Linguistic typography documented?

Question

There are some typographic conventions adhered to in most of linguistic literature where it applies. This includes punctuation affixes

• /phonemic/
• [phonetic] and [±feature]
• ⟨graphemic⟩ or <graphemic>

including quote marks

• ‘semantics’
• “quotation”

or just prefixes

• *wrong, *reconstructed, *ungrammatic

and also italic type

• object language, example

Other possibilities are employed by some scholars (in some fields) but not others, e.g.

typographic effects

• bold face
• serif/roman vs. 𝗌𝖺𝗇𝗌-𝗌𝖾𝗋𝗂𝖿/grotesque vs. cursive/𝓈𝒸𝓇𝒾𝓅𝓉 vs. blackletter/𝔣𝔯𝔞𝔨𝔱𝔲𝔯
• monospaced/fixed-width
• l e t t e r s p a c i n g
• u̲n̲d̲e̲r̲l̲i̲n̲e̲, l̶i̶n̶e̶-t̶h̶r̶o̶u̶g̶h̶, o̅v̅e̅r̅l̅i̅n̅e̅

letter case

• sᴍᴀʟʟ cᴀᴘs
• UPPERCASE
• Initial Caps
• InternalCaps
• camelCase
• snake_case (infix)

paired affixes

• {braces}
• (parentheses), ((doubled))
• ‹guillemets›, «guillemets», ›guillemets‹, »guillemets«

other prefixes

• ?question mark
• #hash sign

or separators (infixes)

• vertical|pipe
• middle·dot

and math-like “operators”

• <, >, =, ≈, …

or arrows

• →, ⇒, ←, ⇐, ↖, ↗, …

I have quite a collection of such semantic styles (sometimes conflicting), but it’s quite unsystematic yet and without authoritative source for the most part.

There are, of course, also conventions for supra/para-textual structures like diagrams, e.g. syntax trees and interlinear glosses, and tables, e.g. mandatory optimality theory tableaux with hand symbol ☞, small-caps in column headers and ugly cell borders. Leipzig Glossing Rules (LGR), for instance, employ single quote marks, space, hyphen, equals sign, uppercase/small-caps, period or underscore, semicolon, colon, backslash, greater than, digits, empty set symbol ∅, parentheses, angle brackets and tilde. Discourse transcriptions in conversion analysis follow complex conventions of their very own (CA, DT, GAT, HIAT, CHAT).

I still have not found meta research that produced an overview of linguistic typography. If there is, I will appreciate pointers very much. Until then, what kinds of typographic conventions, especially inside normal prose, are used in various linguistic fields besides the aforementioned standards?

Disclosure: I’m planning to design and implement a flavor of Markdown (or rather Commonmark / Scholarly Markdown) tailored to linguistics and would like to reuse or at least support as many existing conventions as possible, also see Markdown for linguistics?

Answer 1

There has been no systematic meta-typological study of such conventions in linguistics. The closest that you could get would be journal style sheets. Even then, there are outliers, for example use of bold for linguistic examples in-text, found in NLLT, contrary to the more general practice of using italics. It is interesting that while the Unified Style Sheet reduced the stylistic chaos somewhat, it doesn't touch conventions such as "inline data is to be set as X".

These conventions vary over time. It is obviously relevant that in the old days, people used typewriters which have limited abilities to add fancy features (actual italic became possible in the 70's, but required switching typing elements). The current trend is to separate the author's contribution from the physical implementation, by use of "styles" or similar tagging. Thus an inline example would have the attribute "inline_example" or something like that, which might be cashed out as a point-size difference, different font family, bold vs. regular, depending on the current house specification.

One of the things that we've learned from efforts such as Xlingpaper and ELAN is that we still need to better understand our ontology. An example isn't just an example, it's a complex data structure involving numerous levels of grammatical analysis, sociolinguistic properties (such as speaker-identifiers), or corpus-pointers. A given publication almost certainly does not require encoding all of that information, but it may require representing some of it. For instance, for the first time in 40 years, I'm working on a language with such a high degree of variation and speaker-specificity that I have to notate all examples as to who they come from. This information is represented computationally as a tag on data, but will need to be translated into something visual, like superscripted initials -- unless the publisher objects and wants something else.

Typographic conventions that prevail at a given time can vary massively depending on sub-field. A certain complex standard for using underscore, bold, parentheses and braces emerged in Optimality Theoretic studies of reduplication and morphological structure. These happen to conflict with an emerging standard for Bantu linguistics, where braces and parentheses denote specific stem-sized sub-constituents of words and underlining (single and double) denotes the different derivational sources of H tones. The use of the "outline" font feature in OT studies enjoyed some popularity, as a high-tech insider gadget, but seems to have fallen into desuetude.

I doubt that there will ever be a reasonably-complete meta-typographical study of linguistics, covering the centuries of publication that exist and the myriad sub-fields. It might be more efficient to start with some chronological cutoff point. However, that cutoff would probably by about 2020, since use of inline styles and thinking about ontologies in document- and database-creation is not yet common among workaday linguists.