Macy Halford, of The Book Bench at the New Yorker, wrote to me with a question about "begs the question":
Recently, one of our posts caused quite a stir by misusing the phrase (to mean “raises the question”), and many discussions ensued, the result of which was that we all realized that even though we (kind of) understand the phrase in the Latin, we really don’t understand, etymologically, the English translation—either “begs the question” or “petitions for the principle,” though the latter makes more sense. And we all wondered whether the Latin was used only in the context of formalized debate or argument. It seems like a fairly complex concept, with a complex definition—it makes me wonder why we use or misuse it at all, since the need to use it in everyday speech (or blogging) would seem not very great.
(FWIW, the offending post was apparently this one, with the original "this begs questions like …" quickly amended to "this raises questions like…")
There are four related issues here. First, how did "begging the question" come to be a technical term for (a certain kind of) circular reasoning? Second, do people really need a way to talk about circular reasoning, anyway? Third, why did "begging the question" get re-purposed in common usage to mean "dodging the question" or "raising the question", rather than simply subsiding, along with the rest of the terminology of medieval logic, into the midden heap of obsolete idioms? And fourth, should you go with the flow and use "beg the question" to mean "raise the question", or should you fight for the traditional usage, or what? I'll take up these issues one at a time.
1. The history of "beg the question" — a cavalcade of misleading translations — starts with Aristotle's straightforward presentation of a simple idea. In On Sophistical Refutations, he specified thirteen fallacies, one of which (in the translation by W.A. Pickard-Cambridge) is "assuming the original conclusion". This amounts to arguing that "P, therefore P"; and Aristotle gives various simple examples of how such arguments can be disguised so as to appear persuasive.
There are some subtle logical and epistemological questions lurking in the background here, but the basic idea is not exactly deep or difficult. So how did Aristotle's simple "assuming the conclusion" turn into the opaque phrase "begging the question"?
The wrong turns started with the translation from Greek to Latin.
According to Scott Gregory Schreiber, Aristotle on false reasoning, SUNY Press, 2003 (p. 214)
Aristotle uses as labels for this fallacy τὸ ἐν ἀρχῇ αἰτεῖσθαι and τὸ ἐν ἀρχῇ λαμβάνειν. Literally these would be rendered as "asking the original point" and "assuming the original point."
Liddell and Scott gloss αἰτ-έω as "A. 1. ask, beg; 2. ask for, demand; 3. ask one to do; 4. beg of; 5. in Logic, postulate, assume", suggesting that "postulating" might be a better translation than "asking". And for λαμβάνω they have "I. take, 1. take hold of, grasp, seize; … II. receive, 1. have given one", which in Aristotle's use seems to be like the English take of "I take it that …".
So in both forms of the phrase, Aristotle was talking about arguments that assume or postulate the original point to be proved, in some indirect or disguised form.
Some medieval translator (does anyone know who?) decided to translate Aristotle's "assuming the conclusion" into petitio principii. In classical Latin, petitio meant "an attack, a blow; a requesting, beseeching; a request, petition". But in post-classical Latin petitio was also used to mean "a postulate", in the logical sense of "a fundamental principle to be used as a basis for reasoning", and in particular, it was used as a translation of Aristotle's αἰτεσθαι, to mean something like "postulating". And principium meant "a beginning, commencement, origin" — here translating Aristotle's ἐν ἀρχῇ, and means something like "the start of the argument", or "the original point".
From the 14th century onwards, many writers in English used the Latin term to discuss this rhetorical technique, which remained a very simple and straightforward idea behind its Latin label. Thus Sir Thomas Browne, Pseudodoxia epidemica (1658):
Other waies there are of deceit, which consist not in false apprehension of words, that is, verbal expressions of sentential significations, but fraudulent deductions, or inconsequent illations, from a false conception of things. […]
The first is Petitio principii. Which fallacie is commited, when a question is made a medium, or we assume a medium as granted, whereof we remain as unsatisfied as of the question. Briefly, where that is assumed as a principle, to prove another thing, which is not conceded as true it self.
Or from T. H. Green's General Introduction to Vol. I of Hume's Treatise of Human Nature:
68. The relation of cause and effect, according to Locke's general statement as to relation, must be something 'not contained in the real existence of things, but extraneous and superinduced.' […]
Thus not only does the datum of the process of 'invention' in question, i.e. the observation of change in a thing, involved a derived idea, but a derived idea which presupposes just this process of invention.
Here again it is necessary to guard against the notion that Locke's obvious petitio principii might be avoided by a better statement without essential change in his doctrine of ideas.
But starting in the late 16th century, the English phrase "to beg the question" began to sneak into use as an vulgar alternative to petitio principii. The OED's citation list:
1581 W. CLARKE in Confer. IV. (1584) Ffiij, I say this is still to begge the question. 1687 SETTLE Refl. Dryden 13 Here hee's at his old way of Begging the meaning. 1680 BURNET Rochester (1692) 82 This was to assert or beg the thing in Question. 1788 REID Aristotle's Log. v. §3. 118 Begging the question is when the thing to be proved is assumed in the premises. 1852 ROGERS Ecl. Faith 251 Many say it is begging the point in dispute. 1870 BOWEN Logic ix. 294 The vulgar equivalent for petitio principii is begging the question.
Why begging the question? Well, petitio (from peto) in this context means "assuming" or "postulating", but it has other (and older) meanings, from which the notion of logical postulate or assumption arose: "requesting, beseeching". So rather than use some fancy Latinate term like postulate or assume, people decided to use the plain English word beg[ging] as a sort of calque for the "requesting" sense of petitio. But even in the 16th century, I think, it was a bit odd to warn people against presupposing the end-point of their argument by telling them not to beg their conclusion.
And why begging the question? The OED's first glosses for question are "A point or topic to be investigated or discussed; a problem, or a matter forming the basis of a problem", and "A subject or proposal to be debated, decided, or voted on in a meeting or deliberative assembly". With these meanings, question more or less fits into Aristotle's warning — it's wrong to "assume the question", i.e. to make an argument that presupposes (our conclusion about) the proposal to be decided. But these days, the word question is much more likely to mean "A sentence worded or expressed so as to elicit information from a person; a query, an enquiry." And in this sense, warning someone not to assume the question either means something quite different (say, not to jump to conclusions about what is being asked), or else it makes no sense at all.
2. OK, "begging the the question" is a piss-poor way to say "assuming the conclusion"; But should we continue to care enough about circular arguments to have a shorthand way to refer to them?
First, an aside: circular arguments are not necessarily false arguments. The proposition "A → A" is always true, though not very interesting. And a proposition like "A → B → C → A" may be a perfectly valid and useful argument, for example as a way to show that A, B, and C are all equivalent.
So "assuming the conclusion" is not a fallacy in the sense that the form of argument involved is invalid. The cases that are condemned as petitio principii pretend to reason from accepted premises to a novel conclusion, while actually smuggling a disguised form of the conclusion into the argument's assumptions. Does this actually happen in real life, as opposed to mathematical logic or formal debating?
Well, people certainly think it does. Here's Paul Krugman, a couple of weeks ago, writing about "assuming the conclusion":
As predicted, many of the comments to my Austrian economics post are of the form “Well, of course employment rises when investment is expanding, and falls when the investment is falling — in the first case the economy is booming while in the second it’s slumping.”
As I tried to explain, however, that’s assuming the conclusion; there’s no “of course” about it. Why do periods when the economy is investing more correspond to booms, while periods when it’s investing less correspond to slumps? That’s easy to understand in Keynesian terms — but the whole Austrian claim is that they’re an alternative to Keynesianism. Yet I have never seen a clear explanation of this central point.
And the phrase "circular argument" is fairly common in the news, e.g. in this piece:
I’ve come to call this thoroughly circular argument for Easter’s significance the “Presence Principle.” It implies, in a way that my intellect resists but my heart is willing to entertain, that the terrific annual to-do involving lilies, hymns and dexterous rabbits is, just by virtue of its continued existence, not an absurd, unwarranted phenomenon. A celebration, by my old priest’s reasoning, means that its celebrants must have something to celebrate, and the bigger the celebration, the bigger the something.
By a similar form of logic, the concept of a circular argument must be worth writing about, since people often write about it.
(Of course, it's tricky to reconstruct any natural-language argument in fully explicit logical terms, and even trickier to do so in a way that validates an accusation of circularity. But that doesn't stop people from wanting to make the claim.)
3. Even if "assuming the conclusion" is something that people often want to say, "begging the question" is such a confusing way to say it that only a few pedants understand the phrase in this sense any more.
You can see that this is true by looking at how the phrase is used, even in well-edited sources. For example, if we search the NYT index for recent uses of "beg the question", we find that out of the first 20 hits, 15 use "beg the question" to mean "raise the question" — and of the five that don't, four are usage articles berating people for misusing the phrase!
Turning to the broader and more erratically edited range of material indexed by Google News, a check of the first 50 of the 127 current hits for "beg the question" turns up 49 instances meaning "raise the question" — and one lonely usage column (June Casagrande, "A Word Please", Kilgore New Herald 4/10/2010), which makes this amazing assertion:
But for me, all this raises a more important question. You see, for all the people I’ve heard bemoaning others’ misuse of “beg the question,” I’ve never actually heard someone misuse it.
So all the friends who told me I should do a column on “begs the question” because the term is so widely misused — well, it looks like those people were begging the beg-the-question question.
In my opinion, however, using "beg the question" to mean "raise the question" is only one notch less weird than using it to mean "assume the conclusions". We don't use beg in other contexts with this meaning: "??I'd like to beg the topic of what to have for dinner". So why isn't "begging the question" languishing on a dusty shelf of lexicographical off-site storage, along with other renaissance translations of medieval latin rhetorical terminology like "affirming the consequent" (which has no current Google News hits at all?)
In addition to striking a note of dignity or formality, "beg the question" sounds as though it ought to be a slightly old-fashioned way to say something that you could get it to mean by extending the meaning of beg just a bit, as often happens in archaic expressions. There's no similarly simple move that turns "affirming the consequent" into something that people are likely to want to say.
4. OK, those of you who are still with me, what should we do? Should we join the herd and use "beg the question" to mean "raise the question"? Or should we join the few, proud hold-outs who still use it in the old "assume the conclusion" sense, while complaining about the ignorant rabble who etc.?
In my opinion, those are both bad choices. If you use the phrase to mean "raise the question", some pedants will silently dismiss you as a dunce, while others will complain loudly, thus distracting everyone else from whatever you wanted to say. If you complain about others' "misuse", you come across as an annoying pedant. And if you use the phrase to mean "assume the conclusion", almost no one will understand you.
My recommendation: Never use the phrase yourself — use "assume the conclusion" or "raise the question", depending on what you mean — and cultivate an attitude of serene detachment in the face of its use by others.