I'm not sure this phrase is actually worth using in any guise, especially given that the people using the "correct" meaning are often not using the even more pedantic correct meaning. Many logic books today use "beg the question" as synonymous with "circular reasoning" (some equate them explicitly), but there traditionally were subtle differences. Aristotle and medieval logicians considered circular reasoning to be a formal fallacy, presenting what appears to be a syllogism but is actually not a valid one. But they considered begging the question to be more of an informal fallacy, where an arguer has relied on assumptions that would be unreasonably large to grant, or are too similar to the conclusions, for the argument as a whole to be considered a legitimate or interesting demonstration of the point.<p>In medieval debating, it was sort of an out-of-bounds request more than a fallacy. The difference between on the one hand, putting forth a chain of reasoning that is explicitly circular (and thus fallacious), and on the other hand, asking your opponent to concede a point that, perhaps non-obviously, turns out to be equivalent to what you were arguing for, or at least, more subjectively, concedes too much of the way to your goal. That's not <i>fallacious</i>, because if they concede the point directly, or concede something that amounts to 80% of your point, your conclusion might correctly follow from it, and you have a proper syllogism, where you start with assumptions that all sides agreed on, and through valid reasoning arrive at your result. But it's in some sense cheating at debate, because you've tricked them into conceding the point, rather than having demonstrated the point.<p>(One reason it's subjective and informal is that you're effectively saying that demonstrating B from A is unreasonable, because A already contains too much of B, such that anyone who granted assumption A would in effect have already granted conclusion B, making the argument pointless. But in a strict sense, that's always true with any argument: anyone who grants any set of assumptions is always implicitly granting all conclusions that follow from those assumptions. So what counts as an unreasonable assumption that gives away too much of the conclusion is an outside-of-logic issue.)<p>For any of these meanings, I personally avoid the phrase. If you want the formal circular-reasoning fallacy, you can just say "circular reasoning", and if you want "raises the question", you can say that, or something similar. If you want the meaning I've described above, it's a bit trickier, but you probably will have to explain it anyway using more words, since it's the least common of the three meanings today (but maybe "circular reasoning" will work for that meaning too, in an informal sense of not being strictly circular, but maybe morally equivalent to a circular argument).