Note: This is meant as a question about the historical development of LaTeX conventions, and how they influence Editor design, not as question about how to implement it differently.

One of my pet-peeves about LaTeX, is how editors treat \paragraph as the next level below \subsubsection. While this corresponds just fine to how it is implemented by default with \toclevel@subsubsection=3 and \toclevel@paragraph=4, it strikes me as an odd convention.

To me, the concepts of paragraph and (sub)nsections are entirely distinct, with sectioning defining a top-level structure of the document, and paragraphs/subparagraphs describing a separate layer of structuring within a given (sub)nsection; It doesn't make sense for a \section to contain a \subsubsection directly, but it does make sense for it to contain \paragraphs. This view is mirrored by the standard setting of \paragraphs not appearing in the table of contents, not being numbered, having their title styled entirely different from sections, and simply by having a distinct command-name pattern.

Now, as long as no deeper nesting than subsubsections are needed – and it is probably bad style to have a deeper level of nesting in final documents – this distinction doesn't matter much. It does however start to matter, when editing document structure with editor tools ("promote/demote subtree"), as they tend to convert subsubsections to paragraphs and vice versa. (Depending on the editor this can be changed though.)

So I was wondering, how did that convention come about?

  • 7
    I simply think that Lamport didn't want to go to \subsubsubsection and he (wrongly) chose \paragraph. Bad choice, that's all. – egreg Jun 14 '18 at 9:17
  • 4
    @naphaneal: Well, it should have been \subThreeSection, or something like that… :-) – GuM Jun 14 '18 at 10:16
  • 1
    See also the interview of Leslie Lamport (here or here) (see 2nd column of page 2, continuing onto page 3), where something related is touched upon. (Not exactly an answer to your question.) – ShreevatsaR Jun 15 '18 at 4:57
  • 1
    note that whatever the merits of the command name, \paragraph is the level below \subsubsectionin latex so rather than changing editors not to honour that, it would be better to introduce a new command for formal headed paragraphs that are not part of the section hierarchy if they are needed for some document type. – David Carlisle Jun 16 '18 at 13:12
  • 1
    @kdb The scrjura package implements a concept of clauses ( = headings) and numbered paragraphs. I'm using it for writing contracts according to German standards. – Keks Dose Jun 20 '18 at 11:45

Historically speaking this is inherited from Scribe (designed by Brian Reid in 1979 or so). This system heavily influenced many decisions that Leslie made and you can trace a lot of LaTeX back to choices made by Brian originally.

In the Scribe user manual from 1980 there is


for @Article with and additional @Chapter for @Report and @Book documents.

Notice the name correspondence also for what LaTeX these days called document classes.

So it looks like Leslie invented "subsubsection" as he wanted more heading levels, but that otherwise the names come from Scribe.

  • A bad decision nonetheless – egreg Jun 14 '18 at 13:01
  • 2
    @egrep maybe in hindsight. But on the whole I think Leslie made the right decision in following scribe closely. A) the basic concepts have been good and B) he had a user base to move over --- and we all know what kind of outcry results from us, say, changing a bad but widely used interface . – Frank Mittelbach Jun 14 '18 at 15:01
  • 4
    I know little about Scribe or the background of its authors but this concept of a paragraph as a sub^n section seems to exist in legal work (where it may be numbered) – Chris H Jun 14 '18 at 15:57
  • OK, +1; but this just shifts the question to: Why did Brian Reid treat @Paragraph as a level bwlow @Subsection? – GuM Jun 15 '18 at 10:33
  • 2
    @GuM this is now getting into the guesswork area and I dound you find anybody who knows what went on in Brian's mind at the time (unless there is some information in his PhD thesis about his reasoning on that which I doubt). But I think the possible explanation ChrisH offered is fairly plausible. – Frank Mittelbach Jun 15 '18 at 11:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.