Scholastic regress arguments

Many are familiar with arguments to the effect that an infinite regress of causes is impossible, as Aquinas holds in several of his Five Ways of proving God’s existence.  Fewer correctly understand how the reasoning of such arguments is actually supposed to work in Scholastic writers like Aquinas.  Fewer still are aware that the basic structure of this sort of reasoning has parallels in other Scholastic regress arguments concerning the nature of mind, of knowledge, and of action.  Comparing these sorts of arguments can be illuminating.

Causality

In Aquinas’s First Way, he famously argues that it is impossible for there to be an infinite series of movers or changers, so that a regress of changers must terminate in a first unchanging changer.  As I’ve discussed in many places (such as at pp. 69-73 of my book Aquinas), Aquinas is not claiming here that no causal series of any kind can regress infinitely.  Rather, he has in mind a specific sort of causal series, which commentators sometimes call “essentially ordered series” and sometimes “hierarchical series.”

He illustrates the idea with a stone which is moved by a stick which is in turn moved by a hand.  The stick moves the stone, but not by its own power.  By itself the stick would simply lay on the ground inertly.  It can move the stone only because the hand imparts to it the power to move it.  The hand too, of course, would be unable to move the stick (so that the stick in turn would be unable to move the stone) unless the person whose hand it is deliberately uses the hand to move it. 

The technical way of putting the point is that the stick is a secondary cause in that it has causal efficacy only insofar as it derives or borrows it from something else.  By contrast, the person who uses the stick is a primary cause insofar as his causal efficacy is built-in rather than being derivative or borrowed.  The stick is a moved mover insofar as it moves other things only because it is itself being moved in the process.  The person is an unmoved mover insofar as he can move the stick (and, through it, the stone) without something else having to move him in the process.

Aquinas gives other arguments against an infinite regress in such cases, but this is the one I want to focus on here.  The basic idea is that there cannot be secondary causes without a primary cause, because you cannot have borrowed or derivative causal power without something to borrow it from.  This would be true no matter how long the regress is, and it is important to note that infinity as such is not really what is doing the work here.  Even if we allowed for the sake of argument that the stone in our example was pushed by an infinitely long stick, there would still need to be something outside the stick to impart causal power to it.  For a stick is just not the sort of thing that could, all by itself, move anything else, no matter how long it is. 

Nor, for that matter, would it help if the causal series in this case went around in a circle rather than regressing to infinity – the stone moved by the stick which was moved by another stick which was moved by the stone, say.  For sticks and stones just aren’t the sorts of things that could move anything by themselves, even around in a circle.  There would have to be something outside this circle of movers that introduced motion into it.

Meaning

Now consider a second and at first glance very different sort of argument, which is associated with John of St. Thomas (John Poinsot) and was defended in the twentieth century by Francis Parker and Henry Veatch in their book Logic as a Human Instrument.  It appeals to a distinction between instrumental signs and formal signs.  An instrumental sign is a sign that is also something other than a sign.  Consider, for example, a string of words written in pen on a piece of paper.  It is a sign of the concept or proposition being expressed, but it is also something else, namely a collection of ink splotches.  Now, there is nothing intrinsic to it qua collection of ink splotches that makes it a sign.  By itself, a string of splotches that looks like “The cat is on the mat” is no more meaningful than a string of marks such as “FhjQns jkek$9 (quyea&b.” 

Suppose we say that the string of splotches that looks like “The cat is on the mat” has the meaning it has because of its relations to other strings of splotches, such as the ones we see in a dictionary when we look up the words “cat,” “mat,” etc.  That can hardly give us a complete explanation of how the first set of ink splotches get their meaning, because these new sets of splotches are, considered just by themselves, as meaningless as the first set.  They too have no intrinsic meaning, but have to derive it from something else.

Notice that we have a kind of regress here.  And what the argument says is that this regress must terminate in signs that do not get their meaning from their relations to other signs, but have it intrinsically.  This is what formal signs are.  And unlike instrumental signs, they must be signs that are not also something else – that is to say, they must be signs that are nothing but signs.  With a sign that is also something other than a sign (a set of ink splotches, or sounds, or pictures, or whatever), the meaning and this additional feature can come apart, which opens the door to the question of how the meaning and the additional feature get together.  But a sign that is nothing more than a sign just is its meaning.  Because it just is its meaning, it needn’t derive or borrow its meaning from something else.

The argument goes on to say that these formal signs are to be identified with our concepts and thoughts, which are the source of the meanings that words and sentences have.  This in turn provides the basis for an argument for the mind’s immateriality, as I discuss in Immortal Souls.  The point I want to emphasize for the moment, though, is that we have here an argument that holds that a regress of items having a certain feature in only a derivative way can exist only if there is something having that feature in a built-in or non-derivative way – which is, at a very general level, analogous to the reasoning Aquinas deploys in the First Way.

(As a side note, I’ll point out that John Haldane, in Atheism and Theism, develops a line of reasoning which might be seen as an amalgam of these two arguments.  A person’s potential for concept formation, he says, presupposes fellow members of a linguistic community who actualize this potential by virtue of already possessing concepts themselves.  But their potential to form these concepts requires the preexistence of yet other members of the linguistic community.  This regress can end only in a first member of the series, whose possession of concepts need not depend on actualization by previous members.  This “Prime Thinker” he identifies with God.)

Knowledge

A third line of argument, once again very different at first glance but ultimately similar in structure, concerns epistemic justification.  Consider the “problem of the criterion,” of which Michel de Montaigne gave a famous statement.  In order rationally to justify some knowledge claim, we will need to appeal to some criterion.  But that criterion will in turn have to be justified by reference to some further criterion, and that further criterion by reference to yet some other criterion, and so on ad infinitum.  It seems, then, that no judgment can ever be justified.

As the Neo-Scholastic philosopher Peter Coffey points out in his Epistemology or The Theory of Knowledge, the fallacy in this sort of argument lies in assuming that the justification for a judgment must in every case lie in something extrinsic to the judgment itself.  The skeptical argument fails if there are judgments whose criterion of justification is intrinsic to them. 

Now, suppose it can be established that we cannot fail to have at least some genuine knowledge.  One might argue, for example, that even the skeptic himself cannot coherently raise skeptical doubts without making certain presuppositions (such as the reliability of the rules of inference he deploys in arguing for skepticism).  Then we would have a basis for an argument like the following: We do at least have some knowledge; we could have no knowledge unless there were at least some judgments whose criterion of justification is intrinsic to them; therefore, there must be at least some judgments whose criterion of justification is intrinsic to them.

The point is not to expound or defend this sort of argument here.  The point is rather that such an argument would be a further instance of the general pattern we’ve seen in the other arguments.  In particular, it would be another case in which it is argued that there can be a regress of things having some feature in only a derivative way (in this case, epistemic justification) only if there is something having it in an intrinsic way.

Action

One last example, which concerns action.  Aristotle, and Scholastic writers like Aquinas following him, hold that every action is carried out for a certain good, and that good is often pursued only for the sake of some further good to which it is a means, which is itself pursued for the sake of yet some other good.  The regress this generates can terminate only in some end that is pursued for its own sake, as good in itself.  Naturally, there is a lot more than that to the analysis of action and the good, but the point is to emphasize that once more we see an instance of an argument fitting the same general pattern we’ve seen in other cases.  A regress of items having a certain feature only derivatively (in this case, goodness or desirability as an end) can terminate only in something that has that feature intrinsically.

Here are some features common to such arguments.  First, and to repeat, the basic general pattern is to argue that the existence of items having a certain feature in a borrowed or derivative way presupposes something having that feature in an intrinsic way.  In one case the feature in question is causal power, in another it is meaning, in another it is epistemic justification, and in yet another it is goodness or desirability.  But despite this significant difference in subject matter, the basic structure of the inference is the same.

Second, although the arguments are set up by way of a description of a regress of some kind, the length of the regress is not actually what is doing the key work in the arguments.  In particular, the arguments, on close inspection, are not primarily concerned to rule out infinities.  Rather, they are concerned to make the point that what is derivative ultimately presupposes what is intrinsic or non-derivative.  This would remain the case even if some sort of infinite sequence was allowed for the sake of argument.  For example, even an infinite series of causes having only derivative causal power would presuppose something outside the series which had intrinsic causal power; even an infinite sequence of instrumental signs would presuppose something outside the sequence that was a formal rather than merely instrumental sign; and so on.

Third, the arguments all essentially purport to identify something that must be true of metaphysical necessity.  They are not merely probabilistic in character, or arguments to the best explanation.  The claim is that there could not even in principle be secondary causes without primary causes, instrumental signs without formal signs, and so forth.  The arguments intend to identity the necessary preconditions of there being such a thing as causality, meaning, knowledge, or action. 

Hence, whether one accepts such an argument or not, the claims of empirical science are not going to settle the matter, because the arguments are conducted at a level deeper than empirical science.  The very practice of empirical science presupposes causality, meaning, knowledge, and action.  The arguments in question, since they are about the necessary preconditions of those things, are also about the necessary preconditions of science.  They are paradigmatically philosophical in nature.