Just a vote for adding some kind of automation for this feature at least in the SBL citation style. In keeping with examples like the ones that mbrand, robbarrett, and rschellenberg have mentioned. Despite some other disciplines' preferences for shorter or citation-only footnotes, scholarly writing in biblical studies often makes use of lengthy footnotes to discuss in detail multiple references to secondary literature. So, in my writing too, as in the examples that mbrand, robbarrett, and rschellenberg have mentioned, this kind of embedding of an initial reference to a particular work within a larger parenthetical citation in a footnote actually comes up quite often, and so the need to change parentheses out for square brackets according to CMS, 15th ed., §6.102, actually comes up pretty frequently.

bdarcus, if this kind of automation is indeed not possible with CSL, my apologies for understanding that that conclusion had been reached in the preceding discussion. By and large, your first summary seems correct, though perhaps with one meaningful twist.

By way of trying to provide the kind of example that might help, here is a sentence from a footnote on an essay that I'm currently writing:

Photius of Constantinople and Diodorus of Tarsus seem to refer this text to the Father (John Anthony Cramner, Catenæ Græcorum Patrum in Novum Testamentum (Oxford: Oxford University, 1844), 4:162, lines 22–29; J. P. Migne, ed., Patrologia graeca (Paris, 1857), 102:157b–c).

This citation is correctly formatted per CMS (and SBLHS, which defers to CMS at this point) except that:

(Oxford: Oxford University, 1844) . . . (Paris, 1857)

should be:

[Oxford: Oxford University, 1844] . . . [Paris, 1857]

The (LibreOffice) Zotero field text for the citation at the end of the sentence that I've quoted is as follows:

ZOTERO_ITEM CSL_CITATION {"citationID":"cXqlO3pK","properties":{"formattedCitation":"{\\rtf (John Anthony Cramner, \\i Caten\\uc0\\u230{} Gr\\uc0\\u230{}corum Patrum in Novum Testamentum\\i0{} (Oxford: Oxford University, 1844), 4:162, lines 22\\uc0\\u8211{}29; J. P. Migne, ed., \\i Patrologia graeca\\i0{} (Paris, 1857), 102:157b\\uc0\\u8211{}c).}","plainCitation":"(John Anthony Cramner, Catenæ Græcorum Patrum in Novum Testamentum (Oxford: Oxford University, 1844), 4:162, lines 22–29; J. P. Migne, ed., Patrologia graeca (Paris, 1857), 102:157b–c)."},"citationItems":[{"id":1312,"uris":["http://zotero.org/users/2212/items/D2PWUSVU"],"uri":["http://zotero.org/users/2212/items/D2PWUSVU"],"itemData":{"id":1312,"type":"book","title":"Catenæ Græcorum Patrum in Novum Testamentum","publisher":"Oxford University","publisher-place":"Oxford","number-of-volumes":"8","event-place":"Oxford","note":"http://books.google.com/books?id=Yb4UAAAAQAAJ","shortTitle":"Catenæ Græcorum Patrum","author":[{"family":"Cramner","given":"John Anthony"}],"issued":{"year":1844}},"locator":"4:162, lines 22–29","label":"page","prefix":"("},{"id":17236,"uris":["http://zotero.org/users/2212/items/E95J7EMR"],"uri":["http://zotero.org/users/2212/items/E95J7EMR"],"itemData":{"id":17236,"type":"book","title":"Patrologia graeca","publisher-place":"Paris","number-of-volumes":"162","event-place":"Paris","editor":[{"family":"Migne","given":"J. P."}],"issued":{"year":1857,"season":"1886"}},"locator":"102:157b–c","label":"page","suffix":")"}],"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"} RNDmAdHFDC27j

As has been mentioned previously, parentheses that are simply typed into the word processor and not entered in the citation editor shouldn't be something for which we'd expect Zotero to account automatically. But, since I've here specified the enclosing parentheses as a prefix to the first and a suffix to the second (and last) citation (cf. adamsmith above):

"prefix":"(" . . . "suffix":")"

it would be nice if Zotero would translate any medial parentheses into square brackets. Or, since a compound citation of two or more individual sources might be involved (as in my example here), perhaps what's needed is an additional field or option in the citation editor for applying parentheses to the whole compound citation and then, based on this setting, translating what would be medial parentheses, e.g.,

(Oxford: Oxford University, 1844) . . . (Paris, 1857)

into square brackets, e.g.,

[Oxford: Oxford University, 1844] . . . [Paris, 1857]

These medial parentheses can, of course, already be replaced manually in the citation editor, but making those replacements one-at-a-time over several hundred pages of footnote text makes an automation option desirable.

I hope this example is helpful. Thanks so much for the thoughts.

bdarcus, yes, probably a majority or at least a substantial minority of the citations would occur in a footnote without parentheses around the whole citation (cf. the example of adamsmith above).

In your description, you had summarized that:

[S]ometimes, some users want to include such citations in parentheses, in which case the nested parentheses are wrong, and the inner parenthesis should be converted to brackets. The need for brackets or parenthesis thus depends on context, and so cannot be coded in CSL.

If the "context" mentioned here refers to a situation in which parentheses had been typed directly into the word processor and not into a particular reference's Zotero field through the citation editor—e.g., my example without the

"prefix":"(" . . . "suffix":")"

elements—then (to me at least), it's quite understandable that these parentheses implications "c[ould ]not be coded in CSL."

But, if the "context" mentioned here could refer to instances where a user had added the framing parentheses in the prefix/suffix fields or some other option in the citation editor (perhaps akin to how the suppress author function already works), having framing parentheses and translating what would be medial parentheses into square brackets would seem to be something that could be automated. . . . "could be" in the sense that the user has then "told" Zotero about his or her desire to have these framing parentheses. The feasibility of the nuts and bolts of the code required to perform such an operation may be (is?) another matter entirely.

@dstark,

As has been mentioned previously, parentheses that are simply typed into the word processor and not entered in the citation editor shouldn't be something for which we'd expect Zotero to account automatically. But, since I've here specified the enclosing parentheses as a prefix to the first and a suffix to the second (and last) citation ...

I've just checked in a revision to the citation processor that implements parens-to-bracket flipping, along the lines you suggest. It will take awhile (probably weeks, likely not months) for the revised processor to appear in Zotero. To be on the safe side, I've included a switch to make it easy for the Zotero developers to turn it off, if it causes problems for other users (which I don't expect it will, but you never know). So no guarantees yet, but we have something in the works.

fbennett, thanks so much for the update.

thanks - looks good to me.

As far as I know, the code for this is still in the processor. It would only be triggered when the enclosing parentheses are in the Prefix field of the first item in the citation, and the Suffix field of the last. Parens in the word-processor document text would not be recognized.

It's not part of the style and should work across the board. But to replicate any issues you have: which one are you using?

It's apparently set up to switch to brackets only in note styles. I'll loosen that constraint.

Excellent. Thanks for looking into this.

I have a working solution in a revised draft of the processor that handles nesting correctly across the entire citation, to arbitrary nesting depth.

There is one logic/design issue to settle before it's released. Some citation forms contain square braces. Instances I can think of off the top of my head are square-braces years and square-braced paragraph numbers in some legal cites. There may be others.

So the question is whether those should flip as well (bracket to parens) when they are nested. My inclination would be to say "yes," but I'm not 100% sure that's the right call.

Any thoughts? @adamsmith? @rintzezelle?

No worries, it was good to revisit this end of the code. Whether paens flipping is offered or not, we seem to have an issue with the nesting priority of item prefix and layout prefix. Worth further discussion maybe.

That would be possible, but it would obviously need a flag somewhere in the CSL code of the style.

Not sure what to do in the processor for the time being. The flipping code currently takes effect only on parens inside the layout affixes, which is halfway to the goal, so to speak. Could wind out parens flipping completely, or finish it out completely, or leave it as it is.

Is nested bracket rules something that should be specified in locales, with a style override?

As someone who finds the current parens-flipping helpful, I'd advocate for (1). Chicago does in fact request square brackets within parentheses (see below, copied from earlier in the thread -- though not, apparently, vice versa).

If UK styles prefer parentheses within parentheses, the ideal would be to make parens-flipping style-dependent — be able to turn it on or off in the csl code.

If that is not feasible, it seems that one group of users will have to be making manual changes. So then it would be a matter of weighing the greater inconvenience. Unless you're hearing complaints about current behavior, it would seem a shame to me to lose this function, even if it's not perfect.



"The Chicago Manual of Style" 15th edition (2003), p. 266 has:

"6.102 Parentheses within parentheses. Athough the use of parentheses within parentheses (usually for bibliographic purposes) is permitted in some publications -- especially in law -- Chicago prefers brackets within parentheses (see 6.106)."...

and p. 268:

[under section "Brackets, Square Brackets"]:

"6.106. Within Parentheses. Square brackets are used as parentheses within parentheses, usually for bibliographic purposes."...