When I taught linguistics-for-teachers courses, I spent a significant portion of my class time trying to get students to question their assumptions about language, assumptions that, whether they learned them in school or by general osmosis, are based on premises that linguists know to be incorrect. It always distressed me, therefore, when certain students would make it to the end of the course and drop some comment that made it clear they had internalized little of what I was trying to communicate. I'm not talking here about students who were floundering. I mean students who reached the end of the course with decent competence in the basics of linguistic analysis. They could do broad phonetic transcription, show the syllable structure and morphology of words, and perform a basic syntactic analysis of their own writing—not perfectly, perhaps, but with enough skill to make it clear that they were learning something. And yet some of them still appeared to have the same attitudes about language with which they entered the course. I don't think this represented conscious resistance. In other words, I never had the feeling that they carefully weighed my arguments against those for prescriptive grammar and said to themselves, "No, the other side has the better position." I didn't even sense that they were rejecting my position outright through mere stubbornness ("I know what I think, and nothing you say will change my mind.") Rather, it was as if they had taken the course material into a compartment in their minds without recognizing that it was incompatible with their other attitudes. They deferred to me as an authority figure, accepting what I told them, but when they went into service as teachers, I am sure that they deferred to what the next authority, their textbooks, told them. What made this particularly frustrating was that my primary purpose in these courses was to equip students with the analytical tools to think their way through conflicting statements about language. I was immediately reminded of this frustration when reading the introduction to Robert J. Gula's Nonsense: A Handbook of Logical Fallacies (2002). He presents an excellent list of the habits people have that interfere with logical reasoning. It's a bit lengthy, but worth quoting in full. People, Gula notes,

1. tend to believe what they want to believe.
2. tend to project their own biases or experiences upon situations.
3. tend to generalize from a specific event.
4. tend to get personally involved in the analysis of an issue and tend to let their feelings overcome a sense of objectivity.
5. are not good listeners. They hear selectively. They often hear only what they want to hear.
6. are eager to rationalize.
7. are often unable to distinguish what is relevant from what is irrelevant.
8. are easily diverted from the specific issue at hand.
9. are usually unwilling to explore thoroughly the ramifications of a topic; tend to oversimplify.
10. often judge from appearances. They observe something, misinterpret what they observe, and make terrible errors in judgment.
11. often simply don't know what they are talking about, especially in matters of general discussion. They rarely think carefully before they speak, but they allow their feelings, prejudices, biases, likes, dislikes, hopes, and frustrations to supersede careful thinking.
12. rarely act according to a set of consistent standards. Rarely do they examine evidence and then form a conclusion. Rather, they tend to do whatever they want to do and to believe whatever they want to believe and then find whatever evidence will support their actions or their beliefs. They often think selectively: in evaluating a situation they are eager to find reasons to support what they want to support and they are just as eager to ignore or disregard reasons that don't support what they want.
13. often do not say what they mean and often do not mean what they say.

To these principles, let's add four observations cited by J.A.C. Brown in his Techniques of Persuasion: "Most people want to feel that issues are simple rather than complex, want to have their prejudices confirmed, want to feel that they 'belong' with the implication that others do not, and need to pinpoint an enemy to blame for their frustrations."

These observations crystallized for me what was going on with my students. For that matter, they apply with perfect justice to many prescriptivist writers (particularly points 11 and 12). One reason that linguists have had so little effect upon the idiocy that passes for language commentary among the general public is that prescriptivism is rhetorically effective. It gives people who practice it a chance to belong and to look down on others whose language isn't as "pure" as it should be. It manages to tap people's emotions and reinforce what they want to believe. It's easier to think of language as existing external to its users in some sort of platonic form. It's easier to ignore all the messy variation and believe that every language has a single correct form, and that anything which departs from that form is wrong. Linguists, of course, are not immune from some of these rhetorical tendencies. They too play on the in-group mentality. The difference is that linguistic arguments are backed by rational argument and empirical evidence. If there's any hope for us to change entrenched attitudes, we have to attend to what makes prescriptivism so emotionally appealing and turn those appeals to better ends. For example, I think the think the typical linguistic argument in favor of descriptivism is rhetorically weak, particularly for the crucial audience of language-arts teachers. Descriptivism and prescriptivism are typically presented as a dichotomy. From the perspective of linguists, descriptivism is good and prescriptivism bad. And if you're working as a linguist, of course, that's the right approach. Describing the way language works is what linguists do. Prescription runs contrary to basic scientific method, and so those who are preparing to become linguists need to be disabused of any prescriptive baggage they may be carrying. But look at it from the perspective of a high school English teacher. Her overriding goal is to make her students fluent readers and writers of Standard English. No matter how sensitive to and respectful of students' home language that teachers is, at some point, she has to prescribe: this is the way Standard English is done. If we present descriptivism and prescriptivism in binary conflict, we risk having teachers, who quite rightly see part of their job as prescription, saw the wrong conclusion, saying in effect, "I have to tell students how to do things, so I guess I'm a prescriptivist." And then they turn to traditional materials with a clear conscience. But the real problem with prescriptive material is not laying down particular rules and saying, "This is what you must do to write Standard English." The problem is that the books typically used in K-12 education also contain a series of suspect assumptions and outright factual errors. Those problems are not a logically necessary part of prescriptivism, although they are precisely the reason linguists are so dismissive of traditional grammar books. Rather than posing the issue as a conflict between descriptive and prescriptive approaches to grammar, it might be more productive to talk about textbooks with outdated information. It's easy to imagine a prescriptive grammar based on a sound descriptive account of English. That is what we expect from good foreign-language textbooks, and there's no reason to expect less from works intended for native speakers. No one would accept a physics textbook that ignored the past century of discoveries in physics. And yet the ordinary textbooks that purport to describe English grammar take precisely that approach to their subject. It is an utter scandal, and if there's any place to start trying to change opinions, it's with the people who write (and purchase) those textbooks.