In our ongoing effort to decipher the meaning of grammar (see What Is Grammar?) and distinguish its many varieties (Ten Types of Grammar), we go back almost 60 years to an article by W. Nelson Francis, a linguist at Brown University and one of the co-authors of Computational Analysis of Present-Day American English (1967).

In this excerpt, Francis points to some of the problems that crop up when usage (which he calls "Grammar 3") is confused with either mental grammar ("Grammar 1") or descriptive grammar ("Grammar 2").

Grammar 1
The first thing we mean by "grammar" is "the set of formal patterns in which the words of a language are arranged in order to convey larger meanings." It is not necessary that we be able to discuss these patterns self-consciously in order to be able to use them. In fact, all speakers of a language above the age of five or six know how to use its complex forms of organization with considerable skill; in this sense of the word--call it "Grammar 1"--they are thoroughly familiar with its grammar.

Grammar 2
The second meaning of "grammar"--call it "Grammar 2"--is "the branch of linguistic science which is concerned with the description, analysis, and formulization of formal language patterns." Just as gravity was in full operation before Newton's apple fell, so grammar in the first sense was in full operation before anyone formulated the first rule that began the history of grammar as a study.

Grammar 3
The third sense in which people use the word "grammar" is "linguistic etiquette." This we may call "Grammar 3." The word in this sense is often coupled with a derogatory adjective: we say that the expression "he ain't here" is "bad grammar." . . .

As has already been suggested, much confusion arises from mixing these meanings. One hears a good deal of criticism of teachers of English couched in such terms as "they don't teach grammar any more." Criticism of this sort is based on the wholly unproven assumption that teaching Grammar 2 will improve the student's proficiency in Grammar 1 or improve his manners in Grammar 3. Actually, the form of Grammar 2 which is usually taught is a very inaccurate and misleading analysis of the facts of Grammar 1; and it therefore is of highly questionable value in improving a person's ability to handle the structural patterns of his language.
(W. Nelson Francis, "Revolution in Grammar." Quarterly Journal of Speech, 40 [1954], pages 299-312)

In The Language Instinct (1995), cognitive linguist Steven Pinker makes similar distinctions: "A preschooler's tacit knowledge of grammar," he says, "is more sophisticated than the thickest style manual or the most state-of-the-art computer language system." In other words, grammar (Francis's #1 or #2) shouldn't be confused with "correct" or "proper" ways of speaking and writing (#3).