Argument from Causation

In its most general form, the Argument from Causation holds that every effect requires an efficient cause, that nothing causes itself, and that an infinite regress of caused causes cannot constitute a complete explanation of why there are any caused things at all. From this it concludes that there must exist a first, uncaused efficient cause underlying the causal order, which theists identify with God.

This version is more loosely stated than Aquinas's precise Second Way and can be developed either hierarchically, concerning what presently sustains causal activity, or in a more temporal register, concerning what ultimately originated a causal chain extending into the past.

Cosmological arguments from causation trace back through Aristotle and Aquinas but were also developed independently in Islamic philosophy (al-Kindi, Avicenna) and in early modern rationalism, notably by Leibniz and Samuel Clarke, who emphasized a broader principle of sufficient reason alongside strict causal reasoning.

In the twentieth century, philosophers like Richard Taylor and William Lane Craig revived and refined causal cosmological reasoning, engaging directly with Humean objections and with debates in philosophy of science about causation and explanation.

1. Every caused thing has an efficient cause. 2. Nothing causes itself. 3. Therefore, every caused thing is caused by something other than itself. 4. An infinite regress of caused causes does not constitute a complete explanation of the existence of caused things. 5. Therefore there exists an uncaused first efficient cause, which explains, without itself needing further causal explanation, the existence of the causal order.

Unlike the Kalam argument, this version does not require premise that the universe began to exist in time; the regress in question can be conceived either temporally or hierarchically.

The causal principle in premise one is defended by appeal to the near-universal practice of causal explanation in science and everyday reasoning, even granting Humean skepticism about necessary connections. Premise two, that nothing causes itself, rests on the incoherence of a thing existing prior to itself.

The crucial premise four requires arguing that explaining each member of an infinite causal series by appeal to its immediate predecessor does not thereby explain why the series as a whole exists rather than not, since every member remains contingent and dependent, with no unconditioned member grounding the whole.

The Hume-Edwards objection, developed by Paul Edwards, holds that a complete explanation of an infinite series consists simply in explaining each member by its predecessor; demanding a further explanation of 'the series as a whole' illegitimately treats the series as an additional entity requiring its own separate cause. A second objection, from Hume himself, questions whether the causal principle can be established with the necessity the argument requires, since causal connections may be nothing more than observed regularities.

A third objection notes the ambiguity between hierarchical and temporal readings can allow equivocation, where an arguer defends the temporal premise using considerations that only support the hierarchical version or vice versa.

In reply to Hume-Edwards, defenders argue that explaining each part by its predecessor never introduces anything unconditioned into the explanation; the whole series remains a collection of contingent, dependent members, and asking why there is such a collection at all, rather than nothing, is a distinct and legitimate question that local explanations do not touch.

In reply to Hume, defenders note the argument needs only a modest causal principle, that caused things have causes, which is compatible with regularity theories of causation and does not require the strong metaphysical necessity Hume was attacking.

Even if successful, this argument establishes only that an uncaused first cause exists; it does not by itself demonstrate that this cause is personal, singular, omniscient, or morally perfect, requiring supplementary argument to reach full theism.

It also leaves open whether the first cause is temporally first (as in Kalam-style versions) or ontologically first while the universe is past-eternal (as in Thomistic hierarchical versions), a distinction that must be specified for the argument to be evaluated precisely.

When discussing this argument, first clarify which version is intended, hierarchical or temporal, since conflating them invites objections that do not actually apply to the version being defended.

Be honest that the crucial and most contested step is premise four, concerning whether an infinite regress can be a complete explanation, and be ready to explain why explaining each link does not, in the proponent's view, explain the chain's existence as a whole.