Geoff Pullum's most recent posting on split infinitives noted that handbooks on grammar and usage do not prohibit them, but most say they should be avoided, unless splitting the infinitive would improve clarity. When you think about it, this is decidedly odd advice.
There's some history here, which is well covered in MWDEU, and has now been briefly treated by John McIntyre in his blog You Don't Say. The short version: the split infinitive as a bugaboo lodged itself in "the popular press and folk belief" (MWDEU) in the 19th century, so that the handbooks now say (and have been saying for a century or so), in effect, that there's nothing grammatically wrong with split infinitives, but some people are offended by them, so you should avoid them as much as possible, to avoid giving offense. In McIntyre's words: "the only reason to avoid splitting infinitives is to escape the uninformed censure of people who think that it is a violation of grammar and usage."
In other words, crazies win.
I'll get to the crazies eventually. Hang on.
The handbook advice is about social behavior, not about grammaticality, writing style, grace, or even the exercise of taste in choosing variants. It seems to me that the handbook writers — including the revered Fowler and Strunk & White, by the way — who give this advice have simply abandoned their responsibility as advisers on grammar, usage, and style. If they'd done their job properly, the split-infinitive bugaboo might have died away in the popular mind long ago. Instead, we've had a century of timid advice that treats split infinitives as tainted and so helps to keep the bugaboo alive. In fact, it flourishes in the schools.
On split infinitives in the schools: even if you take Split Infinitive as Last Resort (SILR) to heart as advice about correct English usage, if you're a schoolteacher, you might well end up banning split infinitives entirely (No Split Infinitives — NSI) in your classes anyway.
Here's the reasoning: SILR comes up only infrequently, and teaching kids when a split infinitive is better than the alternatives is not at all easy — if you, in fact, have a grasp on the matter, which the handbooks don't explain well; mostly, they give exemplars and just say that these are cases where the split infinitive "sounds better" or "is clearer". Sometimes they give the advice without the exemplars, as in this passage from Laurie Rozakis's Complete Idiot's Guide to Grammar and Style (2nd ed., p. 133):
While I do not advocate that you go around town splitting infinitives with abandon, there's no point in mangling a sentence just to avoid a split infinitive. Good writers occasionally split infinitives to create emphasis, achieve a natural word order, and avoid confusion. If splitting an infinitive makes it possible for you to achieve the precise shade of meaning you desire, you have my blessing to split away.
(Rozakis, p. 135, also recommends avoiding stranded prepositions unless doing so would sacrifice naturalness and grace, and illustrates her advice with exactly one example, This ticks me off – which does not in fact have a stranded preposition. Sigh.)
In the face of such advice, you might well decide to just ban split infinitives wholesale. Then there's no issue. You might add, Rozakis-style, that there are circumstances where splitting an infinitive is ok, and good writers sometimes do that, but then tell your students that to be safe they should avoid split infinitives, though when they're older and more experienced they'll learn how to use them. Not For Kids.
Whether teachers proscribe split infinitives because they believe in NSI or because they think that's the easy route around SILR, their students grow up thinking that split infinitives are a Bad Thing. If, later in life, they come across SILR in the advice literature, what are they to make of it? What was once banned is now permissible, but in what circumstances? How to know when the alternatives are worse?
The advice givers have tacitly acquired their own tastes in the matter, but they can't assume that their readers (already) share these tastes. In fact, by formulating principles like NSI and SILR, they have failed to teach their readers anything useful about how they might judge different versions of infinitival VPs.
Now let's look at things from a student's point of view. Suppose you've been taught the "permissive" SILR rather than the "strict" NSI. Will you now split infinitives? Probably not, because of the "only when necessary" character of SILR. You can guess that if you split an infinitive, a teacher is likely to demand that you defend it, and your explanation that it just sounds better to you than alternative placements of the adverbial probably won't fly. (Similarly for writers dealing with copyeditors who hold to SILR.)
In general, the advice to use some variant X only when necessary will tend to drive X out of business, because it throws on the user of X the burden of justifying every single use of X.
What I've said so far is that, in practice, SILR is scarcely an improvement over NSI and in fact works to preserve the belief that split infinitives are tainted in some way.
Some usage advisers now say flatly that there's nothing wrong with split infinitives and you should use them whenever they suit you. To my mind, that's the only intellectually justifiable advice. But it bucks against the long and tenacious career of NSI in the popular mind and the unfortunate side effects of SILR in usage manuals.
Every so often, a colleague, scandalized, brings to my attention a split infinitive in the pages of the New York Times. As it happens, NYT writers, including those who write the editorials and the op-ed columns, split infinitives with moderate frequency and apparent abandon. Good for them. But a lot of people — among them, a lot of highly educated people — see these split infinitives as a sign of declining standards, the fall of Western civilization, the forthcoming apocalypse, and all that.
Which brings me to the crazies, who are one step up from those who merely believe, earnestly but incorrectly, that split infinitives are an offense against grammar. (I'll come back to these earnest objectors.)
Crazies are people who INSIST on having their way, who believe so passionately in (or against) something that they cannot let what they see as offenses pass. Crazies don't actually always win, but they're damn hard to deal with. If you have to deal with someone whose face turns purple, who sputters, and who bursts into tears whenever ducks are mentioned (I'm afraid this is not a fanciful invention of mine), then you're probably going to steer clear of duck talk around them. (It was in connection with this case that I first heard the slogan "crazies win" — from Ann Daingerfield Zwicky, some forty years ago. I'm not claiming that the slogan was original with her — surely it was not — only declining to take credit for it myself.)
Grammatical crazies are mostly just appalling nuisances, but if they're in a position of power, we've got a problem. I've met executives who maintained to me (on learning that I was a linguist) that if they got a job application that had a split infinitive in it, they would just throw it away. (Sometimes it's some other damning flaw, like using which as a restrictive relativizer.)
Now people who propose to give advice about grammar, style, and usage have a problem. What is this advice supposed to be good for? One thread of the advice literature is aimed at people who write a good bit, and it tries to push such writers towards greater grace and clarity, mostly by exhibiting flaws in the work of practiced writers (people who "ought to have known better").
Another thread is aimed at relative novices, especially young people and the upwardly mobile, for whom a major goal is achieving a socially acceptable presentation of self outside their usual communities; they'll be concerned, among other things, with avoiding negative judgments about their intelligence and abilities. That is, they don't want to look stupid, ignorant, or uneducated in contexts where intelligence, knowledge, and education are valued.
For this second group, the core of the problem is notably class-marked variants, and the list is not enormous: hisself, 3sg don't (She don't like it), past participle = past tense (I have wrote a letter), demonstrative them (them guys), AAVE invariant be (They be on my case all the time), non-standard double negation, AAVE copula omission (He my best friend), and so on. What's remarkable is how few of these variants are discussed at any length in the advice literature, which tends to focus instead on informal vs. formal variants (to disparage the former), spoken vs. written variants (to disparage the former), variants that are (or are perceived to be) innovative, and arcana like split infinitives, stranded prepositions, which vs. that, sentence-adverbial hopefully, and singular they, where social class associations are a very small part of the picture, but some people judge the disfavored variants to be signs of ignorance or lack of education. The list of these disfavored variants is huge.
People keep writing me to say that it's important that struggling students (of all ages) get advice about how to avoid looking ignorant and uneducated, and as a teacher (and I should add, a child of the working class), I certainly appreciate the point, though I'll add that a lot depends on the context; there are social advantages to be gained by avoiding class-marked variants in certain contexts, but in other contexts, people should be free to speak or write as they will.
Now we're up to an intersection with the topic of grammar crazies: what to say to students who naturally use the disfavored (but not class-marked) variants, like split infinitives. Sally Thomason is on record here on Language Log as recommending that such students should be warned that using these variants will lead some people to judge them negatively. After many years of wrestling with this question — I'd tell students that there were many people who viewed split infinitives as just wrong, and some who were lunatic on the matter, so they might want to take that into consideration — I've decided that the best advice is just to go ahead and do what seems natural to you. There are much more important things to worry about in this life, and if you think you can satisfy the tastes of everyone who reads what you write or hears what you say, you're doomed; a fair number of people are just LOOKING FOR evidence of ignorance and lack of education, and with some application they'll be able to find it, no matter what you do or how hard you try. (Remember that the list of disfavored, but not class-related, variants is huge.)
What you hope for is that the number of grammar crazies is small. So maybe you won't get that particular job — would you really want to work for that grammar crazy?
Back to the earnest objectors. Here, I get some mail that says: if most people think that X is ungrammatical, isn't X by definition ungrammatical? To start with, I'm not conceding that most speakers of English think split infinitives are ungrammatical, though I'll agree that many educated speakers think so.
But even if most people thought that way, that wouldn't make it so, for two reasons: one, beliefs about language use (what we might think of as "meta-facts") are not the same thing as facts about language use; the two things are often at variance. And two, the practices of people who speak and write a standard variety (educated middle-class people, especially those who are practiced writers and speakers) differ; there is considerable variety within the standard.
On point two: as part of this variation, some standard speakers elect — for whatever reasons — not to use (or at least to try not to use) some variants that other standard speakers use naturally. That's their prerogative. No one is telling them they HAVE to use these variants (much less that they have to ALWAYS use them rather than alternatives). There's a problem, though, when people who have chosen to avoid some variant seek to insist that other people (also speakers of a standard variety) adopt their practice.
There's no question that splitting infinitives is a variant within standard English, and has been for a long time. If you believe it's nasty, that's your personal taste, but you're going to have to live with the fact that other (educated and articulate) people do not share your taste, and you should stop beating them with a stick about how their English differs from yours.
(As many people have pointed out, the prejudice against split infinitives is indefensible on rational grounds. For my present purposes, I really don't care about that: you're entitled to your grammatical prejudices, however irrational, but you're not entitled to tell me, and other educated and articulate people, what to do.)
On point one: the disconnect between folk belief and fact — between "social fact" and "objective fact" — is a commonplace of social-science research, and indeed of research in science in general. (There's a fascinating literature on folk physics, and also some research on folk biology, folk logic, folk statistics, folk psychology, and more, not to mention folk beliefs about historical events. And, of course, folk linguistics.) This disconnect is a serious problem when researchers try to communicate findings to the general public: how to tell people that some of what they believe, and believe to be confirmed by their experiences, is seriously skewed, or just flat wrong? (No, I don't have a solution.)
The objective fact is that split infinitives are standard English. So my advice is: split an infinitive if it suits you (or don't, if that suits you). Good writers do it. And you don't even have to have a defense for it; do it because it sounds right for you. Don't let the crazies win.