Tag Archives: Geoffrey Pullum

The Negative Canon: ‘Split Infinitive’

This post is one in a series about The Negative Canon.

It really shouldn’t be necessary any longer to state that there is no rule of English grammar that prohibits the placing of anything between the particle to and the infinitive of the verb, yet there still seem to be those who believe there is. When this does occur they speak of a ‘split infinitive’, as if to was part of the verb. It isn’t, so there’s nothing to split. This applies to English just as much as it did to Latin, from which this strange notion is said to derive. As R L Trask writes in ‘Mind the Gaffe’,this traditional term is a misnomer, since, in this construction, nothing is split, least of all the infinitive, which is a single word’. Or, as Geoffrey Pullum has said on Language Log (my emphasis):

The misnamed ‘split infinitive’ construction, where a modifier is placed immediately before the verb of an infinitival complement, has never been ungrammatical at any stage in the history of English, and no confident writer of English prose has any problems with it at all. (As the grammarian George O. Curme pointed out in 1930, it’s actually the minor writers and nervous nellies, the easily intimidated, who seem to worry about it.)

Merriam-Webster’s Dictionary of English Usage’ suggests that ‘the enduring popularity of the split infinitive is due to its catchy name’ and quotes Ambrose Beirce (he of ‘The Devil’s Dictionary’) as more or less saying so in 1909:

Condemnation of the split infinitive is now pretty general, but it is only recently that any one seems to have thought of it. Our forefathers and we elder writers of this generation used it freely and without shame – perhaps because it had not a name, and our crime could not be pointed out without too much explanation.

Just to be clear about it, all verbs have a plain form, that is, the basic word that you look up in a dictionary. All other forms derive from it. Walk, for example, is the plain form of a regular verb, and from it we get the forms walk, walks, walking and walked. Walk is used for all forms of the present tense except the third person singular. It is also used as the infinitive, and is found in clauses such as:

(1) I like to walk


(2) I might walk.

Example (2) provides evidence, if any is needed, that it is walk that is the infinitive, not to walk.

No serious authority that I know of claims that the construction is ungrammatical. On the contrary, not to use it can have infelicitous results. As Trask writes his article:

Desperate attempts at avoiding ‘split infinitives’ often produce English which is not only tortured and unnatural but even ambiguous and misleading.

Even Harry Blamires is not greatly disturbed by it:

. . . it is not difficult to assemble instances of the split infinitive which are unobjectionable.

In fact, the question is really more to do with the placing of adverbs than anything else. We might write, for example:

3a. In heavy rain it’s wise to drive slowly.

rather than:

3b. ?In heavy rain it’s wise to slowly drive.

If placing slowly between to and drive is ungrammatical, it’s not because of any zombie rule about infinitives, but because that’s not where adverbs of manner normally occur. In other cases, the position of an adverb can change the emphasis of a sentence. It is certainly possible to write:

4a. The government have decided to increase the allowance paid to the unemployed gradually.

Here, the terminal position foregrounds gradually, and that is something the government might not want to do. They might want instead to emphasise the fact of the increase, and to minimise the fact that it is to be introduced in stages. In that case, losing the adverb between to and the verb does the job:

4b. The government have decided to gradually increase the allowance paid to the unemployed.

Pointing out that there is no rule of English grammar that prohibits the placing of anything between to and the verb doesn’t mean that it is always effective to do so. It was an unwise placing of an adverb that led Geoffrey Pullum to take an incompetent writer to task on Language Log for producing this sentence:

On this occasion we’ve been forced to take the decision to regrettably cancel your flight.

What the writer meant was that the airline regretted that the flight was cancelled, not that it was cancelled, as the sentence says, in a regrettable manner. It’s so poorly constructed that moving the regrettably to before to or after cancel isn’t enough to put it right. Pullum recast it this way:

On this occasion we’ve regrettably been forced to take the decision to cancel your flight.

There is also a pragmatic aspect. As Peter Harvey says in his ‘A Guide to English Language Usage’:

. . . it must be said that there is still a considerable feeling among English speakers that a split infinitive is wrong

and he explains his position more fully on his blog:

My job is to teach and advise people who need to use English for their business and professional purposes. I also translate texts, sometimes for publication, for people who are demanding in their requirements. I simply cannot afford to have a student come back to me with a complaint that something that I have taught or tolerated has been criticised in no uncertain manner as a grammatical solecism by a native speaker; nor can I afford to have an argument with a translation client on the same matter.

A similar comment, of course, might be made about the negative canon as a whole. Individual writers must decide what words and structures are in their best communicative interest. All I say is that those features of the language that I suggest should be in the negative canon are not ungrammatical. Whether or not it is effective to use them in any particular context is a matter of style, not grammar, a distinction which I made in an earlier post.


Filed under The Negative Canon

The Negative Canon: [Noun Phrase] and ‘I’/‘Me’

This post is one in a series about The Negative Canon.

The English personal pronouns are I/me, you/you, he/him, she/ her, it/it, we/us and they/them. The first form of each pair is, in traditional terminology, in the nominative case, and the second is in the accusative case, where the nominative case is used for the subject of a clause and the accusative case is used for the object of a clause and following a preposition. This means that we say, at least in Standard English, She saw him and He saw her rather than *Her saw he or *Him saw she, and He came towards me rather than *He came towards I.

So far, so clear. However, variation can occur when the first person (I/me) is coordinated with a noun or with another pronoun, as in:

(1) They invited my wife and [I/me] to lunch.


(2) There’s a great deal in common between you and [I/me].

I imagine the majority of those who read this will say that what happens when I/me occurs alone must also apply in other circumstances too. We wouldn’t say They invited I, so we can’t say They invited my wife and I. We wouldn’t follow a preposition with I, so we can’t follow a preposition with <…> and I. This argument is put forward by both Harry Blamires in ‘The Penguin Guide to Plain English’ and R L Trask in ‘Mind the Gaffe’. Blamires cites this example from the former British Prime Minister John Major, commenting that we may see from it ‘how deeply this error has corrupted us’:

It may be that many people would like to invent divisions between he and I, but there are none.

I have myself heard David Miliband, a former British Foreign Secretary and the brother of the Leader of the Labour Party say on BBC radio:

. . . invite Michael Howard and I . .

Trask makes the point about not using I in those contexts where it occurs alone and continues:

The presence of that ‘and’ seems to throw many writers into a panic, a panic which inevitably leads to the insertion of an impossible ‘I’ where only ‘me’ is possible.

However, the OED records I ‘as object of a verb or preposition’ and notes that it is:

Used for the objective case after a verb or preposition when separated from the governing word by other words (esp. in coordinate constructions with another pronoun and and).This has been common at various times (especially towards the end of the 16th and in the 17th century, and from the mid 20th century onwards); it has been considered ungrammatical since the 18th century.

There you are, those 18th century grammarians again. Hypercorrection is the usual explanation, but, as Peter Harvey has commented:

If the usage was ‘very frequent’ 400 years ago, hypercorrection cannot really be the whole reason for its use nowadays.

It could even be argued that and me rather than and I is itself a form of hypercorrection.

The authors of ‘The Cambridge Grammar of the English Language’ (p. 9) challenge the view that what happens when a pronoun occurs in isolation must determine what happens when it has company:

But why should we simply assume that the grammatical rules for case assignment cannot differentiate between a coordinated and a non-coordinated pronoun?

They cite another example of where the rules make just such a differentiation. We can say I don’t know if you’re eligible, but we can’t say I don’t know if she and you’re eligible. They comment that the sequence you are can be reduced to you’re in the first sentence, where you is subject, but not in the second where the subject has the form of a coordination of pronouns.

This shows us not only that a rule of English could apply differently to pronouns and coordinated pronouns, but that one rule actually does . . .The argument from analogy is illegitimate.

Elsewhere they write that constructions such as (1) are (my emphasis):

. . . used by many highly educated people with social prestige in the community; it should therefore be regarded as a variant Standard English form.

‘Merriam-Webster’s Dictionary of English Usage’ (MWDEU) records Henry Sweet’s suggestion in his ‘New English Grammar’ of 1892 that you and I occurs so frequently that it comes to be seen as an invariable group in all contexts. This comes close to Chomsky,mentioned later in the MWDEU article, when he claimed that :

. . . [‘between’]can assign the objective case agreement only to the whole compound, which cannot be declined, and the individual words in the phrase are free to be nominative or objective or even be reflexives.

The preference for between you and me over between you and I, at least, seems strong, as shown by the following figures:

British National Corpus Corpus of Contemporary American English
Between you and me 43 183
Between you and I 2 17

Still, the preference is much less marked when the search is narrowed to spoken English, and the two examples of between you and I in the British National Corpus occur in sentences which must be considered as Standard English. The first is from a radio programme:

But at a fundamental level there’s really little difference between you and I struggling to count on our fingers, and the most modern and sophisticated piece of computing wizardry.

The second is from a piece of dialogue in a Mills and Boon novel:

I felt as if I knew you, and you knew me — almost from the beginning of time. It was as if, between you and I, the ordinary processes of — what shall I call it… courtship? — were totally superfluous.

If you think a citation from a Mills and Boon novel too down-market to count, Shakespeare had Bassanio say in ‘The Merchant of Venice’:

Sweet Bassanio, my ships have all miscarried, my creditors grow cruel, my estate is very low, my bond to the Jew is forfeit. And since in paying it, it is impossible I should live, all debts are cleared between you and I if I might but see you at my death.

We can take from all of this that both you and me and you and I in object position and following a preposition, particularly between, occur in the speech of speakers of Standard English, but that it might be unwise to use the latter variant in formal prose. As Pam Peters very sensibly comments in ‘The Cambridge Guide to English Usage’:

The vacillation over ‘me/I’ is symptomatic of shifting case relations among pronouns generally . . . But because between you and I seems to have become a shibboleth it’s to be avoided in writing. In fact a confidential between you and I is unlikely to occur to anyone writing a formal document, because of the impersonal nature of the style that goes with it.

MWDEU comes to a similar conclusion:

You are probably safe in retaining ‘between you and I’ in your casual speech, if it exists there already, and you would be true to life in placing it in the mouths of fictional characters. But you had better avoid it in essays and other works of a discursive nature.

We could say the same about the constructions in sentences (1) and (4). More adventurous writers might like to use and I, but they must be prepared to have attention paid to the grammar of their writing at the expense of its content. There is, however, no reason to chastise its use in speech, or even in speeches. As is often the case, two forms exist side by side until one ousts the other. As Jean Aitchison once wrote:

Language lamenters mostly haven’t understood how language works. In language change, new variants grow up alongside existing ones, often as stylistic alternatives.

Differences between the forms of pronouns are less distinct than they once were (look at who and whom). That is only natural. Just as the three numbers, four cases and, in the third person, the three genders, of the Old English personal pronouns have eroded to what we have now, our present system seems to be in the process of reducing further.


Filed under The Negative Canon

The Negative Canon: Introduction

A fellow contributor to the LinkedIn group Grammar Geeks, Bessel Dekker, has suggested the term negative canon for those features of English that are frequently the object of the attention of those in various Facebook groups and elsewhere who seek to tell us how to use our own language. They complain about current developments in the language, oblivious to the fact that such developments are sometimes far from new, and that English contains features that have come about through the type of changes in the past that they condemn in the present. We are asked to accept what they say because that’s their opinion or what they’ve always been taught.

A prominent characteristic of the negative canon is that it consists of a relatively small number of features which come up time and again, and are often of the kind that Geoffrey Pullum has called ‘the most intellectually trivial details of standard written English’.  As an example, Bessel has listed the following salient features of Afro-American Vernacular English:

(1) the multiple negative.
(2) nonstandard forms such as aks.
(3) absence the forms of be when it is a copula.
(4) uninflected be when it is an existential verb.
(5) word order which is different from that in Standard English.

His point is that of these five, only the first two are habitually seized upon. Other instances include sentence adverbs, where hopefully is often disdained, but all the others are ignored. Similarly, the pronunciation mischievious is thought to be an indication of moral decline and the end of civilisation as we know it, but, as Bessel has said, it is not the first, let alone the only, case were epenthesis has caused language change.

Bessel has therefore made an important distinction between the negative canon and what he has called the negative pool. While the latter consists of all instances that could be objected against, the former comprises the usual suspects which are in fact singled out. Thus, sadly as a sentence adverb, for example, would be in the negative pool only, while hopefully has made the negative canon. The composition of the negative canon, it seems, is quite arbitrary, which suggests that no that no systematic study underlies it.

Bessel has suggested a few criteria which give an item a good chance of being included in the negative canon. They are:

(a) the intelligibility criterion: objections against it should be easy to understand.
(b) the beaten path criterion: it should have been discussed before, the more frequently the better.
(c) the authority criterion: it should have been rejected or challenged by such writers as Lowth, Murray, Fowler, or Strunk.

(d) the criterion from logic: it should be vulnerable to the argument that it flies in the face of ‘logic’.
(e) the criterion from semantics, including etymology: it should be demonstrably subject to meaning shift (the ‘etymological fallacy’).
(f) the criterion from syntax: it should be demonstrably at variance with syntactic behaviour elsewhere.

It might, as Bessel has also suggested, be illuminating to compile an inventory of items that make up the negative canon and to trace each to its first source. That may be a little ambitious for Caxton, but in a series of forthcoming posts I plan to examine some of the usages in question in an attempt to disentangle fact from fiction, truth from prejudice, reality from fantasy. They can for convenience be considered under the headings of grammar, vocabulary, and orthography and punctuation. Negative canon posts will not be in any particular order, and not necessarily uninterrupted by unrelated posts.


Filed under Language, The Negative Canon

Grammar and Style Again

‘Articles about English grammar in UK newspapers tend to exhibit an almost incredible degree of stupidity.’ Geoffrey Pullum on Language Log today.

Leave a comment

Filed under English Language