How far back should one go in citing classic works?

Question

I once jokingly included in a thesis:

The equations of motion [1] for a rigid body can be reformulated as…

[1] I. Newton, Philosophiæ Naturalis Principia Mathematica, 1687.

While citing Newton's work for his equations of motion might be over the top, there is a valid underlying question: what criteria can one use to know when a work is “such a classic” that it doesn't warrant citation anymore.

I'll give example in my field:

Ewald summation is a neat trick, and very widely used by people who perform molecular simulation of charged species. It was initially published by Paul Ewald in 1921.
Density functional theory in the Kohn-Sham formalism, published in 1964 (Hohenberg & Kohn) and 1965 (Kohn & Sham).
Nosé–Hoover thermostat, 1984

I would cite anything that is not taught at Master level (or below), regardless of the date of publication. — , Nov 21 '12 at 09:27
I once cited Darwin and one of the reviewers thought it was funny — Leon palafox, Nov 21 '12 at 10:02
I think Newton himself gives you a rule: "If I have seen further it is by standing on the shoulders of giants". Citations are meant to refer to works you used (explicitly) in your research. Otherwise most papers would cite Cantor and Dedekind for constructing the real numbers. — , Nov 22 '12 at 11:10
@Procrastinator this seems to be somewhat field-dependent. While I've never been explicitly told this, my experience in reading and writing high-energy physics papers is that citations are as much a way to acknowledge the originator of an idea as they are a way to demonstrate which resources you used. — David Z, Nov 27 '12 at 01:52
@DavidZaslavsky I see your point but there is always a "backward" limit in any area. (A bit tongue-in-cheek, again) I am sure Leicippus and Democritus are not highly cited in your area despite pioneering the atomism. — , Nov 27 '12 at 09:04

score 24 · Accepted Answer · answered Nov 21 '12 at 12:44

The ultimate question is what would the average well-informed bachelor's or master's degree-holder in your field be expected to know.

Something that comes up in an undergraduate textbook, or is a commonly known and easily demonstrated fact (for example, the definition of the error function), probably doesn't need to be cited, because everyone in your field would be expected to know it. On the other hand, if you're publishing in a "general interest" journal, then you might want to assume a lower "base" of knowledge, and cite a commonly available source.

However, if it's something that's a very specific tool or fact (or a derivation, etc.), then it probably merits a citation in any case.

score 8 · Answer 2 · edited Dec 29 '13 at 14:31

What they taught me, is that the main criterion is:

is there a standard textbook covering the topic I want to cite?

If there is, then rather point to it than to the original paper. And only don't point to it, when you are sure, that the reader is familiar with the topic (e.g. equations of motions).

However, when you want to point to a very specific observation on discovery, still you might like to point to the original paper.

How far back should one go in citing classic works?

2 Answers2

Linked

Related