12

I've been away from Phil.SE for a while, aside from the occasional lurking, and I recently noticed something. There are (more than I remember) a large number of answers that don't attempt to provide any external sources.

This other meta-question covers some related issues, and suggests there's a policy of encouraging links to external resources, with Keelan's answer quoting the help center as saying that "Links to external resources are encouraged,...". So, we seem to have a policy of encouraging well-sourced answers (which is a policy I love; I might even support a stronger policy of banning unsourced answers entirely, except for "show your work" kind of answers, e.g., a logic problem).

So my question is, assuming we want to encourage well-sourced answers, how might we best go about it? I've been downvoting (since they often also drift firmly into the "subjective answer" category) and leaving comments explaining the reason for the downvote and trying to encourage them to expand upon their answer with references to some of the relevant work.

Is there anything else we could do?

  • 4
    This is a great thing for us to be discussing. – virmaior Oct 14 '15 at 5:34
  • Seems close to to this question? – Joseph Weissman Oct 14 '15 at 12:42
  • 1
    @JosephWeissman Definitely, missed that one. I guess I'm just focused on a particular type of low-quality answer that may be more avoidable. If unsourced answers are either deemed subjective or heavily downvoted, then (hopefully) people will get the hint and start including sources. I'd hope this would ratchet up the quality of answers without veering too close to an "only academics can answer" policy. I really just want to avoid answers that read like the poster is simply opining ex nihilo. – Dennis Oct 14 '15 at 16:37
  • @JosephWeissman I edited the title to make clear that the focus of this question is low-quality in the sense of lacking citations. – Dennis Oct 14 '15 at 16:40
  • 2
    The comments are sometimes full of sources and references while the answer is a wall of text. I think a problem is the work behind a well backupped answer. This work, if the sources are not present, is what's preventing non-academics from delivering a good answer with sources. So again, how normatively strict can one become without excluding the people a SE should reach? The others have sources of their own, hopefully. – Philip Klöcking Oct 14 '15 at 19:26
  • 2
    Skeptics SE has a pretty stringent policy about sources; might we worth reviewing what they do. skeptics.stackexchange.com/questions – Dave Oct 15 '15 at 13:01
  • What comes to my mind as I think of the questions and possibilities to answer of today: The best way to provide well-sourced answers are well-sourced questions. Simple as that. – Philip Klöcking Oct 15 '15 at 20:54
8

On this site more than anywhere else I (we) seem to have the problem of users who

  1. Do not follow our policy
  2. Have conflicting ideas about what is a good answer
  3. Are absolutely convinced their ideas are correct
  4. Refuse to engage in a mature discussion about their ideas
  5. Refuse to adhere to our policy

As a recent example, take this 'discussion'.

This poses a problem, and makes, in my experience, this one of the hardest sites to moderate across the SE network.

I have the impression that this confusion of mostly new users is because there is no summary of our policy. There is the help center, and the tour, but neither really go into details about how we think this particular site is to be moderated. There are meta questions, but when are they authoritative, and which question should I look at?

I am thus in favour of a policy like Skeptics.SE. In this policy it should be very clear (a) when references are needed, and (b) that answers that need references will be removed when the author fails to provide references within a reasonable time limit (say, a day?).

Ultimately, if users keep refusing to adhere to that policy, they should be banned. It is not a very good use of the time of any of us to keep deleting answers of the same users over and over again.

What should be in our policy?

I'd say, at least the following:

  • Answers should be objective.

    Questions like "Is it ethical to...?" should be answered in the format "A [Kantian] would say ..., because ... (see ...)".

    Questions about things as formal logic don't necessarily need references, because they're often easily backed by any introductory textbook or can at least be easily verified (by the OP and the community, through voting).

    Answers that are not objective or not verifiable will be removed.

  • Questions that stimulate subjectiveness will be edited or closed.

  • Comments should be posted in the same spirit as questions and answers.

  • We strive for high quality here. It is not necessary to speak perfect English, but chat abbreviations is frowned upon and punctuation is encouraged. Write as if you were writing something to publish.

This is a just a start. If there is sufficient support for a policy / welcome message in general, it would perhaps be good to open a separate meta question to flesh that out, and then a final meta question to actually have that policy exposed (preferably written by a moderator, to show some authority).

  • 4
    I certainly am with you here. This issue is something that has driven me away from Phil.SE at times. Anecdotally, it's the reason several experts I know who participate online at, e.g., the PhilPapers forum -- and who use SE sites like TeX.SE -- refuse to use Phil.SE. The questions and answers are seen as too untethered from any literature to be worth engaging with. It really reminds me of the self-confident math crackpots who come out with their purported disproofs of, e.g., the fact that 1=0.9999.... While I respect the right of the users to steer policy, I really hope this is where it goes. – Dennis Oct 26 '15 at 20:45
  • It does seem to be the direction we have been going in in the last couple of months. What makes it difficult however, for me at least, is that the users who disagree don't want to discuss on meta. But somehow you want to agree on something together. – Keelan Oct 26 '15 at 20:47
  • We do have an FAQ that gathers together a lot of the most important policy questions we've had on meta – Joseph Weissman Oct 28 '15 at 22:18
  • 2
    @Dennis I wish I knew how to fix things but I feel like half the user base wants something that doesn't interest me at all but goes under the name "philosophy". I'm not quite sure how to accommodate everyone... – virmaior Dec 9 '15 at 5:16
  • 2
    Personally I don't see any problem in requiring referenced answers, I'm an occasional participant (quite recently joined in), and of my 10 answers all but 2 make essential use of papers or books published by actual academic philosophers. One of those 2 answers was also partly based on a paper but I didn't mention it, the other is a baby logic question. It took no effort to use references, in fact it made answering easier. – Johannes Dec 9 '15 at 16:53
7

We should have a policy like Skeptics.SE

Properly sourced answers are better-factored for StackExchange. An answer that "shows its work" has greater pedagogical value. It serves as a kind of minimum quality filter, and may help focus discussion around particular thinkers and texts.

  • Just to try to flesh out this position a little bit! Not necessarily advocating this etc – Joseph Weissman Oct 18 '15 at 16:38
  • I'm assuming you mean this policy? – Keelan Oct 25 '15 at 11:44
  • @Keelan yeah, in particular the "answers need references" component of the policy – Joseph Weissman Oct 25 '15 at 16:40
3

Providing sources obviously takes more time, but it requires .. and reflects .. additional rigor. I appreciate it when I see it, and I find I am fairly quickly motivated to return the favor.

Maybe the best approach is threefold:

  • prominent FAQ & policy guidance
  • providing positive comments (aka "thank you's") as reinforcement when useful references are provided
  • leadership by example

Some shorthand simplifies the task, if allowed. Maybe a few title keywords in italics, and the year? Chasing down section refs, page numbers and publishers can consume hard to find cycles.

Confession & disclaimer: some of my answers/comments are penned at the coffeeshop. Dog earred they may be, but I don't always have Aristotle & Kant in tow :)

2

Good question!

If a good percentage of the community referenced questions and answers then it might be easier to suggest users should do so, without driving people away.

It does take some work to reference a good answer; especially if it's to a book; which is why I often opt for the SEP.

It might be a little easier if there were standard places where one could go for citations.

Some canonical texts have standard ways of citation, I've noticed. Like Kant, Wittgenstein, Aristotle - it would be nice if one could simply type in [Phys:217a3] and have that link directly to a standard citation service point - it's saves looking things up and typing them out.

Having said that, no doubt doubt we'll get walls of citations, rather than walls of text - as a form of textual abuse!

You must log in to answer this question.

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