The canonical form of academic communication is the oral delivery of a written text. PowerPoint may be increasingly used, but the general idea remains the same: academics think out loud by following a script. We take this sort of activity as indicative of deliberate thought, careful reasoning and, not least, respect for the audience.

But perhaps all it shows is that academics literally don't know what they're talking about. We are the ugly autocuties. More to the point, it may prove that we are not very intelligent - or at least don't know how to recognise or value intelligence. Truly intelligent people know how to improvise well. Academics don't.

One would think that a life spent single-mindedly immersed in a subject would make academics at least as adept performers as professional actors, who spend much less time ranging over many more roles - and possibly leaving a much more lasting impression. On the contrary, the one aspect of academic life in which such skills might be cultivated, classroom lecturing, tends to be treated with considerable suspicion, especially if it elicits a strong response, either positive or negative, from student audiences.

Education may well play second fiddle to research in academia today. But even those who claim to defend education quite happily dump the lecture as a medium of instruction. I suppose they think that if the score is all that matters, then the difference between a virtuoso and a novice performer doesn't matter, as long as the right notes are hit most of the time.

³ÉÈËVRÊÓƵ

Alas, improvisation is a skill that goes unrewarded, if not actively despised, in academia. When in doubt, we quickly reach for the "bullshit detector" to belittle the improviser. Yet too much of what passes for intellectual activity in our own quarters is little more than meeting well-defined expectations, if not sheer template-matching. We decry rote memorisation merely because it is old-fashioned. But aren't today's cut-and-paste PowerPoint presentations even dumber? No wonder hope springs eternal for mechanical models of artificial intelligence.

Moreover, administrators are not to blame for this behaviour. If anything, they might be thanked for rewarding tendencies that academics already spontaneously display. But perhaps our gratitude is already expressed in the ease with which we adapt to each new academic audit regime.

³ÉÈËVRÊÓƵ

We hate it when students ask what counts as a good answer, but only because they realise that exam-taking is little more than literary marksmanship. At a slightly more sophisticated level, shadows may be cast over an academic's entire body of work, especially within one's own discipline, by showing that he or she fails to say or do particular things just the right way. To use the crypto-moral, pseudo-rigorous term that academics prefer on these occasions, such a colleague is "unreliable".

But why should failure to conform in detail be taken as indicative of anything deeper than minor errors? No known theory of rationality supports a harsher judgment. But then theories of rationality tend to value accuracy only in so far as it bears on something relevant, while academics value relevance in so far as it accurately hits their expertise.

There was a time - say, 500 or more years ago - when it made sense for academics to, quite literally, "do things by the book". Back then, reading from a prepared text provided students with an opportunity to copy something required for their studies that they would otherwise not possess. Of course, this was before the printing press, personal libraries, cheap paperbacks, let alone the world wide web. Yet those bygone days of primitive knowledge transmission persist not merely in our bad lecturing habits but also in the very fact that we call it a "lecture", which derives from the Latin for reading out loud.

Don't misunderstand: there's nothing wrong with an academic using a pile of books as a prop, from which he or she pulls one "at random" to read a few sentences sonorously and then launch into a riff that is the intellectual equivalent of virtuoso jazz or John Sessions at his best on Whose Line Is It Anyway? The attractiveness of this practice, more so in Franco-German than in English circles, rests on the idea that imitation may be the sincerest form of domination.

When old-school humanists complain that today's students read more literary criticism than literature, they are not merely bemoaning the search for intellectual quick fixes. After all, much literary criticism is at least as difficult as the literature it claims to criticise. Rather, the humanists begrudge the talents of the academic improviser whose own performance can make the original text look much less interesting by comparison - should one then bother to read it.

Indeed, one literary critic, Harold Bloom, made his career at Yale University by using a little Freud to generalise this point into a strategy for achieving literary greatness. In The Anxiety of Influence (1972), Bloom argued that all poets live in the fear that their dead intellectual fathers - that is, the people whose work they have plagiarised - will be discovered. The great poets turn out to be the ones whose improvisational skills allow them to get away with murder.

In the case of academic culture, the recent metastasis of the plagiarism taboo into a full-blown moral panic over the security of intellectual property rights suggests that behind the disdain for improvisation may lurk an admission of intellectual weakness. After all, would the stealing of text and ideas appear so criminal if we took seriously that, regardless of its provenance, any product of the mind is open to multiple novel uses and that, in any case, there are always many more things to think and say?

Here we catch sight of the sin that improvisers commit: they refuse to say exactly what they know or know exactly what they say. On the one hand, they present the established as if it were novel. That's plagiarism. On the other, they present the novel as if it were established. That's bullshit. From this unholy alliance of plagiarism and bullshit, the improviser conspires to make a virtue out of unreliability.

³ÉÈËVRÊÓƵ

A more relaxed view towards matters of reliability can often be found among those outside the academic's home discipline, a safe distance away from its guild pieties. To be sure, this can result in quite literally "fabulous" judgments of the intellectual merit of people's work. A good case in point is Thomas Kuhn, the populariser of "paradigm" and "scientific revolution", who was the most influential theorist of science in the second half of the 20th century.

³ÉÈËVRÊÓƵ

Kuhn is celebrated as an historian, philosopher, sociologist and even psychologist of science. However, as someone trained only in physics, he skated on thin ice in all these fields and, by practitioner standards, often fell through the cracks. Yet he is hailed as a genius.

And Kuhn would have been a genius, if what he achieved was what he had set out to do. He would have been a great improviser who started from clear themes in various disciplines, taking them in novel directions that ignored or contradicted actual developments in those disciplines. Like the great actors and musicians given to improvisation, his genius would have then rested on having demonstrated enough technical mastery to allow his audience to ignore any remaining deficiency. Why? The freedom permitted by that deficiency resulted in something superior to what would have resulted from merely doing things right.

But Kuhn proved to be more idiot than savant. He was completely flummoxed by his reception and refused to play along with it. Indeed, his subsequent work left the impression that his magnum opus, The Structure of Scientific Revolutions (1962), had been a fluke. The true improviser, if he admits to bluffing at all, would do so only at the end of his career.

Consider Kuhn's French contemporary, Louis Althusser, the most influential Western Marxist of the 1970s, who admitted in his memoirs that most of what he knew about Marx came from hearsay and from his students' essays. To be sure, his star has fallen much more precipitously than Kuhn's. Even worse than an idiot is a savant who makes everyone else look like an idiot.

But that trait alone need not be damning, if others eventually come along and play your role straight, with everything laid out in good academic order. Unfortunately, Althusser's improvisational skills were devoted to intellectually engaging defences of Marxism's most totalitarian tendencies. Had the Communists won the Cold War, he might still be considered a genius.

Lest I unduly discourage would-be improvisers, let me turn now to the grandmaster of intellectual improvisers, someone whose exquisite sense of world-historic timing allowed him to stay one step ahead of his doubters, while remaining one step behind his vindicators. I refer to that icon of the scientific revolution, Galileo Galilei.

Galileo was full of what the Italian Renaissance called sprezzatura, a broad term for the improviser's art that runs the gamut from "making it up on the spot" to "thinking for oneself". The data of his famed experiments were massaged, when not outright manufactured. He claimed great powers for a telescope whose images were blurry and whose optics he could not explain. On top of all that, Galileo did not merely propose an alternative view of the cosmos. Unlike his timid precursor Copernicus, he explicitly said that his view should replace the Church's.

No doubt, had the next two generations of scientists not been inspired to make good on Galileo's improvisations, his reputation would now be languishing either in Purgatory, alongside Kuhn's, or in Hell, alongside Althusser's.

The stories of Kuhn, Althusser and Galileo - and their varying fates - highlight improvisation's unique brand of intelligence. Its genius can be maintained only by sustained engagement with the audience, since it is quite literally an art that is made up as one goes along. Often that interaction consists in reassuring the audience that each strange turn is indeed for the better, as the suspension of disbelief is rewarded by a pleasant surprise that causes one to reassess all that had previously transpired. In that respect, as we saw with Galileo, the improviser is always playing for more time.

So what would an improvisation-friendly academia look like? Certainly standards of public performance would shift. We would become more tolerant of people who speak crudely without notes, if they can improve as they take questions from the audience. But we would equally become less tolerant of people who refuse to take questions simply because they stray from their carefully prepared presentation. Instead of "sloppy/rigorous", we would apply the binary "expansive/ limited" to describe the respective intellects of these people.

This shift in standards may have some interesting knock-on effects. Academics might question the "value added" of inviting a speaker who does little more than rehearse well-publicised writings. They might become more assertive in proposing to speakers topics outside their intellectual comfort zones, perhaps even as a condition of their invitation. Overall, these developments might convert the lecture from the usual rehydration of freeze-dried thought into an occasion for full-blooded inquiry.

³ÉÈËVRÊÓƵ

My first taste of academic improvisation came in maths class in a Jesuit high school. Students were invited to complete a geometric proof. Even if the first student got it right, the teacher would continue to ask: "Does anyone think otherwise?" Without fail, many of us would propose all sorts of ingenious - usually wrong - alternatives. The teacher would wait for us to run dry and then break the bad news. But he would go on to show that most, if not all, of those alternatives had a sound basis that worked in other contexts. None of us knew what to expect of each other in such classroom encounters, yet they almost always worked. After all, the cause of inquiry is better served by being interestingly wrong than being reliably right.