The necessary references, I: Some guidelines

In the series Typesetting TeXniques, we share our expertise as professional copyeditors and typesetters of mathematics. We frequently hear from authors that our work makes their papers better. This is how we do it.

MSP’s production editors aren’t only LaTeX typesetting pros. Most have advanced backgrounds in mathematics, many have published work themselves, and all are familiar with the standards of mathematical writing and, in particular, good practices for citing references. Here are some guidelines they hope to see authors follow.

Stay relevant

#

An item should appear in an article’s list of references only if

1. the item is cited somewhere in the text,
2. the item is relevant, and
3. the text makes the item’s relevance clear.

If the final condition can’t be met, then the item probably isn’t relevant. For example, if it’s only possible to include a reference in a long string of undifferentiated citations, the reference probably isn’t relevant and shouldn’t be cited.

On the other hand, not all items that could conform to the three conditions above should be included in an article’s list of references. Items such as these may be relevant to a work yet not merit inclusion:

• Personal communications.
• Items that are not publicly available.
• References for standard results.

Readers won’t be able to find items of the first two types, and those of the third type are most likely superfluous. Mentions of personal communications, when needed, should be kept in the body of the paper, rather than given a spot in the list of references.

Point the way

#

Within a reference

#

Citations should include coordinates to guide readers to the relevant portion of a reference:

• Theorem 1.2 in [3] establishes that…
• Motivated by the discussion on page 32 of [6], we…

To a reference

#

Of course, before readers can examine the relevant portion of a reference, they must find the reference itself. The list of references should provide the information needed for readers to do so reliably, and as easily as possible.

• For any item that has been published, the work that is cited and included in the list of references should be the published version, not a preprint. The published version is the version of record.
• Every preprint that must appear in the list of references should include the year it went into circulation and, if applicable, a version number.
• Every item should include complete bibliographic details. Some of the most common are:
  • author, title, year
  • journal, volume, number, pages (for articles)
  • publisher, edition (for books)
  • institution (for theses)
  • arXiv number (for preprints that have not been published)
  • MR number (MathSciNet), Zbl number (zbMATH Open)
  • DOI or URL (if the work does not have a DOI)

We all make mistakes

#

Occasionally, authors inadvertently include an item in their article’s list of references that they have not cited in the text. Sometimes, a reference has incorrect or incomplete bibliographic details. Other times, authors intend to update a preprint to its published version and forget to do so. MSP’s production editors catch and correct these types of mistakes or oversights. In the sequel to this post, we’ll show how they process lists of references for publication.

Header image by You Le, available from Unsplash. Free to use under the Unsplash license.