Few modern slogans in popular atheism have the enduring appeal of Hitchens’ Razor:

What can be asserted without evidence can also be dismissed without evidence.

It is elegant and sharp. And it is an effective epistemic heuristic—a rough and dirty rule that gets things right often enough. So, at first glance, the razor looks like an epistemic principle for distinguishing rational belief from mere assertion. Unfortunately, despite its pith, if taken as an epistemic norm, it is garbage. This article is meant as a warning against taking Hitchens’ Razor too seriously, even while acknowledging that there are indeed cases where the lack of evidence is sufficient for rejecting something. Taken as a hard and fast epistemic rule, it fails for two reasons.

I. Not All Rational Beliefs Require Evidence

The first mistake is the assumption that every rational belief must be supported by evidence. That view is intuitively appealing—especially to empiricists—but it cannot be sustained without circularity.

Start with a simple case: the belief, “I have a belief.” That belief is rational, and indeed undeniable, even though there’s no evidence for it in the ordinary sense. You don’t infer it from experience. You don’t need to check behavioral outputs or introspectively test hypotheses. You simply believe, and in believing, know that you do.

That belief isn’t self-verifying—it could, in principle, be formed in a dream or under illusion—but it is self-intimating: rational to hold by virtue of the very state it expresses. The belief’s rationality doesn’t depend on an evidential relation to something external; it’s built into the structure of consciousness itself.

This example generalizes. Consider beliefs like, “Something exists,” “There is a world beyond my private imagination,” “The future will resemble the past,” “I am thinking,” or “2 + 2 = 4.” None of these are rational because of evidence. They are rational because they constitute or underwrite the very possibility of evidence. Evidence itself presupposes them.

To demand evidence for these beliefs is to commit what philosophers sometimes call epistemic bootstrapping: using the very mechanism whose reliability is in question to justify itself. If all rational beliefs require evidence, we would need evidence that our evidential processes are reliable. But any such evidence would be generated by those same processes. The result is regress or circularity.

So epistemologists since Descartes have recognized the need for a class of properly basic beliefs: beliefs rationally held without evidential support because they form the foundation of all justification. Alvin Plantinga revived this idea in contemporary epistemology, arguing that belief in God could (controversially) be properly basic, but the point doesn’t depend on the theology. One could easily be a staunch atheist and still recognize the epistemic need for properly basic beliefs. That is because without basic beliefs about things like perception and memory—that they are reliable enough to be relied upon, for example—no system of evidence justification could get off the ground. There would be an endless need to justify each belief into an infinite regress.

This isn’t to say that Hitchens’ rhetorical point—about who bears the burden of proof—is wrong. In ordinary debate, the person making a positive assertion regularly, though not always, bears the evidential burden. However, this is more of a practical point about rules for conversation rather an epistemic norm. So the razor extends too far to convince properly basic beliefs when that would be the wrong verdict. It extends too far in other ways too, namely to itself.

II. The Razor Itself Fails Its Own Test

The second problem is that the razor itself lacks obvious evidence. And worse still, it is unclear what sort of evidence could support it.

First, suppose one inquires into the evidential basis of the razor itself. Or, to put the problem a little differently, what reasons or evidence is there in favor of adopting the razor as an epistemic norm compared to other, competing epistemic norms, e.g. one that allows for basic beliefs. The razor itself is not an empirical claim, instead it is a normative claim about what we ought to do. That cannot be empirically investigated. Nor is the razor an implication of a broader epistemic norm that does have evidence in its favor.

So where does the epistemic authority of the razor derive? At best, from intuition where people seemingly intuit that the razor is correct. One of the mistakes here is that the intuition appears to back the razor as an epistemic heuristic—that is a quick and dirty rule that is sometimes wrong, but right often enough—for an epistemic norm where people ought to always follow it whenever it applies and are epistemically irrational, in some form or fashion, for not following it in their epistemic lives. It is easy to confuse the two, even though they make for very different epistemology.

Second, the intuitions-as-evidence view is a highly contested topic in analytic philosophy and philosophical methodology. Earlenbaugh and Molyneux (2007) argue that we mostly do not treat intuitions as evidence:

One is willing to treat what other people seemed to see, what they seemed to hear, and what they seem to remember, as evidence, in the sense that one will base one’s beliefs upon it. But one is not willing to base one’s philosophical beliefs on the intuitions of another […] we are hardly affected by the intuitions of others, no matter how strongly they are professed and no matter how many others have the intuition … [We do not treat] intuitions as basic evidential states.

It is odd that of all the things we treat as evidence from third-parties, we do not treat the intuitions of others as evidence—though this has been highly disputed by other philosophers—which is an odd difference. It implies that we treat intuitions as something other than conventional evidence from third-parties like testimony in other areas like perception or memory.

Of course, defenders might respond that the razor is not itself an assertion but a rule of inference or a pragmatic maxim—something like, “Don’t believe things without reason.” That only relocates the problem. Rules of inference, too, require justification for the simple reason that they compete with other rules of inference for adoption, some of which conflict.

Perhaps we are to take the razor as self-evident? If so, it fails its own evidential demand. Are we to treat it as empirically grounded? Then we need empirical evidence that unevidenced assertions are always dismissible—a proposition that’s not only false but conceptually incoherent, since “unevidenced assertions” include foundational beliefs, mathematical truths and subjective reports of experience.

Hitchens’ Razor fails as an epistemic norm because it is asserted without evidence, and so it self-applies, especially since it is not empirically supported or analytically true (think “2+2=4”), not derivable from more basic premises, and not intuitively uncontroversial. It is, in the final analysis, an assertion without evidence.

“What can be asserted without evidence can also be dismissed without evidence” is a good epistemic heuristic and a reminder that most times belief carries an obligation to offer evidential support. Taken only that way, it would be fine. But when people take it as an epistemic norm or principle, it collapses under scrutiny.

Not all rational beliefs require evidence. Some are rational because they constitute the framework of evidence itself. And the razor, ironically, is one of those unevidenced assertions that—by its own rule—ought to be dismissed.