When I split an infinitive, God damn it [...] it will stay split

« previous post | next post »

In the spirit of Geoff Pullum's lyrical prescriptive poppycock offering, I can offer some Raymond Chandler in verse and letter. And this being Language Log, I will follow it with a light dessert of cheap science. Here's a small sample of Chandler's 1947 poem Lines to a Lady With an Unsplit Infinitive for your edification:

There ain't no grammar that equals a hammer
To nail down a cut-rate wit.

And the verb 'to be' as employed by me
Is often and lightly split.

A lot of my style (so-called) is vile
For I learned to write in a bar.

The marriage of thought to words was wrought
With many a strong sidecar.

The poem was at the expense of an overzealous proofreader at the Atlantic Monthly, and has been making the blog-and-other-social-media rounds for a few months, after Letters of Note picked up on it — see the full post here. And here's the letter to the magazine's editor which the poem accompanied (text taken from Letters of Note). Be careful, as it might cut through the glass on your ipad:

6005 Camino de la Costa La, Jolla, California Jan. 18th, 1947

Dear Mr. Weeks:

I'm afraid you've thrown me for a loss. I thought "Juju Worship in Hollywood" was a perfectly good title. I don't see why it has to be linked up with crime and mystery. But you're the Boss. When I wrote about writers this did not occur to you. I've thought of various titles such as Bank Night in Hollywood, Sutter's Last Stand, The Golden Peepshow, All it Needs is Elephants, The Hot Shop Handicap, Where Vaudeville Went it Died, and rot like that. But nothing that smacks you in the kisser. By the way, would you convey my compliments to the purist who reads your proofs and tell him or her that I write in a sort of broken-down patois which is something like the way a Swiss waiter talks, and that when I split an infinitive, God damn it, I split it so it will stay split, and when I interrupt the velvety smoothness of my more or less literate syntax with a few sudden words of barroom vernacular, this is done with the eyes wide open and the mind relaxed but attentive. The method may not be perfect, but it is all I have. I think your proofreader is kindly attempting to steady me on my feet, but much as I appreciate the solicitude, I am really able to steer a fairly clear course, provided I get both sidewalks and the street between.

If I think of anything, I'll wire you.

Kindest Regards,


Of course, Language Log not just a place for acerbic poetry and prose takedowns of prescriptivists. So here is the cheap science I promised,  a Google n-gram graph showing a broad rise over the last 60 years in the rate of split infinitives in books (red and blue lines) as against a fall in competing un-split forms: split infinitive rise

I used trigrams like "him to always", as in "I wanted him to always love me". The graph starts in 1950, and the green and yellow lines show a decrease (as a % of all trigrams) in the use of the trigrams "him always to" and "him quickly to", respectively, through the second half of the twentieth century. The decline was apparently arrested as the new millenium began. A significant part in that late 20th century decrease was offset by increases in the split infinitive forms "him to always" (blue) and "him to quickly", which are now vastly more common in books than they were 60 years ago.

We can play the same game with slightly different n-grams.  Here are Google n-gram comparisons of "always to be"/"to always be"/"to be always", and  "always to have"/"to always have"/"to have always". With a bunch of different searches, I've generally seen the same pattern, namely clear increase in the split infinitive form (though for some verb / adverb combinations, the split infinitive remains very rare in this book sample) over the last 60 years, sometimes with a leveling off or drop at the end of the sample period. Of course, nothing else is quite as dramatic as the graph for "to boldly go".

OK, last graph. The blue line here shows what's been happening to the use in books of the term "split infinitive" from 1800 until 2008: the term \ You see that people started using the term at the end of the 19th century. It was climbing well before Strunk first published "The Elements of Style" in 1918, and reached its peak a few years later. For comparison, I've also included "start with and" (green), "who or whom" (red) and "stranded preposition" (yellow). So we're well past the heyday of split infinitive discussion. It's been relatively stable for the last 40 years as a perennial book topic, though whether the books are pro- or anti- will have to await some reader's future studies.

As regards the actual usage of split infinitives, is it possible that they're now in decline? Unfortunately, the n-gram corpus stops at 2008, so if you want to know the answer, you'll have to watch this space for another decade, or maybe a couple. For the moment, my take-home message is that whatever infinitives Chandler split may well have been daring when he split them, at least for books, if not for the bars he frequented. But it looks to me like infinitives are indeed likely to remain split for quite a while yet.



  1. Andy Averill said,

    September 9, 2012 @ 4:07 pm

    A quick flip through The Big Sleep doesn't turn up any split infinitives. Maybe that's not too surprising considering that adverbs don't really fit in well with his hard-boiled/Hemingwayesque style.

  2. Stan said,

    September 9, 2012 @ 4:08 pm

    'infinitives are indeed likely to remain split for quite a while yet'

    For many people, alas, they will remain unsplit at all costs for quite a while yet too. I've collected many examples from books, some of them modern, where avoidance of the split form leads writers (or their editors) to resort to phrasing that's unnecessarily awkward or ambiguous. The influence this superstition continues to wield is quite amazing.

    I like Chandler's poem.

  3. David J. Littleboy said,

    September 9, 2012 @ 4:45 pm

    Perhaps the increase in unsplit infinitives post 2000 is due to MS Word's grammar checker being turned on as the default and people doing what it says instead of turning it off. (Memory has it that the grammar checkers didn't exist in 1990; maybe it's just that they exist at all.) Or editors getting lazy and insisting that writers respond to its whining.

  4. Andrew (not the same one) said,

    September 9, 2012 @ 5:16 pm

    It was climbing well before Strunk first published "The Elements of Style" in 1918.

    So far as I can see, Strunk does not use it in any case. (White does, though not with complete disapprobation.)

  5. Mark F. said,

    September 9, 2012 @ 6:45 pm

    I think it's been mentioned here that there's a fairly substantial shift in the mix of books that occurs around 2000, which is why Google by default stops at that year. I very often notice trends appearing to reverse direction at the year 2000 and I generally suspect these reversals of being artifacts.

    [(dib) Yes, I agree. However, it's worth noting that if you perform the search with books that Google says are fiction, you get roughly the same result. I haven't researched how Google selects its fiction sub-corpus, and can't say whether there's any interesting sense in which the make-up of books has changed.]

  6. Carl Offner said,

    September 9, 2012 @ 7:51 pm

    From the poem:
    O dear Miss Mutch, put down your crutch,
    and leave us to crack a bottle.

    A guy like I weren't meant to die
    On the grave of Aristotle.
    A couple of brilliant rhymes there, I think.

  7. Thom said,

    September 9, 2012 @ 8:32 pm

    It is interesting to note that the numbers on the second chart start picking up sometime in the 1870's. It is interesting to note that the Kalamazoo Case of 1875 (I'm finding dates range from 1872-1875) established the use of taxes to support public education. Is this a coincidence, or could it be that the taxes helped build the system for textbook companies to be established, which is why we see such a growth in prescriptivist discussion?

    [(myl) An interesting speculation -- but the underlying facts are not clear. A lot of the "start|starts with and" hits are things like "I will give you ten dollars to start with, and if you find him and bring him back, I will give you forty more." I checked all the pre-1900 Google Books hits for this pattern, and found that none of them are part of a "prescriptivist discussion". So you'd want to use a more carefully-curated source of data before making a connection to possible causes.]

    [(dib) It is an interesting idea, but two reasons other than Mark's to be dubious about there being any such causal connection, reasons that you can easily check with your own n-gram searches, are: (i) the pattern for British English is broadly comparable to that for US English, and (ii) other educational terms (grammar, logic, rhetoric, physics) show very gradual increase right across this period, with no significant bump around 1875.]

  8. Joel said,

    September 9, 2012 @ 9:44 pm

    It seems to me that "him always to" and "him quickly to" are syntactically ambiguous fragments. Could not "to" be a locative as in "I took him quickly to the hospital"? Or the "to" be infinitive, but the adverb adjoining the preceding verb, as in "I avoid him always to maintain my sanity"?

    [(dib) Yup. I chose those strings after quickly checking a page of ordinary google results, and finding that all the hits were to+always+infinitive. Extrapolating, I guessed most of the trigrams would be the ones I wanted. But of course that could be wrong, and that's why I also looked at the same question with strings like "to always/quickly/suddenly be/have/go". The whole process took me only about 15 minutes, and if my conclusions were incorrect, I'm guessing you'll need to invest at least that much time to prove it.]

  9. Mark Mandel said,

    September 9, 2012 @ 10:07 pm

    Of course, nothing else is quite as dramatic as the graph for "to boldly go".

    Not as dramatic as I was expecting:

    Not Found
    The requested URL /nll/wp-admin/boldly to go,to boldly go,to go boldly was not found on this server.

    [(dib) Thanks. Corrected. Gremlins.]

  10. L said,

    September 10, 2012 @ 8:55 am

    What I notice is this:

    The sum of the four traces declines fairly steadily until around 1990, where it bottoms out and climbs until at 2010 it's almost back to the 1950 level.

    I take this to mean that use of the infinitive overall declined and then recovered.

    I have no speculations as to why.

  11. Ray Dillinger said,

    September 10, 2012 @ 11:45 am

    I recall some prescriptivist advice about avoiding the infinitive in older books on effective rhetoric and persuasive writing. It was considered a poor "weak" style, cousin to the much-maligned passive voice.

    Also, I think a lot of what we're seeing post-1950 is not so much a change in everyday usage as a decline in the influence of editors on published work – particularly on published fiction. There was a time when works of fiction were gone over by editors, and even the most uneducated thugs from the hinterlands and foreign characters who had only recently learned English were forced to speak with "perfect" grammar, according to the ideal of perfection held by that particular editor and publisher.

    This is precisely the kind of editing that Chandler was upset about when he penned the letter above, and mangled many an otherwise-brilliant characterization. I recall some discussion about how Samuel Clemens had to fight to get his editors to pass on his characters' use of dialect, but the results "speak" for themselves. Sometime in the 1950s editors got out of the business of "correcting" character quotes, and sometime in the 1980s or thereabouts they got out of the business of editing fiction at all for any but the most egregious errors of word selection and spelling.


  12. HP said,

    September 10, 2012 @ 10:47 pm

    Are you sure about that, Ray? My experience with 19th and early-20th c. literature is that working-class and ethnically marked characters speak in affected eye-dialect so thick that the resulting prose is very nearly unreadable.

  13. Janet said,

    September 19, 2012 @ 10:49 am

    Has anybody ever thought to ask ME how I feel about being split? Sheesh! Read my "Message from the Infinitive" on janet's website http://www.medlinguistics.com/Infinitive.asp (See link "Infinitive" on left sidebar)

    I'm on in years, so a few facts have escaped me, but I've remembered the important ones.

    From The Infinitive

RSS feed for comments on this post

Leave a Comment