From my own experience, one of the main cost factors in editing a book is the enforcement of a uniform text-structure style (I have (co-)edited nine books, some of them big multi-volume works, so I have quite a bit of experience).
I am not talking about esthetic issues of typesetting here (font size, line spacing, style of headings, etc), but about the kinds of things that the authors do when they type a text: use of italics, boldface, quotation marks, in-text citations, bibliographic style, inclusion of abstract, section numbering, etc. This is what I call text-structure style here (in contrast to typesetting style, which will always differ among publications).
As we all know, different journals and publishers have different style requirements, so when submitting a paper to a journal, you are supposed to read and conform to the specific stylesheet of the venue. I don’t know how many people actually do read the stylesheet and take it seriously, but either way, the system is clearly inefficient: If you have to reformat your paper to submit it to some journal, that’s extra work, and if you don’t do it, someone else will have to do it – maybe the journal’s copy-editor: Some publishers still think that they can afford these, and the copy-editors do not only do useful things (such as catching typos), but they also insert “Oxford commas” and make other non-essential or unhelpful changes.
I recently had the unpleasant experience of seeing a paper of mine copy-edited in such a way that it was sometimes difficult to see what I had meant: The copy-editor removed all highlighting of special terms (even though it was a survey article where explanation of key terms was crucial), inserted “example” in front of all example numbers (even though some of the items were not examples in the strict sense), and mutilated author names (leaving only the first letter of the given names) and journal names (e.g. “Linguist. Typol.” for the journal Linguistic Typology). The paper thus conformed to the publisher’s house style, but not at all to what is usual practice in linguistics.
Clearly, having the same text-structure style for all linguistics publications would benefit everyone: authors, publishers and readers. Linguistics students could get used to the general style from early onwards, and there would not be any need to reformat anything.
Would some publishers resist such a move? Perhaps, because some of them think of these things as their area of expertise, and they want to impose a uniform style on all their journals. (For instance, Elsevier has a redundant comma in in-text citations in all their journals, even the linguistics journals, although linguists normally don’t use this comma.)
But ultimately, publishers will do what the community of scholars wants. Some very promising steps toward discipline-wide style rules have already been taken by linguists: Since 2007, there has been a standard for bibstyles (called “Unified Style Sheet for Linguistics”) which has been fairly widely adopted. And for glossed examples, the Leipzig Glossing Rules have been adopted almost universally.
Thus, there are good chances that linguists will also be happy to adopt a set of discipline-wide style rules for other aspects of their texts. One proposal for such a set of rules is called “Generic Style Rules for Linguistics” (download here).
I hope that the Generic Style Rules (or some other set of general rules that is not journal-specific or publisher-specific) will be adopted by more and more journals and book series in the future. This would make our lives easier, and editors would be able to focus on the content of the articles and books. Especially for smaller, low-budget journals (of which there are more and more, generally in open access), this would be a significant improvement.
Of course, it will be difficult to agree on every individual point, but fortunately, there is already fairly wide agreement in the styles of the major linguistics publishers De Gruyter and Benjamins, and the Generic Style Rules can be seen as a compromise between these. Some people may want to regulate even more things (such as the use or non-use of serial commas, the date format, the use of hyphens and capitalization of compass points), but it seems to me that in highly specialized publications such as scholarly articles and books in linguistics, we don’t need more regulation than the 14 pages of the Generic Style Rules. Everything else can be left to the author.