9

There's no way to evade having a lot of acronyms in a scientific paper, often already in the title.

I use the package glossaries to automatically expand the first use of every acronym (except usage in the title), but for a scientific paper, I do not include a table of acronyms (for my thesis I do). In a long paper, if someone does not remember all acronyms or does not read sequentially, it could be difficult to look up the meaning.

If I use a not very well known acronym in, say, line 100, and again in line 500, should I expand it again, or is it a better style to stick with expanding only and exactly once?

superuser0
  • 976
  • 2
  • 8
  • 28
gerrit
  • 42,782
  • 14
  • 131
  • 215

4 Answers4

12

First, try to avoid acronyms unless you repeat it frequently.

In the case you mention, occurring twice, an acronym is not warranted. New acronyms may be useful if it is likely that it will be used over and over again, both in your paper and in subsequent ones. It is difficult to say how many repetitions is needed before an acronym may become a viable option.

Acronyms generally make a paper less readable. We all know some like DNA or UN and particularly in the first case the acronym is easier to remember than the actual words. So use acronyms sparingly, be careful about what you abbreviate, and avoid publishing acronyms that only your research group uses unless it starts to become a standard (identified by other people also using it in presentations etc.). I think many acronyms start by it becoming jargon first. Trying to be restrictive is the best measure.

Gandalf the White
  • 3
  • 3
Peter Jansson
  • 73,217
  • 12
  • 206
  • 344
  • 1
    In addition to universally known acronyms like DNA and UN, there are also acronyms that have become so common that they're regarded as ordinary words and many people aren't even aware that they are (or were?) acronyms. Examples include "laser" and "scuba". There's surely no need to expand these. – Andreas Blass Nov 11 '14 at 02:02
  • 4
    I rented a satellite phone where the manual talked about the Subscriber Identity Module card. I had to think a few seconds before I realised what they were referring to. – gerrit Oct 30 '15 at 12:12
6
  • (I've been using a table with abbreviations also in a paper, that was no problem at all)
  • For longer texts (thesis) I give long version and abbreviation again if the abbreviation has not been used for "a while" of the text. E.g. introduced in the introduction, but next used in chapter 7 (400 lines ≈ 10-11 pages, that I'd consider probably close enough for not giving the long form again).
  • It may be a good idea to have the long forms in parts of the text that are expected to be read without the rest (obviously in the introduction, but also in summary and outlook)
  • I also consider how widely-used and how ambiguous the abbreviation is: Does PCA mean principal component analysis or polymerase chain reaction? Is LDA linear discriminant analysis or latent dirichlet allocation?
cbeleites unhappy with SX
  • 23,007
  • 1
  • 44
  • 91
2

Ask the journal or conference where you're going to publish the paper. Some specify that it is, indeed, exactly once at the first use. (Shorter-format journals especially prefer this.) Others prefer "upon first use and where necessary for clarity" instead.

Regardless, in a long work with many acronyms that may be referenced out of order, there is no substitute for a table. Likewise if you define specialized symbols, a table of symbols greatly simplifies out-of-order reading. It's often worth devoting a table to such things even if your allowed number of figures and tables is limited.

Rex Kerr
  • 3,834
  • 16
  • 18
0

It appears that other answers have confirmed that the definition of an abbreviation can be used multiple times. In that case, I suggest defining the abbreviation wherever you think the readers may read it individually.

For instance, I typically read most papers by only reading their abstracts, figures, tables, and conclusions. As a result, I redefine or avoid abbreviations in these sections to make the reading experience more efficient.

chongkai Lu
  • 101