We all know how important proper attribution of ideas is. At the same time, certain things have become basic enough that citing the paper where they were first discussed is overkill: to give an extreme example, if you need to do some differentiation in your math/physics paper, you don't need to go and cite Newton and Leibniz. Now, on occasion students ask me how one can determine if a piece of knowledge is common enough that they can forgo a citation. The rule of thumb I give them is:

If it is something that is explained in a standard first-year undergrad textbook, then anybody who is going to read your papers knows about it and you don't need to provide a citation.

[Here I want to emphasize that I give this to them as a rule of thumb, and I always tell them to ignore it and provide the relevant citation if they think it is necessary to do so in a specific case]

Are there better or alternative ways of drawing the line?

  • 3
    It probably depends on your audience. Do they "know about it" or would they find a reference useful?
    – henning
    Jun 25, 2015 at 9:01
  • Very closely related question :Almost everything we know is taken from someone else, so what do I cite in a paper?. I am almost for certain it's a duplicate. Please take a look at it.
    – Nobody
    Jun 25, 2015 at 9:17
  • I frequently see references to textbooks in articles, they are less "overkill" than the original paper and they usually include remarks about history, similar results (which you don't necessarily learn of when you are first shown the result). Plus, one good textbook line in the bibliography can cover many "common knowledge" references throughout the text.
    – T. Verron
    Jun 25, 2015 at 10:01
  • @scaaahu The linked question is not asking for advice for academic publications, but for a student paper which will consist (almost) only of bibliographic references. But the answer does seem applicable to this question too.
    – T. Verron
    Jun 25, 2015 at 10:04

2 Answers 2


I find Latour and Woolgar's spectrum of "facticity" a useful tool for thinking about these questions (a nice summary can be found at this link). It breaks scientific statements into five rough categories by level of certainty in the assertion:

  1. Speculations - don't have to be backed by anything
  2. Descriptions - not established, so need to be directly backed by evidence
  3. Tentatively established - need to be backed by citations
  4. Well accepted - should be stated, but don't need evidence or citations
  5. Tacit - should not even be stated

Where exactly a fact lies on this spectrum depends on the community and state in discussion. In general, the broader the audience, the less well accepted facts should be assumed to be. I think the notion "Should everybody reading this have been taught in a class?" is a good one, though undergraduate is not necessarily the stopping point. For a machine learning audience, for example, you should assume everybody has had graduate level machine learning courses, while for a biology audience you should not assume they have even had undergraduate computer science.

  • 1
    Since the OP mentioned an example from math, I must say that I don't understand where mathematical theorems fit into this system. The numbered list seems to contrast "well accepted" with tentative or speculative; in that case, proven theorems are definitely "well accepted". On the other hand, the paragraph seems to conflate "well accepted" with "widely known". But there are plenty of specialised theorems that are familiar to an extremely small group! Jun 25, 2015 at 23:20
  • @ArtiePrendergast-Smith I think the notion of "proven" is less absolute than it appears at first glance. "Proven" means that a mathematical argument has been found convincing (and therefore accepted) by some community of reference. Consider that there are plenty of people who assert that a paper they have written contains a proof resolving the P=NP question. You probably don't consider P=NP to be "proven," though, because neither you nor people whose opinions you trust have read and been convinced by their proofs. Acceptance, I would claim, is just as valid a concept in math as in medicine.
    – jakebeal
    Jun 26, 2015 at 0:06
  • There's something in what you say, but I think it's not realistic to claim that there's no distinction between, say, math and medicine in this respect. However, this is beside the point of my original comment: what I am saying is that your spectrum implies that anything that requires a reference is tentative or speculative, which is not the case. Jun 26, 2015 at 9:23

Let me quote Kate Williams and Jude Carroll (2009, pp. 26–27) with a good overview on this topic:

"You need to reference when you:

  • use facts, figures or specific details you pick from somewhere to support a point you’re making – you report
  • use a framework or model another author has devised. Let’s say you ‘acknowledge’
  • use the exact words of your source – you quote
  • restate in your own words a specific point, finding or argument an author has made – you paraphrase
  • sum up in a phrase or a few sentences a whole article or chapter, a key finding/conclusion, or a section – you summarise.

You don’t need to reference if you:

  • believe that what you are writing is widely known and accepted by all as ‘fact’. This is usually called ‘common knowledge’
  • can honestly say, ‘I didn’t have to research anything to know that!’.


If finding it out did take effort, show the reader the research you did by referencing it."

Williams, K. & Carroll, J. (2009). Referencing & Understanding Plagiarism. Basingstoke:Palgrave Macmillian.

You must log in to answer this question.

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