The world is full of "bad books"; not just uninteresting, or ill-informed, or morally repugnant books, but books that set out to present or defend positions that are insupportable in logic. I speak here not of books such as Hitler's Mein Kampf but of books that include Erich von Daniken's Chariots of the Gods? (1968), which presents "proof" of visits to Earth by extraterrestrials, or of Barry Fell's America B.C. (1976), which "proves" that ancient Celts reached North America before the time of Christ, or The Holy Blood and the Holy Grail (1982), in which Michael Baigent, Richard Leigh and Henry Lincoln purport to prove that lineal descendants of Jesus (and his wife, Mary Magdalene) walk among us. The Holy Blood has the additional distinction of having been the inspiration for Dan Brown's best-seller The Da Vinci Code (2003). Often these bad books become quite popular, and frequently gain a wider audience than good books on the same subjects. In discouraging my students from relying on such bad books, I began to wonder why they are popular. Few are models of prose style, although most provide a brisk enough narrative. Most of them are long, between 300 and 500 pages. Are we seeing here just the literary equivalent of Gresham's law, or is there something else going on?

As a Celticist, I have been particularly plagued by the books of the late Barry Fell. Students, members of the general public and news reporters seem to find his work irresistible. His books are even for sale in museum gift shops. Early one morning I telephoned a local news radio station in San Francisco to attempt to explain why it was most unlikely that, as they had gleefully reported, St Brendan and his crew had left extensive inscriptions in a cave in West Virginia in the 7th century, according to Fell. "But he's a Harvard professor!" argumentative enthusiasts would explain to me. "Of marine palaeontology," I would point out. "He's published dozens of articles on Libyan ogham (sic) in the New World!" "But all of them were published in his own journal, whose only peer reviewer is himself." And so on. In self-defence, as much as anything else, I began to look at how Fell and his ilk presented their arguments, and the more closely I looked, the more a set of rhetorical characteristics began to emerge that seemed to be common to many of these "bad books", regardless of their subject matter. In 2003, I decided to offer as part of an undergraduate course a freshman seminar on the rhetoric of bad books, "Bad books and how to spot them".

The freshman seminars are often an oblique introduction to the instructor's department, and it seemed that having bright and eager freshmen do close reading of even bad books was a step in the right direction appropriate to the department of rhetoric in which I teach. The bad news was that most of these students had actually never really dealt with books as books. Most of their school reading was from anthologies and few of them regarded books - codices - as more than physical frames for content, the context of which remained unexamined.

Their very ignorance, however, turned out to be an ideal way to approach the problem of recognising "badness" in books and a test for my contention that a careful reader, familiar with the conventions of publishing and argument, can learn to spot a bad book without knowing anything about the subject that the book purports to elucidate. The first book we looked at was The Alphabet Versus the Goddess: The Conflict Between Word and Image (1998) by Leonard Shlain. Shlain's general thesis in the book is that the introduction of writing caused the shift from matriarchy to patriarchy and from goddess worship to god worship in all the cultures around the Mediterranean. The first task I set my students was to look at the information surrounding the text. When was the book first published? Who published the book, and where? Is there an index in the back? Is there a bibliography, footnotes, endnotes? Are there maps, illustrations, an introduction? Is there biographical information about the author? When we began to look at the book itself, the students learned that the volume they were reading had originally been published in 1998, but that they were holding the 1999 first printing of the paperback version. The book has an index, and, as it happens, an extensive and useful one. There is also a bibliography, and it is here that a hint of potential "badness" first surfaced. Most of the bibliographical entries concerning anthropology, physiology, neuroscience and the like were reasonably up to date (S. F. Witelson, "Hand and sex differences in the isthmus and anterior commissure [sic] of the human corpus callosum: a postmortem morphological study", Brain, vol. 112, 1989: 799-835, for example), but a large number of works cited in the areas of culture and history seemed curiously antique (H. Schoenfeld, Women of the Teutonic Nations, Philadelphia, 1908). Granted, there is nothing wrong with consulting classic works from the past, Edward Gibbon's The History of the Decline and Fall of the Roman Empire, for example, but the bibliography seemed to show that the author either chose to consult and cite seriously old-fashioned scholarly works because more recent ones failed to support his argument, or that he simply did not know that recent and superior work had been done in these areas, or that he did not care. None of these explanations bodes well for the quality of the argument of the book itself.

³ÉÈËVRÊÓƵ

? And thus we can state The First Characteristic of Bad Books: the bibliography has strange features. It contains, for example, a large number of items that seem antiquated, unscientific or off the subject. The author cites himself repeatedly, or cites a very small number of sources. By way of comparison, in America B.C., Fell cites his own articles in the journal of the Epigraphic Society (which he edited himself) extensively and the work of recognised linguistic scholars of Celtic and Native American languages almost never.

Turning to the front of The Alphabet Versus the Goddess, we find an extremely enthusiastic preface in which the author describes his excitement while on a tour of the Mediterranean at the large number of shrines to goddesses (from several different cultures) which had fallen into desuetude in ancient times. Speculating on the reasons for this change, he hits on a "neuroanatomical hypothesis" to explain it. So we have a kind of mystery, uniting physiology, anthropology, history and mythology. He then promises shocking revelations: "My hypothesis will ask readers to reconsider many closely held beliefs and open themselves up to entirely new ways of looking at familiar events." We also learn from the preface that Dr Shlain is head of his surgery department at a major medical centre and experienced in operations involving the carotid arteries that supply blood to the brain.

³ÉÈËVRÊÓƵ

? The Second Characteristic of Bad Books: overenthusiastic prefaces by autodidacts. The front matter in bad books usually includes a preface with the following characteristics: the author has had his eyes suddenly opened to the existence of a mystery in a field in which he is not an expert (although he may be an expert in other fields) and the author has discovered the key to unravelling this mystery, a key that has been overlooked, disregarded or suppressed by experts. The introduction to Gavin Menzies' 1421: The Year China Discovered America (2002) opens with the sentence, "Over ten years ago I stumbled upon an incredible discovery, a clue hidden in an ancient map which, though it did not lead to buried treasure, suggested that the history of the world as it has been known and handed down for centuries would have to be radically revised." Menzies, by the way, is a retired naval officer.

Good books, of course, may also be written by autodidacts; Michael Ventris' decipherment of Linear B and Alexander Marshack's work on Palaeolithic drawing techniques come to mind. Neither of those independent scholars, however, offers the kind of breathless preface usually found in bad books, even though their discoveries actually did change "the history of the world as it has been known".

But what of the text itself? Here, too, there are rhetorical commonalities. The texts of these books all continue in the same excited first-person voice. They often introduce vague, undefined or invented terms. Shlain speaks of a "group psyche" and divides, quite arbitrarily, the mind into "three realms: inner, outer, and supernatural". Fell has invented "Libyan ogham", and the authors of The Holy Blood and the Holy Grail declare as an "indisputable historical fact" that there was a secret order behind the Knights Templar known as the Prieure de Sion which continues to operate in the present.

? The Third Characteristic of Bad Books: their central arguments depend on special definitions or special knowledge peculiar to the author.

A persistent rhetorical sequence in bad books is "assumption creep". Things described in early chapters as speculation or conjecture soon become likely, and are then taken as established facts. The question "Could the Cathar 'treasure' like the 'treasure' Sauniere discovered, have consisted primarily of a secret? Could that secret have been related in some unimaginable way to something that became known as the Holy Grail?" in chapter two has become by chapter nine "For if the Templars are indeed guardians of the Grail...the Grail existed not only in Arthurian times, but also during the Crusades." In 1421 we learn on page 75 that "three of these great fleets were placed under the command of Grand Eunuch Hong Bao, Eunuch Zhou Man and Eunuch Zhou Wen". By page 135, we learn that "Admiral Hong Bao's designated task was to chart the world eastwards from the fixed reference point established at the Falkland Islands", a conclusion based not on any surviving Chinese documents, but on a long chain of suppositions concerning a 15th-century European map of the Atlantic.

? The Fourth Characteristic of Bad Books: assumption creep. Confident conclusions are often the result of chains of circumstance and supposition so long that even remembering their origin points while reading the books is difficult.

Allied to assumption creep is the apparent abolition of Occam's razor. Bad books create arguments based not only on dubious premises and shaky data, but almost never resort to the simplest explanation available. The Holy Blood and the Holy Grail, for instance, asks why the Crusades (starting in the 11th century), the Cathar heresy (12th and 13th centuries) and the Grail legend (late 12th century) should arise "together". Coincidence is ruled out, although it seems the most probable answer.

? The Fifth Characteristic of Bad Books: rejection of the simplest and most logical explanations for observed phenomena, Occam's beard.

³ÉÈËVRÊÓƵ

Another interesting feature of many bad books is the half-comparison. Fell, for instance, compares the discovery of Irish ogham writing to Cotton Mather's discovery of some strange pre-16th century markings on the so-called Dighton Rock in Massachusetts. He proceeds at length to discuss and "decipher" the Dighton Rock inscriptions as Celtic ogham, but never actually completes his comparison. In addition to Fell, I seem to be haunted by the Dighton Rock. As a child, I was taken to see the stone when visiting my grandparents who lived nearby. So I was amused to see it brought forward as "evidence" by Fell, but dumbfounded when it turned up again in 1421, this time as "evidence" for a visit to Massachusetts by the Chinese fleet in the 1420s. I am tempted to suggest that any book using Dighton Rock as part of its evidence ought to be regarded as a bad book on that basis alone.

? The Sixth Characteristic of Bad Books: use of the half-comparison. Similarly, data or proof are often said to be coming in later chapters, but in the event do not appear. 1421, for example, refers readers to its website for data and proof.

But why do people seem to prefer books peddling snake oil to books peddling antibiotics? My guess is that it is exactly those rhetorical features signalling the books' "badness" that account for their popularity. Copious bibliographies and footnotes provide credentials for the author's gravitas. Breathlessly enthusiastic prefaces and claims of unveiling secrets make the reader look for further exciting revelations; and "outsider" status, somewhat paradoxically, can be taken as evidence of the writer's lack of bias. Assumption creep seems to be understood as the writer's growing confidence in his conclusions rather than as the definitional sleight of hand it actually is. Unfulfilled promises of information-to-come in later chapters move the narrative forward while obscuring weaknesses in data, although half-comparisons and the failure to apply Occam's razor probably ought to be red flags to almost anybody.

I fear that most of us are susceptible to confident and daring assertions about the unknown and the counter-intuitive, however, so I am betting that bad books are here to stay. I just want to warn my students.

What makes a book 'unpickupable?'

Susan Bassnett, professor of comparative literature in the department of English at the University of Warwick

³ÉÈËVRÊÓƵ

What makes a bad book? Well, sometimes it isn't the book itself, it is where we as readers happen to be at the time we encounter it. We might rate a book bad and then years later reassess our views, or vice versa. I recently gave away a pile of books I had never been able to finish, all of which had "postmodernism" or "postcolonial" somewhere in the title, because in the 1990s those were fashionable buzzwords.

Somewhere on my shelves is a book about aliens visiting the Earth and drawing the Nazca Lines (in Peru), which attracted a global readership and must have made the writer a lot of money.

High on my list of Really Bad Books are two best-sellers: Dan Brown's The Da Vinci Code and Hilary Mantel's Wolf Hall, both of which I rate as dreadfully badly written. Brown wrote to a computer game formula: solve one level and move on to the next, whereas Mantel just wrote and wrote and wrote. I have yet to meet anyone outside the Booker panel who managed to get to the end of this tedious tome. God forbid there might be a sequel, which I fear is on the horizon.

Tom Palaima, professor of Classics at the University of Texas at Austin

A prime example of a bad book is Victor Davis Hanson and John Heath's Who Killed ³ÉÈËVRÊÓƵr? The Demise of Classical Education and the Recovery of Greek Wisdom (2001). US universities had already become "employment credentialing stations" when Dick Cheney's "pet classicist" and favourite Iraq War champion, Victor Davis Hanson, and the classics professor John Heath took 300 pages to pillory classics scholars for metaphorically killing ³ÉÈËVRÊÓƵr and "fail[ing] the country". There were more weapons of mass destruction in Saddam Hussein's Iraq than sound arguments in Who Killed ³ÉÈËVRÊÓƵr? One telltale sign to look for in a bad book: when authors tell you in their introduction what they have not done in the book they have written, you know what you will find in the book you are about to read.

Valerie Sanders, professor of English at the University of Hull

Bad books are unreadable books, in my view, irrespective of what they say. Hard-core literary theory of the 1980s must be responsible for some of the most impenetrable and jargon-ridden prose of the past quarter century: "The center is at the center of the totality, and yet, since the center does not belong to the totality (is not part of the totality), the totality has its center elsewhere" [sic]. This is actually from one of Jacques Derrida's essays, Structure, Sign and Play in the Discourse of the Human Sciences, but any of his books would do - along with any author who dares to call George Eliot's Middlemarch "an autonomous signifying practice". As for novels, J.R.R. Tolkien's The Lord of the Rings every time ("'Ware! Ware!' cried Damrod to his companion. 'May the Valar turn him aside! M?mak! M?mak!'").

Thomas Docherty, professor of English and comparative literature, University of Warwick

I don't think that the world is full of bad books, but I do think that it is full of books that have not yet found a reader adequate to the task of reading them. And I am, of course, one such reader. The real task is not to make ungenerous judgement from a position of critical superiority but rather to find a critical humility that allows for the possibility of reading, for the necessity of re-reading and, above all, to respond to the great call from Rainer Maria Rilke that, when faced with art, "you must change your life".

Steve Fuller, professor of sociology at the University of Warwick

These three books - popular in their day - each exemplify how academics are all too willing and able to play the "ignorance is strength" card:

Paul Gross and Norman Levitt's Higher Superstition: The Academic Left and Its Quarrels with Science (1994). This book set the tone for PoMo-bashing that inspired the Sokal Affair (a publishing hoax). The mere fact that a mathematician and a biologist can't make easy sense out of what humanists say about science apparently licenses an endless stream of out-of-context quotes bathed in invective.

Steven Pinker's The Blank Slate: The Modern Denial of Human Nature (2002). Pinker argues like a high-school debater who hides a superficially attractive thesis in mounds of data that have been intellectually asset-stripped, so as to hide any theoretical or methodological doubts that might be raised about the thesis.

Richard Dawkins' The God Delusion (2006). This book demonstrates the asymmetry of academic standards. A book by a theologian speaking this ignorantly about biology would never have been published, let alone become a best-seller. Unfortunately, it doesn't seem to work the other way round.

Roger Luckhurst, professor in modern and contemporary literature at Birkbeck, University of London

³ÉÈËVRÊÓƵ

Working in the weird borderlands of literary and critical theory, one can come across some pretty bad books. When a theorist becomes fashionable, presses rush to publish their flimsiest peelings (which is not necessarily their fault). There are quite a few expensive little books by the likes of Jacques Derrida or Giorgio Agamben that the world could probably get along well without. But the worst book? Surely my own first attempt, The Angle Between Two Walls: The Fiction of J.G. Ballard (1997). There's some truly horrible, obscurantist prose in there.