What Is Begging the Question? An Argument That Assumes Its Own Conclusion

A circular argument is formally valid, which is what makes it interesting, since the conclusion genuinely follows from the premises. Its failure is not logical but dialectical: it cannot do the work an argument is supposed to do, which is to move somebody from what they accept to what they do not. If a premise is only acceptable to somebody who already accepts the conclusion, the argument gives them no reason to change their mind. The obvious cases are trivially recognisable, as when someone argues that a text is reliable because it says so and what it says is reliable. The interesting cases are subtle, where the circularity is concealed by different vocabulary in the premise and the conclusion, or spread across several steps so that the circle is large enough not to be visible. Detecting it requires asking of each premise whether somebody who rejected the conclusion could accept it, which is a more useful test than looking for repeated words.

Several argumentative failures are neighbours and worth distinguishing:

• •Circular reasoning in the narrow sense, where the conclusion appears among the premises, possibly reworded
• •Question-begging terms, where a loaded word builds the disputed claim into the description, so calling a policy a reform or a scheme settles the argument by vocabulary
• •The complex question, which presupposes something disputed inside the question itself, so that any direct answer concedes it
• •Assuming the antecedent, where an argument proceeds from a premise that is precisely what is in dispute without acknowledging it
• •The no true Scotsman move, which redefines a term mid-argument to protect a generalisation, which is a cousin rather than the same thing
• •Legitimate circularity in some foundational contexts, since arguments for the reliability of reason or of the senses cannot avoid using them, which philosophers treat as a genuine problem rather than a simple error

In ordinary modern usage the phrase overwhelmingly means to raise the question or to invite the question, as in a situation that begs the question of what happens next, and that usage is now far more common than the logical one. Prescriptivists object that this destroys a useful technical term, and the objection has some force since no convenient replacement exists and the logical sense has no other name in general use. Descriptivists point out that the technical sense is itself a poor translation, since the original Latin phrase is better rendered as assuming the initial point, and that the English phrase has never transparently meant what logicians use it for, which is a large part of why the drift happened. The practical position, adopted by most careful writers, is to avoid the phrase entirely and say either assumes what it sets out to prove or raises the question, both of which are clearer than the ambiguous original and neither of which requires anyone to lose an argument about usage.

Concealed circularity appears in serious argument frequently. Definitional disputes are the commonest venue, since defining a contested term in a way that settles the dispute is a question-begging move dressed as clarification, and arguments about what counts as violence, as art, as science or as a person routinely proceed this way. Statistical and evidential arguments can be circular when the data used to test a hypothesis was selected or coded using assumptions derived from the hypothesis. Legal reasoning is vulnerable where a standard is applied that presupposes the conclusion, which is why appellate arguments frequently attack the framing rather than the application. Scientific reasoning encounters the problem in calibration, where an instrument is validated against a standard that was itself established using the instrument. In each case the remedy is the same and is easier to state than to apply: identify the premise doing the work and ask whether an opponent could grant it while still denying the conclusion.

A circular argument is valid and useless, since a premise only acceptable to somebody who already agrees gives nobody a reason to change their mind. The test is whether an opponent could grant each premise while denying the conclusion. Loaded vocabulary and presupposing questions are neighbouring moves. The phrase now usually means raises the question, so careful writers avoid it and say what they mean instead.