By authority of Goold Brown, "When more than two words or terms are connected in the same construction, or in a joint dependence on some other term, by conjunctions expressed or understood, the comma should be inserted after every one of them but the last."
Fowler went a step further, insisting on a comma after the final term of an enumeration as well, in order to resolve such ambiguities as he found in the following two quotations.
But the general purport is the same—the blindness, the degrading passions, the short-sighted greed by which the economic unity of Europe has been broken ....If the loss of unity is to be attributed to blindness and passions and not merely to greed, then, Fowler argues, greed must be followed by a comma.
Nothing had been allowed to be published except books, pamphlets, and papers which had secured the approval of the Communist party.Was prior approval required only for papers, or for books and pamphlets as well? This is indeed unclear, but the example perhaps better serves another of Fowler's causes, namely his proscription of the use of which to introduce a restrictive clause. A comma after papers, added to the sentence as written, would lead most readers to take the subordinate clause to be nonrestrictive, thereby compounding the ambiguity. But if both pieces of Fowler's advice were followed, that is, if which were replaced by that with a preceding comma, then this interpretation would not be possible and the problem would be solved.
While Fowler's distinction between that and which has gained fairly wide support (at least among American writers), his liberal use of commas has not. In fact, the modern trend is in the opposite direction, marked by the common omission of the so-called serial comma—the one that belongs between the penultimate term of an enumeration and the conjunction that follows it. As Brown observed, "This is certainly a great error. It gives us such punctuation as comports neither with the sense of three or more words in the same construction, nor with the pauses which they require in reading. ... 'John, James and Thomas are here,' is a sentence which plainly tells John that James and Thomas are here; and which, if read according to this pointing, cannot possibly have any other meaning." Similarly, the apocryphal book dedication "To my parents, Ayn Rand and God" would be found only in the work of a demigod.
Nonetheless, a number of arguments have been put forth in support of this practice. It is popularly held, for example, that the purpose of a comma in an enumeration is to represent an omitted conjunction and that it is therefore redundant when the conjunction is present. Gowers, for example, in yet another corruption of the original Fowler, rejects the serial comma in all cases "because with and it would be otiose."
Another consideration, I am told, is aesthetics. When I attempted to edit a missing comma into a business card that my wife had asked me to proofread, I learned that the principles of graphic design supersede those of punctuation. Apparently, this accounts for the generally accepted practice of omitting the serial comma from names of businesses such as Fish, Richardson and Neave, regardless of context.
A more practical concern is the conservation of space, which is of particular interest to newspapers and other publications that contain narrow columns of text. As an example of the fervor with which copy editors routinely reject the comma, consider the following sentence:
Safire is the author of 14 books on grammar and usage and the author of four novels, the dictionary, The New Language of Politics and an anthology of great speeches, Lend Me Your Ears. *Once we know that The New Language of Politics is a dictionary, it is clear that the title is intended here as appositional phrase. Thus, the deletion of the comma after Politics, undoubtedly the act of an overzealous editor, has the further unfortunate effect of leaving the preceding comma unmatched. The result is a triumph of perversity. Unfortunately, in this case the reinstatement of the comma is not entirely satisfactory, as the reader is left to wonder whether the dictionary and The New Language of Politics are distinct elements of the list. But since the latter phrase is restrictive with respect to the former, the comma that separates them is the one that should have been deleted. Once that is done, the sentence is finally readable.
Some newspaper style guides, such as The Washington Post Deskbook on Style and The Times (of London) Style and Usage Guide, unequivocally forbid any instance of the serial comma. But then, ignoring their own advice, even these publications consistently use it before etc. I hesitate to point this out, however, as a veteran of the war against split infinitives, for fear of being reminded once again that English and Latin are governed by different sets of rules.
Other such guides, including The Associated Press Stylebook and Libel Manual and The New York Times Manual of Style and Usage, allow the comma in certain instances where it is clearly required for clarity. Unfortunately, none of these concessions is sufficient to prevent constructions like the following, which appeared in a New York Times report on a Peter Ustinov documentary:
Highlights of his global tour include encounters with Nelson Mandela, an 800-year-old demigod * and a dildo collector.Mr. Mandela may want to file this description alongside George W. Bush's characterization of him as a "great African-American".
Book publishers, especially university presses, take a very different view of the matter, hence the aliases Oxford comma and Harvard comma. The same is true of most respectable commentators (although I confess that this is one of my benchmarks for respectability), who usually base their defense on the interest of avoiding ambiguity. According to Follett, "Omission always tends to confusion; inclusion can never confuse. Writers who wish for airtight rules will find one here." Garner agrees that the controversy is "easily answered in favor of inclusion because omitting the final comma may cause ambiguities, whereas including it never will—e.g., 'A and B, C and D, E and F[,] and G and H.'"
The flaw in this argument is that its premise, that the comma never creates ambiguity, is simply false. If the book dedication quoted earlier were modified to read "To my mother, Ayn Rand and God", however distracting the punctuation may be, the author's intent to honor three distinct entities would be unmistakable. But if the missing comma were inserted, some extrinsic knowledge would be required to determine whether Rand and the author's mother are the same person. If ambiguity is our primary concern, then should the comma be omitted in this case?
The best defense of the serial comma, as expressed most clearly by Brown, is not that its omission may result in ambiguity, but rather that it necessarily results in error. To the popular press, I say hav it ur way, and I certainly know better than to attempt to infringe on artistic license. But the proper reply to Gowers is that his charge of redundancy betrays a basic misunderstanding of the purpose of punctuation, which Bishop Lowth defined as "the art of marking in writing the feveral paufes, or refts, between fentences, and the parts of fentences, according to their proper quantity or proportion, as they are expressed in a just and accurate pronunciation." In particular, the function of a comma is to separate, while that of a conjunction is to join, and hence their juxtaposition is in no way redundant. The natural reading of an enumeration includes a pause of equal duration after every term except (possibly) the last, each of which demands equal prosodic representation.