For Want of an Oxford Comma


The minor yet highly controversial issue of the Oxford Comma (or serial comma) arises solely in one very restricted context: what is known in classical grammar as a monosyndetic multiple coordination, where there is just a single coordinator (a word like and or or) before the last of three or more coordinated items. Should you write You need celery, apples, walnuts, and grapes (which has the so-called Oxford comma), or alternatively You need celery, apples, walnuts and grapes?

In binary coordinations like walnuts and grapes no one advocates a comma. And everyone uses commas in asyndetic coordinations like You need celery, apples, walnuts, grapes. In polysyndetic multiple coordination, with the coordinator repeated before each noninitial coordinate, it’s optional: you can either eschew commas (You need celery and apples and walnuts and grapes) or slow things down a bit by adding commas all the way through (You need celery, and apples, and walnuts, and grapes). Only monosyndetic multiple coordinations occasion bitter dispute.

Ideally, it should be the writer’s choice. In rare cases the Oxford comma is best omitted and in others it crucially prevents misparsing (as in the famous imaginary book-dedication inscription “To my parents, Ayn Rand and God” — the atheist novelist and the omnipotent deity are a truly unlikely breeding pair).

Nonetheless, some publishers and newspapers shun Oxford commas. The Economist is one. Its style guide bans them unless either the final or the penultimate coordinate item independently includes a coordinator. But that rigid rule lands them in trouble.

The Johnson column on language in the November 12 edition includes this sentence:

The world has been shaken by the election of Donald Trump to the American presidency because he has been saying things for a year and a half that seem to commit America to radical new policies: an abandonment of NATO allies who do not pay more for their protection, an end to free trade and the killing of terrorists’ families.

In the print edition the words free trade and the killing of terrorists’ families actually occur after a column break, making it even easier to misparse that sequence as a binary syndetic noun-phrase coordination. I read it thus at first, as if abandonment of NATO allies was the first policy and the rest of the sentence defined the second. But the writer’s intent was (of course) to cite three policies, expressed in a monosyndetic multiple coordination.

Of course I knew that Trump opposes free trade but has advocated killing terrorists’ families (which would be a war crime, and is not current policy), so from my own background knowledge I could figure out that I had misread. But that did not prevent the momentary double-take and does not excuse the unclarity!

Johnson may not be responsible: The Economist appears to systematically erase Oxford commas when its inadequate style-guide rule is violated. Robert Ayers pointed out to me an even worse instance (here) two weeks later:

[Italy's constitution of 1948] gave significant autonomy to four culturally distinct regions — Sicily, Sardinia, Valle d’Aosta, bordering France and the partly German-speaking region of Trentino-Alto Adige/Südtirol.

Ayers went back over the passage, baffled. It seemed to promise four regions but deliver only three: (1) Sicily, (2) Sardinia, and (3) Valle d’Aosta (bordering both France and Trentino-Alto Adige/Südtirol).

Fetch out the atlas and you will see that this cannot be right: Valle d’Aosta does not abut Trentino-Alto Adige/Südtirol (parts of Piedmont and Lombardy intervene). But for heaven’s sake, we shouldn’t need to pore over maps to discern what the article is saying!

Besides, there are actually two reasons why there should be a comma after bordering France. Signaling that France and the partly German-speaking region of Trentino-Alto Adige/Südtirol is not to be misparsed as a coordination is important enough, but there is a second, stronger reason: The phrase bordering France is a parenthetical interruption, which is why it has a comma indicating where it begins, so we need another comma to signal its end, and that is obligatory. The style-guide edict banning the Oxford comma has quite wrongly overridden two independent motivations for a comma, and needlessly created ungrammaticality.

Our needs as readers seeking clarity should take priority over rigid enforcement of an arbitrarily stipulated rule of punctuational table manners.

This is not a radical view. Institutions whose style guides recommend the Oxford comma include the American Psychological Association, the American Medical Association, the American Association for Medical Transcription, the Council of Science Editors, the University of Chicago Press, the Modern Language Association, Oxford University Press, and the U.S. Government Printing Office. And in case that’s not enough, let me note that it has additional support from usage-guide authors Wilson Follett and Bryan Garner, and (as if it mattered) even those shameless old weasels Strunk and White.

Return to Top