Citation Style Language

Proposal: new CSL variable multivolume-title (or similar) rather than volume-title

I’d like to invite everyone to reconsider: I have come to the conclusion
that introducing volume-title (see is not the
optimal solution.

Reason: For a book (a single volume that is part of a multivolume
monograph) with a title and a volume-title, the title – quite unlike
all other titles – will not be the title of the smallest, most specific
unit involved. Also, title-short will no longer be a convenient, i.e.,
specific shorthand to refer to this item – the shortest unique from will
always have to include a volume number, e.g. “Collected Works, Vol. 3”.

My revised proposal is to introduce a new CSL variable for holding the
title of the multivolume monograph (e.g. “Collected Works”), and keep the
CSL title variable for the the title of the single volume (e.g.,
Tragedies”). Assuming we call the new CSL variable collective-title,
as many library catalogues do, this is what it would look like in CSL:

  • title (e.g. “Tragedies”)
  • collective-title (e.g., “Collected Works”)
  • collection-title (e.g., “Oxbridge Classical Texts”)

and for chapters:

  • title (e.g. “Macbeth”)
  • container-title (e.g. “Tragedies”)
  • collective-title (e.g., “Collected Works”)
  • collection-title (e.g., “Oxbridge Classical Texts”)

(collective-title of course is very similar to collection-title; other
candidates are main-title, or multivolume-title.)

This, by the way, matches the approach taken by biblatex, which has, for

  • title (e.g. “Tragedies”)
  • maintitle (e.g., “Collected Works”)
  • series (e.g., “Oxbridge Classical Texts”)

and for chapters:

  • title (e.g. “Macbeth”)
  • booktitle (e.g. “Tragedies”)
  • maintitle (e.g., “Collected Works”)
  • series (e.g., “Oxbridge Classical Texts”)

I would still introduce issue-title, as originally proposed, for journal
articles. A whole issue, such as a special issue, however, again should
have its title in the title variable, rather than in issue-title.

This would always be the hierarchy, from specific to broad? Do you have any
links to library catalog entries on hand, that demonstrate the point?


I don’t have a clear opinion on this yet, but two downsides that come to

  1. It will make introducing this in styles much more code intensive:
    instead of just adding the volume-title to the citation, we’d have to test
    for the presence of volume title and if it exists, reverse the order of
    elements in the style.
    This may not become entirely clear with the Macbeth example, but if you
    look at the CMoS example that motivated this:
    Pelikan, Jaroslav. The Christian Tradition: A History of the Development of
    Doctrine. Vol. 1, The Emergence of the Catholic Tradition (100–600).
    Chicago: University of Chicago Press, 1971.

“The Christian Tradition: A History of the Development of Doctrine” would
now be the collective title
which leads to

  1. It would require all upstream users to re-enter (rather than complement)
    their data entry (moving the title to volume title).

I’m particularly concerned about 1. I don’t have a strong opinion on this
either way, but let’s at least be aware of that.


Downsides: fair enough. But the Chicago Manual, 16e, 14.124, actually lists
two formats, and for the second one – the one I’d favour –, your first
concern would not apply.

Pelikan, Jaroslav. The Christian Tradition: A History of the Development
of Doctrine.
Vol. 1, The Emergence of the Catholic Tradition (100–600).
Chicago: University of Chicago Press, 1971.
Pelikan, Jaroslav. The Emergence of the Catholic Tradition (100–600).
Vol. 1 of The Christian Tradition: A History of the Development of
Chicago: University of Chicago Press, 1971.

What I am worried about most, by contrast, is sorting, and the handling of
short titles. For one of the Knuth books, e.g.,

Knuth, Donald E. 1986. METAFONT: The Program. Vol. D of Computers &
. Reading, Mass.: Addison-Wesley.

we’d certainly want “METAFONT” as the short title in notes, and I don’t
really see how to do this without even worse complications in the code if
we do not use “METAFONT: The Program” as the title and “METAFONT” as the
short title.

For the British Library’s definition of “collective-title” (“An inclusive
title for an item containing several works or a title used to collocate the
publications of an author, composer, or corporate body containing several
works, or extracts etc. from several works, e.g. Complete works.”), see
For one example of a library catalogue using “title” and
"collective-title", see

Finally, on “This would always be the hierarchy, from specific to broad?” –
Yes, from a systematic point of view this would be best. This also seems to
have been the consensus when I asked the MODS mailing list,
MODS@LISTSERV.LOC.GOV, mostly frequented by librarians, how to best set up
such title hierarchies in MODS.

fair enough re: Chicago. Could we maybe check a couple more? Maybe APA,
MLA, NLM, and MHRA – I also just want to make sure we’re covered.

Just quickly - am not entirely clear on collection vs collective
distinction. Other than one being a noun and the other an adjective, they
seem the same idea. A periodical title, for example, is a collection title
in the CSL logic.

a periodical title is a container-title, at least currently it is.
As for the distinction, it’s the distinction between a multi-volume work
(the proposed collective-title) and a series (the current
collection-title), which can and do exist coexist.

We can argue about naming of the variable, but there are only so many words

I’m just asking that the distinction be clear.

you mean conceptually or in the variable name? Because conceptually I think
it is. If not, it would help if you could expand on how you think it’s not.

In terms of the variable name, the idea would be (and I’m assuming here,
Nick can correct me) that one is “the title of a collection of works” (such
as a book series), while the other is “the collective title of several
works that make up a whole” (such as the title of a multi-volume work).

I don’t particularly care about the naming, to be honest, as long as it
makes some type of sense. People will have to look at specs and/or mapping
charts in the end anyway.

APA has a style more similar to the first format for Chicago:

Zedeck, S. (Ed.). (2011). /APA handbook of industrial and organizational
psychology: Vol. 1. Building and developing the organization/. Washington,
DC: American Psychological Association.

Schmitt, N. W., & Highhouse, S. (Eds.). (2012). /Handbook of psychology:
Vol. 12. Industrial and organizational psychology /(2nd ed.). Hoboken, NJ:
John Wiley & Sons.

In general form:

Editors. (Year). /Collective title: Vol. number. Volume title./ Place:
Publisher. doi

The thing that I would want to cite or look for when thinking about these
works is “The Handbook of Industrial and Organizational Psychology”, not
"Building and developing the organization". It would be counterintuitive for
me (and I imagine most psychologists) to have to store “Building and
developing the organization” in the title field. It would make scanning the
library and finding the item difficult. It would also mean having to rewrite
a lot of translators for psychology sources, where “The Handbook of
Industrial and Organizational Psychology” would be offered in the metadata
as the title.

In short, I would be really opposed to Nick Bart’s proposal.–
View this message in context:
Sent from the xbiblio-devel mailing list archive at

Thank you for your comment. In fact, I’m not wedded to the
‘collective-title’ proposal, ‘volume-title’ would solve the problem of the
one missing title variable as well. I’m probably more influenced by
blblatex, the British Library, and MODS, and tend to see the individual
volume as the most relevant unit. I’d also find it a little inconvenient
having twelve or more ‘Handbook of psychology’ entries in the middle pane,
and having to add either volume or volume-title columns to be able to tell
them apart. ‘volume-title’ on the other hand has the advantage of
MLZ/citeproc-js and pandoc-citeproc (and MLZ and some inofficial CSL
styles) supporting it since years.

So, if a majority on this forum prefers ‘volume-title’, I certainly won’t
stand in the way.

I’d find it much more important to reach a decision at all, and adopt
either ‘collective-title’ or ‘volume-title’ (but of course not both) as
official CSL variables, so they can – eventually – be used in official CSL