The true history of the split verb rule

« previous post | next post »

The "split verb rule" says that an adverb must not be placed between an auxiliary and the following verb. On this account, you should never write "you should never write", but rather "you never should write". In an earlier post, I followed (what I thought was) the lead of James Lindgren ("Fear of Writing", California Law Review 78(6):1677-1702, 1990) in attributing this bizarre idea to The Texas Law Review Manual on Style. But in a comment this evening, Jon Weinberg cited Allen Black, "Judge Wisdom, the Great Teacher and Careful Writer", 109 Yale L.J. 1267 (1999-2000):

He was death on split infinitives and split verbs. A sentence such as "The burdened vessel was slowly proceeding down river at the time of the collision" would never survive.

Since John Minor Wisdom would have learned his attitudes towards such things in the 1920s, and the Texas Law Review's Manual does not seem to have appeared until the 1950s, Jon suggested that we need to look elsewhere for the source of this peculiar prejudice. And indeed, a quick Google Books search turns up a more promising source — Henry Watson Fowler and Francis George Fowler, The King's English, 1908, section 46 on "'Split' Auxiliaries":

Some writers, holding that there is the same objection to split compound verbs as to split infinitives, prefer to place any adverb or qualifying phrase not between the auxiliary and the other component, but before both.

The Fowler brothers don't specify who "some writers" are, and Google Books turns up no earlier hits for "split auxiliaries" or "split verbs". Perhaps some readers can track this one down further.

In any case, the Fowler brothers reject the rejection of "split auxiliaries", instead making some interesting remarks about interaction with information structure and stress:

Provided that the adverb is then separated from the auxiliary, no harm is done : 'Evidently he was mistaken' is often as good as 'He was evidently mistaken', and suits all requirements of accentuation. But the placing of the adverb immediately before or after the auxiliary depends, according to established usage, upon the relative importance of the two components. When the main accent is to fall upon the second component, the normal place of the adverb is between the two; it is only when the same verb is repeated with a change in the tense or mood of the auxiliary, that the adverb should come first. 'He evidently was deceived' implies, or should imply, that the verb deceived has been used before, and that the point of the sentence depends upon the emphatic auxiliary ; accordingly we should write 'The possibility of his being deceived had never occurred to me ; but he evidently was deceived', but 'I relied implicitly on his knowledge of the facts ; but he was evidently deceived'.

They then make free to decide that Burke, Beaconsfield, Ruskin, Charlotte Bronte, and the London Times all often did it wrong — because for the Fowlers, placing an adverb between an auxiliary and the following verbal material is not only not forbidden, it's actually obligatory, except in certain circumstances:

In our first two examples below the adverb is rightly placed first to secure the emphasis on the auxiliary : in all the others the above principle of accentuation is violated. The same order of words is required by the copula with whatever kind of complement.

I recognize this truth, and always have recognized it.
Refined policy ever has been the parent of confusion, and ever will be so, as long as the world endures. — Burke.
They never are suffered to succeed in their opposition. — Burke.
She had received the homage of . . . and occasionally had deigned to breathe forth . . .— Beaconsfield.
He ordered breakfast as calmly as if he never had left his home.— Beaconsfield.
Miss Becky, whose sympathetic powers never had been called into action before. — Ferrier.
They now were bent on taking the work into their own hands.— Morley.
There may have been a time when a king was a god, but he now is pretty much on a level with his subjects. — Jowett.
They both are contradicted by all positive evidence. — W. H. Mallock.
Religious art at once complete and sincere never yet has existed. — Ruskin.
Not mere empty ideas, but what were once realities, and that I long have thought decayed. — C. Bronte.
So that he might assist at a Bible class, from which he never had been absent .— Beaconsfield.
If we would write an essay, we necessarily must have something to say. — Bygott & Jones.
The protectionists lately have been affirming that the autumn session will be devoted to railway questions. — Times.
Visitors no longer can drive in open carriages along the littoral. — Times.
It still is the fact that his mind . . . was essentially the mind of a poet. — Times.
To whom in any case its style would have not appealed. — Times.

To go wrong with not is an achievement possible only with triple compounds, where the principal division is of course between the finite (would) and the infinitive with participle (have appealed). ' Would not have appealed ' must be written, though at an enormous sacrifice of 'distinction'.

This enhanced value of old English silver may be due partly to the increase in the number of collectors ; but it also has been largely influenced by the publication . . . — Times.
Mr. Fry showed to a very great extent his power of defence . . . To-day, if runs are to be of importance, he very likely will show his powers of hitting.— Times.

This is a canonical case of a self-appointed authority inventing a grammatical theory, observing that elite writers routinely violate the theory, and concluding not that the theory is wrong or incomplete, but that the writers are in error. Still, we're not talking about the grammar of English, but about the history of mistaken ideas about the grammar of English. And in this case, the Fowlers' argument is against a "split verb rule" rather than for it.

So this leaves us with two possibilities. Perhaps the "split verb rule" started with the Fowlers' "some writers", whoever they were, and spread through the legal profession despite the Fowlers' best efforts to quash it. Alternatively, a generation of spectacularly careless readers of The King's English understood as a "rule" a pattern that the Fowlers explicitly rejected as a violation of the "principle of accentuation".

Update — Compare Austin Osgood Hubbard, Elements of English Grammar (1827), whose Rule 16th, Obs. 2 asserts that

Those adverbs which qualify compound verbs, are usually placed after the auxiliary.

Or the section "Of the syntax of adverbs" in A Complete English Grammar, Compiled from Louth, Johnson, L. Murray, Cobbett, &c., with Additions, Corrections, and Improvements (1836), which begins

Adverbs, though they have no government of case, tense, &c., require an appropriate situation in the sentence, viz., for the most part before adjectives, after verbs active or neuter, and frequently between the auxiliary and the verb; as, "He made a very sensible discourse ;" "He spoke unaffectedly and forcibly, and was attentively heard by the whole assembly."

Or Brandon Turner, A New English Grammar (1840), which observes that

For the placing of adverbs, no definite general rule can be given. Those which relate to adjectives, immediately precede them; and those which belong to compound verbs, are commonly place after the first auxiliary.

Or many other sources from the first half of the 19th century, which uniformly and straightforwardly assume that the common practice, then and now, is also the correct practice. It would be interesting to understand why and how this sensible discourse was transmuted into the absurd "split verb rule", accepted a century later as gospel by someone as intelligent as John Minor Wisdom seems to have been.

MWDEU notes that

Copperud 1970, 1980 talks about an erroneous idea widespread among newspaper journalists that adverbs should not separate auxiliaries from their main verbs (as in "you can easily see" or "they must be heartily congratulated"). This bugaboo, commentators agree, seems to have sprung from fear of the dread split infinitive. Copperud cites five commentators on the subject, all of whom see no harm in placing an adverb between the parts of a verb, and one of whom (Fowler 1965) prescribes such placement. […] Since dividing the auxiliary from the verb with an adverb has been approved at least since Lindley Murray 1795, it would seem that Fowler is justified in calling the avoidance a superstition.

In his various works on English usage, Bryan Garner agrees, calling the split verb rule "nonsense" and citing various authorities in favor of normal adverb placement. But the only clue that he gives us about the source of the nonsense is an indirect allusion to the Texas Law Review Manual of Style:

A fairly well-known manual on legal style long cautioned its readers to avoid splitting verb phrases with adverbs …

So perhaps the split verb rule was purely a folk superstition, elevated to a point of elite principle in the early 20th century by some now-forgotten law professor in Texas or Louisiana. But it seems unlikely that the Fowler brothers' "some writers" came from the U.S. gulf coast. The history remains a mystery.


  1. Emily said,

    December 23, 2012 @ 10:08 pm

    I believe I remember reading about the origin of the rule in either Crystal's Stories of English or The Fight for English–if I recall correctly he had traced it to some sort of vicar or clerk's grammar book from either the 1880s or 1890s. I don't have the books handy, or I'd hunt down the exact quotation for you.

  2. J.W. Brewer said,

    December 24, 2012 @ 10:48 am

    Even if the Tex. L. Rev. Manual didn't actually originate this particular zombie rule, it probably played a culpable role in spreading it. I think I may be able to claim credit for the first pejorative reference to that particularly dreadful manual on this site: (NB: Bryan Garner is himself an alumnus of UT-Austin's law school, where he was an editor of the Tex. L. Rev., so he may feel a filial obligation to avoid naming the manual when criticizing it.)

  3. Ralph Hickok said,

    December 24, 2012 @ 11:55 am

    In the quote from Austin Osgood Hubbard, "quality" should be "qualify".

    [(myl) Thanks — fixed now.]

  4. Chris said,

    December 24, 2012 @ 1:07 pm

    As is the case for your previous discussions of this issue, every follower of this zombie rule that I've encountered has sincerely believed that violating it constitutes "splitting infinitives."

    I did manage to break one of my writers of the habit, but only by issuing a direct order. She still thinks it's wrong.

  5. Arrant Pedantry » Blog Archive » Relative Pronoun Redux said,

    December 24, 2012 @ 1:27 pm

    […] then has been a rationalization to continue to support a flawed rule. Mark Liberman said it well on Language Log yesterday: This is a canonical case of a self-appointed authority inventing a grammatical theory, observing […]

  6. Andy Averill said,

    December 24, 2012 @ 2:43 pm

    In any event, isn't "mistaken" an adjective in "he was evidently mistaken"? In which case it's not a very good example to illustrate a verb with an auxiliary.

    "Mistaken" would be a verb in, eg, "he was mistaken so often for his twin brother that he grew a mustache".

  7. Hitchcock said,

    December 24, 2012 @ 3:44 pm

    Here are two early 20th-century references to the no-verb-splitting "rule," the second of which seems to parallel the Fowlers' point of view (and treat it as an extension of the split-infinitives case):

    First: Sara Elizabeth Husted Lockwood & Mary Alice Emerson, Composition & Rhetoric for Higher Schools (1904), section 209, item 6, p. 293: "An adverb should not separate the parts of a verb phrase if it can be avoided."

    Lockwood & Emerson's first example is a split infinitive, but their second example makes it clear that they're talking about the larger question of splitting verb phrases generally & not just infinitives:

    "What has never been said has never has to be recalled.
    "Better: What has never been said never has to be recalled."

    (Google books:

    Second: Mary Hall Leonard, Grammar and Its Reasons (1907), p. 217, in a paragraph immediately following a reference to the "rule" against split infinitives:

    "Some critics have also objected to the placing of the adverb between the parts of a compound tense, preferring "probably will go," "has searched carefully," etc., to "will probably go" and "has carefully searched." It is frequently better that the adverb should precede or follow the entire phrase, but there are many instances in which the middle of the verb phrase seems to be the required place for the adverb."

    Google books:

  8. Steve Treuer said,

    December 24, 2012 @ 7:15 pm

    William Cobbett quotes this "specimen of false grammar" from Samuel Johnson: "I hope not much to tire those whom I shall not happen to please." Cobbett then argues that "not much" belongs between the verbs in preferred sentence: "He did not mean he did not much hope, but that he hoped not to tire much. 'I hope I shall not much tire those whom I may not happen to please.' This was what he meant; but he does not say it."
    1923 edition of A Grammar of the English Language, originally published in 1819. It fits one of his themes:"The station, or place, of the adverb is a great matter." He has a whole section on the "wrong placing of words."

  9. Andy Averill said,

    December 24, 2012 @ 9:07 pm

    @Hitchcock, the example from Lockwood & Emerson is a little confusing. Isn't "has been said" a verb phrase? But they approve of sticking never between has and been.

  10. Hitchcock said,

    December 25, 2012 @ 1:15 am

    @Andy Averill: My mistake in transcribing the example. The "Better" version should read "What never has been said never has to be recalled."

  11. Sven said,

    December 26, 2012 @ 11:35 am

    If this "rule" is popular with lawyers, that may explain why CJ Roberts messed up the presidential oath of office – it was against his religion to say "will faithfully execute".

    [(myl) Indeed.]

  12. J.W. Brewer said,

    December 26, 2012 @ 3:02 pm

    One could (given more time than I have today) rummage through the massive corpus of Judge Wisdom's published opinions to see if he managed to follow the taboo in practice as strictly as his former clerk (writing through the haze of the passage of decades) recalls him advocating in theory. (Tulane has his papers, which probably contain less edited types of prose which might or might follow the same norms as his polished-for-publication formal opinions, although I doubt those papers have been digitized and put online to facilitate easy corpus-linguistics work.)

    I have fwiw been told by someone or other that two of the present nine members of the Supreme Court seem to follow the no-split-infinitives shibboleth fairly strictly (i.e. it is allegedly difficult-to-impossible to find examples of "violations" in their published opinions because they have either internalized the norm or at least rigorously catch "violations" in the editing process) whereas the other seven do split infinitives from time to time as circumstances or whimsy may dictate. Chief Justice Roberts is not alleged by my source to be one of the two alleged sticklers on this point. It certainly remains possible that even if Chief Justice Roberts is not himself a conscious adherent to or proponent of the no-split-verbs zombie-rule he might nonetheless have had enough encounters with True Believers early in his career to occasionally still be left in a state of Nervous Cluelessness(tm) when confronted by a text which does not conform to that zombie-rule.

    [(myl) I don't have time for a systematic Justice-wise count either, but I observe that an overall search of the SCOTUS materials at turns up 761 instances of "have consistently held" to 7 for "consistently have held" and 4 for "have held consistently".

    While we're at it, there seem to be 3,690 hits for "one which", vs. 4,010 for "one that". These are reasonable (though noisy) proxy searches for restrictive relative clauses, suggesting that the justices have historically flipped an approximately fair coin when it comes to the which/that choice. ]

  13. Nathan Myers said,

    December 26, 2012 @ 8:48 pm

    @Steve Treuer

    William Cobbett does not seem to be saying that Samuel Johnson's grammar was wrong, as such, but only that what Johnson wrote did not plausibly express accurately what he felt. I.e., the meaning unambiguously implied by Johnson's chosen word order was not what Cobbett (reasonably) concluded Johnson must have meant. Re-expressing it with the same words, and accurately, the word order comes out different.

    This is similar to criticisms of "All that glitters is not gold", or Tolkein's "All that is gold does not glitter". There's nothing technically ungrammatical there, but the meaning the grammar suggests is nonsense, leaving us to construct a sensible interpretation by shifting word associations around.

    Was Johnson (like Tolkein) being puckish, or was it so common that people would have no difficulty extracting the intended meaning? Was there ever a puckish phase to this practice? Did the Romans have a name for it?

  14. J.W. Brewer said,

    December 27, 2012 @ 7:11 pm

    Interestingly enough, one can find the precise "split-verb" sequence of words "was slowly proceeding" that was deprecated above in an actual federal judicial decision involving the same context as that example (a collision between two vessels on a river near the Gulf Coast), albeit a decision issued before Judge Wisdom was born. It's in The Mary S. Blees, 120 F. 44 (S.D. Ala. 1902). The opinion was authored by the Hon. Harry Theophilus Toulmin (1838-1916).

  15. E W Gilman said,

    December 28, 2012 @ 11:59 am

    My superficial reading suggests to me that the examples of Johnson and the writers the Fowlers listed that the nonuse of the split verb was an 18th century idiom that fell into disuse (and disrepute) in the 19th.

  16. Lane said,

    January 1, 2013 @ 12:10 pm

    Mark, JW, in my debate with Bryan Garner, Garner noted that he changed his co-author Scalia's restrictive "whiches" to "thats," with Scalia's slightly grudging acquiescence. Garner:

    My most recent writing on “that” vs. “which” appears in “Reading Law: The Interpretation of Legal Texts.” My co-author, Justice Antonin Scalia, softened my words there because he sometimes (when I’m not around) uses “which” restrictively. When I tell him that’s a literary failing, he harrumphs. Fortunately, he has allowed me, in both our books, to change all his restrictive “whiches” to “thats.” It makes the style so much better.

    That surprised me, given that Scalia's word is literally law… But Garner has softened his own position, now considering restrictive "which" more of a style gaffe than an error.

  17. Richard C. Reynolds said,

    January 20, 2014 @ 1:15 pm

    I think about the last source I would check for grammatical precision would be a journal written by lawyers, who are notorious for wordiness and abstract language, and for whose legal work the rules of grammar are often suspended, as a grammatical error is not cause for the negation of legality of any document, including an indictment. Nevertheless, I think the main point here is that when an adverb is not necessary, it should probably be eliminated. Most adverbs are superfluous and are used as false intensifiers, as is the case with words such as "actually," "literally," and "absolutely," to name three that are in the common parlance. A sentence can be strengthen or intensified without resorting to adverbial enhancement by rendering it as a declarative statement. But to go to the heart of the matter, anyone who is familiar with the old system of sentence diagramming will be able to see how a split construction poses a significant problem. In a phrase, for example, such as "can actually do," the question is whether the adverb modifies the axillary verb or the main verb? It cannot modify both in that position; but if it's placed properly–"actually can do"–then it modifies the entire verb phrase; if it's dropped–"can do"–then the declaration is solid and definite.

RSS feed for comments on this post