What is the most appropriate way to refer to the preceding/succeeding (sub)sections within a paper? In my case, the paper style does not allow enumeration of the (sub)sections.
I would recommend simply saying it directly, like, "in the previous section" or "in the following section." You can also refer to a section by content, e.g., "in the review of prior work" or "in our discussion of results."
I agree with jakebeal's answer, but if you have sections or subsections with short titles, you can even refer to the section's title:
As described in the Introduction, etc.
I actually disagree with @jakebeal's advice on this. I would recommend not to use "previous" and "following" qualifiers. The rationale for my advice is rather simple - writing artifacts and, especially, academic writing ones, are naturally almost always subject to rearranging sections and smaller chunks of material by authors (between revisions and within ones). Thus, these rearrangements risk breaking the correct navigation in case of using relative qualifiers (actually, the same applies to using section numbers for reference).
Therefore, I would advise to use section titles as more stable identifiers. Having said that, I realize that it is possible to use auto-fields or similar features, which somewhat alleviate the problem, but still present risk of having inconsistent document in the end (i.e., due to auto-updating issues).
Make it a section with a number.
If it's so important that you want to refer to it, why not make it a section? I don't think section numbers are always and necessarily structured logically. If the subsection is important (like "an important example") it could be a section.
Reference the Example/Theorem/Remark/Note/Notation/equation number/[whatever]. You should have something referenceable in the vicinity of your subsection.
Reference the section (not the subsection).
Use "above" or "previous section", but only if it's half a page before what you wrote.
Refer to the subsection by name. (In my opinion least preferable. How am I going to find it?)
Whatever you do, make it easily findable! I almost never read papers start to finish on a first reading. In fact, I often read it backwards from the main part of the paper. I read the interesting bits and try to follow the argument. I hate coming across references that are hard to find. Don't assume every reader is going to read your paper cover to cover on a first reading.