The Coherence of Theism
The pages on omnipotence, omniscience, eternity, and goodness, simplicity, and aseity each examined a single attribute and its internal strains. This capstone asks the joint question: are the divine attributes mutually consistent — can one being possess all of them together? This is the deepest form of the incoherence-of-attributes strategy, and it cuts beneath the arguments for and against God's existence. For if the concept of God is incoherent — if "omnipotent, omniscient, perfectly good, eternal, simple, necessary being" describes an impossible object, like a round square — then God's non-existence is established a priori, no argument from evil or empirical inquiry required, and the theistic arguments (which would establish an impossible being) are unsound at a stroke.
This page frames the coherence question, surveys the main incompatible-properties arguments, sets out Swinburne's program of defending coherence, and explains why the issue is logically prior to the existence debate.
Why coherence comes first
The dialectical situation is stark and is well captured by the modal structure of the ontological argument. A maximally great being, if it exists, exists necessarily; so it exists in every possible world or none. Hence:
If the concept of God is coherent (possibly instantiated), the theist is most of the way home; if it is incoherent (not possibly instantiated), the atheist is all the way home — God is impossible, not merely absent. So the possibility of God is the hinge, and an incompatible-properties argument tries to settle the question without ever weighing evidence: it aims to show directly, by deriving a contradiction from the divine attributes alone.
This makes the coherence question a powerful atheological strategy (Drange, Martin, Everitt) — and equally a precondition the theist must secure, since every argument for theism presupposes that there is a coherent something for the argument to establish.
Incompatible-properties arguments
These come in two grades, within a single attribute and between attributes.
Within an attribute
- Omnipotence: the stone paradox and the can-God-sin? puzzle — largely defused by restricting omnipotence to the logically possible and compossible with God's nature.
- Omniscience: can a being that cannot learn, doubt, or err know what it is like to learn, doubt, or sin? Can a timeless being know what time it is now? The first is met by distinguishing propositional from experiential knowledge (or denying the latter is genuine knowledge-that); the second is the live tensed-knowledge problem.
Between attributes
The more serious arguments allege that pairs of attributes cannot be jointly held:
- Omniscience vs. immutability (Kretzmann's argument). An omniscient being knows what time it is now; what time it is changes; so an omniscient being's knowledge changes; so an omniscient being is not immutable. Conversely a strictly immutable being cannot track the moving present, so is not omniscient. The classical escape is timelessness plus a B-theory of time (there is no objective "now" to track) — which is why these attributes form a package (see eternity).
- Omniscience vs. human freedom — the foreknowledge problem: infallible foreknowledge of free acts seems to remove the freedom. (Strictly a tension between an attribute and a further commitment — libertarian free will — rather than between two attributes, but it functions as a coherence worry for the package "omniscient creator of free agents.")
- Omnipotence + omniscience vs. omnibenevolence — the problem of evil in its logical form: the inconsistent triad. Now widely conceded to fail as a strict-inconsistency argument (Plantinga's free-will defense), but it is precisely an incompatible-properties argument, and the most important.
- Perfect justice vs. perfect mercy. Justice gives each exactly what is deserved; mercy withholds deserved punishment. A being cannot, for the same act, both exact and remit the deserved penalty — so perfect justice and perfect mercy seem incompatible. The reply distinguishes spheres or appeals to atonement, but the tension is real and old.
- Transcendence vs. omnipresence/immanence; eternity vs. agency; simplicity vs. the plurality of attributes. Each pair pulls against the other and is addressed in the eternity and simplicity pages. The simplicity objections are the most radical, since DDS itself (God is his attributes) is alleged to be incoherent — meaning the classical theist's chief tool for harmonising the attributes is accused of being the deepest incoherence of all.
Swinburne's defense of coherence
Richard Swinburne's The Coherence of Theism (1977) is the major sustained attempt to show the attributes are jointly consistent. His strategy is telling:
- Define the attributes carefully and in a mutually adjusted way. Each attribute is glossed so as to be individually coherent and jointly compossible: omnipotence as the power to do whatever is logically possible and consistent with God's other properties; omniscience as knowing every true proposition it is logically possible to know; perfect goodness as always doing the best (or a good) action; and so on. Apparent contradictions often trace to careless definitions (unrestricted omnipotence, omniscience including the unknowable), which refinement removes.
- Where classical attributes resist coherent joint definition, revise them. Swinburne is willing to drop or weaken the most problematic classical commitments — he rejects divine simplicity (in the strong identity sense) and timelessness, opting for an everlasting, temporally-engaged, personal God whose omniscience does not extend to the free future. This theistic personalism trades metaphysical austerity for coherence and religious adequacy.
- Net result: a defensible concept of God — not the full classical package, but a coherent maximal-being concept that the theistic arguments can then aim to establish and to which the cumulative case applies.
The cost, as classical theists note, is that securing coherence by abandoning simplicity, timelessness, and exhaustive foreknowledge yields a God arguably less than the id quo maius cogitari nequit of perfect-being theology — so the coherence debate doubles as a debate over which concept of God is the right target.
The shape of the verdict
No incompatible-properties argument has won general assent as a decisive disproof, for a structural reason: the theist can almost always refine the offending attribute (restrict omnipotence, relocate "now"-knowledge, distinguish senses of goodness) to block the derivation. But refinement has a price — each adjustment moves the concept further from the simple, maximal, classical ideal, and the adjustments must be mutually consistent, which is hard to verify across the whole set. So the realistic options are:
- Classical theism (Aquinas): keep simplicity, timelessness, immutability, and exhaustive providence as an interlocking package, and defend each tension (Cambridge change, ET-simultaneity, the truthmaker reading of simplicity) — maximal coherence-by-mystery, contested at the simplicity joint.
- Theistic personalism (Swinburne, Plantinga): drop the most paradox-prone classical attributes for a coherent, everlasting, non-simple personal God — coherence purchased by revision.
- Atheological verdict (Martin, Drange): the un-revised classical concept is incoherent, and the revisions either fail or yield something that is no longer God.
Where coherence sits
The coherence of theism is logically prior to the entire existence debate: a coherent God is possibly instantiated and, given the modal point, necessarily instantiated if at all, whereas an incoherent God is impossible and no evidence could establish him. The incompatible-properties arguments — within attributes (the omnipotence paradoxes), and especially between them (omniscience vs. freedom and immutability, power-and-goodness vs. evil, justice vs. mercy, simplicity vs. plurality) — none of which has proven a decisive disproof, because the theist can refine the attributes; but each refinement migrates the concept from the classical package toward a more modest personal deity. The standing question, which every later topic inherits, is therefore twofold: not only whether God exists, but which coherent concept of God — austere-classical or revised-personal — is the one whose existence is at issue. With the nature of God mapped, the section turns from metaphysics to the epistemology: granted a coherent God, how could anyone know he exists?