Bindweed is a thing of evil genius. Try to dig it up and it breaks into pieces, each of which will grow into a new plant. More than just robust or flexible, it actively thrives on disruption. Yet we lack a word for this property  which Nassim Nicholas Taleb describes as the exact opposite of fragility. So he has given it one: antifragility.

Taleb argues that antifragility is essential for dealing with the subject of the book that made his name: black swans. These are rare, unpredictable events for which we can never be fully prepared. What we can do, however, is locate the fragility in the system and reduce it. "Not seeing a tsunami or an economic event coming is excusable," he writes; "building something fragile to them is not."

Taleb is also right to point out that many things require some kind of stress or irregularity to function well, such as our bodies. If the environment is too disinfected, we lose the ability to resist infection. Vaccines introduce a little of what is harmful to build resistance to disease. True, but that does not make us the opposite of fragile. On the contrary, we are all-too fragile, vulnerable to being destroyed by one dose of the wrong poison or virus, one smash by the wrong vehicle at the wrong speed. What Taleb calls antifragile only loves volatility "up to a point": even bindweed dies if you stress it too much. But there is no "up to a point" about fragility's hatred of volatility: a vase will not refuse to shatter if you drop it from high enough.

Nothing is the exact opposite of fragile, and that is why "There is no word for 'antifragility' in the main known languages, ancient, colloquial or slang." Taleb should have heeded his own warning against the hubris of thinking we are wiser than thousands of years of human culture.

This is far from the only time when Taleb overstates his case, committing errors he attributes to others. His overconfidence, ironically, makes his arguments more fragile to refutation than they needed to be. This is a shame, because his book is full of important warnings and insights. For instance, greater efficiency is making us more fragile: ever-leaner supply chains and production methods mean no slack is built into the system. The self-employed are more, not less, financially secure than the employed, because they can constantly adjust their offering to respond to changing demand, whereas the employed lose everything if their jobs go. He also argues that the exact same theory, dismissed as trivial or common sense, will be lauded as brilliant if presented in the form of a complicated mathematical formula. And that the rule of unintended, unforeseen consequences means that you do not need positive evidence of harm to be wary of major health interventions. And that we tend to place too much weight on there being a good theory to explain why something works, rather than simply looking at whether it does so or not.

Despite its many virtues, Taleb's book is hard to like, since while its author endorses Nietzsche's maxim "What is not intelligible to me is not necessarily unintelligent" he writes according to the opposite principle, dismissing with a sneer all with which he disagrees, such as those in the "phony professions" of journalism and academia. If my conclusion seems harsh, remember Taleb himself believes that what he would say about someone after a third bottle of wine, he is obliged to say in print. Believe me, I haven't even pulled the second cork.

The core idea behind this book is simple and quite enticing. Nassim Nicholas Taleb divides the world and all that's in it (people, things, institutions, ways of life) into three categories: the fragile, the robust and the antifragile. You are fragile if you avoid disorder and disruption for fear of the mess they might make of your life: you think you are keeping safe, but really you are making yourself vulnerable to the shock that will tear everything apart. You are robust if you can stand up to shocks without flinching and without changing who you are. But you are antifragile if shocks and disruptions make you stronger and more creative, better able to adapt to each new challenge you face. Taleb thinks we should all try to be antifragile.

If the idea is nice and neat, however, the book that houses it is just the opposite. It is a big, baggy, sprawling mess. Taleb seems to have decided not just to explain his idea but also to try to exemplify it. One of his bugbears is the fragility of most of what passes for "knowledge"  especially the kind produced by academics  which he thinks is so hung up on order and completeness that it falls apart at the first breath of disruption. So he has gone for deliberate disorder: Antifragile jumps around from aphorism to anecdote to technical analysis, interspersed with a certain amount of hectoring encouragement to the reader to keep up. The aim, apparently, is to show how much more interesting an argument can be if it resists being pinned down.

There are two problems with this. First, the book is very hard going. Everything is taken to link to everything else but nothing is ever followed through. Taleb despises mere "theorists" but still aspires to produce a theory of everything. So what we get are lots of personal reminiscences buttressed by the ideas of the few thinkers he respects, almost all of whom happen to be his friends. The result is both solipsistic and ultimately dispiriting. Reading this book is the intellectual equivalent of having to sit patiently while someone shows you their holiday snaps.

The other difficulty is that too many of the ideas contained here appear thin and brittle rather than rich and flexible: fragile rather than antifragile. Taleb is keen on "heuristics"  shortcuts to wisdom that encapsulate human experience  but often these seem simply to reflect his own prejudices. To take just one example: Taleb thinks modern states become fragile when they get into debt, and that a prerequisite of political antifragility is rigid fiscal conservatism. This is nonsense. Eschewing debt makes states just as fragile as having too much of it. The durability of both the British and American states throughout their history has depended on their ability to use public debt to adapt to different challenges. As political analysis, Taleb's heuristic  "when you don't have debt you don't care about your reputation  and somehow it's only when you don't care about your reputation that you tend to have a good one"  is glib and unconvincing.

Antifragile is trying to be two things at once: a philosophical treatise and a how-to guide for living. Taleb's two previous books  Fooled by Randomness and The Black Swan  drew their appeal from being more narrowly focused on the failures of economists and financial traders to understand the game they were in. Their enormous success derived in part from his apparently being proved right by the financial crash of 2007-08. But now Taleb wants more than just vindication: he wants long-term intellectual respect. He makes a great play in this book of denigrating those earlier volumes as somehow lesser versions of his big idea. He says Antifragile, along with a technical treatise he published before he became famous, are by far his favourite pieces of writing. If I may be forgiven a heuristic of my own, it is a very bad sign when authors start to look down on the books that connected them to their audience: it means they are now irredeemably up themselves.

As a how-to guide Antifragile is a mixture of the pretentious and the banal. Some of this is deliberate provocation: we are told that real scholarship depends on having a private library rather than learning in the classroom. But much of the advice is just a warmed-up anti-health-and-safety rant with a bit of Nietzsche thrown in. Relying on gyms and doctors make us ill. We all eat too much: better to avoid breakfast. Our kids are being cosseted into fragility by "soccer mom" parenting: we need to let them toughen up. The childrearing implications of Taleb's argument illustrate some of its limitations. Being a parent is an inherently fragile business, given the permanent possibility of something going disastrously wrong. Of course, one way to avoid that would be to live in a world where people are accustomed to their children dying young. Taleb is deeply and depressingly nostalgic for the virtues of the ancients, with their stoicism and tolerance for suffering. To want to return to the miseries of a world that requires such virtues strikes me as ridiculous.

Antifragile is not all bad: it has flashes of wit and insight. Taleb says the least antifragile state in the world at the moment is Saudi Arabia, a plausible claim though one it would be nice to see argued out. He is good at knockabout invective, laying into "fragilista" economists such as Joseph Stiglitz and self-aggrandising journalists such as "the vile and harmful" Thomas Friedman, who was an apologist for the Iraq war. And there are some nice lines: "We practitioners and quants aren't too fazed by remarks on the part of academics  it would be like prostitutes listening to technical commentary by nuns." The problem, though, is that Taleb no longer writes from the perspective of the practitioner but of someone who has crossed over to live among the academics and wants to tell them what they are missing. He is now more like the nun with a racy past who lectures the rest of the convent about the meaning of sex: not much fun for anyone.

He says books and their authors should be antifragile too. That means that negative reviews should be welcomed: "Criticism for a book is a truthful, unfaked badge of attention, signalling that it is not boring." He even specifies a bad review given by an academic to a popular author in these pages recently (Glen Bowersock's review of Tom Holland's In the Shadow of the Sword) as a reason for going out to buy the book. At the same time, he suggests that a useful heuristic for book readers is not to bother with anything less than 10 years old, since most recently published books will soon reveal themselves to be worthless. He thinks you should abandon a book as soon as it starts to bore you (so don't be a book reviewer: I had to plough on with this one to the bitter end). He thinks all criticism of his work is essentially ad hominem (though with a book as self-referential as this it's hard to know what else there is to do). All in all, Taleb is not going to care about anything I say here.

Still, this book should be approached with caution. We do live in a fragile world, vulnerable to extreme shocks. But antifragility is not the solution. It is too crass an idea, and Taleb, for all his vaunted intellectual curiosity, is not really curious about the lives of anyone who doesn't live like him. He says it's better to be a taxi driver than a stockbroker, because you are less exposed to the whims of others. Let him try it. He thinks it's better to be a mafia hard man than a tenured academic. Again, let him try it. The problem with Antifragile is that it is a deeply antisocial book. I am pretty sure people will still be reading Taleb's two previous books in 10 years' time. But I'd be surprised if they are still reading this one.

 David Runciman's Political Hypocrisy is published by Princeton.