Read Shakespeare's Richard III fully cover to cover and you'll find five uses of the verb infer, only one of which seems to obey the aggressively touted dictum that infer can only mean 'to take from what is being said' or 'to deduce from a premise.'
To cite just one example, in Act 3, Scene 5, Richard of Gloucester is giving talking points to his fixer Buckingham, who is to follow the mayor to Guildhall and:
There, at your meetest vantage of the time,
Infer the bastardy of Edward’s children.
Tell them how Edward put to death a citizen
Only for saying he would make his son
Heir to the Crown
One can imagine the comments-section brawl such a usage would spawn today: "Like, duh, imply the bastardy of Edward’s children. How can we trust anything this idiot says?! What this bozo has done to the language of Shakespeare is unforgivable!"
According to the Merriam-Webster Dictionary of English Usage, however, it is only in the last hundred years that usage guides and schoolteachers have been insisting that infer is, in a way, the opposite of imply, or as the 1957 Dictionary of Contemporary American Usage put it: "The speaker implies, the hearer infers."
However recent, the campaign to build a Great Wall between imply and infer appears to have been a success, and the rule is now sacrosanct among adult educated users of Standard English. Once mastered by the careful speaker, it is rarely violated, or at least that’s my impression. As an editor, I doubt that I have ever had to correct for it.
The purported difference between imply and infer was the source of some controversy in 1961, when Webster’s Third used forms of each word to define the other:
imply: "to indicate or call for recognition of as existent, present, or related not by express statement but by logical inference …"
infer: "to derive by reasoning or implication …"
In the critical storm that followed the dictionary’s release (which I chronicle in The Story of Ain’t), Life magazine, the New York Times, the American Bar Association Journal, and Dwight Macdonald in the New Yorker all cited the definitions for imply and infer as symbolic of the carelessness and perfidy that (in their telling) made Webster’s Third the "permissive" dictionary of "anything goes."
But the words are clearly related and exist in a semantic gray area that, by definition, brings with it a certain amount of ambiguity. Even critics of Webster’s Third are forced to agree. Take William Morris, editor of the first edition of the American Heritage Dictionary, which appeared in 1969 and was marketed as an alternative to Webster’s Third. While English-language dictionaries have long relied on specialists, the American Heritage Dictionary was the first to formally establish a panel of outside literary generalists to help the lexicographer adjudicate usage disputes.
In 1985, William and his wife Mary convened a similar such panel when they published the Harper Dictionary of Contemporary Usage, a lighthearted selection of disputed usages commented on by a selection of 165 people, including a handful of critics of Webster’s Third (Nero Wolf creator Rex Stout, Dwight Macdonald, Sheridan Baker), numerous editors, and many writers, though precious few linguists. Still, if there were a simple consensus among users of Standard English about the difference between imply and infer, certainly this group would know about it.
The question to the usage panel read:
A brochure of the Magazine Publishers Association recently quoted Dr. Samuel Johnson as having written: 'Advertisements are now so numerous that they are very negligently perused.' Making its point, the copy continued: 'Sam then went on to infer that reach and frequency were really of no importance unless there was perception of the advertising as well.' Dictionaries define infer as 'to deduce from evidence,' while imply is defined as 'to express indirectly, to suggest.' Which do you think Johnson did?
Which indeed, for it seems possible to argue on behalf of either. Before he could have implied (stating indirectly) that perception of an advertisement was key to its success, Johnson surely inferred this from the evidence of his own experience with advertisements. Forced to choose, and given the structure of the sentence—"… reach and frequency were really of no importance unless there was perception"—I would vote for infer. The sentence seems to be emphasizing a logical deduction rather than any kind of utterance. But without more information, I would only be guessing. Apparently, I'm not alone. Twenty-eight percent of the usage panel thought the correct answer was infer. Seventy-two percent thought it was imply.
Curiously, the editors of the Harper Dictionary of Contemporary Usage did not offer their own view. They said only that the distinction between imply and infer can be confusing even to the "most erudite scholars." Fine then. Call me erudite.