The “rule” against splitting infinitives is one of most-remembered but least understood of all the rules in our language. Striving against the “rule” is made all the more difficult because it is taken so seriously by those who know the rule – because they learned it in primary school, and found it was insisted on by others who had learned it the same way and had never questioned it.

But the “rule” against splitting infinitives is recent.

To boldly go breaches the injunction against splitting infinitives and has no advantage of emphasis or clarity over to go boldly.  By contrast, to cheerfully sing again conveys clearly what is an ambiguous possibility in to sing cheerfully again: in adhering to the rule, the second statement leaves the hearer uncertain whether the cheerfulness attaches to the act of singing or the fact of repetition.

What is odd about the English horror of split infinitives is that it is based on the observation that Latin infinitives could not be split, with the conclusion that English infinitives must not be split.  Latin infinitives could not split because they were in the form of a single word: amare to love, habere to have, cantare to sing, etc.  In consequence with the early grammarians’ unwavering adherence to the conventions of Latin, we daily wrestle with tensions created by the rule.

The OED defines infinitive as follows:

“The name of that form of a verb which expresses simply the notion of the verb without predicating it of any subject. Usually classed as a ‘mood’, though strictly a substantive with certain verbal functions, esp. those of governing an object, and being qualified by an adverb…”

Johnson is at once briefer and more opaque:

“In grammar, the infinitive affirm, or intimates the intention of affirming, which is one use of the indicative; but then it does not do so absolutely”

Webster is a bit less opaque:

“an  infinite verb form normally identical in English with the first person singular that performs certain functions of a noun and at the same time displays certain characteristics (as association with objects and adverbial modifiers) of a verb and is used with to  (as in to err is human, I asked him to go)…”

It is significant that Webster says the infinitive “is used with to”, which suggests that, in common with the Indo-European pattern, the infinitive in English is a single word even if its status as infinitive is generally marked by the word to.  For example, Latin and Greek infinitives are a single word, but so they are also in French and German.  So, to run is currere in Latin, correr in Spanish,  courir in French, trexo in Greek and laufen in German.

In English, where to is separated from the infinitive form of the verb, the result is called a split infinitive.  But it is fairly clear that to is not part of the infinitive at all.  Putting to one side Webster’s definition of infinitive, Oxford English: A Guide to the Language (1986) says :

“The split infinitive is the name given to the separation of to from the infinitive by means of an adverb…” (emphasis added)

The OED sheds some light on the matter in its entry for the word to as preposition, conjunction and adverb.  It’s a long entry: about 24,000 words.  At about the 13,000 word mark, it says “to before an infinitive”, which clearly suggests that to is not part of the infinitive.

By convention, where an infinitive is preceded by to, there is  typically no word between to and the infinitive. So, to go boldly is generally thought to be better English than to boldly go. In ordinary usage, the second form is referred to as a split infinitive.  But the “rule”which dictates that infinitives should not be split emerged very late, and seems to be based on the observation that, in languages like Latin, the infinitive was never split (because it was a single word, and could not be split).

In 1834 a letter to the editor of the New England Magazine declared that infinitives should not be split.  The author was identified only as “P”.  The writer declared that split infinitives were only used by “uneducated persons” and in “newspapers where the editors have not had the advantage of a good education.”

In 2004, the Cambridge Guide to English Usage repeated the rule in modified form: “Don’t split an infinitive if the result is an inelegant sentence”.

Given that the “rule” was only proposed (or invented) in 1834, it is not surprising that it was ignored by most English writers.  Shakespeare did it a lot.  In 1931 a study found split infinitives in English literature from every century: beginning with the fourteenth-century epic poem Sir Gawain and the Green Knight, and including William Tyndale, Oliver Cromwell, Samuel Pepys, Daniel Defoe, John Donne, Benjamin Franklin, Samuel Johnson, Edmund Burke, Samuel Taylor Coleridge, Elizabeth Barrett Browning, and others.

Because the notion of not separating to from the infinitive has been entrenched in the language for almost 200 years, care needs to be taken.  Even arbitrary rules can gain a veneer of significance by virtue of longevity.  Fowler recognised this.

He begins his article on split infinitives this way:

“The English-speaking world may be divided into

  1. Those who neither know nor care what a split infinitive is;
  2. Those who do not know, but care very much;
  3. Those who know and condemn;
  4. Those who know and approve;
  5. Those who know and distinguish”

He then analyses each group, and comments that

“Those who neither know nor care are the vast majority & are a happy folk to be envied by the minority classes; ‘to really understand’ comes readier to their lips and pens than ‘really to understand’, they see no reason why they should not say it (small blame to them, seeing that reasons are not their critics’ strong point), & they do say it, to the discomfort of some among us, but not to their own.”

As for the second group (those who do not know, but care very much) “who would as soon be caught putting their knives in their mouths as splitting an infinitive”.  Fowler comments that “These people betray by their praactice that their aversion to the split infinitive springs not from instinctive good taste, but from tame acceptance of the misinterpreted opinion of others…”

As to the fifth group (those who know and distinguish), Fowler clearly includes himself in this group and notes: “We maintain, however, that a real s. i., though not desirable in itself, if preferable to either of two other things, to real ambiguity, & to patent artificiality”

So to go boldly is probably better grammar than to boldly go because, even though they convey identical meanings, the Star Trek version draws attention to itself: just as a diner would if they were to hold their knife in the left hand and their fork in the right hand.

See how conventions ossify into rules.