Although not much encountered these days, the original meaning of the phrase “to beg the question” refers to a piece of faulty logic, which H.W. Fowler’s Dictionary Of Modern English Usage (1926) defines as:

The fallacy of founding a conclusion on a basis that as much needs to be proved as the conclusion itself

Fowler offered two examples that could have fallen straight out of the “Jeeves and Wooster” stories of P.G. Wodehouse:

That foxhunting is not cruel, since the fox enjoys the fun, & that one must keep servants, since all respectable people do so

These have the (very slight) kudos of attempting to justify one claim by at least offering a different unsubstantiated claim in its support. But we can also (with tedious frequency) encounter examples that are just a closed loop, in which the speaker attempts to justify a claim by referring to the same claim dressed up in different language. Here’s one concocted by Madsen Pirie for his splendid guide, The Book Of The Fallacy (1985):

Justice requires higher wages because it is right that people should earn more

So that’s the earliest usage of “to beg the question”, going back to the sixteenth century, according to the Oxford English Dictionary. But (as I imagine you’ll agree) the phrase itself seems to have absolutely nothing to do with the concept it describes. Little wonder, then, that people have taken the stock phrase, puzzled over it, and given it new meanings.

The most common current usage is as a synonym for “to prompt the question”:

If this law is not enforced, it begs the question, “What’s the point of making it a law in the first place?”

The other (in my experience much rarer) usage is synonymous with “to dodge the question”. That’s how Howard Jacobson used it in his novel Redback (1987):

However, our tearful fight […] had nothing at all to do with politics. I’d say it had to do with women, except that’s not true either. Let’s settle for its having something to do with love, and beg the question of just who was in love with whom.

But pretty much no-one uses it to refer to a logical fallacy, except in discussions of logical fallacies, or in rants about how no-one speaks English properly any more.

So how did a fairly straightforward example of woolly thinking (or deliberate deception) acquire such an opaque label? It was probably an ill-considered translation of an equally opaque Latin tag.

The name of this logical fallacy in Latin is petitio principii. Now, in my compact Latin dictionary, petitio has meanings involving attacking or pleading—it’s the origin of our word petition. And principii is the genitive case of principium, which means things like “foremost”, “origin” and “beginning”. Principium is related to our words prince and principle. In the plural it is principia, which features in the Latin title of one of the most famous scientific works of all time, Newton’s Philosophiae Naturalis Principia Mathematica (“Mathematical Principles of Natural Philosophy”), often referred to in familiar terms as just “The Principia”. So that would make petitio principii translate as “petition of the beginning”, which isn’t any clearer than “to beg the question”, really.

But petitio acquired a new meaning in post-Classical Latin—beyond the scope and remit of my little dictionary, but available in the Dictionary Of Medieval Latin From British Sources. When mediaeval scholars were discussing mathematics in Latin (as they were wont to do), they used petitio to mean “postulate”. In mathematical terms, postulates are the assumptions on which all further reasoning is based—Euclid built the whole of his understanding of geometry systematically from just five famous postulates. And I can’t offer a better definition of “postulate” than the one given by Sir Henry Billingsley in 1570, in the first English-language translation of Euclid’s mathematical textbook, the Elements.

[…] certain general sentences, so plain, and so perspicuous, that they are perceived to be true as soon as they are uttered

For instance, Euclid’s First Postulate is just:

A straight line can be drawn joining any two points

Back to petitio. The late-Latin meaning, involving a fundamental assumption on which further reasoning is based, was so solidly adopted that Billingsley happily used the word “petition” instead of “postulate” in his translation of Euclid—the definition I offered above is actually his definition of the word “petition”, as he is using it.

So now we’re making progress. Petitio principii, in post-Classical Latin, meant “assumption of the beginning”. While it is okay (and indeed absolutely necessary) to make some initial assumptions in mathematics, cogent arguments in the real world need to be founded on observations made in the real world. So the Latin phrase petitio principii is saying, “Oops, this argument is based on nothing more than an assumption at the beginning!”

But how did that end up as “to beg the question” in English? It seems that the mathematical and scientific usage of petitio was lost in translation at some point, and people went back to the Classical Latin meanings relating to pleading—hence, “to beg”. And the “question” bit comes from an unusual usage of the word, only tangentially related to queries, but still with us in various stock phrases. In its long entry on “question”, the OED offers the following definition:

A subject for discussion, a proposal to be debated or voted on […]

We use this meaning when we say, “That’s completely out of the question!” And the bit at the start of an argument, which leads to subsequent discussion, is “the matter in question“.

So petitio became “to beg” because of a confusion between Classical Latin and specialist post-Classical Latin usages; and principii became “the question” as a reference to the fundamental matter which leads to the subsequent discussion.

One can hardly blame people for forgetting the original meaning and coming up with new ones.