I know, the issue of our policy on reference requests was discussed (too) many times:

a. Allowing some reference requests: here, here, here, here, here.

b. The idea of a chat-room serving as a reference-desk: Praised but, aparently, abandoned; also a suggestion to send reference-requests to chat here; the idea of "select reference-lists" and a call for a CW references list; T.E.D.'s elaborate proposal; Mark C. Wallace's elaborate proposal: No idea what happened to these.

c. An abandoned call to renew the discussion here; an inconclusive discussion here.

From what I understand, some limited reference requests are OK, but what about other proposals (such as selected lists, CW, chatroom): Did we try to implement these but it did not work?

I understand why we have a policy against general reference requests, but I do not feel good about just closing such questions: Where should people ask if not here? Answering in comments is not right either. Telling people to read Wikipedia articles is not right as well, since the number of references there can be overwhelming and most are irrelevant for a newbie.

Since, I assume, the other proposed ideas did not work, did we try channeling reference requests to Meta-HSE? Of course, Meta is not meant for this, but answering on Meta would resolve at least some issues related to reference requests (such as gaining reputation by posting trivial answers, or attracting low quality answers). Unlike chatrooms, Meta is easily searchable; Meta-votes mean something quite different from votes on main; Meta-questions need not admit definitive answers: disagreement on Meta is common (by design).

OK, now to the question of why am I reopening the old issue which was discussed at length before:

  1. I am very hesitant reviving at once many old threads (primarily items in b and c lists), but curious to find out what (if anything) happened.

  2. The meta-based proposal for reference lists appears to be new (but I might have easily missed something).

Lastly: It is quite possible that nothing can be done regarding reference requests since people most active on this site are too busy or/and tired of discussing this issue, or mostly are opposed to reference requests of any type. That would be a fine answer as well.

Edit. To answer mod's question asked in a comment to one of the answers:

1. I am active on three other stackexchange sites: Mathoverflow (MO), Math.stackexchange (MSE) and History of Science and Mathematics (HSM). All three allow for reference requests; to give few examples from HSM (since it is closer to HSE):

a. "What are some good references elucidating the discovery/creation of Fourier Series?"

b. "Literature on Mayan mathematics".

c. "Request for good resources on 'history of infinity' topics".

d. "Books on the History of Physics": I think this request was not shot down because the body of the question made it clear that the questioner was primarily interested in history of Quantum Mechanics.

e. But this reference request was too broad and closed for that reason:

What are some good books on history of mathematical thought?.

2. Two more stack exchange sites I am familiar with (but not active at): matheducators and physics.

(i) Matheducators allow for very broadly-stated reference requests such as:

What are some great books for exploring mathematics? (not kids' books and not textbooks). It was well-received and had many answers.

(ii) The current situation on physics stack exchange is less clear, but they created (via CW) a large list of book references for variety of topics in physics, Book recommendations [closed], with the preamble:

Every once in a while, we get a question asking for a book or other educational reference on a particular topic at a particular level. This is a meta-question that collects all those links together. If you're looking for book recommendations, this is probably the place to start.

However, of course, each stackexchange site has its own culture regarding what questions are acceptable.

Edit. So far, looking at the votes, it appears that there is nothing close to a solid majority agreement on anything regarding references (the split is about 50-50). But maybe the sample is too small.

  • 2
    Agreeing with Mark here. For example, I'm pretty sure we had a meta discussion at one point where it was agreed that certain types of requests for references were on-bounds ("What is the canonical reference for..." q's IIRC)
    – T.E.D. Mod
    Jun 8, 2020 at 12:59
  • 1
    I think we should try to close some of the "old" topics that didn't receive answers and point them to a more recent one so it would be easier to see the "duplicate" list.
    – gktscrk
    Jun 9, 2020 at 8:00

4 Answers 4


We should not make book recommendations. They rely on the users opinions. Sometimes that might be good, sometimes its bad. Not all our users have 'good' intent. Not all users have equal levels of understanding. None of our users are free from bias.

Those of you who have been around a while will have noted how many bad recommendations/sources can appear or be cited. I recall one person asking why a downvote and the response was that the 'source' was a neo-nazi Hitler apologist (actual phrasing might have varied a little). We have had other individuals who self-publish their 'theories', with no academic review, and then link to those publications, calling them sources. My point is that 'those individuals' can make a book recommendation right alongside any of the rest of us.

Opinions are not answers. Our site is built on strong answers with facts, not opinions, to support those answers. Book recommendations do not fit the structure of this site. The process of the historical method is built on the fact that sources are not equal, secondary sources even more so. To make recommendations such as this, and to allow random anonymous votes to assign legitimacy to the responses, is asking for trouble.

  • 2
    This a a very valid point. Recommendations from users who think, for example, that Benjamin Freedman (Per Wikipedia an "American businessman, Holocaust denier,and vocal anti-Zionist"), and Sisly Huddleston (a Vichy apologist ) are "reliable" sources, may not be the sort of books that most users of History:SE would like to see recommended here. Jun 10, 2020 at 20:38

The analysis above - both the question and the top voted answer - is superlative. The money quote is:

I understand why we have a policy against general reference requests, but I do not feel good about just closing such questions. . . MoisheKohan

Summary: let's treat this as an X:Y question and fix the question.

I think most of the time, we can answer the underlying question by helping the OP to revise the question to fit within our guidelines. Everything below assumes that answers on H:SE should include citations (there are some exceptions, but even in those cases, answers with citations will generally rise above those that omit them). I'm going to revise the problem statement and then offer some suggested solutions. Building on @gtkscrk's analysis; When I see a question that requests references, I think the actual question falls into one of the following categories

  1. I want to know X; I've checked the secondary sources and I need to verify with a primary - I believe this is/should be permitted on our site. I've added the tag to allow people to very specifically request primary source answers. This is a fairly rare pattern, but it happens.

  2. Where can I read about X? If there is a canonical source, then this is a corollary to the previous case; we can provide it. This is permitted on our site.

  3. I want to learn about X; I don't know enough to know where to begin to look_? I'm sympathetic to these questions, and sympathetic to those who point out that answering these questions encourages the type of inquiry that resulted in all of us being here. But there are several fundamental problems. First and foremost, questions of this type should be closed as vague and unfocused. Second, it is impossible to provide an authoritative answer; all answers are equally valid and equally useless. Third, these are the kind of questions that are more likely to produce arguments/discussion than answers (for precisely the reasons identified in the second problem). These questions are likely to produce opinionated answers. Whigs will offer suggestions that are diametrically at odds with Labor historians, and then the other disciplines will join in with their analytical frameworks and it is only a matter of time before bricks are thrown.

    a. In this case, I would suggest that we close the question, refer the individual to Wikipedia, and ask them to come back with a specific question. (Optionally point out that Wikipedia has references, that are probably as good as you'll get from H:SE) If the question is "Where can I learn about the Riders of Rohan?", close the question, refer them to Rohan, and ask them to come back with a specific question, e.g. "Were the horses of the eored gaited?". H:SE is far better suited to answer the specific question than the general (yes, I know that any questions about Rohan are out of scope but I wanted an example that was less likely to distract people with real history) This is a change from our site procedures; I'm open to feedback.

    b. Sometimes the response is "I've checked Wikipedia and I don't understand X. . . " - to which the correct reply is (a) fantastic, (b) Then let's revise your question to focus on "I don't understand X" - that is what you want to know right now - that is the narrow question that H:SE exists to answer and (c) please always include that preliminary research in the question.

    c. Alternatively ask in comments, "Yeah, X is fascinating, but it is a broad topic; what interests you about X?", or invite the OP to chat where you can offer all the suggestions that you want.

    d. I've seen people offer "standard" (not quite canonical, but the acknowledged first reference on X) references in comments, and I think that's fine.
    e. I'd be happy to try the chatroom again

  4. I want to understand X, but my current research is leading me in circles (example 'Why did East Germany fail? And why did DDR need hard currency 1986-1989?') The example didn't explicitly ask for references, but this is another case where I think we need to be as welcoming as possible; OP is doing the research, but for some reason (frame challenge, obscure technical definitions, arcane assumptions), OP can't move forward. This is the kind of thing that H:SE does extremely well We need to look past the reference request and revise the question to focus on what OP really needs to know (in this case they don't know what they need to know, because of the frame problem). (this is a minor variation of other cases, but is worth discussing separately).

  5. I want to learn about X (e.g. "Are there primary sources about Jinnah's vision for Pakistan?") - examining the question, this isn't a request for primary sources - this is a request for information about Jinnah's vision for Pakistan.

    a. Quickly revise the question to focus on the knowledge sought. "What was Jinnah's vision for Pakistan?" The request for a source is a red herring; while I'm reluctant to ascribe motivations, my hypothesis is that these are questions from new people who don't know that they don't need to ask for a source; the source will be provided.

Thoughts? Feedback? Can anyone extend the list of tactics to provide answers to the actual question while retaining our focus on providing authoritative answers?

In a different context, here is Scifi's policy

  • 1
    Comments space 2small: Revise our guides to more clearly distinguish between primary and secondary srcs / Prime importance, if there is an underlying question, isolate it & focus (this is the 'XY-problem' of ref-reqs? We might help with it more easily) 2. Anything coming close to doubt needs fast close and then, if revised, ample help to RO. "Want to learn" is imo always "too broad". The general approach should be to get a concrete Q – without src reqs! – but that we insist on providing srcs like these *in answer en passant/by def anyway. (Alas, your last "1!* is theoretical here.) Mar 14, 2021 at 12:41
  • 1
    That leaves me with 2 observations: Our own defs are suboptimal, leading among to short-circuited reactions/comments ("ref-reqs are off-topic", even if Q is fixable) Our own repeated failure to insist on proper refs in As necessitates this very thread multiple times over. If standard policy includes "A must have refs, or they're not proper As", much of the above would vanish. Point 1 needs more elaboration and thoughts, but point 2 should have been implemented long ago and without any exceptions. We despise "bar argument Qs" but routinely see barren As of that style upvoted. Mar 14, 2021 at 12:52
  • Dear Mark: Thank you for revisiting the question. IMHO, anything is better than the current state of affairs regarding reference requests, including a chatroom. (The latter is suboptimal, comparing to the Meta-option, since it is hard to search) Few thoughts: My impression is that most (likely, over 90%) requests for references that we close are rather broad (like "Can you recommend a book on the history of medieval England"). Your suggestions (as I understand it) would be to send the questioner to W. Personally, I dislike this for the reasons I stated above. TBC below. Mar 19, 2021 at 21:02
  • In particular, mostly, there is no way to tell who edited the given W article and how reliable they are (in contrast, HSE provides some quality control in the form of reputation count). Personally, I would not want to buy a handful of books recommended by a W history article just to discover that they are so opinionated, I cannot stand any of these. As for HSE unable to choose between different book suggestions (say, Whig vs Marxist), I do not see a problem here: Surely somebody comment under such an answer "this book has a heavy Whig/Marxist bias, use at your own risk." TBC below. Mar 19, 2021 at 21:09
  • As for people promoting, say, Nazi or Stalinist apology books: I trust mods (and I really do!) to delete such an "answer" very quickly. Another thing: From what I understand, people are asking broad reference requests simply because they want to learn basics about the subject (or use it for teaching basic history to their kids; I was in such situation myself). For this, it does not really matter (afaiac) if a book is moderately biased (again, short of Nazi/Stalinist apology). Even Zinn's book (not that I would ever use it) will get basic facts right. But, likely, I am a minority here. TBC Mar 19, 2021 at 21:17
  • Lastly, about us unable to choose the "right" answer when we upvote: There is a plenty of opinionated answers at HSE, so we vote on these according to our personal biases. In the ideal world, SE is designed to have just one right answer. Reality is obviously different. Mar 19, 2021 at 21:21
  • I would prefer that all answers on H:SE are supported by research. More research, less opinion. If you want to learn basics, go to google or Wikipedia. I understand that you're not thrilled with Wikipedia; I respect that opinion, but don't share it.
    – MCW Mod
    Mar 20, 2021 at 14:46

I like the idea of source requests being permissible, both for primary as well as secondary sources. No complications arise from primary source requests as I see it as these can be evaluated objectively from the point of view of whether they cover the subject matter. With regards to secondary sources, the main complicating factor is the difficulty that we have in assessing the answer (much as has been written before). Therefore, the question, as also indicated in the past, relates to "how" these requests can be made acceptable in the SE format.

I would be very interested in seeing answers relating to secondary sources as it would no doubt enhance our perspective. Quite possibly, a good answer along the lines I've brought out below will force us to be better and more thorough in our own future reading which is always a good thing.


I think the policy should work along the lines that when we propose a book we should also make the case—the biggest possible case—against that book. We should also include and link to reviews in reputable sources if such are available—and we can even comment on these reviews!


Every author has their strengths as well as weaknesses; for someone to be able to determine if that answer suits their question (let's get to that later) the answer should highlight those weaknesses and try to give an objective measure of that author. Hence, answers should be of the type:

Gibbon's 'Decline and Fall' is outdated [reference], exhibits bias against the Greeks [reference], misses out on important recent developments on historiography [reference & examples] but is eminently readable and still provides a good understanding of the general trends if the reader can compensate for the above weaknesses.

Weir's biography of Isabella the Fair is weak, because the majority of evidence is inconclusive, the author indulges in frequent hyperbole to make a case, relevant research or proof is rarely brought, most statements are obvious conjecture ("We may have to assume..."/"It is likely that..."). Nevertheless, the author has brought together all the material that is relevant to Isabella which is presently available. Even if the book is poor, it is a notch above previous volumes that consider the topic, and the female perspective from the author is a considerable improvement over previous Victorian historians.

At the same time, an answer along the lines of the below would be unacceptable:

Read Gibbon because he's the classical source on Rome who constructed the theories that are being used today.

Weir's 'Isabella the Fair' is good because it's written by a woman.

The references that we bring could either be external or specific citations from the book that can be refuted easily. Hence, if an author was a Holocaust-denier, it would be quite easy to reveal that fact.

Caveat: Admittedly, however, I am far happier with source requests that concern Medieval England or Ancient Carthage as I also try (and fail) to stay away from the WW2 and more recent topics (or for even more recent things, they might be too recent). It is clear that there's plenty of controversy that some groups knowingly try to create about this time; therefore, it might make sense to have a list of 'hot' topics which would cause any answers to be 'community wiki' and subject to harsher requirements.


This also moves some onus to the questions:

Nevertheless, a charitable reading can always get past the first part. Especially given canonical resources are already allowed, we should read the above question as:

  • What is the canonical resource on the history of England?

We can always direct the poster to why we are making an edit along these lines for this specific point. Of course, the same difficulty arises in assessing what makes a resource canonical and therefore I'd treat all book answers as I've described mine above.

We are left, at this point, with issue #2. This is still unsolvable and this is something we should enquire about:

  • Is the OP interested in:
    1. Political history;
    2. Military history;
    3. Economic history;
    4. Cultural history.

This would enable to pick a more specific volume that would fit the OP's request as well as create a topic that others could refer to for similar questions.

There is one alternative to generic history questions which would entail a more community-based solution. Namely, we could also find that the question "What is the canonical resource on the history of a country?" would always be answered by the same publisher's books (e.g., Oxford's 'History of X'). The potential downside for this is the advertising factor this would entail and while this sounded a better idea to me previously, that a book is released by Oxford or Cambridge doesn't mean we shouldn't read it with a critical eye.

  • 2
    Wouldn't the strengths and weaknesses of an author also be a matter of opinion? For example, I'm told that David Irving's biography of Hermann Goring is thorough & detailed, with a bibliography that provides a clear overview of the available sources anyone wishes to take their research further. I've even heard it described as "the standard biography". The obvious downside is that Irving is a Holocaust denier and Hitler apologist, and those views often show in his writing. Now, I'd call that a major weakness, but I doubt if our Nazi trolls would agree with me. Jun 10, 2020 at 20:57
  • 1
    @sempaiscuba: Surely that's a generic weakness in the author you've brought out? I don't think it's any more of an opinion than in the answers we already assess based on our sources—answers without sources get comments, questions; answers with poor sources do similarly. Books can act in the same way (links to some reviews, excerpts brought out). My emphasis was also on making the biggest case against the book one is recommending.
    – gktscrk
    Jun 11, 2020 at 6:16
  • 1
    On the contrary, my point here is that asking the person recommending the book to also make the case against it is inherently problematic. Jun 11, 2020 at 7:10

On my opinion reference requests and books recommendations should be ALLOWED. This is the case on other SE sites, and asking reference is an important type of question.

Certainly any book recommendation is a question of opinion. But what is wrong with this? Many answers express the opinion of the person who answers. How can this be otherwise?

Here is an example of recently closed question: Good books on Kievan Rus era? and I would like someone to explain me what is wrong with this question and/or with the answer. Yes, the subject itself is controversial, but does it follow that discussion of such subjects must be banned?

If expressing opinions is not allowed then what is this site for?

  • I ask the person who downvoted to write a comment.
    – Alex
    Jun 10, 2020 at 21:38
  • 1
    I'd be interested to know which SE sites allow book recommendations? Most sites that I've seen class them with shopping recommendations, which are also generally off-topic on the SE network? Jun 10, 2020 at 22:41
  • @sempaiscuba: See the edit to my question. Jun 10, 2020 at 23:42
  • @MoisheKohan Interesting. The sites I use generally either quote the answer from Meta:SE that I linked above, or have their own explicit reason (e.g. Robotics "Life Questions are off-topic. Questions about choosing how to spend your time (what book to read, which class to take ...)". Moreover, whenever I've seen this discussed on Meta:SE, the response has been pretty much the same: "They're off topic because answers will all be equally valid". Jun 11, 2020 at 0:01
  • 1
    @sempaiscuba: 2 more stackexchanges that allow for reference requests are physics and matheducators, for instance this extremely broad request was well-received. My take on the different culture of these 2 and the other 3 sites I listed, is that they all regard their primary purpose as research (Mathoverflow) and education (the rest). Given this, anybody familiar with a particular topic (math, history, whatever) knows that not all books are created equal. Jun 11, 2020 at 0:42
  • Can someone explain me a justification of this "no books no reference" policy. Yes, not all books are created equal. So what? And what is the goal of this site, really? If not research and education?
    – Alex
    Jun 11, 2020 at 2:29
  • @Alex All good questions, but since you've been a member of the site far longer than I have, presumably you were here when those decisions were being made? Jun 11, 2020 at 3:51
  • @semipaiscuba: No. When I joined the site was old and established.
    – Alex
    Jun 11, 2020 at 4:12

You must log in to answer this question.

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