Four Problems with the First Way

It will be convenient to break the claim about the finiteness of causal chains into two parts: first, the impossibility of circular causation, and second, the impossibility of noncyclical infinite regresses. To establish that circular causation is impossible, we need two assumptions: that causation is transitive, and that no change causes itself.

Here’s the argument:

(1a.1) There are changes.

(1a.2) Every change has a cause. (Seriality)

(1a.3) Causation* of change is transitive. If change x causes change y, and change y causes change z, then x causes z. (By the definition of cause*, the transitive closure of cause.)

(1a.4) No change can cause* itself. (Irreflexivity, a consequence of Finitude)

(1.a5) No change has a non-cyclical infinite chain of causes*.  (Well-foundedness, another consequence of Finitude)

(1a.6) Therefore, every change is caused* by some non-change, and this non-change causes at least one change. (From 1a.1-1.4, identical to Theorem 1.1)

(1.7) A non-change that causes change must be the state of a thing that has the power to cause change without thereby undergoing any change.

(1a.8) Therefore, there is at least one thing with the power to cause change without thereby undergoing any change.

There are four problems with the simple version. We take premise 1a.1 to be unassailable. Everyone will concede that some things change. So, no problem there.

Problem 1. (Re premises 1a.2 and 1a.7) Why must every change have a cause? And why must this cause be a mover (i.e., something exercising a power to cause change, as premise 1a.7 requires)? Even if we accept a Principle of Sufficient Reason, requiring that there be a reason or explanation for every instance of change, couldn’t there be some change whose sufficient reason involves something other than the action of a changer?

Problem 2. (Re premise 1a.4) Why couldn’t there be cases of spontaneous change? Why, for example, couldn’t some creatures cause themselves to change through an act of free will?

It’s true that St. Thomas seems to give us a subsidiary argument for premise 4 in the First Way. Here is a standard formulation of that argument:

1a.4.1 Nothing can be both actual and potential in the same respect at the same time (whether an instant or a period).

1a.4.2 If x changes y at time t, then there is some respect R in which x is actually R at t and y is only potentially R at t (or, perhaps, there is some interval (t*-t), not including t, such that x is actually R throughout (t*-t) and y is potentially R throughout (t*-t)).

1a.4. Nothing changes itself.

But premise 1a.4.2 is itself problematic. When something actually F thereby causes something else to be F, Thomas calls the causation involved “univocal.” But he recognized that not all causation is univocal. Many instances of causation are “equivocal.” Suppose that x changes y at t, in such a way that y becomes F for the first time at t. Condition F could be, by way of example, (i) occupying a certain location L, (ii) having a certain temperature E, or (iii) having a certain velocity of motion V. In none of these representative examples is it necessary that x (the changer or agent) be actually F at or before time t. The causation could be equivocal. The changer or agent x could cause y to move to location L without being in location L itself, or to reach a certain temperature E without being that temperature itself, or reach a velocity V without having velocity V.

In fact, the argument of the First Way depends on rejecting 1a.4.2 in its full generality, since otherwise it would be impossible for an unmoved mover to cause motion without being in motion itself.

Therefore, the argument from 1a.4.1 and 1a.4.2 cannot reflect Thomas’s actual reasons for accepting 1a.4. So, what were his reasons?

Problem 3. (Re premise 1a.5) Why couldn’t an infinite series of causes, extending infinitely far into the past, sufficient to explain present motion? As is well known, both Aristotle and Thomas grant that such historical infinite regress is possible—in fact, for Aristotle, this is the actual situation. If such a regress were actual, why couldn’t we appeal to some past mover’s causing some kind of persisting impetus or inertia as the cause of present motion?

In fact, we don’t have to bring in scientific innovations like Buridan’s impetus or Newton’s inertia to create this problem. The problem already exists within Aristotle’s own cosmology, since the heavenly spheres rotate by a kind of self-contained, natural “inertia,” and sub-lunar bodies fall and rise naturally by a kind of intrinsic disposition. Both Aristotle and Thomas seem to speak of sublunar bodies in natural motion as being set in motion by their original generator, the thing that originally generated their fiery or earth substance. Such an original generator can cease to exist before the resulting motion ends. So, it seems that the cause of present motion need not be acting in the present. If that’s right, then it seems that a temporally extended infinite regress (which both Aristotle and Thomas admit to be possible) would be sufficient to explain present motion, in contradiction to premise 5.

Problem 4. (Re premise 1a.8) The gap problem. Why must an unmoved mover have any of the characteristics of God? Why must it be eternal/timeless, necessary, lacking accidents, radically simple?

I will take up the four problems in the future posts. I will offer a new interpretation of the First Way that resolves these problems, and U will argue that our new interpretation is well supported by the textual evidence in both Aristotelian and Thomistic works.