Why "God Exists" is an Unintelligible Statement

This is deliberately a provocative title. I am not trying to prove atheism here, and I am not claiming that every sophisticated theological system is literally incapable of constructing a coherent proposition involving a deity. My target is the ordinary assertion “God exists” as it is commonly presented in religious discussion: as though it were a straightforward, self-evident factual statement whose meaning is already settled and whose denial therefore requires a defense.

My claim is harsher than “I am unconvinced.” I often do not know what proposition I am being asked to believe. The terms God and exists shift across speakers, traditions, and even stages of the same conversation. The subject is described as personal and impersonal, temporal and timeless, present everywhere and nowhere locatable, causally active but outside causation, a being and the “ground of being.” Meanwhile, exists can mean ordinary spatiotemporal existence, abstract existence, necessary existence, metaphysical reality, psychological presence, or something closer to “is ultimately real.” These are not minor variations around one fixed proposition. They can amount to different propositions entirely.

So my objection is semantic before it is metaphysical. Before I can ask whether “God exists” is true, I need enough stability to know what would make it true, what would make it false, what “God” identifies, what kind of existence is being asserted, and what follows from the assertion. If the key terms remain fluid, or if every attempt to clarify them simply moves the claim into a new semantic register, then the sentence may retain grammatical form and rhetorical force while failing to function as a well-formed assertion in critical discussion.

This position is close to ignosticism: the demand that the concept of God be made intelligible before belief, disbelief, proof, or disproof can even get started. What follows is an attempt to explain why that demand matters, why “God exists” often fails it, why the phrase can nevertheless feel perfectly meaningful inside a religious community, and why conversations about it so easily become frustrating and evasive.

1. What Would Make “God Exists” a Meaningful Claim?

It is easy to treat meaning as though words simply contain it. They do not. Words contribute meaning, but what an utterance communicates also depends on reference, context, presupposition, pragmatic purpose, grammatical structure, and the inferential commitments the speaker is prepared to accept. For this essay I do not need a complete course in semantics and pragmatics. I need only a few distinctions that directly bear on the phrase “God exists.”

The first is the distinction between semantics and pragmatics. Semantics concerns the relatively stable content contributed by words and structures: what an expression refers to, what conditions would make a proposition true, what it entails, and what it presupposes. Pragmatics concerns what an utterance is doing in a particular conversation: signaling commitment, invoking common ground, shifting a burden, expressing reverence, coordinating a group, or performing some other social act.

That distinction matters because a sentence can be pragmatically successful while remaining semantically unstable. A slogan can unite a community even if its literal content is disputed. A ritual formula can be deeply meaningful to participants without functioning like an empirical report. Conversely, a sentence can be grammatically ordinary yet fail to produce a determinate proposition if its key terms do not have stable application conditions.

The second distinction is between sense, reference, and truth conditions. Roughly, reference concerns what an expression picks out; sense concerns the mode under which it is presented; and truth conditions concern what would have to be the case for the resulting proposition to be true. If I say “the current president,” the phrase can refer only after a context supplies a relevant country and time. If I say “the tallest person in this room,” the description gives criteria that in principle determine a referent. If no one fits the description, we know what has failed.

The third issue is presupposition and inferential role. Assertions do not merely attach labels to reality. They commit speakers to consequences. “She stopped smoking” presupposes that she used to smoke. “He killed the king” entails that the king died. To understand a predicate is partly to understand what follows from applying it, what would conflict with it, and what circumstances would license or defeat its use.

This matters enormously for theological predicates. If “God knows,” “God acts,” or “God is present” are asserted as ordinary propositions, we naturally inherit at least some of the inferential roles associated with knowing, acting, and being present. We expect there to be something known, some distinction between action and inaction, and some intelligible contrast between presence and absence. The more those contrasts are denied, the less obvious it becomes that the original predicates are still functioning normally. A word can survive syntactically while its ordinary inferential structure has disappeared.

Presupposition also helps explain why “God exists” can seem effortless in one setting and bizarre in another. In a church, prayer group, or theological discussion among believers, the referent of “God” may be treated as part of the common ground before anybody speaks. The sentence therefore arrives after a large amount of semantic work has already been socially performed. Outside that setting, the hearer may not grant any of those presuppositions. What looks like needless pedantry to the insider can therefore be the outsider's first opportunity to ask questions the community answered implicitly rather than explicitly.

Finally, tense, modality, and indexicality matter because theological language constantly trades on them. Words such as now, present, responds, acts, knows, and creates carry temporal or perspectival structure in ordinary usage. If God is then declared timeless, changeless, or outside spacetime, the speaker has to explain which parts of that ordinary structure survive and which do not. Merely keeping the same words does not guarantee that the same concepts are still in play.

So the relevant meaning test is not complicated. When somebody asserts “God exists,” I want to know: What does God refer to, or what description determines its extension? What sense of exists is being used? What identity conditions tell us whether we are speaking about the same entity across contexts? Which predicates are literal and which are analogical? What follows if the claim is true? What would conflict with it? What would count against it? These are not exotic demands. They are the minimum structure required for an assertion to become assessable rather than merely repeatable.

2. The Semantic Problem with “God Exists”

What Does “God” Pick Out?

The word God looks deceptively simple. Grammatically, it often behaves like a proper name. In ordinary conversation people say “God knows,” “God wants,” “God created,” or “God exists” in much the same syntactic form as they might say “Alice knows” or “John arrived.” That surface form encourages the impression that the reference has already been fixed: there is some particular entity called God, and the only remaining question is which predicates apply to it.

But once the term is challenged, the apparent simplicity disappears. God may mean a personal creator, a tri-omni being, a deistic first cause, the ground of being, a maximally great being, a universal consciousness, an impersonal absolute, a higher power, a sustaining cause, or simply the particular deity presupposed by a tradition. These descriptions are not interchangeable. They make different claims, carry different commitments, and can generate different answers to questions about personality, causation, time, morality, uniqueness, and intervention.

This creates a basic reference-fixing problem. If God is functioning as a proper name, what fixes its referent? There is no uncontested causal-historical chain comparable to ordinary naming, no agreed observational encounter, and no stable public identity criterion shared across traditions. If it is functioning as a description, then which description? If it is a cluster concept, how much of the cluster is necessary before something counts as God? Could two beings satisfy it? Could an impersonal ground and a personal creator both satisfy the same term? If the answer changes from speaker to speaker, then two people can utter the same sentence while asserting materially different propositions.

That is already enough to complicate the slogan “God exists.” Before existence can be assessed, the subject of the existential claim has to be sufficiently determinate to enter the assessment. Otherwise the sentence is not one proposition awaiting a yes-or-no verdict. It is a syntactic shell into which different ontologies can be loaded.

The ambiguity is not merely denominational. Even within a single tradition, the term can shift according to conversational need. In devotional language God may be intensely person-like: a father who hears, loves, grieves, commands, forgives, and responds. In philosophical theology the same God may become immutable, timeless, simple, impassible, non-spatial, and categorically unlike any creature. Those may be parts of a carefully worked-out doctrine, but in ordinary conversation the bridge between them is often assumed rather than explained. The name remains constant while the descriptive content around it changes radically.

This is why asking “which God?” is not a cheap rhetorical move. It is a request to stabilize the predicate before quantification. “A creator exists,” “a necessary ground exists,” “a personal moral governor exists,” and “an impersonal absolute exists” may all be interesting claims, but success for one does not automatically establish the others. If the word God is allowed to inherit the evidential support of whichever description is easiest to defend and then transfer that support to all the other descriptions, the argument has equivocated before it has begun.

Russell and the Illusion of Reference

Bertrand Russell’s theory of descriptions makes this problem much sharper. A phrase such as “the creator” looks as though it refers directly to an object. Russell’s analysis shows that definite descriptions of the form “the F” can instead be represented as quantified claims. A sentence of the form “The F is G” can be analyzed, roughly, as saying that there exists an F, there is only one F, and that unique F is G:

$$ \exists x \big(F(x) \wedge \forall y(F(y) \rightarrow y=x) \wedge G(x)\big) $$

This matters because grammatical smoothness does not guarantee successful reference. “The present King of France is bald” is perfectly intelligible English even though there is no present King of France. On a Russellian analysis, the sentence can be false because its existential condition fails; we do not have to posit some mysterious non-existent object to which the description nevertheless refers.

Now replace the present King of France with the creator of the universe. If “God” is shorthand for a definite description such as “the unique creator of the universe,” the existence claim becomes something like:

$$ \exists x \big(Creator(x) \wedge \forall y(Creator(y) \rightarrow y=x)\big) $$

Immediately the hidden commitments become visible. What domain is being created? What counts as creating? Does creation imply temporal priority, causal production, dependence, or something else? Why exactly one creator? Could multiple entities satisfy the relevant description? If the alleged creator is timeless, what does the predicate create mean when ordinary creation involves a transition from a state in which something does not exist to a state in which it does?

If “God” instead abbreviates “the tri-omni creator,” “the necessary ground of being,” or “the maximally great being,” then the existential claim inherits an entirely different conjunction of predicates. Different theists may therefore say “God exists” while silently quantifying over different property bundles. The same surface sentence does not guarantee the same logical content.

Strawson famously objected to parts of Russell’s treatment of failed descriptions, emphasizing presupposition rather than simple falsity in some cases. That disagreement does not rescue the self-evidence of “God exists.” If Russell is right, a description can fail because its existential or uniqueness conditions are unsatisfied. If a Strawsonian analysis is preferred, the utterance may fail to receive an ordinary truth value when the relevant presuppositions are not secured. Either way, grammatical “aboutness” does not magically provide a referent.

What Does “Exists” Mean?

The second unstable term is exists. In formal semantics, existence is not simply an ordinary property like redness. An existential sentence is usually represented through quantification: “Unicorns exist” becomes something like $\exists x\, Unicorn(x)$. “God exists” can therefore be represented schematically as $\exists x\, God(x)$—there is at least one entity in the relevant domain that satisfies the predicate God.

That formal shell is thin. To interpret it, we still need a domain and criteria governing the predicate. In ordinary English, existential talk also shifts across contexts. Physical objects exist. Numbers are often said to exist. Characters exist in stories. Institutions exist. Possible worlds are invoked in modal semantics. A theorem can establish that some mathematical structure exists without locating it anywhere in space. So the problem is not that every meaningful use of exists must reduce to physical detection.

The problem is that the sense has to be fixed strongly enough for the claim to have discriminatory power. If “God exists” means physical or causal existence, then ordinary empirical and causal questions become relevant. If it means abstract existence, the comparison class shifts toward numbers or propositions, and then agency predicates become much harder to interpret. If it means necessary existence—existence in all possible worlds—then a modal framework is being invoked and must be defended. If it means “ground of being,” then we need to know what positive content that phrase contributes and what it excludes.

“Outside space and time” is especially problematic because it is usually offered as clarification while mostly supplying negations. Not spatial. Not temporal. Not material. Perhaps not causally constrained in the ordinary way. But what positive application conditions remain? What would it mean for something of that kind not to exist? What identifies one such entity rather than two? What differentiates a real non-spatiotemporal being from an empty description with no satisfier?

Once exists is detached from enough of the contrasts and inferential expectations that make existential claims useful, it risks becoming an honorific: “God is maximally real,” “God is ontologically ultimate,” “God cannot fail to be.” Those may be meaningful technical theses inside a specified metaphysical system, but they cannot simply inherit the clarity of ordinary existential grammar. The speaker has changed the semantic job of the word and now owes an account of the new one.

This is also why I resist the reply that existence is simply “being in the ontological inventory.” That phrase may be harmless inside a formal ontology, but an inventory still requires rules of membership. What kinds of things populate the domain? What makes two entries identical or distinct? What predicates are permitted of them? What is the difference between asserting an object in the ontology and introducing a symbol into the language? Without such rules, “ontological inventory” merely redescribes the question in more impressive vocabulary.

The modal version has the same issue. “God exists necessarily” sounds stronger than “God exists,” but necessity does not fix an unclear concept. To say that some x exists in every possible world first requires a sufficiently stable x-description to track across those worlds and a modal system in which the quantification is meaningful. Otherwise necessity can amplify an ambiguity rather than resolve it. A vague entity does not become clearer by being declared necessary.

Timelessness, Agency, and Inferential Breakdown

The tension becomes sharper when the metaphysical description of God is combined with ordinary agency language. God is said to create, choose, intend, know, respond, intervene, love, judge, answer prayer, and have purposes. In ordinary language these verbs are not empty labels. Their meanings are partly constituted by temporal, causal, informational, and counterfactual relations.

Create ordinarily suggests some kind of dependence and often a before-and-after structure: something comes to be, or at least depends for its being on an act or cause. Respond suggests an input and a subsequent response. Choose normally distinguishes alternatives. Intervene implies a process into which something enters or alters a course. Know that it is now t contains an indexical tied to temporal perspective. Even where philosophers construct non-temporal analyses of some of these predicates, the burden is precisely to explain what inferential structure replaces the ordinary one.

So the phrase “timeless creator” is not clarified merely by saying that timelessness is mysterious. There is a real semantic pressure. Either theological predicates retain much of their ordinary inferential role, in which case timelessness and changelessness strain against them, or their meanings are revised so heavily that the familiar verbs become analogical placeholders. In the latter case the speaker must tell us which inferences remain valid. Otherwise the sentence continues to borrow the force of ordinary agency language after evacuating the conditions that made the language intelligible.

The present tense in “God exists” adds another small but revealing tension. English present tense can express an episodic present (“he is here now”), a generic truth (“water boils at 100°C”), or something treated as timeless (“triangles have three sides”). A theist who means necessary or atemporal existence is asking the hearer to reinterpret the surface tense accordingly. That is perfectly possible as a pragmatic adjustment. But surrounding religious language often shifts straight back to temporal indexicals: God is “with us now,” “answers,” “acts,” “enters history,” or “intervenes.” The issue is not any single metaphor. It is the uncontrolled movement among semantic registers while the surface vocabulary remains unchanged.

3. How the Claim Gets Repaired

Once these problems are pressed, the conversation often moves into repair mode. The initial claim sounded ordinary: “God exists.” The challenge asks what the sentence means. The response then introduces a more specialized vocabulary: God is outside spacetime, a necessary being, the ground of being, pure actuality, a non-physical mind, Being itself, or something that does not “exist” in the same sense creatures exist.

A clarification is not automatically an evasion. A serious metaphysical position may genuinely require technical concepts. The issue is whether the repair produces greater semantic discipline or merely moves the target beyond the pressure being applied. A successful clarification should fix the sense, state its inferential commitments, explain what would conflict with it, and remain stable when the conversation turns to evidence, agency, causation, or reference.

In practice, many repairs fail in one of two directions. The first is vagueness. “Ground of being,” “outside space and time,” or “higher power” may sound philosophically substantial while remaining too indeterminate to decide what satisfies the description. If I cannot state what would count as an instance, what would count as a non-instance, or what consequences follow from applying the concept, then the phrase has not yet repaired the original problem.

The second failure is stipulation or arbitrariness. A definition can become suspiciously tailored to preserve the conclusion: when empirical constraints become inconvenient, existence becomes non-empirical; when causal efficacy is demanded, God becomes active in the world; when temporal tensions appear, action becomes timeless; when literal meaning is pressed, the predicates become analogical. Each move may have a theological rationale in isolation, but if the governing criterion changes primarily in response to whatever would otherwise count against the claim, the result is not clarification. It is semantic immunization.

This is why “it is metaphorical” does not automatically rescue the assertion. Metaphor is not semantic anarchy. In familiar conceptual metaphors such as ARGUMENT IS WAR or TIME IS MONEY, we can identify a source domain, a target domain, and the relational structure being projected. We know why “defend your position” works and why some other inference would be a misuse of the metaphor.

If “God” or theological agency language is metaphorical, then the same discipline is required. Metaphorical of what? Which human or experiential domain supplies the structure? What is the target? Which inferences are licensed? Which are not? If “God loves” is analogical, what remains of ordinary love and what is explicitly discarded? If “God acts” is not action in any temporal or causal sense, what positive structure justifies retaining the verb?

Often the direction of projection is actually the reverse of what the defense suggests. We take familiar human predicates—father, king, mind, designer, judge, shepherd, agent—and project them onto an initially obscure posit. The posit then inherits enough human structure to become imaginatively vivid, after which the resulting picture is treated as though it independently anchored the vocabulary. That may be psychologically or religiously powerful, but it does not by itself provide a stable referent or a determinate existential proposition.

The semantic challenge therefore remains simple: if the repaired claim has a stable ontology, explicit analogy rules, identity conditions, inferential constraints, and recognizable defeaters, then we finally have something to argue about. If every repair instead strips away another ordinary application condition while preserving the original assertion’s force, the sentence becomes harder—not easier—to understand.

There is an important fork here. One possibility is that the theological vocabulary is genuinely technical. If so, it should be treated like technical vocabulary: defined with enough care that participants know what follows from it. The other possibility is that the vocabulary is primarily expressive, analogical, or apophatic—designed less to describe an entity than to gesture toward what cannot be captured in ordinary concepts. That can have religious value, but it weakens the right to insist that the utterance is simultaneously a straightforward factual claim. You cannot indefinitely move between “this is literally true” and “human concepts cannot apply” without explaining what survives the move.

Apophatic language makes the tension especially clear. Saying what God is not can protect a tradition from crude anthropomorphism: not finite, not material, not temporal, not composite. But a growing list of negations does not automatically yield a positive object of reference. If every familiar category is denied and the remaining positive predicates are declared analogical beyond ordinary comprehension, the sentence “God exists” can end up with enormous rhetorical force and remarkably little publicly recoverable content. At some point mystery stops being an explanation of the proposition and becomes part of the problem of identifying it.

4. Why the Phrase Still Works

Pragmatic Function

None of the semantic criticism implies that the utterance “God exists” is useless. In many contexts it is highly effective. The key is that its pragmatic function can be much more stable than its literal truth conditions.

Saying “God exists” can mark identity: I am one of us. It can set a framework: we are now reasoning inside a theistic picture. It can authorize norms: certain moral obligations are not merely preferences. It can express dependence, trust, reverence, gratitude, hope, or submission. It can organize narrative: suffering has a frame, life has a purpose, history is going somewhere. It can coordinate testimony by licensing scripture, church, tradition, or spiritual experience as sources within the discourse.

In those settings, treating “God exists” as though it were merely a primitive empirical hypothesis can miss what the utterance is doing. It may be closer to a framework declaration or commitment marker than to a laboratory claim. That does not make it true. It explains why it remains socially and psychologically meaningful even when an outsider cannot extract a stable proposition from it.

This also explains why disagreements can become pseudo-disagreements. One participant hears “God exists” as a factual assertion about reality and asks for reference, truth conditions, and defeaters. The other is partly performing allegiance, worldview orientation, or testimony. Both think they are debating the same sentence, but they are participating in different conversational games. The grammatical posture is assertoric; the pragmatic job may be devotional, social, or normative.

The distinction also clarifies why the phrase can shift burdens so effectively. Once “God exists” is installed as conversational common ground, later statements can presuppose it rather than defend it: God wants this, God forbids that, God designed humanity for a purpose, God explains why anything exists at all. The existential commitment disappears into the background while downstream claims accumulate on top of it. A dissenter who questions the foundation can then appear to be derailing the discussion because everyone else is operating several inferential steps later. Pragmatically, the framework is already running; semantically, the outsider is still asking whether the initial symbol ever received a stable interpretation.

There is nothing uniquely religious about this mechanism. Communities constantly coordinate around presupposed concepts, specialized vocabularies, and background norms. The difference is that “God exists” is frequently asked to perform two jobs at once: internal framework axiom and external ontological report. The first job can succeed through shared practice. The second requires a publicly assessable proposition. Confusing the success of the first for success of the second is one of the recurrent problems in the debate.

The Taken-for-Granted Background

In many religious subcultures, “God exists” is not first encountered as a hypothesis. It appears in prayer, stories, ritual, moral instruction, family life, music, communal identity, and ordinary explanations long before it is isolated as a proposition requiring defense. It becomes part of the background against which other claims are interpreted.

That background can stabilize meaning pragmatically without stabilizing it philosophically. Within the community, a cluster of concepts—creator, father, judge, loving presence, ground of meaning, sovereign, savior, higher power—may be “good enough” because participants share the stories and practices that tell them roughly how the term is used. They do not have to stop every time the word God appears and specify a necessary-and-sufficient definition. The surrounding form of life supplies the missing context.

From the inside, this feels like shared understanding. From the outside, it can look like semantic slipperiness. The outsider does not automatically inherit the stories, presuppositions, authority structures, or interpretive habits that tell the insider how to repair ambiguity. So when the phrase is transported from devotional life into a critical discussion and presented as “obvious,” the community’s internal common ground is being imported into a context where it is no longer common.

This is one reason “God exists” can feel self-evident without having been established as self-evident. What has become cognitively and socially default is mistaken for what has become semantically and epistemically unavoidable. Repetition and presupposition can make a proposition feel like the starting point of thought even when an outsider is still asking what proposition has been stated.

Why It Can Feel Obvious

The original draft explored several psychological mechanisms that may help explain why sparse theological specifications can nevertheless feel richly meaningful. I think those observations fit better here, as an explanation of the insider/outsider gap rather than as a separate argument against theism.

Familiarity and cognitive fluency matter: repeatedly heard claims become easier to process, and ease of processing can contribute to a feeling of obviousness. Human cognition is also highly receptive to agency and purpose-based explanations. Once a situation is interpreted in agentive terms, events can be read as intentions, messages, responses, or plans. Teleological explanations likewise make “what is this for?” feel natural even where a purely mechanistic description would be possible.

Minimally counterintuitive religious concepts may also be memorable precisely because they combine familiar human structure with a small number of extraordinary properties: an agent who is invisible, immortal, omniscient, or otherwise unlike ordinary agents. Add communal repetition, costly commitment, shared ritual, and common-ground management, and an expression can acquire extraordinary pragmatic density without becoming correspondingly precise in truth-conditional terms.

The point is not “religious people are irrational.” The point is almost the opposite: ordinary human cognition is extremely good at reconstructing meaning from sparse social context. That is usually an advantage. A community can communicate efficiently because participants fill in enormous amounts of unstated background. The difficulty appears when that socially reconstructed meaning is treated as though it were a publicly specified proposition that should be equally transparent to someone who does not share the background.

There is also a relevance effect. Hearers generally assume that a speaker is trying to communicate something worth processing, so they search for an interpretation that makes the utterance coherent and relevant. The vaguer and more socially important the phrase, the more interpretive labor a sympathetic listener may contribute. In religious contexts that labor is guided by sermons, stories, symbols, and shared expectations. What feels like content contained in the sentence may therefore be content supplied by the interpretive environment.

This helps explain why the same phrase can feel profound to one person and empty to another without either person merely failing to hear the words. The believer may have decades of associations compressed into the token God; the outsider may hear an unresolved noun. The believer may understand exists through an entire metaphysical and devotional framework; the outsider may search for ordinary existence conditions and find them repeatedly withdrawn. The disagreement is partly over reality, but it is also over how much hidden structure the sentence is entitled to borrow.

The Map and the Territory

The map-versus-territory distinction ties the pragmatic and semantic sides together. Religious traditions do not merely hand people isolated propositions. They provide narratives, symbols, roles, moral expectations, rituals, identities, hopes, explanations, and practices. These structures function like maps: systems of representation and orientation through which people make sense of life.

Within such a map, “God exists” can function as a structural axiom. It organizes the rest of the system. Prayer presupposes an addressee. Providence presupposes an agent or order capable of directing events. Divine command presupposes authority. Salvation narratives presuppose the relevant divine roles. In that sense, the God-concept “exists” very robustly inside the interpretive system: it is a stable node around which enormous amounts of practice and narrative are organized.

But map-existence is not automatically territory-existence. A concept can be indispensable to a framework without the framework thereby proving that the corresponding entity exists independently of it. The key slide is this:

God is indispensable within our narrative and practical framework; therefore God exists as an entity in reality.

That conclusion requires bridging premises. Perhaps the argument is that the best explanation for the map’s success is that its central posit refers. Perhaps the claim is that certain experiences are best explained by a real divine agent. Perhaps a metaphysical argument is supposed to connect the structure of reality to a necessary being. Fine—but those are arguments. The structural necessity of God within a religious map cannot substitute for them.

This framework also explains a familiar oscillation. When challenged as a territory-level posit, God becomes difficult to locate: transcendent, outside nature, outside spacetime, not an object among objects. When the religion describes lived consequences, God becomes richly agentive: present, loving, commanding, answering, judging, intervening. The same symbol moves between the map’s narrative role and an asserted external referent. Unless the transition is explicitly defended, the map can lend vividness and apparent obviousness to a territory claim it has not actually established.

A simple illustration is an annotated tourist map. The map may contain a smiling bear, a stylized redwood, a little sun hovering over a beach, and exaggerated landmarks. Those features can be useful and meaningful without corresponding literally to objects in the territory in the form depicted. A competent map user understands the representational convention. The mistake would be to arrive in California and ask where the giant smiling bear from the map is located. The representation was doing orientation-work, not straightforward object-reporting.

Religious symbols are vastly richer than cartoon icons, but the logical distinction is the same. A concept can organize a representation without its representational role settling what kind of external referent it has. “God” can be indispensable to a story about creation, sin, obligation, suffering, redemption, or cosmic purpose and still require a separate argument before we infer a corresponding entity with determinate identity conditions. The map gives us reasons to understand how the symbol functions; it does not by itself give us the ontology of the territory.

None of this means maps are “mere” or that myths cannot encode truths. Maps can correspond to territory; narratives can disclose patterns that flat description misses. The point is narrower and more damaging: correspondence cannot simply be assumed because the map is meaningful, useful, ancient, emotionally powerful, or structurally coherent.

5. What Goes Wrong in Argument

The semantic instability becomes most visible when “God exists” leaves the devotional or communal setting and enters a critical discussion. In that setting an assertion takes on responsibilities. The speaker is not merely expressing allegiance; the speaker is presenting something as true and therefore answerable to questions about meaning, reasons, defeaters, and consistency.

This is where calling the proposition “self-evident” becomes particularly problematic. Self-evidence cannot be a substitute for specification. Before a claim can be trivially true, obvious, properly basic, or universally known, interlocutors at least have to know what claim is being assigned that status. If God still floats among creator, higher power, tri-omni person, ground of being, and impersonal absolute, while exists floats among physical, abstract, necessary, and sui generis metaphysical existence, then “self-evident” is doing rhetorical work before semantic work has been completed.

A Walton-style way of putting the point is that assertions in a critical discussion are defeasible moves governed by burdens and critical questions. If the proponent presents “God exists” as presumptively established, the opponent need not immediately accept the role “disprove God.” A prior challenge is legitimate: fix the verbal criterion, identify the referent or defining description, specify the relevant domain, and tell us what would count for or against the classification.

When those requests are repeatedly answered by changing the criteria, the interaction begins to feel like a moving target. Ask where God is, and God is beyond location. Ask how God acts, and God acts in the world. Ask how timeless agency works, and “act” becomes analogical. Ask what evidence would count against the claim, and the claim becomes beyond empirical adjudication. Ask what difference God makes, and concrete intervention returns. The assertion becomes maximally ordinary when it wants practical force and maximally extraordinary when ordinary standards threaten it.

This is the phenomenon I originally described as feeling like gaslighting. I still think the word captures something about the experience, even though it should not be confused with a clinical or moral accusation that every theist is deliberately manipulating an interlocutor. The conversational shape can resemble gaslighting: the listener asks for stable criteria, receives successive “clarifications” that do not remain stable, notices the inconsistency, and is then treated as irrational, spiritually deficient, overly literal, or incapable of understanding something supposedly obvious.

Philosophically, the sharper description is assertoric pressure plus semantic evasiveness plus blame reversal. The speaker wants the force of a factual assertion without consistently accepting the responsibilities that make factual assertions answerable. The problem is not that every theological revision is dishonest. The problem is that uncontrolled shifts among literal, analogical, metaphysical, devotional, and psychological registers make it impossible to know which proposition survives the entire conversation.

A useful way to see the argumentative structure is as a two-stage challenge. First comes vagueness: the verbal criterion is too underspecified to support the classification “God is an existent.” The appropriate response is to withhold the classification until the criteria are fixed. If the proponent then supplies a succession of increasingly specialized definitions, a second challenge becomes available: arbitrariness. Why this criterion? Was it chosen because it captures a stable concept with explanatory value, or because it blocks the most recent objection? A definition that changes mainly to preserve the conclusion is not doing ordinary clarificatory work.

This is also where burden shifting enters. “God exists” is sometimes asserted as though it starts with presumptive status and the skeptic must now produce a disproof. But if the sentence is still semantically unstable, the burden has been assigned too early. The skeptic cannot be obligated to refute every possible referent and every possible sense of existence that might later be attached to the words. The proponent first has to say which proposition is on the table. Otherwise the target expands whenever a particular formulation becomes vulnerable.

Consider “God is everywhere.” It can mean at least three different things. It can mean metaphysical sustaining: everything depends on God. It can mean spatial omnipresence: God is literally present at every location. Or it can mean psychological or normative presence: God is experienced in conscience, community, or orientation. All three uses can be meaningful. The problem begins when one sense supplies the emotional and rhetorical force while another supplies immunity from criticism.

The same applies to literal versus analogical predication. “God loves,” “God responds,” “God enters history,” and “God is present with you” sound richly personal and temporally structured. “God is immutable, timeless, impassible, and beyond creaturely categories” sounds metaphysically insulated. A coherent theology may have a theory connecting these registers. But ordinary conversation often uses the registers opportunistically without stating the rules. That is not merely mysterious. It is a semantic failure in the dialogue.

6. A Precision-First Analogy: Software Engineering

Part of why this bothers me so much is professional deformation. I work in a field where ambiguity eventually breaks something. Software engineering trains you to care about bindings, types, interfaces, schemas, contracts, invariants, namespaces, data models, protocols, and boundary conditions. You can build extremely abstract systems, but abstraction does not eliminate specification. It makes specification more important.

In a programming language, a grammatical expression is not automatically meaningful. Symbols have to bind. Operators have signatures. Types constrain legal operations. Evaluation happens in some environment. Interfaces specify what interactions are possible. A function call such as exists(God) is not informative until we know what God denotes and which implementation of exists is being invoked.

Natural language is obviously not source code, and I am not arguing that every human statement needs a formal specification. The analogy is diagnostic, not reductive. It makes visible exactly what frustrates me about “God exists”: the sentence is often presented as a clean Boolean while both arguments are effectively unresolved.

God behaves like an identifier whose binding changes by namespace: God@classical_theism, God@trinitarian_christianity, God@deism, God@pantheism, God@higher_power. The token is shared, but the underlying object model is not. Meanwhile, exists behaves like an overloaded function whose implementation is chosen after the test begins: physical existence under one question, necessary existence under another, abstract existence under another, “real in my life” under another.

The timeless-agency problem then looks like an interface mismatch. The system declares that the object is outside time and change while exposing methods whose ordinary semantics involve temporal sequence and causal interaction: create(), respond(), intervene(), choose(). Theology may reply that these methods are analogical. Fine—but then the interface documentation matters. What are the preconditions? What behavior is preserved? Which ordinary inferences are invalid? What counts as an exception or contradiction?

The important point is that abstractness is not the problem. Computing and mathematics are full of non-physical objects—types, numbers, graphs, protocols, formal languages, complexity classes—and claims about them can be perfectly determinate because their identity conditions and inferential rules are explicit. I do not reject God-talk because God is “not measurable.” I reject the expectation that an expression should count as a determinate assertion while its referent, operator, constraints, and failure conditions remain unsettled.

From this precision-first perspective, the recurring failures are easy to name: unbound identifier (“God” lacks stable reference), overloaded operator (“exists” changes sense), missing specification (truth conditions and defeaters are unclear), interface mismatch (timelessness combined with time-structured agency predicates), and non-portable semantics (the phrase acquires much of its meaning from a community-specific runtime). That does not prove that no God exists. It explains why the bare claim so often fails to compile as a proposition.

Knowledge graphs provide another useful analogy. You cannot meaningfully assert arbitrary relations among entities without an ontology that defines entity classes, relation types, constraints, and permitted inferences. An edge labeled acts_on means something only because the schema tells us what kinds of nodes can participate and what the relation licenses. “God acts on the world” is often asserted without an equivalent shared ontology. Is God an agent, an abstract ground, a substance, a mind, a relation, or something outside every category? What kind of edge is acts on? Does it encode causation, dependence, intention, or metaphorical significance? If the ontology changes with each query, the graph has stopped representing a stable model.

Requirements engineering is perhaps the closest analogy. A requirement that cannot be tested because every failure is reclassified as acceptable is not a robust requirement. A specification that says a system must be “perfectly secure” but supplies no threat model, success criteria, scope, or failure conditions sounds strong while being practically useless. “God exists” often strikes me in the same way: enormous apparent content, insufficient specification. The response “you know what I mean” may work socially, but it is exactly what a precision discipline trains you not to accept when correctness matters.

7. A Practical Meaning Challenge

The practical consequence of all this is not “therefore atheism.” It is: do not grant the proposition before it has been specified. When someone says “God exists,” I do not need to begin by offering a theory of the universe without God. I can ask what exactly has been asserted.

A useful critical discussion can be organized around a small set of questions rather than an endless list of theological objections:

Reference: What do you mean by God here? Give the defining description or criteria that distinguish this concept from alternatives.
Existence: What sense of exists are you using—spatiotemporal, abstract, necessary, or something else? What are its positive application conditions?
Identity: What makes this God the same entity across situations or possible worlds? Could two beings satisfy your definition?
Predication: Which claims about God are literal and which are analogical or metaphorical? If analogical, what inferences are licensed?
Agency: If God is outside time and space, what do create, act, respond, intervene, and know mean without simply importing their ordinary temporal and causal structure?
Consequences: What does “God exists” rule out? Name something that would be different, inconsistent, or impossible if the claim were true.
Defeaters: What could count against the claim? If nothing could, why should the utterance be treated as an ordinary assertion rather than a framework commitment?

These questions do not rig the debate against theism. A sophisticated theist can answer them by supplying an ontology, a theory of predication, a modal framework, an account of divine action, and criteria for rational assessment. Classical theologies have tried to do precisely that. But once those answers are supplied, we are no longer debating the bare slogan “God exists.” We are debating a much richer and much more specific metaphysical system.

Two common replies illustrate how the challenge works. If the response is “God is the necessary ground of being, so ordinary spatiotemporal existence is the wrong category,” I can provisionally grant the redescription and ask what follows from it. What does ground mean here? What does it explain that an alternative ontology does not? What would be incompatible with such a ground? How do personal predicates such as willing, knowing, commanding, or loving attach to it? If there are stable answers, the conversation advances. If “ground of being” merely means “whatever ultimately explains everything” and no conceivable state of affairs counts against it, the phrase is functioning more like a placeholder for explanatory ultimacy than a determinate entity-description.

If the response is instead “God's existence is self-evident; everyone knows it,” the challenge moves to common ground and burden. Self-evident to whom, and in what sense? Psychologically compelling? Culturally familiar? Impossible to deny once the terms are understood? Those are different claims. Radical disagreement among sincere speakers about what God is already weakens the idea that the bare sentence carries one universally transparent content. A claim cannot gain precision merely by being pronounced obvious.

That distinction is the point. I object to the way the unspecific slogan is often granted the prestige of the sophisticated theory without carrying its conceptual machinery. The ordinary hearer is expected to assent to two words, while any challenge is answered by importing whichever technical apparatus is useful at that moment. If the technical apparatus is necessary to make the proposition intelligible, then state it up front and let that—not the slogan—be the object of debate.

Until then, the proper response is not disbelief in a clearly specified proposition. It is refusal to pretend that specification has already occurred.

8. Conclusion: Clarification Before Assent or Denial

I began with a deliberately provocative claim: “God exists” is an unintelligible statement. I mean that seriously, but not simplistically. I am not claiming that the English sentence contains no recognizable words or that nobody ever understands anything by it. I am claiming that, as a purported stand-alone factual assertion in ordinary critical discussion, it often fails to provide a stable enough semantic object for rational evaluation.

“God” may function as a name without an agreed reference-fixing story, a description whose predicates differ across speakers, or a culturally stabilized cluster whose boundaries remain vague. “Exists” may shift among physical, abstract, modal, metaphysical, psychological, or honorific senses. The traditional attributes attached to God can then strain against one another: timelessness and immutability sit uneasily beside acting, responding, creating, and intervening unless an explicit theory of analogical predication is supplied.

When challenged, the claim is frequently repaired. Sometimes that repair really does produce a more sophisticated and coherent theory. But often the repair simply moves the criterion: ordinary existence becomes non-spatiotemporal existence; ordinary action becomes analogical action; evidential vulnerability becomes transcendence; concrete religious consequences return when practical significance is needed. The sentence retains assertoric force while the standards by which it could be assessed keep changing.

At the same time, the phrase can be pragmatically rich. Inside a religious form of life it signals identity, frames morality, organizes narrative, coordinates trust, expresses dependence, and anchors communal practice. Familiarity, agency-oriented cognition, teleological intuitions, ritual, and common ground can make it feel not merely meaningful but obvious. The map supplies an enormous amount of structure. The philosophical mistake occurs when indispensability inside the map is treated as though it automatically established a corresponding entity in the territory.

That is also why the software analogy matters to me. I am accustomed to symbols whose bindings can be resolved, operators whose semantics are known, interfaces whose constraints are explicit, and abstractions whose identity conditions can be stated. “God exists” is commonly offered with the confidence of a Boolean expression while behaving like an unresolved identifier passed to an overloaded operator inside an unspecified runtime. Human conversation is forgiving enough to make that feel meaningful. Critical inquiry should be less forgiving.

So I am not asking to be shown that the proposition is false before I reject it. I am asking for the proposition first. Tell me what God means, what exists means, how the predicates work, what follows, and what would count against the claim. If those things can be fixed coherently, then we finally have something that can be true or false and something worth arguing about.

Until then, I am not rejecting a clear claim. I am rejecting the expectation that I should assent to a sentence that has not been specified enough to become one.

Search This Blog

CriticalThinkingAcademy