Concept
Metaphysical vs Logical Modality
kinds of modality, types of necessity, metaphysical necessity, logical necessity, nomological necessity, physical necessity, epistemic possibility, what does possible mean in philosophy, metaphysical possibility vs logical possibility, necessary a posteriori, rigid designationIntro
The word "possible" means at least four different things, and almost every bad argument about God, minds, or miracles runs on swapping one for another mid-sentence.
Say someone tells you time travel is possible. They might mean any of these. That the idea is not self-contradictory, the way "married bachelor" is. That nothing about the nature of time rules it out. That the laws of physics permit it. Or simply that for all they know it might turn out to be true. Those are four separate claims, and one can be true while the others are false.
Philosophers give them names. Logical possibility is about contradiction. Metaphysical possibility is about what things really are. Nomological possibility is about the laws of nature. Epistemic possibility is about the state of your information.
Why this earns a page: two of the strongest arguments for God turn on which one you are using. The contingency argument needs the universe to be metaphysically contingent, and an opponent who demands a logical contradiction has answered a question nobody asked. The modal ontological argument needs it to be metaphysically possible that God exists, and an atheist who shrugs and says "sure, possible" almost always means it epistemically, which grants nothing. Getting this right is not pedantry. It is the difference between an argument that works and one that only appears to.
See this page's graph and discover linked topics
In full
Modality concerns necessity, possibility, contingency and impossibility. The formal apparatus is treated at Modal Logic; what is at stake here is the prior question of what kind of necessity a given modal operator is being used to express, since the same formalism serves several distinct modalities with different truth conditions and different epistemologies.
The standard contemporary taxonomy distinguishes logical modality (grounded in form), metaphysical modality (grounded in the natures or essences of things), nomological or physical modality (grounded in the laws of nature), and epistemic modality (grounded in an agent's information, and arguably not a genuine modality of the same family at all). The taxonomy in its current form is downstream of Saul Kripke's Naming and Necessity (lectures 1970, published 1980), which severed the necessary from the a priori and thereby made metaphysical necessity a category distinct from both logical and conceptual necessity. Kit Fine's "Essence and Modality" (1994) subsequently challenged the direction of explanation, arguing that essence is prior to modality rather than reducible to it.
The apologetic payoff is precision about which modality a premise asserts and which one an objection attacks. Most modal equivocation in live debate is unintentional, and naming it defuses the objection faster than answering it on its own terms.
The four modalities
| Modality | Grounded in | Necessary example | Impossible example |
|---|---|---|---|
| Logical | Logical form alone | Either it is raining or it is not | A square circle; P and not-P |
| Metaphysical | The natures or essences of things | Water is H2O; Hesperus is Phosphorus | Water is XYZ; this oak was born of human parents |
| Nomological | The laws of nature | Nothing accelerates past the speed of light | A body reaching light speed under thrust |
| Epistemic | An agent's information | Whatever the agent knows | Whatever the agent's evidence excludes |
Two clarifications the table cannot carry.
Conceptual or analytic modality is sometimes listed as a fifth, covering truths held in virtue of meanings, such as "all bachelors are unmarried." Denying one is not a formal contradiction but is incoherent given the concepts. Some philosophers fold this into broadly logical necessity; others keep it separate. The distinction that matters for apologetics is between conceptual and metaphysical, because "water is H2O" is metaphysically necessary and not conceptually necessary, which is precisely Kripke's point.
Epistemic modality is the odd one out. "The keys might be in the car" reports the speaker's ignorance, not a feature of the world. It varies with who is speaking and what they have checked; the other three do not. Treating it as a member of the family is the single most common source of modal error in argument.
How the spaces nest
Order the modalities by how much they rule out. Logical possibility rules out the least, so its space of possible worlds is the largest; nomological possibility rules out the most, so its space is the smallest.
nomologically possible < metaphysically possible < logically possible
Necessity runs the other direction, because a claim true in every world of a larger space is automatically true in every world of a smaller one:
logically necessary => metaphysically necessary => nomologically necessary
Worked examples, which are more use than the diagram:
- Faster-than-light travel. Metaphysically possible (there is no incoherence in a world whose laws permit it), nomologically impossible (our laws do not).
- Water being XYZ. Logically possible (no formal contradiction), metaphysically impossible (nothing that is not H2O is water).
- A married bachelor. Impossible in every sense.
- A miracle. Nomologically impossible, metaphysically possible. This is the whole reply to one common form of the Humean objection; see Miracles and Argument from Miracles.
One live dissent worth knowing. Timothy Williamson and others resist the neat nesting, holding that metaphysical possibility simply is the widest genuine possibility, and that "logical possibility" in the narrow syntactic sense is a property of sentences rather than a modality of the world. On that view the top layer of the diagram is not a bigger space but a different subject. This is not a fringe position, and it does not damage any argument below; it tightens them, since it removes the weakest layer.
Kripke and the break with the empiricist picture
Before 1970 the dominant picture, running from Hume through the logical positivists, treated three distinctions as one:
necessary = a priori = analytic, and contingent = a posteriori = synthetic
Kripke separated them by populating two cells that had been thought empty.
Necessary a posteriori. "Hesperus is Phosphorus." "Gold is the element with atomic number 79." "Water is H2O." Each was discovered empirically, so knowledge of it is a posteriori. Each is necessarily true, because the terms are rigid designators: they pick out the same thing in every possible world in which that thing exists. If Hesperus is Phosphorus, there is no world in which they differ. There are only worlds in which some other body occupies the evening sky, which is a different claim.
Contingent a priori. Kripke's case is the standard meter bar. Whoever fixes the reference by stipulating that stick S is one meter long knows that truth a priori, without measuring. It is nonetheless contingent, since S could have been heated and been longer.
The result reorganizes the field: a priori and a posteriori is epistemology, necessary and contingent is metaphysics, and analytic and synthetic is semantics. Three independent axes, not one. Nearly every careless modal argument assumes they still line up.
Note what this does not license. Kripke did not show that we can read metaphysical necessity off linguistic intuition. He showed that the necessity of "water is H2O" is discovered rather than stipulated, which makes metaphysical modality a substantive subject matter with its own difficult epistemology rather than a branch of semantics.
Essence and the direction of explanation
The standard modern gloss defines essence in modal terms: a property is essential to a thing if the thing has it in every world in which it exists. Kit Fine argued in "Essence and Modality" (1994) that this is backwards, with a counterexample that is easy to state and hard to answer.
Socrates is necessarily a member of the set whose sole member is Socrates. There is no world containing Socrates in which he is not in that set. Yet membership in a set is no part of what Socrates is; the set depends on him, not he on it. So there are modal truths about a thing that are not essential truths about it, and essence cannot be defined as necessity-of-property.
Fine's proposal reverses the order: essence is primitive, and metaphysical necessity is what follows from the essences of things. This matters downstream for Divine Simplicity and for Aseity, where the claim is not merely that God exists in every possible world but that God exists in virtue of what God is, which is a claim about essence and only derivatively about worlds.
How anyone knows a modal claim
The hardest problem in this area is epistemology, and it should be conceded rather than bluffed.
- Conceivability as defeasible evidence. The mainstream working principle, developed by Stephen Yablo and David Chalmers, is that what is coherently conceivable on reflection is evidence of what is metaphysically possible, defeasible by specific arguments for incoherence. Chalmers's two-dimensional framework distinguishes prima facie from ideal conceivability and primary from secondary intensions in order to explain why "water is XYZ" seems conceivable while being impossible.
- Williamson's counterfactual account. Modal knowledge is a byproduct of ordinary counterfactual reasoning, the same capacity used to judge that the vase would have broken if dropped. This has the advantage of making modal knowledge unmysterious and the disadvantage of extending less comfortably to exotic cases.
- The honest limit. Conceivability is not proof. Mathematicians have found conceivable propositions to be necessarily false. Any argument leaning on a possibility premise should defend it as a defeasible default and invite the opponent to supply a defeater, rather than treating conceivability as demonstration. Modal Ontological Argument handles this correctly in its second rebuttal to P1, and the practice should be copied wherever a possibility premise appears.
Equivocation traps
Five patterns, with the correction in each case.
- Epistemic read as metaphysical. The opponent says "sure, God's existence is possible," meaning "I have not ruled it out." An argument that needs metaphysical possibility has been handed nothing. The correction cuts both ways: a theist must not bank the concession, because the same reading licenses "possibly, no maximally great being exists," which by S5 delivers atheism just as fast.
- Logical read as metaphysical. "You cannot derive a contradiction from it, so it is possible." Absence of formal contradiction gives logical possibility only. "Water is XYZ" clears that bar and is still metaphysically impossible. What is wanted is a positive case that the described state of affairs is compatible with the natures of the things involved.
- Metaphysical read as nomological. "The laws of physics forbid a resurrection, so it is impossible." The laws settle nomological modality and say nothing about metaphysical modality, and the theist's claim was never that resurrection is a natural process. See Miracles.
- Nomological read as metaphysical. The mirror error, made by theists as often as by anyone: treating the fine-tuned constants as though a different value were unthinkable. The point of Fine-Tuning Argument depends on the constants being metaphysically contingent, which is what makes their values call for explanation.
- Conceivability read as demonstration. Dropping the defeasibility rider turns a reasonable evidential claim into an overreach that a competent opponent will punish. Say "defeasible evidence, absent a defeater," and then ask for the defeater.
Where this bites in the codex arguments
- Contingency Argument. The premise that the universe is contingent is metaphysical, not logical. Nobody claims "the universe does not exist" is a formal contradiction, so an objection demanding one is aimed at the wrong target. Symmetrically, the argument's Necessary Being is metaphysically necessary, existing in virtue of its own nature, not a logical truth. The page's P2-P3 block already argues this correctly; the vocabulary here is what makes the move explicit under pressure.
- Modal Ontological Argument. The whole argument lives or dies on P1 being a claim of metaphysical possibility. Trap 1 above is the standard failure. S5 is the right system for metaphysical modality, which is why "possibly necessarily p" yields "necessarily p"; an opponent's principled route of resistance is to deny S5 for this modality rather than to fight the later premises.
- Free Will Defense and the logical problem of evil. Mackie alleged a formal inconsistency in the theist's set of beliefs. Plantinga defeated it by supplying a merely possible auxiliary proposition, which is why plausibility was never required and why complaints that transworld depravity is implausible miss the structure of the reply. That defense is only intelligible if logical consistency and metaphysical plausibility are kept apart.
- Modal Argument from Mind. Runs on conceivability yielding metaphysical possibility, and it is the codex argument most exposed on modal epistemology, since Kripke's own apparatus was turned against exactly this style of argument in the philosophy of mind.
- Kalam Cosmological Argument. Distinguish the metaphysical impossibility of an actual infinite from its mere nomological or practical impossibility. The argument needs the stronger claim.
- Divine Simplicity and Aseity. Both are essence claims in Fine's sense before they are modal claims about worlds.
Objections to metaphysical modality itself
Quine. "Two Dogmas of Empiricism" and "Reference and Modality" hold that de re modality is unintelligible and that essentialism is invidious, since a thing counts as necessarily rational or necessarily bipedal only relative to how it is described. This was the dominant view for a generation and lost ground badly after Kripke supplied both a semantics and an account of rigid designation that dissolves the opacity puzzle. Treated at length in Quinean Rejection of Modal Logic Objection Defeater.
Modal deflationism. Amie Thomasson's modal normativism holds that modal claims express semantic rules in the indicative mood rather than describing a realm of modal fact. Ted Sider grants modal truths while denying that modality is metaphysically fundamental. These are serious contemporary positions and are not answered by pointing at Kripke, since neither denies the formal apparatus.
The honest response is that the theistic arguments here need modal realism about metaphysical modality only in a modest sense: that there are objective facts about what could and could not have been, independent of our conventions. That is compatible with several accounts of what grounds those facts, and it is a much lighter commitment than David Lewis's modal realism, on which possible worlds are concrete existing universes. Lewis's version is in tension with theism for independent reasons; the actualist framework of Alvin Plantinga, on which worlds are maximal consistent states of affairs and are abstract, is the standard theistic setting.
See also
- Modal Logic, the formal apparatus, operators, possible-world semantics, and the systems T, S4, S5
- Metaphysics, the parent domain hub
- Necessary vs Contingent Being, the distinction as applied to existence
- Contingency Argument, which stands on metaphysical contingency
- Modal Ontological Argument, which stands on metaphysical possibility
- Modal Argument from Mind, conceivability and metaphysical possibility in philosophy of mind
- Quinean Rejection of Modal Logic Objection Defeater, the standing objection to de re modality
- Principle of Sufficient Reason, the explanatory principle the contingency argument pairs with modality
- Free Will Defense, where logical consistency and plausibility must be kept apart
- Miracles, nomologically impossible and metaphysically possible
- Equivocation, the fallacy this page exists to prevent
- Laws of Logic, the ground of logical modality
- Aseity, existence in virtue of one's own nature
- Divine Simplicity, an essence claim before a modal one
- Meta-Ontology, the prior question of what existence questions are asking
- Alvin Plantinga, the actualist framework standard in theistic modal metaphysics
Common questions this page answers
Q: What is the difference between logical and metaphysical possibility?
Logical possibility is freedom from formal contradiction: if a description does not entail a contradiction, it is logically possible. Metaphysical possibility is compatibility with the natures of the things involved. "Water is XYZ rather than H2O" contains no contradiction, so it is logically possible, but it is metaphysically impossible, because nothing that is not H2O counts as water. Metaphysical possibility is therefore the narrower and more demanding notion.
Q: What does it mean to say God is a necessary being?
It means God exists in every metaphysically possible world, and does so in virtue of God's own nature rather than depending on anything else. This is a metaphysical claim, not the narrow logical claim that "God does not exist" is a formal contradiction. The traditional term for it is aseity, existence from oneself.
Q: Can something be necessary but only known through experience?
Yes, and demonstrating this was Kripke's main achievement. "Water is H2O" and "Hesperus is Phosphorus" were discovered empirically, so knowledge of them is a posteriori, yet both are necessarily true. Before Kripke, philosophers generally assumed necessary truths had to be knowable a priori. He showed that how we know something and whether it could have been otherwise are separate questions.
Q: What is a rigid designator?
A term that picks out the same object in every possible world where that object exists. Proper names are rigid: "Nixon" refers to the same man in every world. Most definite descriptions are not: "the winner of the 1968 US presidential election" could have picked out someone else. Rigid designation is what makes identity statements between names necessarily true when they are true at all.
Q: Are miracles impossible?
Not metaphysically. A miracle is nomologically impossible, meaning the laws of nature by themselves will not produce it, which is exactly what makes it a miracle rather than an unusual natural event. Nothing about the natures of the things involved rules out an agent acting from outside the system. An objection that a miracle "violates the laws of physics" establishes only nomological impossibility and leaves the theological claim untouched.
Q: Why does it matter which kind of possibility an argument uses?
Because arguments are valid only when the same modality runs through every premise. The modal ontological argument needs it to be metaphysically possible that a maximally great being exists; an atheist who concedes possibility in the everyday epistemic sense of "I have not ruled it out" has conceded nothing the argument can use. Silently swapping one modality for another is equivocation, and it is the most common way modal arguments fail.
Q: Is conceivability a reliable guide to possibility?
Only as defeasible evidence. The mainstream position is that what remains coherently conceivable under careful reflection counts as evidence of metaphysical possibility, but the evidence can be defeated by a specific argument that the described state of affairs is incoherent. Propositions that seemed conceivable have turned out to be necessarily false, particularly in mathematics, so conceivability should never be presented as demonstration.
Q: Did Quine refute modal logic?
He argued that de re modality, necessity attaching to things rather than to sentences, is unintelligible, and that essentialism depends improperly on how an object is described. The position was influential through the 1950s and 1960s and lost the mainstream after Kripke supplied a formal semantics and an account of rigid designation that answers the description-relativity worry. Modal deflationists today make subtler arguments that do not depend on Quine's.