This post is one in a series about The Negative Canon.
In ‘A Guide to English Language Usage’, Peter Harvey states the standard position on the use of imply and infer:
To imply something is to suggest it indirectly.
To infer something is to deduce or conclude it from evidence.
Native speakers sometimes use infer to mean imply.
That is a fair enough comment in a guide destined for foreign learners. By contrast, both Harry Blamiers’s ‘Penguin Guide to Plain English’ and R L Trask’s ‘Mind The Gaffe have native speakers in mind, and their approach is rather more dogmatic. The first baldly states:
What goes generally wrong is that the word infer is used as though it meant the same as imply.
Trask writes, with little further comment:
When you imply something, you suggest indirectly that it is true. When you infer something, you conclude that it is true.
The OED, to be sure, gives the ‘imply’ sense as its fourth definition of infer, but cautions:
This use is widely considered to be incorrect, especially with a person as the subject.
‘Merriam-Webster’s Dictionary of English Usage’ (MWDEU) article on the subject is comprehensive. Drawing on the OED, it records that in 1528 Sir Thomas More introduced infer with the meaning given by Harvey, Blamires and Trask. Five years later, however, he used it to mean ‘imply’, but not with a human subject. Both uses continued, apparently without disapproval, for around 400 years. The OED’s citations supporting this use include those from Milton, Scott and Mervyn Peake. MWDEU also gives citations from James Boswell and Jane Austen.
The use of infer to mean ‘imply’ with a human subject first occurred, according to MWDEU, in 1896. Then, and subsequently, its use was oral. It was only when, in the 1950s, that infer came to be used in print with a human subject that the controversy appears to have arisen. In ‘The Cambridge Guide to English Usage’, Pam Peters also discusses the case when the subject of infer is impersonal, concluding that:
The shift from nonpersonal use of infer (“indicate”) to personal use as “imply” is no great move.
Her comment is, as usual, balanced and pragmatic:
In conversation and debate many people do not distinguish between these constructions; and in context it’s usually quite clear whether infer is intended to mean making an active suggestion (= “imply”), or a deduction made from something else. As often, the distinction is more important in writing, and writers may be reassured . . . that the word they need most of the time is imply. Like other shibboleths of language, the issue needs to be defused.
Corpus evidence confirms the claim that imply is indeed most often what is needed. Imply is nearly three and a half times more frequent than infer in the Corpus of Contemporary American English, four and a half times more frequent in the British National Corpus, and just over four times more frequent in the Corpus of Global Web-Based English.
In fact, the problem, if there is one, may be more imagined than real. MWDEU observes that:
. . . our evidence shows a marked decline of occurrence of personal infer in edited prose . . . It has been the chiefly oral use of infer with a personal subject that has been under attack all along, and that seems not to pose much of a problem for writers.
Careful writers will still, no doubt, maintain the distinction, but criticism of the use in speech and, these days, in informal prose, of personal infer where others might prefer imply is more likely to be a sign of bigotry than of learning. MWDEU describes one of the objections to personal infer as social in that ‘the personal infer has been associated with uncultured persons.’ That might be said of many usages in the Negative Canon. If so, it ignores the fact that the language of ‘uncultured persons’ is just as capable of serving a particular communicative purpose at a particular time and in a particular place as that of the culturally sophisticated.