In Defence of the Material Conditional
Alexander V. Gheorghiu · 2026
Read In Defence of the Material Conditional by Alexander V. Gheorghiu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-13066.txt.
Introduction
Alexander Gheorghiu defends the material conditional against the classic charge of paradox. Defined by truth values alone, it makes any conditional with a false start or true end come out true. Critics say such a connective cannot match plain if-then speech. He argues the clash comes from cognitive bias and misreading, not from meaning. The conditional belongs with indicative claims that track co-variation, not counterfactual ones. Read as marking entailment under a theory, it serves science well.
Catalog record
| CREATOR | Alexander V. Gheorghiu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 19837 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.13066
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
In Defence of the Material Conditional Alexander V. Gheorghiu Abstract . The material conditional has long been charged with paradox. Defined truth-functionally, it renders true any conditional whose antecedent is false or consequent true — hence, seemingly absurd statements such as ‘If unicorns exist, then 2 + 2 = 4 2+2=4 ’. This has been taken as proof that the connective cannot capture the meaning of ordinary if–then sentences, which appear to imply a causal or evidential link. I argue, by contrast, that the paradoxes arise from a confusion of what it expresses caused by cognitive biases. The material conditional properly belongs to the class of indicative, not subjunctive, conditionals — those that register patterns of co-variation rather than counterfactual dependence. When understood as a formal device marking entailment under a background theory, it faithfully represents a mode of reasoning essential to science itself: correlation without causation. The faults ascribed to it, therefore, are not flaws in meaning or standard use, but as misapplication and misreading. Few constructions in logic have provoked as much suspicion as the material conditional, A\supset B . This simple device, introduced by 3 , has the virtue of formal elegance but the vice of indifference: in its truth-functional reading, it renders true many if–then statements that seem intuitively absurd. 17 observed that it is equivalent to the statement ‘not- A or B ’. 10 critiques this as an unpalatable result that ‘a true proposition is implied by any proposition, and a false proposition implies any proposition (the so-called paradoxes of material implication)’ an outcome which, though tautologically correct, offends the ordinary sense in which if connects antecedent and consequent. On such grounds, 8 dismissed the connective as unfit to express genuine implication, proposing instead his modal notion of strict implication, under which ‘if A , then B ’ holds only if B follows of necessity from A . Others have pressed the same objection from a pragmatic direction. 14 argued that an if-sentence conventionally suggests that its antecedent provides some reason for its consequent, and that, for this reason, we hesitate to call such statements true or false in abstraction. 11 , in a similar spirit, remarked that if one already knows the antecedent to be false, such conditionals ‘cease to mean anything’ — the question whether ‘if A , then B ?’ simply does not arise. But the material conditional declares all such if-then true ‘by vacuity’. 4 sought to defend it by relegating the divergence to pragmatics: it captures the literal truth-conditions, while our reluctance to assert conditionals with false antecedents arises from conversational implicature. 2 and others have noted, this defence sits uneasily with how we actually evaluate conditionals. Were material implication our ordinary standard we would be ‘intellectually disabled’, unable to distinguish believable from unbelievable if-then statements. I would like to re-examine these long-standing objections. Rather than rehearse them, I observe that they fail to understand what the material implication corresponds to in natural language. For despite its apparent paradoxes, the connective remains indispensable in formal reasoning. I shall argue that many of its supposed defects arise from a confusion between two notions of truth. Once that distinction is kept clear, the alleged paradoxes simply dissolve. First, I argue that the material conditional does capture a standard and legitimate use of if–then constructions. In both scientific and everyday reasoning, we often use conditionals to express patterns rather than causal relations. A familiar example comes from 5 : If the barometer needle falls, then there will be a storm. No one imagines that the barometer causes the storm. The barometer falls because the air pressure drops; the same drop in pressure also produces the storm. The conditional thus expresses a systematic correlation under common conditions, not a causal chain. Now suppose that on some particular day we are seated before a barometer and the needle is steady. The antecedent of the conditional is false, but does that fact bear on the truth of the statement? Of course does not: the conditional remains true because it expresses a general relation between pressure and weather, not a particular episode. Scientific instruments and diagnostic tests routinely rely on this very pattern: a variable A co-occurs with B because both are effects of a common cause C . When A does not occur, the conditional A\supset B remains true, since its purpose is to record the correlation, not the mechanism. Against Ramsey, then, we may note that such conditionals do not ‘cease to mean anything’ simply when the antecedent is false. In these cases, the implication is a test: we detect A to infer B , because both arise in the same context. We do not expect our tests to be causes; to reject the material conditional because it fails to encode causation is therefore to overlook the very abstraction that makes scientific generalisation possible. Science students are taught that ‘correlation does not imply causation’, perhaps textbooks on classical logic should add similar clarification. In linguistic terms, these kind of if-then statements are indicative conditionals . They express a standing or recurrent connection between antecedent and consequent. By contrast, subjunctive conditionals describe how things would stand under an imagined or counterfactual situation — for example, ‘If the barometer were to fall, then there would be a storm’. The distinction between indicative and subjunctive conditionals has, of course, been central since Stalnaker’s work ( 12 ; 13 ) in which he formalised the contrast by introducing a possible-world semantics in which the truth of an indicative conditional depends on what holds at the closest contextually relevant world. 9 extended this framework to counterfactuals, giving the subjunctive reading its now-canonical semantics. On such approaches, every conditional receives a modal analysis, and the material conditional appears as a degenerate or impoverished fragment of a richer logic. Later accounts such as Jackson’s ( 6 ; 7 ) approached the indicative–subjunctive divide in terms of assertibility and contextual relevance, emphasising that our willingness to assert a conditional depends on the evidential relation between antecedent and consequent. However, a more fine-grained linguistic analysis suggests a narrower correspondence. These scientific examples match a subset of indicative conditionals often described as lawlike or generic — sentences that express regular correlations or generalisations over events. As 1 observes in his comprehensive taxonomy, lawlike indicatives constitute a distinct and well-established class of conditionals. A lawlike indicative conditional of the form ‘if A , then B ’ is typically paraphrasable as ‘ B whenever A ’. My claim is that the material conditional abstracts these lawlike indicatives in the sense that its semantics captures their meaning. Here I have only shown that such conditionals are pervasive in ordinary language; below I show how the material conditional abstracts them. Before turning to the arguments, I will first consider why certain constructions appear absurd. Second, I argue that the apparent absurdities of the material conditional arise from misreading lawlike indicative conditionals as saying something other than intended. These misreadings are often justified by attempting to slot statements whose context is unclear into usual modes of discourse; but they are misreadings nonetheless. As a representative example, consider the apparently nonsensical statement: If unicorns exist, then 2+2=4 . This example is