Why An Infinite Regress Won't Work

Chris Redfield on YouTube asks: “I've always been confused on why exactly an infinite regress of causes or ontological dependence can't work. People give examples where a first cause or ground is removed from a hierarchical series that originally had one and I get why that would be flawed but why is removal of a first cause equivalent to an infinite regress of causes or dependence?”

Recall that Thomists don’t maintain there could never be an infinite causal regress, only that there could never be an infinite causal regresses when considering a per se — or essentially ordered — series.

In an essentially ordered series of X → Y —> Z, Y borrows the power from X to cause Z. The stock example of an essentially ordered series is the hand using the stick to push the stone, where the causal property within the series is derivatively had by the instrumental members. No first mover — that is, nothing with the in-built power to move (like the mental agent) — then no motion.

In an accidentally ordered series X —> Y —> Z, Y does not borrow the power from X to cause Z. The stock example of an accidentally ordered series is Bob begetting Frank and then Frank begging Harry. Frank can beget Harry after Bob is dead because Frank has the inbuilt power to do so.

The reason there cannot be an infinite regress of per se ordered causes is purely a logical one. If we have shown that Thumper’s existence is for Thumper to be dependent upon something extrinsic — call that something, A — we say Thumper exists only if A exists. Now, if A is existentially dependent like Thumper, the series extends — a series, that is, of conditionals; a series of “only ifs”.

Conditionals can be piled up to whatever extent — including ad infinitum — without producing a categorical conclusion. Just think: from the conditional “P if Q” we want to infer P, we require the categorical “Q” apart from the conditional. It is the lack of the categorical inherent within the postulation of an infinite regress that makes it impossible (logically) to assert categorically that Thumper exists. But since Thumper exists (there he is, after all, stealing carrots) and we can draw the categorical conclusion, our series cannot be comprised of only conditionals. There must be at least one categorical — “Z exists” — which means the series ends with an existentially independent entity.

Returning to your point about removing a first cause, such an illustration is typically meant to convey the conditional nature articulated above. Take a room with the property of illumination, which it receives from the moon outside. If, however, we are right in thinking the moon can only reflect light then we continue searching for what introduced that causal property into the series, knowing it could not have been the moon. Increasing the number of moons within the system fails to alleviate the issue, no matter how many moons we posit (finite or infinite), no matter how the moons are arranged (linearly, in a circle, etc.). At some point — either within or without the system of moons — there must be something that is either inherently luminous or otherwise has the power to produce light, else the causal property would not be there in the first place.

And of course, it won’t due for the critic to just say it’s a brute fact that this or that causal property is within an infinite causal chain absent a primary cause, for that doesn’t support the infinite regress objection as an alternative adequate explanation, it just denies explanation altogether.

For more on the Infinite Regress objection, I recommend Michael Rota’s paper.