We've recently had John Rennie asking "canonical" questions which he's answered himself:

What is time, does it flow, and if so what defines its direction?

What is time dilation really?

What is the proper way to explain the twin paradox?

He says his answers are intended to be definitive and authoritative, and that he welcomes comments. But he doesn’t. He said "why not post your view as an answer?" but he deleted my point-by-point view without discussion. There were also polite but challenging comments from another poster, followed by John Rennie saying "let's continue this in chat", but now those comments have vanished without trace. And more importantly, those "definitive", "authoritative" answers are not correct. Time really isn't "just a coordinate". A clock really isn't something that "measures distances". The resolution to the twins paradox really isn't "the metric you use to do the calculation is not the same as the metric I use to do the calculation". Whilst those answers have received an awful lot of upvotes, I must point out that science is not a democracy. All the upvotes in the world won't make a wrong answer right.

All in all , I think this canonical sort of thing is fraught with problems. It's the sort of thing that will harm physics stack exchange by saddling it with bad-science lies-to-children answers that can’t be corrected. It is in essence saying “mine is the right answer and all other answers aren’t worth looking at”. It’s imposing an orthodoxy, and it’s the sort of hubristic mindset that gets in the way of scientific progress. It’s more suited to a medieval theocracy than 21st Century physics. IMHO there's already problems with closevotes wherein the alleged duplicate isn't really a duplicate, and/or all the answers are wrong, and/or the closevoter is the author of one of them. This is only going to make it worse.

So I think we should discourage "canonical" questions with "canonical" answers.

What do you think?

  • 7
    $\begingroup$ What points were deleted and how? I don't see anything in the edit history and only mods can delete comments (and he isn't one). The same is true about chat rooms I believe. I also don't know that harm will come by "saddling it with bad-science lies-to-children answers" -- even if all the answers on those questions were horrible, we're talking about like... 3 questions of the 12,250 total on the site. I'm sure we can find way worse things that would do more harm. Like homework questions. $\endgroup$
    – tpg2114
    Commented Mar 18, 2016 at 14:21
  • 12
    $\begingroup$ Yes, we want canonical answers and questions. $\endgroup$
    – ACuriousMind Mod
    Commented Mar 18, 2016 at 14:27
  • 2
    $\begingroup$ I'm fine with discussing whether we like an idea after we tried it -- that post was a year and a half ago and we now have a few questions to base some opinions on. But I don't feel like the argument against it was made very clearly here and it reads as a thinly veiled way to express unhappiness with a particular user (and that unhappiness may be misappropriated without knowing what comments/chat rooms went missing when/where). I could be convinced it is not a great idea to have canonical Q&A, but I'm not yet. $\endgroup$
    – tpg2114
    Commented Mar 18, 2016 at 14:31
  • 1
    $\begingroup$ It's not applying an orthodoxy if the matter is settled, it's stating the truth of the matter. That's the point of these canonical Q&A, staying what the truth of the answer is. You just happen to be in the strange position of disagreeing with the settled position (DGF the reasons or to extend matters, just stating how I see it) $\endgroup$
    – Kyle Kanos
    Commented Mar 18, 2016 at 15:59
  • $\begingroup$ @tpg2114 : it's only three questions now, but that could increase. Imagine if it turned into thirty. Or three hundred. I'm not unhappy with JR so much as I'm unhappy with the whole canonical idea. The idea that "it's stating the truth of the matter" when it isn't. $\endgroup$ Commented Mar 19, 2016 at 12:12
  • 2
    $\begingroup$ Actual it's six questions not three! See this, this and this, all heavily upvoted :-) $\endgroup$ Commented Mar 19, 2016 at 12:20
  • $\begingroup$ @John Rennie: heavily upvoted despite stuff like this: "So at the Big Bang we have the very odd situation where the spacing between every point in the universe is zero, but the universe is still infinite". $\endgroup$ Commented Mar 19, 2016 at 13:10
  • 2
    $\begingroup$ @JohnDuffield: how very singular $\endgroup$ Commented Mar 19, 2016 at 13:34
  • $\begingroup$ @John Rennie : the important point there is that our assumption that the universe is homogeneous doesn't provide us with any scientific evidence that the universe is infinite or closed. $\endgroup$ Commented Mar 19, 2016 at 13:52
  • 2
    $\begingroup$ I should make a list of all the times I've pointed out an error in some "definitive" "authoritative" answer and the poster refuses to discuss it. $\endgroup$ Commented Mar 21, 2016 at 8:17

1 Answer 1


We have always had canonical questions/answers, it’s just that until Danu’s post here in the Meta we didn’t call them canonical. A canonical question/answer is one where the question is so apposite and the answer so clearly explained it ends up being cited many times. Sorting the questions by votes will quickly reveal many such questions. My own favourites include (to list just a few):

Though of course everyone will have their own favourites.

Now, the Physics SE isn’t a blog and shouldn’t be treated like one. If we all started writing what amounted to blog posts just because there was some subject that we found fascinating that would quickly alienate the site members. Setting out to deliberately write a canonical Q/A (as opposed to achieving it as a side effect, like the three examples above) is something that shouldn’t be undertaken lightly.

While I want to keep things general and avoid specific issues let’s take the three Q/As of mine that John mentions as an example.

As the most prolific poster of answers on this site (note that I’m claiming no special merit here, unless you regard fecundity as a virtue :-) I have seen lots and lots of questions and answers - I’ve probably at least glanced at every question on the site for the last five years. As a result I get a feel for when the same topic is being questioned again and again, and the question What is time in its various forms has been asked countless times. When this finally annoyed me once too often I posted on the Meta asking:

Canonical question about the flow of time

and got the general impression that a deliberate attempt at a canonical Q/A would be well received. Even then it took me a month to decide how best to write one. So this was not a spur of the moment impulse because I was bored. It was a considered decision based on consultation. In the end I wrote three questions, but they’re all one Q/A really - I just split them into three to make them more digestible.

Whether or not my three recent Q/As are a long term benefit to this community remains to be seen. I’ve already used them several times as duplicates, or I’ve posted them in comments as a related question, but I haven’t attempted to analyse how many times others have linked them. They have ended up with a middling number of upvotes, but as John says the number of upvotes is a poor guide to quality, especially when a question makes the Hot Questions list as at least one of mine did.

However the question of canonical Q/As is broader than just the three questions mentioned in John’s question. I have linked my three favourite Q/As mentioned above many, many times, and the site would be a considerably poorer place without them. But of course those questions weren’t deliberately posted to be canonical, they just happened to hit the target, and I think the crux of John’s question is:

Should site members deliberately post questions and answers intended to be canonical?

Perhaps surprisingly my answer to this is that no they should not. Just because a subject seems interesting to you is no guarantee that it will be interesting to anyone else, and no guarantee that an attempted canonical Q/A on the subject will be useful. You should at the very least discuss it in the Meta before posting, and perhaps bring it up at the fortnightly physics chats before setting pen to paper.

A last point, and again this is a general point although I’ll address the specific issue of the Q/As mentioned by John. Posting a canonical question and answer does not mean no-one else can answer the question. Indeed, there are frequently many different perspectives on a complicated issue and ideally a canonical question will attract answers from several site members who want to post different perspectives. Indeed this is what has happened with all three of the questions that John mentions. Naturally I think my own answers are the best :-), and they have attracted the most upvotes, but anyone add their own answer if they think they can do a better job.

  • 1
    $\begingroup$ Well, supposedly all posts will be canonical. $\endgroup$
    – Kyle Kanos
    Commented Mar 18, 2016 at 16:43
  • 1
    $\begingroup$ @Kyle Kanos: and as Dirk said: "Eventually what it will become is a place where anyone who asks any question will be referred to a previous answer before their question is closed down". I don't want that. John: all points noted. $\endgroup$ Commented Mar 19, 2016 at 12:17
  • 2
    $\begingroup$ @JohnDuffield: Oddly, you manage to completely ignore my comment on Dirks post... $\endgroup$
    – Kyle Kanos
    Commented Mar 19, 2016 at 12:23
  • 1
    $\begingroup$ @Kyle Kanos : I didn't ignore it, I just didn't think it was anything I should mention. $\endgroup$ Commented Mar 19, 2016 at 13:06
  • 1
    $\begingroup$ @JohnD: well it is worth noting because it'd (strongly) suggest the joke I was making. $\endgroup$
    – Kyle Kanos
    Commented Mar 19, 2016 at 14:33

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .