Plato’s affinity argument

In an earlier post I suggested that the arguments of great philosophers of the past need to be understood, not only in the context of their times, but also in light of how later thinkers built on them. For an argument can contain, inchoately, real insights which only later thinkers were able to spell out adequately; and we will miss these insights if, overly fearful of anachronism, we insist pedantically on reading the argument in isolation from this later tradition. What ultimately matters in philosophy is not exactly who said exactly what, exactly when and exactly how. What matters is what is true, and whether an argument is likely to lead us to it. Anachronism, then, while a danger, is a less serious danger than loss of truth. To think otherwise is to abandon philosophy for mere scholarship. (Scholarship has its place, of course. But its place is to serve the ends of philosophy.)

What immediately prompted that post were some reflections on Plato’s Phaedo, on which I had been lecturing at the time. The Phaedo is famously concerned with the immortality of the soul, and Socrates is represented as putting forward four main lines of argument in its defense – or rather, as commentators these days often emphasize, four sub-arguments in the course of what should be understood as a single long, complex argument. Modern readers of the Phaedo often dismiss the arguments as manifestly bad. David Stove, in The Plato Cult, describes them as “so contemptible that, on their own merits, they could hardly ever have imposed upon a child of eight” (pp. 100-1). This sort of judgment is, I think, completely unwarranted. When understood both in the context of Plato’s philosophy as a whole and in light of the ongoing classical tradition of thinking about the soul which Plato inaugurated, his arguments can be seen to be very interesting indeed. That is not to say either that Plato’s arguments are adequate as they stand or that all of them can at the end of the day be defended. It is to say instead that the arguments embody serious considerations in favor of the soul’s immortality, and that at least some of these considerations were developed by later writers into more complete and compelling arguments. Modern readers often fail to see this because – as with their treatment of Aquinas’s Five Ways – they too frequently lack familiarity with the metaphysical presuppositions of the arguments, and read into them modern assumptions which classical writers did not make and often would have rejected.

The four arguments in question are usually labeled the cyclical argument, the recollection argument, the affinity argument, and the final argument. Let us begin by briefly surveying the arguments. It is evident from the dialogue that the cyclical and recollection arguments are intended to work in tandem. Putting aside the details of the examples Socrates uses in stating them (e.g. the long discussion about the Form of Equality) we might summarize the case that these first two arguments are together intended to make as follows:

1. We have knowledge (of Forms and of mathematical truths) that could not have come from sensory experience or in any other way been learned during this life.

2. So it must have been acquired during an existence prior to this one.

3. So our souls must have pre-existed their embodiment in this life.

4. But things arise out of their opposites in a cyclical pattern, such as sleeping from waking and waking from sleeping.

5. So, just as souls pre-exist their embodiment during life, so too must they continue on disembodied after death.

One way this line of reasoning might be challenged is via the suggestion that even if our souls pre-existed this life, perhaps their embodiment has altered them to such an extent that they will not survive our deaths, the “cycle” being thereby broken. Answering such an objection seems to be the point of the affinity argument, the core of which might be summarized as follows:

1. The soul knows the Forms, which are eternal, whereas the senses know material things, which pass away.

2. But each of these faculties is like the thing it knows (e.g. the senses are material, the soul is invisible).

3. Thus it is because the senses are like the things they know that they too pass away.

4. So the soul, since it is like the Forms that it knows, must not pass away.

Following Socrates’ presentation of the affinity argument in the Phaedo, his interlocutors Simmias and Cebes raise some further objections, including Simmias’s famous suggestion that the soul may be a mere harmony or attunement of the body’s components, like the harmony of a lyre. This begins a long discussion which culminates in the final argument, which might be summarized as follows:

1. The soul is the principle by which a thing is alive.

2. So it participates in the Form of Life.

3. But a thing cannot participate in contrary Forms (e.g. fire, which participates in the Form of Hot, cannot participate in the Form of Cold).

4. So the soul cannot participate in the Form of Death.

5. So the soul cannot perish.

Now I have no intention of exploring each of these arguments in any detail. Indeed, the line of reasoning enshrined in the combination of the cyclical and recollection arguments is one which I – as a Thomist who takes the soul to be the form of the body – do not think can succeed at the end of the day. One obvious objection to it is that its first premise is false if (as I would maintain) an Aristotelian analysis of concept formation is correct. Another is that step 2 doesn’t follow from step 1 all by itself: Even if our knowledge of forms and of mathematical truths cannot have come from the senses, it may be that it was stamped into our intellects by God when He created our souls (as per a rationalist theory of innate ideas) or that it results from a kind of divine illumination (as per St. Augustine). And of course, the stuff about things arising out of their opposites needs to be tightened up, to say the least. All the same, I’m sure a serious Platonist could give the argument a run for its money, and it would be a useful philosophical exercise to see how far it could be defended against objections. Much farther than most contemporary readers would suppose, I would bet.

Anyway, it is the second two arguments that I think are the most interesting. In the case of the final argument, this is not because it is sufficient to establish personal immortality. It is not. But what it does do, it seems to me, is foreshadow the Aristotelian-Thomistic insight that while particular composites of matter and form – that is, individual material substances – are generated and pass away, form itself (and thus the soul) is not susceptible of perishing. For a material substance’s perishing just amounts to its matter losing its form, and a form can’t coherently be said to lose its form (because it is a form). But then, since the soul is (on the A-T analysis) just a kind of form, the soul is not susceptible of perishing.

The reason this does not by itself establish personal immortality is that the forms of material things, considered by themselves, are in general mere abstractions. What exist concretely are individual material substances, and thus form and matter together. For this reason, matter too, considered apart from form, is a mere abstraction – speaking concretely, matter always exists with some form or other. And thus, if the soul qua form were imperishable merely in the sense just described, it would be no more imperishable than matter is. As Aquinas says in On the Principles of Nature, “prime matter [i.e. matter without form], and even form, are neither generated nor corrupted… properly speaking, only composites are generated” (2.15). A thing’s matter carries on after its destruction in the sense that the matter simply takes on a new form; its form carries on in the sense that some new substance with the very same form can always come into being. It is only the individual concrete substance that the form and matter together compose which comes into being and passes away. Since it passes away, though, the survival of its form – its soul, in the case of a living thing – merely qua something which another, future substance might take on, does not entail that the substance itself survives in any sense. When a particular rose bush dies, its form carries on in the sense that new rose bushes can always come into being, but that rose bush is gone for good.

For the imperishability of the human soul to ensure personal immortality, then – the survival not merely of the abstract form of man, but of your form or soul specifically – it would have to be what the forms of other material things in general are not: a kind of subsistent form, something whose operations are not, or at least are not entirely, dependent on matter. That is to say, it would have to be something which operates as a kind of immaterial particular thing even when it informs the matter of the body. Only then would its survival count as the survival of the particular human being whose soul it was. Or rather, it would count as the survival of the chief part of that particular human being; for the whole human being to come back into existence, the soul would have to be reunited with the matter that made up its body, which in Aquinas’s view it does at the resurrection. But the survival of the soul as a kind of immaterial particular at least makes this possible in a way it would not be if the human soul were like every other form.

I spell out Aquinas’s position in detail in chapter 4 of Aquinas. The point for now is that the reasons why the soul should be understood as something which operates independently of matter are not to be found in Plato’s final argument. But they are hinted at in Plato’s remaining argument for the soul’s immortality, namely the affinity argument.

As it happens, the affinity argument is the one some commentators seem to regard as the worst of Plato’s arguments for the soul’s immortality. One reason for this is that they often interpret it as an argument from analogy. That is to say, they think Plato is arguing along something like the following lines: If the soul is like the Forms in one respect – namely, being invisible to the senses – then it is probably like them in another respect as well, viz. in being imperishable. The trouble with the argument, then, is obvious: The analogy is simply too weak and undeveloped to support the conclusion.

It is understandable why a modern reader would read the argument this way. Plato does make use of an analogy, after all. And modern readers are used to thinking of metaphysical arguments as quasi-empirical hypotheses a la Paley’s “design argument” for God’s existence. But as every Thomist knows, not every philosophical use of analogy constitutes an “argument from analogy” in the modern sense, and metaphysical arguments (or good metaphysical arguments, anyway) are not quasi-scientific empirical ones. A failure to make these distinctions is what leads so many modern readers to misread Aquinas’s Fifth Way as a precursor of Paley. It is also what leads them to misread Plato as offering a lame argument from analogy for the immortality of the soul.

In fact, as Michael Pakaluk has pointed out, the affinity argument is not an “argument from analogy” at all, but rather “an argument about the nature of things.” Plato is not saying: “The soul is like the Forms in one way, so there is some significant probability that it is like them in this other way too.” Rather, he is saying something like: “The soul of its nature is X. But as we know from the example of the Forms, which are also X, things that are X are imperishable. So the soul is imperishable.”

What is X? As Lloyd Gerson suggests in his book Knowing Persons: A Study of Plato, what Plato seems to be emphasizing here is that the soul, like the Forms, is immaterial (p. 86). What the argument is (arguably) saying, then, is that whereas the senses pass away just as the things they know pass away, because both are material, the soul by contrast must be as imperishable as the things it knows – namely the Forms – because like the Forms, it is immaterial. In other words, it is, on this interpretation, the immaterial nature of the Forms that makes them imperishable, so that something that shares that nature – as (the argument claims) the soul does – must be equally imperishable. Notice, again, that this is not a probabilistic argument from analogy, but, in effect, an attempt at a proof. The argument is not:

1. The Forms are imperishable.

2. The soul is in some other respects analogous to the Forms.

3. So it is probably like them in being imperishable too.

That would indeed be a bad argument. Rather, the argument is (I am suggesting) to be understood along something like the following lines:

1. What is immaterial is imperishable.

2. The Forms are immaterial.

3. So the Forms are imperishable.

4. But since the soul knows the Forms, it must be as immaterial as they are.

5. So the soul must be imperishable.

Say what you will about this argument, it is not an argument from analogy, and it is not probabilistic. (Leave such weak tea arguments to the moderns. The ancient and medieval philosophers preferred the strong drink of metaphysical demonstration.)

OK, but what of the crucial premise 4? If it does not rest on an argument from analogy, what does it rest on? I would suggest that what Plato is gesturing at here is a line of argument that would later be developed more thoroughly and carefully by writers like Aristotle and Aquinas. The basic idea is that the intellect’s grasp of an abstraction like triangularity (for example) would simply not be possible unless it were as immaterial as triangularity itself is. There are three main considerations in favor of this judgment:

A. For a parcel of matter to take on a form is for it to become a thing of the kind the form is a form of; for example, for it to take on the form of triangularity is just for it to become a triangle. Now for the intellect to grasp the nature of a thing is just for it to take on the form of that thing. And in that case, if the intellect were material, it would become a thing of the kind that it grasps; for instance, it would become triangular when it grasps the form of triangularity. But this is obviously absurd. So the intellect is not material.

B. Forms and our thoughts about them are precise, exact, or determinate in a way no material thing can be even in principle; hence a thought cannot possibly be anything material.

C. Forms are universal while material representations are necessarily particular; hence to grasp a form cannot in principle be to have a material representation of any sort.

Obviously these arguments need spelling out, and I have said much more about them in several places. All of them are discussed in The Last Superstition (see especially pp. 123-126) and at greater length in Aquinas (see especially pp. 151-159), and both books set out in detail the metaphysical background apart from which the arguments cannot properly be understood. Arguments of the sort represented by B and C are discussed in more detail in the “Intentionality” chapter of my book Philosophy of Mind. Argument B in particular was the subject of an earlier post, and has in contemporary philosophy been developed most thoroughly by James Ross (see e.g. his article “Immaterial Aspects of Thought”). The point for now is that there is a family resemblance between what Plato was apparently up to in the affinity argument, and later arguments in the broad classical tradition that attempt to show, not by means of probabilistic arguments from analogy, but on the basis of an analysis of the nature of thought itself, that the soul (or its intellectual powers, in any event) cannot even in principle be material.

Apart from their misguided tendency to interpret classical metaphysical arguments as quasi-empirical hypotheses, another reason modern interpreters often badly misunderstand the affinity argument may be that, as Gerson notes, they tend to take for granted a post-Cartesian “representationalist” conception of knowledge (Knowing Persons, pp. 81-2). But the classical and medieval approach to knowledge was in general not representationalist. The great pre-modern writers did not regard the mind as a veil of “representations” which needed somehow to be correlated with external objects and events. Rather, for them knowledge involved a kind of union – not a correlation, but an identity of sorts – of the knower and what is known, insofar as when we know, one and the same essence or nature exists simultaneously in the intellect and in the world. (I have discussed this idea at greater length in an earlier post.) Naturally, then, if the object of knowledge is immaterial, that with which it is (in some sense) identical – the state of the knower – was regarded by them as immaterial too.

Here, though, as so often happens when modern readers encounter the arguments of ancient and medieval philosophers, the clearing up of one misunderstanding is likely only to lead to another. “How can the mind possibly be identical with what it knows?” the modern reader is bound to ask incredulously. “What absurdity!” Just as they often assume (with little or no argument, other than an appeal to the authority of Frege) that the concept of existence is entirely captured by the existential quantifier, so too might contemporary readers assume that the concept if identity is entirely captured by the identity relation as it is understood in modern logic. But this would be a mere prejudice. As Gyula Klima has emphasized in a series of important papers, the metaphysical doctrines of pre-modern philosophers typically presuppose semantic and logical doctrines that are very different from, but every bit as sophisticated and defensible as, those taken for granted by most contemporary philosophers. (What I have to say in the earlier post just linked to, and what I have said in recent posts about the Thomistic doctrine of analogy, may provide at least a hint of some of the differences. But this is a gigantic topic of its own.)

In general, while contemporary readers know that the difference between their view of the world and that of the ancient and medieval philosophers is radical, they generally do not know just how radical it is. Indeed, their ignorance of the differences is so great that they typically do not so much as even understand what the ancients and medievals were saying – hence (for example) they regard Aristotle as having put forward a “functionalist” “philosophy of mind,” Plato as having given an “argument from analogy,” and Aquinas as having foreshadowed Paley. In short, the moderns insist on reading the great figures of pre-modern philosophy as if they were all essentially just less well-informed versions of themselves – a double insult.

I suggested in an earlier post that Descartes’ “clear and distinct perception” argument for dualism represents a corrupted version of earlier and better Scholastic arguments. What is true in it isn’t new, and what’s new isn’t true. But if Descartes marks a decline, Plato’s affinity argument marks the beginning of an ascent. It is a mark of our own philosophical impoverishment that we can no longer recognize either the ascent or decline for what they were – much less the power of the Aristotelian-Scholastic arguments to which the ascent led and away from which the decline has taken us.

Of course, you might not agree with me about that. So at least agree with the more measured words of Gerson, with which I’ll conclude: “[The affinity] argument, like the others, is not supposed to stand on its own. And at its core there is an argument which, far from being inconsequential, is the origin of a family of immensely influential arguments for the immateriality of the person. These arguments are refined and elaborated upon by countless later Platonically inspired philosophers. They are still, in my view, worthy of interest.” (Knowing Persons, p. 79)