The Modal Ontological Argument

In prior posts, I've written about how absurd I think this argument is. Its an argument that, if you have no familiarity with philosophical logic, it appears obvious if you are already a theist. If you have any familiarity with this kind of argumentation, it's quite dubious. This polarizing nature is interesting to me because I think it highlights problems with using philosophical argumentation to persuade anyone of anything under conditions of deep disagreement (and generally, the uses of argumentation by rational agents with differing hinge commitments). The paper "Wittgenstein and the logic of deep disagreement" is where my thoughts are currently with regard to this question; specifically with respect to arguments for the existence of a god. Are we fundamentally disagreeing about arational beliefs, and if so, what is the relationship between hinge commitments and trust? Are theistic beliefs knowledge-apt beliefs or folk-beliefs? This is an active area of inquiry within philosophy. My objective here is not to investigate the causes of deep disagreement, but instead I want to specifically show why this argument is significantly lacking. So first we will steelman the argument in argumentation schema form (because I find this structure to be an effective pedagogical tool). Then, at a high level, I will discuss the particular aspects I find ridiculous.

Below is a steelmaned (i.e., strongest, most charitable) modal ontological argument (MOA) cast in Walton-style argumentation scheme form, with (i) explicit definitions, (ii) hidden assumptions made visible, and (iii) a broad, in-depth set of critical questions (CQs) you can use to evaluate any instance of the argument.

Contents

Key terms and notation
The steelman MOA as a Walton-style scheme
Hidden assumptions needed for the argument to seem cogent
Comprehensive critical questions (Walton-style CQs)
A compact map of where the MOA lives or dies

Key terms and notation

Modal operators

□p: p is necessary (true in all metaphysically possible worlds).
◇p: p is possible (true in at least one metaphysically possible world).
Impossibility: p is impossible iff ¬◇p iff □¬p.

Possible-world semantics

A possible world is a maximal way reality might have been (metaphysically).
Metaphysical possibility is stronger than “I can imagine it” and stronger than “consistent with what we know.”

God / Maximally Great Being

Let G(x) abbreviate “x is a maximally great being.”

A common Plantinga-style definition:

Maximal excellence (ME): a cluster of “perfections” such as omnipotence, omniscience, and moral perfection (details vary).
Maximal greatness (MG): maximal excellence in every possible world.

So:

G(x) iff □ME(x) (or equivalently: x has maximal excellence essentially / necessarily).

Necessary existence

NE(x): x exists in every possible world: □E(x) (where E(x) is “x exists”).
In MOA presentations, NE is treated as a “great-making” feature or as built into maximal greatness (via □ME(x)).

S5 principle (the rocket booster)

In S5, ◇□p → □p.

Informally: “If something is possibly necessary, then it is necessary.”

The steelman MOA as a Walton-style scheme

The MOA is best seen as a compound of two schemes:

a modal transfer scheme (from ◇□p to □p), plus
a definition/essence scheme connecting “God” to necessary existence.

Scheme A: Modal Transfer (S5) Scheme

Premise A1 (Modal principle): In the relevant modality (metaphysical), ◇□p → □p.

Premise A2 (Possibly necessary claim): ◇□(∃x G(x)).

Conclusion A (Necessity): □(∃x G(x)).

Bridge to actuality: From □(∃x G(x)), infer ∃x G(x) (if it’s necessary, it’s actual).

Defeaters (typical):

Reject S5 for metaphysical modality.
Deny A2 (the crucial possibility premise).
Deny that the modality used is metaphysical (equivocation to epistemic/imagined possibility).

Scheme B: Definition/Essence Scheme (God ⇒ necessary existence)

Premise B1 (Definitional/essential claim): By “God” we mean a being such that if it exists, it exists necessarily (or equivalently: maximal greatness entails necessary existence).

Formally: G(x) → □E(x) (or G(x) → NE(x)).

Premise B2 (Coherence claim): The property-set constituting maximal greatness is internally consistent and compossible.

Conclusion B: The concept “possibly necessarily existing maximally great being” is not obviously incoherent.

This doesn’t yet yield existence—but it’s what’s used to motivate A2 as “live.”

Hidden assumptions needed for the argument to seem cogent

These are often left implicit; making them explicit is where the real evaluation happens.

Modal-framework assumptions

Metaphysical modality is well-modeled by possible worlds (at least enough to support □/◇ reasoning).
S5 is the correct logic for metaphysical necessity (or a principle strong enough to get ◇□p → □p).
No equivocation among modalities (conceivable/epistemic/practical vs metaphysical).

Quantification/existence assumptions

Cross-world quantification is coherent (talking about “exists in every world” isn’t nonsense).
The domain assumptions are fixed appropriately (constant domain vs varying domain; actualism vs possibilism).
The step from □∃xG(x) to ∃xG(x) is licensed (i.e., “necessary implies actual” under the intended semantics).

Essence/property assumptions

Essences / essential properties are legitimate (there are de re necessities: some things have properties necessarily).
Maximal greatness is a well-defined property, not a vague honorific.
The “great-making” set is objective and non-ad-hoc (or at least defensible by independent theory).
Compossibility: omnipotence/omniscience/perfect goodness (etc.) can co-exist without contradiction.
Uniqueness (often assumed): there can’t be two distinct maximally great beings (or else “maximal” is unstable).

Epistemic/dialectical assumptions

We have some warrant for A2 (◇□∃xG(x)) that isn’t mere “I can’t see a contradiction.”
The burden of proof: it’s legitimate to treat the possibility premise as “default plausible” absent a demonstrated contradiction.
No parody premises with equal standing (e.g., “possibly necessarily no MGB exists”) are equally warranted.

Comprehensive critical questions

I’ll group them so you can target the exact weak joint the proponent is relying on.

Group 1: Clarifying the claim and the modality

CQ1. Which modality is meant—metaphysical, logical, epistemic, nomological (physical-law), or conceptual?
CQ2. Are you switching modalities mid-argument (e.g., “conceivable” → “metaphysically possible”)?
CQ3. What semantics for worlds/necessity are you assuming (and why is that the right model here)?

Group 2: Logic choice and the S5 step

CQ4. Why should metaphysical modality satisfy S5 in particular, rather than S4 or weaker?
CQ5. What is the argument for the accessibility assumptions that validate ◇□p → □p?
CQ6. If S5 is rejected, is there an alternative route to the conclusion, or does the argument fail?
CQ7. Are there independent reasons to accept S5 that don’t covertly assume the conclusion (e.g., “metaphysical necessity is absolute”)?

Group 3: The pivotal premise (◇□∃xG(x))

CQ8. What is your warrant for ◇□∃xG(x) specifically (not merely ◇∃xG(x))?
CQ9. Is your warrant more than “I can conceive it” or “it seems coherent”? If so, what is it?
CQ10. What would count as evidence against this possibility premise?
CQ11. Why isn’t the symmetrical premise ◇□¬∃xG(x) (or similar) equally warranted?
CQ12. If both “possibly necessarily G” and “possibly necessarily not-G” look prima facie conceivable, what adjudicates between them?

Group 4: Definition of God / maximal greatness

CQ13. What exactly is maximal greatness? List its component properties.
CQ14. Are these properties defined univocally (same sense) across worlds and contexts?
CQ15. Are these properties coherent as a set (no hidden contradictions)?
CQ16. Do you have an independent reason to treat “necessary existence” as a perfection rather than a status label?
CQ17. Does the definition risk triviality: “God = necessarily existing being” so the real work is smuggled into the possibility premise?
CQ18. Why prefer this definition over nearby rivals (maximally powerful but contingent; morally perfect but not omniscient; etc.)?

Group 5: Great-making properties and value theory

CQ19. What theory of “great-making” are you using (axiology/ordering of properties)?
CQ20. Is the ordering objective or relative to evaluators? If objective, what grounds it?
CQ21. Are “perfections” constrained by compossibility and metaphysical structure, or can you stipulate them?
CQ22. Does this invite Euthyphro-style concerns (is greatness independent, or defined by God)?

Group 6: Existence, quantification, and cross-world identity

CQ23. What does it mean for an individual to “exist in another world” on your metaphysics?
CQ24. Are you assuming transworld identity (the same being across worlds) or counterparts? Does the argument survive either way?
CQ25. Do you assume constant domains? If domains vary, does □∃xG(x) still entail an actual instance?
CQ26. Are you assuming actualism (everything quantified over exists actually) or possibilism? How does that affect “exists in every world”?

Group 7: Dialectical and methodological issues

CQ27. Does refusing the possibility premise commit one to “God is impossible,” or is suspension of judgment allowed?
CQ28. What is the proper burden of proof for a premise that—together with S5—yields an existential conclusion?
CQ29. Are there independent reasons (outside the MOA) to think the possibility premise is more plausible than its negation?
CQ30. If “plausible” is invoked, what are the criteria—simplicity, coherence, explanatory power, fit with other commitments?
CQ31. Are those criteria shared by both parties, or are they background-belief dependent (culturally/ideologically loaded)?
CQ32. If the argument is sound only relative to a thick package of background metaphysics (S5 + essences + objective perfections), is it still persuasive to someone who doesn’t already buy that package?

Group 8: Robustness tests (stress tests / defeaters)

CQ33. Can you construct a parallel argument for an incompatible “maximally great” entity (or necessary nonexistence) using the same method?
CQ34. Does the argument risk “modal collapse” style consequences (too much becoming necessary) given your modal/essentialist commitments?
CQ35. If maximal greatness is possible, why doesn’t it entail other exotic necessities (necessary angels, necessary universes, etc.) via similar definitions?

Now that we have seen the argument, I'll discuss what I see to be aspects that call into question its persuasive force.

“Defining God as necessary” isn’t a proof—it’s a constraint

I’m not granting that God must be necessary. God could be contingent. And if the reply is, “Sure, some godlike being could be contingent, but God (the traditional perfect being) includes aseity/independence, so contingency disqualifies,” then that just exposes what’s really happening. At that point, the MOA isn’t proving “a god.” It’s arguing: if the perfect-being concept is coherent/possible, then it’s actual. That’s a very different claim—and it’s exactly why the whole thing feels like “defining into existence.” Necessity is baked into the definition, and then the premise asserts possibility. So the real question isn’t “did you run the modal logic correctly?” The real question is whether the concept “a necessarily-existing maximally great being” is genuinely possible/coherent—or whether it’s more like “a necessarily-existing married bachelor.”
Rejecting S5 is not a cheap move

Rejecting S5 is not a dodge. S5 is a strong commitment about the structure of metaphysical modality—roughly, that what’s possible doesn’t vary from world to world in the relevant way. Plenty of philosophers think weaker systems (like S4) make more sense depending on what modality even is. And if the pushback is “S5 is very plausible for metaphysical necessity,” my response is simple: plausible isn’t free; that’s exactly what’s at issue. If the argument needs S5, then the theist owes an account of modality on which S5 is actually warranted.
The “symmetry/parody” worry is real—and it targets the possibility premise

The method invites a symmetry problem. If I can license the “possibly necessary” move for God, then I can also run a parallel move in the opposite direction:
- ◇□(God exists) → □(God exists)
- ◇□(God does not exist) → □(God does not exist)
Defenders will say, “Fine—but both can’t be possible; one possibility premise is false.” Exactly. That’s my point: the MOA doesn’t establish possibility; it assumes it. So the argument is conditional on a premise that’s almost as contentious as the conclusion. Assuming the modal logic isn’t the problem; the “possibly God” premise is where the theology is smuggled in.
“Being” equivocation: “a being” vs “Being itself”

This is a real problem in a lot of popular presentations, because they slide between two very different ideas:
- Plantinga-style MOA is about a being (an individual entity).
- Thomistic/classical-theist talk about Being itself (ipsum esse subsistens) is a different metaphysical project and doesn’t neatly map onto “one more entity” in possible-world semantics.
When someone argues with the MOA but then starts talking like they’ve proven “Being itself,” that’s an equivocation. If they want the MOA, I’m keeping them pinned to “a being” in possible-world semantics—unless they can explicitly bridge the gap rather than just gesturing at it.
Kant: “existence isn’t a predicate” vs modal versions

Kant’s classic point targets the move where “existence” gets treated like an ordinary property that makes something greater. Modal defenders often respond: “We’re not adding bare existence as a perfection; we’re talking about necessary existence, a modal status.”

But that doesn’t end the problem—it just relocates it. I still want answers to the deeper questions:
- Even if “necessary existence” is a coherent modal property, why think it’s great-making?
- And even if it is, that only matters once the perfect-being framework is already accepted.
So the MOA may dodge Kant’s exact formulation, but it inherits the deeper worry: conceptual analysis doesn’t get you reality unless the possibility claim is independently justified.
“Maximal greatness” and “great-making properties” look stipulative

The “maximal greatness” package often looks like it’s built to guarantee the desired result. The underlying problem is basically a modernized Euthyphro-style dilemma:
- If “great-making” just means “whatever God has,” it’s circular.
- If “great-making” is independent, then God is being measured by an external standard—and that standard needs explanation.
Yes, defenders try to ground “great-making” in intrinsic value, perfections (knowledge/power/goodness), or rational idealization (“better to have than not have”). But outside a fairly specific value theory, “maximal greatness” still looks like a philosopher’s caricature of greatness.

At minimum, I’m owed a non-ad-hoc account of which properties are genuine perfections, which are compossible, and why. Without that, “MGB” starts to look like a hand-built key designed to fit the lock.

So is it just “P1 God; therefore God exists”?

Not formally. Formally, it’s closer to this structure:

P1: Possibly, a necessarily-existing maximally great being exists.
P2: S5 (or equivalent modal principles about metaphysical necessity).
C: Therefore, such a being exists.

But my “question-begging” diagnosis is still aimed at the real center of gravity: P1 does almost all the substantive work. And it’s hard to see why anyone who doesn’t already lean theistic should grant it. So my conclusion is straightforward: the MOA can be valid while still being dialectically ineffective. Its key premise isn’t more plausible than the conclusion; it often just restates the conclusion in modal clothing.

Whenever I criticize the modal ontological argument, I keep getting the same rhetorical comeback: “So you actually think it’s impossible that God exists?” And that’s the move I’m rejecting. It’s designed to make my position sound extreme, but it does that by quietly confusing the modal options on the table.

Here’s the basic point: the statements “it is possible that God exists” and “it is impossible that God exists” are not the right “symmetric pair.” The symmetric pair is:

“it is possible that God exists”
“it is possible that God does not exist”

Those are the genuinely balanced counterparts—because they apply the same modal operator (“possibly”) to opposite propositions. The “possible vs impossible” framing swaps in a different relation and loads the exchange in one direction.

The modal sleight-of-hand

What’s happening is that people blur together two different kinds of “opposites”:

Contradictories: one must be true and the other must be false.
Contraries: they can both be false, but they can’t both be true.

In modal logic, possibility and impossibility are contradictories. But two possibility claims—one about a proposition and one about its negation—are not contradictories in that same way.

Why the two “possibly…” claims are the real symmetry

Let G mean “God exists.” Then the genuinely symmetric pair is:

◇G (“possibly, God exists”)
◇¬G (“possibly, God does not exist”)

These really are symmetric: same operator, opposite content. And the crucial point is that they can both be true. That happens whenever G is contingent—true in some possible worlds and false in others. In that case:

God exists in at least one possible world ⇒ ◇G
God does not exist in at least one possible world ⇒ ◇¬G

So this pair lines up exactly with the ordinary, non-dogmatic posture: maybe God exists, maybe God doesn’t.

Why “possibly G” vs “impossible G” is dialectically loaded

Now compare that with the usual trap question. “Impossible that God exists” isn’t the symmetric counterpart of “possibly God exists.” The real opposite of ◇G is:

not possibly G (¬◇G)

And in standard modal equivalences, ¬◇G is equivalent to □¬G (“necessarily not-G”), which is what people mean by “impossible G.”

So yes: ◇G and □¬G are contradictories—one must be true. But notice what the defender is doing in conversation: they’re swapping the discussion from the balanced pair (◇G vs ◇¬G) into a maximally loaded pair (◇G vs □¬G) so it sounds like I must be claiming God is ruled out by logic.

The missing middle position everyone tries to erase

The modal ontological argument doesn’t merely need “possibly God exists.” It typically needs something stronger, like:

◇□G (“it is possible that God exists necessarily”)

And my critique is often: I’m not granting that premise.

But then comes the bait-and-switch: “So you think God is impossible?” That tries to force my refusal to accept their premise into an acceptance of its negation. In other words, it tries to turn:

“I don’t accept ◇□G”

into:

“I accept ¬◇□G”

And that’s not a legitimate move in ordinary argumentation. Withholding assent is not the same thing as asserting the negation.

Even more importantly, there’s an entirely coherent middle position that the loaded framing tries to delete:

I might accept ◇G (maybe God exists)
while rejecting ◇□G (maybe God exists necessarily)

That’s just the “God, if real, could be contingent” stance. It’s not extreme. It’s not a claim of impossibility. It’s simply a refusal to let someone smuggle necessity in under the cover of “possibility.”

So what’s the real false dichotomy here?

Purely as a matter of semantics, ◇G vs □¬G is exhaustive: one of them is true. Fine. But that’s not the relevant point in a debate about whether an argument works.

The real false dichotomy is dialectical. It’s the attempt to frame the discussion as if my only options are:

(1) “I accept your possibility premise,” or
(2) “I assert God is impossible.”

There’s a huge rational middle space:

I’m not convinced your concept is coherent, your modal framework is correct, or your possibility premise is justified—so I’m withholding.
And more precisely: I’m not granting ◇□G. That doesn’t commit me to □¬G.

So no, I’m not claiming □¬G. I’m saying you haven’t justified ◇□G (or even ◇G, depending on the version). Withholding assent to your key premise isn’t the same as asserting its negation. Your argument needs “possibly necessary.” I’m not granting that. That doesn’t mean I’m claiming “necessarily not.”

When I point out the false dichotomy in how people defend this argument, the next line I often hear is: “Well, isn’t it plausible that God exists?” And I’m not buying that as a substitute for doing the actual work the modal premise requires.

What’s usually happening here is a quiet slide from conceivability (“it seems coherent to me” / “I can imagine it”) to metaphysical possibility (“there is a genuine possible world where this obtains”). And I’m arguing that this slide is not reliable—and often not even valid.

Why “conceivable ⇒ metaphysically possible” can’t be the foundation

There is a principle some philosophers flirt with: if something is conceivable, then it’s metaphysically possible. But I’m rejecting that as a general bridge, because we have strong reasons to think conceivability generates modal illusions—cases where something feels coherent from the armchair but turns out to be impossible once you factor in necessities we don’t see.

That’s why I lean on a Kripke-style lesson: we can “conceive” scenarios that are metaphysically impossible because we’re ignorant of necessary identities or essences. In other words, my imagination isn’t a special access point to metaphysical space—it’s a fallible heuristic that can be fooled. And if it can be fooled in those cases, then it can’t be the engine propping up a premise as strong as the one the MOA needs.

Same basic issue with Putnam-style externalist themes: what I feel like I’m imagining may just reflect semantic or epistemic gaps—mislabeling a different scenario as “the same thing,” rather than accessing a real metaphysical alternative. So even if something feels “conceivable,” that doesn’t automatically tell me it’s metaphysically possible in the robust sense.

And this is exactly where the MOA needs more than vibes. The argument doesn’t just want “God seems coherent to me.” It typically needs something like:

◇□G (“it is possible that God exists necessarily”)

But conceivability-based support—at best—usually buys you something closer to epistemic openness (“for all I know, maybe”), not the metaphysical possibility the argument is trying to cash out into actuality.

If the defender’s “plausible” just means “not obviously ruled out,” fine. But that doesn’t get them to the kind of possibility premise the MOA actually needs.

The deeper problem: “plausibility” is doing disguised rhetorical work

Even setting conceivability aside, I think “plausible” is a deeply unstable lever in this debate. There isn’t a principled, shared scoring rule for plausibility the way there is for probability—unless someone explicitly supplies a framework. And in practice, plausibility is highly relative to:

your background metaphysics (naturalism vs theism, etc.),
your explanatory tastes (simplicity, depth, scope),
what your intellectual community treats as “live options,”
and the broader cultural inheritance that shapes your default intuitions and priors.

So when someone says “isn’t it plausible that God exists?”, I hear something like: “given my background package, this doesn’t feel crazy.” But that’s not an intersubjective justification—especially not one strong enough to underwrite a premise that’s meant to deliver necessity and existence from modal machinery.

When someone appeals to “plausibility,” I don’t let it float as a magic word. I force specificity:

Plausible in what sense? Epistemically live (not ruled out), or metaphysically possible in the strong sense the argument needs?
By what standard? What would actually raise or lower plausibility here—what evidence or theory-choice rules are you using?
Which premise are you defending? ◇G (“possibly God exists”) or ◇□G (“possibly God exists necessarily”)?

Because that’s the point: the MOA doesn’t get to treat “plausible” as a free pass. If the argument needs possibly necessary, then “I can imagine it” and “it feels plausible to me” are not doing the job.

It might be epistemically open that God exists. But the argument needs metaphysical possibility—often “possibly necessary.” Conceivability or intuitive plausibility doesn’t establish that, and we have plenty of reason to doubt that conceivability reliably tracks metaphysical possibility.

Optional references

Finally, this brings me back to my opening remarks about deep disagreement. The success (or persuasive force) of this argument solely relies on whether we accept certain definitions and rules of inference. The "plausibility" of the premise fundamentally rests upon pre-theoretic, a-rational hinge commitments. Selecting S5 over S4 also seems to be highly biased by these pre-theoretic assumptions. Mathematicians do have rules of thumb for selecting certain systems over others but there is no strict decision criteria to my knowledge that binds someone to a specific formal system, unless it fails to meet very basic expectations such as consistency or if the axioms are overtly ad-hoc. I think analyzing this argument makes it pretty clear that philosophical argumentation is somewhat problematic for establishing claims whether a god exists. Hence, my agnosticism; because empirically based arguments are even more dubious. But, we could just be like Plantinga, and claim that opponents to his arguments are just fundamentally corrupted by sin.