognize that there is why I call a certain set of perceptions “mine” in order to distinguish them from yours. Hume seems to be saying that ‘person’ can’t be a fiction after all since all these connections hold. But, he admits that he cannot account for how they are connected and how we know about the connections.
If one held the first interpretation, he might believe that Hume is admitting only that he cannot find the right relations that we posit when positing personal identity. He would believe that the two results mentioned above are incompatible insofar as he cannot discover the ‘non-real’ connections, i.e. those which we merely posit, between the distinct perceptions. Understanding the earlier passage in the way I have described allows us to see that Hume understood both ways to look at personal identity and anticipated the problems to both.
1 Quotations are taken from Section 6 of Part IV of Book I of the treatise as printed in Personal Identity. Ed. John Perry. University of California Press. 2008. Pages 161-172.