OttoBib → Why Citation Styles Disagree

Why citation styles disagree

MLA, APA, Chicago and Vancouver disagree because their disciplines cite different kinds of thing. Almost every surface difference falls out of that.

Asked why MLA puts a page number in the parenthesis and APA puts a year, most guides answer that this is simply the convention in that field. That is true and it explains nothing.

The styles disagree because their disciplines cite different kinds of thing. Almost every surface difference (what goes in the parenthesis, whether there are footnotes, whether the date is early or late) falls out of that one fact.

What each field is pointing at

Literary scholarship cites pages of fixed texts. The object of study is the text itself, and its date is either irrelevant or misleading: Moby-Dick is 1851, but the copy in your hand is a 1988 scholarly edition with different pagination and a different last chapter of apparatus. What the reader needs is the location — which page, which line. So MLA gives you (Melville 42) and spends its parenthesis on position rather than year.

Experimental psychology cites datable findings, where age is part of the claim. A 1962 result and a 2022 result are different kinds of evidence about the same question, and a reader has to know which one is being leaned on without turning to the back of the paper. So APA spends the parenthesis on the year: (Smith, 2019).

History cites archives. A letter from Jefferson to Adams dated 3 May 1812, held in a named collection at a named society, has no author-date shape at all — there is no volume, no publisher, and often no title. It needs a sentence. It also frequently needs a remark: this letter contradicts the one from the following week, and the disagreement is the point. A parenthesis cannot hold that; a footnote can. Chicago's notes system exists because the humanities argue in the margin.

Medicine cites densely. A clinical review may carry sixty references, several to a paragraph, and (Smith, 2019; Jones, 2020; Patel, 2021) three times a page would push the prose off the edge. So the Vancouver style, agreed by a handful of journal editors who met in that city in 1978 and now maintained by the ICMJE, compresses each to a bracketed number. The cost is total: [14] tells you nothing. You cannot see whether it is a 1970 case report or a trial published last year without leaving the sentence.

Four different objects, four different parentheses. None of them is a style decision in the sense of taste.

They started as printers' documents

The manuals now read like standards bodies. They did not begin that way, and their origins show through.

The University of Chicago Press opened in 1891 with its own composing room, setting scientific notation and fonts as awkward as Hebrew and Ethiopic. A proofreader wrote a single sheet of typographic fundamentals to keep the shop consistent. The sheet became a pamphlet and in 1906 the pamphlet became a book, published under a title that leaves no doubt what it was for: Manual of Style: Being a compilation of the typographical rules in force at the University of Chicago Press. Two hundred pages, fifty cents.

APA style began in 1929 as a seven-page article in Psychological Bulletin, written by a committee of psychologists, anthropologists and journal business managers. The presence of the business managers is a fair indication of what problem was being solved. The authors described what they were proposing as a standard of procedure to which exceptions would doubtless be necessary. The first full Publication Manual came in 1952, as a sixty-one-page supplement to the same journal.

MLA's first style sheet appeared in 1951: thirty-four pages, printed inside an issue of PMLA rather than sold separately. The freestanding MLA Handbook, aimed at students rather than at scholars preparing manuscripts, did not arrive until 1977.

All three began as instructions for getting a manuscript through a print shop. The apparatus that now gets graded is downstream of a typesetting problem.

The rule-per-medium trap, and the one real fix

For most of a century the manuals worked by template: a pattern for books, another for journal articles, another for a chapter in an edited volume. Then film, then broadcast, then the web, then a rule for a tweet and a rule for a podcast episode and a rule for a video with no named author. Every new medium needed a new entry, and the manual was permanently one format behind whatever people were actually citing.

MLA's eighth edition, in 2016, abandoned the approach outright. In its place is a list of nine elements in fixed order (author, title of source, title of container, other contributors, version, number, publisher, publication date, location) with two rules: omit what does not apply, and apply the list recursively. A story sits in an anthology; an episode sits in a series which sits in a streaming service. Each nesting is another container, described by the same nine slots.

This is the most substantial idea in modern citation practice, and it is not really a style change. It is the difference between a list of examples and a description of structure. A template collection has to be amended every time the world produces a new kind of object; a recursive element list does not.

What the identifier took away

Look at what a traditional reference entry is doing and most of it is retrieval instructions. Place of publication, publisher, edition statement, volume and issue: these tell a reader how to walk to a shelf and come back with the right physical object. That was a real problem, and the entry solved it.

It is not the problem any more. A DOI resolves to one article. An ISBN identifies one edition of one book, and identifies it more precisely than "New York: Random House, 2004" ever did, because two different books can share that description and cannot share an ISBN.

APA conceded the point in its seventh edition, in 2020, by deleting publisher location from the reference entry altogether. The city was there because you might once have needed to know which office to order from. Nobody has needed that for a long time, and it took until 2020 for a manual to say so.

The same edition raised the author cut-off from six to twenty before the list is elided, and made singular they the required generic pronoun. Both are the manual catching up with practice rather than leading it.

What is left is judgment, not retrieval

Strip out everything an identifier now does better and a citation still has two jobs it cannot delegate. It has to let a reader judge a source at the moment the claim is made, meaning who wrote it, when, and where it appeared. And it has to give credit.

Judged on the first job, the author-date parenthesis is the strongest of the four. Author and year at the point of use is the most information any of these styles delivers without making the reader leave the sentence, and it is why the convention spread well beyond psychology. Numbered styles are cheaper to write and strictly worse to read; they are a reasonable trade only at the citation density that produced them.

Judged on the second, the conventions are weaker than they look. APA's initials-only names were a typesetting economy from a period when column inches were the binding constraint, and they have the side effect of obscuring who did a piece of work. That is a real cost, and no longer a necessary one.

There is one more thing worth saying plainly and then leaving alone: the mechanical parts of formatting are graded, and the labor of complying falls hardest on the people with least practice at it. Getting a comma in the right place is not a proxy for the quality of an argument, and it is treated as one.

Which is best

One book, four styles. Thomas Kuhn's The Structure of Scientific Revolutions, in the University of Chicago Press printing of 2012:

MLA        Kuhn, Thomas S. The Structure of Scientific Revolutions.
           University of Chicago Press, 2012.
APA        Kuhn, T. S. (2012). The structure of scientific revolutions.
           University of Chicago Press.
Chicago    Kuhn, Thomas S. The Structure of Scientific Revolutions.
           Chicago: University of Chicago Press, 2012.
Vancouver  1. Kuhn TS. The structure of scientific revolutions.
           Chicago: University of Chicago Press; 2012.

The entries are nearly the same. The real divergence is at the point of use, which is where the reader actually is:

MLA        (Kuhn 64)
APA        (Kuhn, 2012)
Chicago    a footnote, which can also say what it thinks
Vancouver  [1]

Note what the year is doing here, because this book makes the problem visible. Kuhn published in 1962. The object in your hands says 2012. APA's parenthesis reports the printing you consulted, which is exactly wrong for placing the argument in its own history, and exactly right for telling a reader which pagination you used. MLA sidesteps it by not printing a year at all. Neither is a mistake; they are answering different questions.

If you are choosing for yourself rather than following an instruction, the honest answer is that the best arrangement is not one of the four. It is the author-date parenthesis for what the reader sees, MLA's container model for the structure underneath, a persistent identifier doing all the retrieval work, and full names rather than initials for credit. Written out, it looks like this:

Kuhn, Thomas S. (1962; this printing 2012). The Structure of Scientific
  Revolutions. University of Chicago Press. ISBN 9780226458113.

Every field there is doing a job. The full first name credits a person rather than an initial. Both years are present because both are true and they answer different questions. The ISBN identifies this printing exactly, so the city and the edition statement have nothing left to contribute and are gone. Chicago's author-date system is the nearest thing anyone actually ships.

In practice the argument has been quietly settled by software. Citation Style Language is an XML format that describes how to render a reference, and the Zotero style repository holds well over ten thousand styles written in it. The consequence is that the same underlying record, the same author and year and title and identifier, renders to MLA or APA or Chicago or a single journal's house variant without anyone retyping anything. Hand-formatting a bibliography is now a thing people do because they were told to, not because it is how it gets done.

Which makes the interesting question not which style but which facts. Get the author, the year, the title, the edition and the identifier right, and the style is a rendering decision you can change afterwards in a dropdown. Get them wrong and no amount of correct punctuation will save the entry.

If you want the mechanical comparison rather than the argument, the five-style guide sets the same book out in each. If you have a citation you did not write yourself and want to know whether it describes a real book, that is a different test.