Should we give explicit numbers in headings?


@vimala @blake @brahmali

So in the Digha, many suttas have a series of headings. These are a feature of the MS text, and are not found in the manuscripts, nor in many translations. I have been translating them, though: I love a structured text!

The Pali text gives explicit numbers for them. I’m not sure if this is a good idea, though. Since the numbering is not really intrinsic to the text, normally numbers are added automatically. This has advantages:

  • keeping the text clean,
  • ensuring consistency,
  • allowing the numbers to be formatted, (though this could be emulated by regexing them into a span)
  • allowing them to be displayed or not according to context. (ditto)

It is in principle the right thing to do.

However, there are arguments for the other side, too.

  • Hard coded numbers are more robust,
  • the source text has them (but we could remove them)
  • many translations use them, often inconsistently (should we respect the translator’s choice?)
  • Due to elision, it is not always going to be the case that the heading numbers can be inferred; i.e. some implied headings will be omitted but should still be counted. This will mostly be in obscure Abhidhamma texts, but still.

So far I have not been adding them to my translation. But that last point is bugging me: I am really not sure how to address it.

If we are to move to not having heading numbers, we should remove them from all texts to ensure consistency.

All in all, I’m leaning to including them in the new translations, and regexing them into a span through all texts that have them. It’s not pretty, but it should work okay.


I’m in favor of having numbers because it’s a big pain not having them and trying to figure out where something is, especially on long texts like DN.


@sujato I’m also in favor of adding number! Thanks for all your good work, Bhante.


I agree that it’s better to have them. Explicit numbering is better than implicit from a robustness point of view.

Technically numbers can also be used as # bookmarks, altough there is also an argument for using the heading text itself for the bookmark for a readable URL.


Yes, I think you’re probably right.

At some point we should look at ensuring consistency through all our texts. In fact it probably would be a bad idea to review the corpus as a whole and polish up the consistency. I mean, it’s not too bad, but I’m sure there’s room for improvement.

Given that the main bookmarks will be the segment numbers, I’m inclined to think that a sanitized version of the heading text is preferable. That can also serve as a subtle reminder in some cases; you can see what heading you’re under while reading.


I’ve just come to this in my efforts to catch up a bit.

I am not sure exactly what you are suggesting here. Do you propose to do this exactly as in the Mahāsaṅgīti? For DN 16 you would then have to number the sections 3.1, 3.2, etc. This would work well for the Khandhakas, too. For the Suttavibaṅga, however, the number system is much more detailed, with up to four numbers separated by three dots. The whole thing will start to look a bit like a university text book. In general, I am concerned sequential numbering is going to make the text seem more academic and less reader friendly.


The point is not about whether such numbers are displayed, but how they are coded. Should they be literally written in the text, or generated automatically? Display is another question, which can be handled contextually. Whichever way we choose—and I think we will go the hard-coded route—we should be able to decide whether to display them or not in any specific context.


Ok. What does this mean in practice? Should I add these numbers at this stage, and if so, how should they be marked up?


Don’t worry, just keep doing what you have been doing. It’s easy for us to add them later.