Continued from yesterday. Here is the first half of Part 3 of my new essay on writing and publishing. Please remember that, if at any point you decide that you would like to have all of this essay available in one lump, you can go here and download the whole thing as a PDF file. You can then print it out, pass it to friends, et cetera.
Part 3: The experiment with rats when applied to publishing
Applicability and relevance
Taleb’s ideas about randomness have proved useful in facilitating clear thinking in a number of fields of activity: for example, Card Player magazine tells us that they are applicable to the game of poker. It will therefore be useful to go through each of the points made in Part 1 and to examine their relevance to the book trade in general and the writer’s position in particular.
The first thought that struck me, on reading Taleb’s draft chapter for his book on black swans, was that the experiment with rats is closely analogous to the process of selecting books from the slush pile. And, just in case there is anyone reading this who doesn’t know what a slush pile is, let me explain.
Defining the slush pile
Unknown and unpublished writers tend to be, as remarked in Part 1, wildly ambitious and eager for fame, money, and literary reputation. To achieve these objectives they have to get published. And to get published they have to arrange, as a first step, for their work to be offered to publishers.
In the past, it was common for novelists to submit their completed manuscripts to publishers themselves. Every day the postman would deliver, to every publisher in the land, a pile of ten or twenty manuscripts. These unsolicited submissions are known in the book trade, throughout the English-speaking world, as the slush pile.
The term ‘slush pile’ gives a clear flavour of the contempt in which unsolicited submissions are held. It is widely agreed in publishing circles (on the basis of countless years of experience) that many of these manuscripts will be unreadable, unpublishable junk. But it is also the case (as history demonstrates) that the slush pile will occasionally contain a black swan.
One point to note is that every writer, and every novel, is at some point in someone’s slush pile. With absolutely no exceptions.
At some stage, and possibly at many different stages, decisions have to be made on whether to continue to consider a book for publication, or to send it back to its author with a rejection slip. This iterated process has its parallels with the rats-in-the-vat experiment. The rats which were submitted to radiation included every type of rat: fat, thin, strong, weak, young, old. Similarly the slush pile contains writers and manuscripts covering the whole range of ability and quality, from masterpieces to illiterate rubbish.
The role of literary agents
For over a hundred years there have been individuals within the book trade who undertake to handle the business side of writers’ affairs for them. These literary agents, as they are called, will submit manuscripts to publishers, negotiate a contract, and check royalty statements; they may well give advice on market demands, provide detailed comment on content, undertake editing, and generally act as an intermediary between writer and publisher when things go wrong (as they all too frequently do). In return for these services, an agent will receive an agreed percentage of a writer’s income.
The unknown writer, let us say a single mother living on a council estate in Gateshead, will these days find it impossible to submit a novel to a major publisher; the publisher will simply send it back to her, unread, accompanied by the advice that she should try to find an agent to represent her. So the unknown young woman from Gateshead will end up in an agent’s slush pile rather than a publisher’s.
If, at some point, the writer is accepted as a client by an agent, the agent will then offer the book to a publisher, usually to an editor with whom the agent is on first-name, let’s-do-lunch terms. The book is then part of the editor’s slush pile.
If, in the course of further time, our young lady from Gateshead happens to generate a black swan which amazes the entire universe with its brilliance, she will still find that her next novel will still end up in the editor’s slush pile, in the sense that its publication will have to be subject to a conscious decision. The new book may rise immediately to the top of the editor’s reading pile, and the decision to go ahead with publication of the second novel may be uncontested, but a decision will have to be made, none the less.
And by the way, publication of book number two, or number twenty-two, even to follow a big success, may not be uncontested; it may be a matter of considerable debate. In 1986, after Dean Koontz had published fifty-four novels, he appeared on the US hardcover bestseller list with Strangers. He then wrote Lightning, which involved him in a bitter struggle with his editor, who prophesied the end of his career if it was published. Koontz insisted that publication should proceed, and in due course he was proved right, because Lightning became another hardcover bestseller. The editor concerned was Phyllis Grann, then at Putnam.
What the slush-pile process is designed to do
The purpose of the experiment with rats was to find the ‘strongest’ rats – strength being regarded, by the designers of the experiment, as the most desirable of all possible characteristics. To this end, increasingly high doses of radiation were administered until in the end there were only a few rats left.
But what is the slush-pile process – whether undertaken by agents or publishers – designed to do?
Everyone in the book trade is anxious to find the ‘best’ books. Different participants in the trade will have differing definitions of ‘best’. For some it will mean the books which generate the most income. For others it will mean the books which get the most favourable reviews from the highbrow critics. But if the submission and selection process has any purpose at all it is to select the ‘best’ books from the point of view of the organisation conducting that process. In particular, it is surely the hope of most parties that the process of selecting books from the slush pile will throw up an occasional black swan.
How the slush pile is dealt with
If, every day, the postman brings even as few as ten manuscripts into an agent’s office, the agent must assume (if she is willing to consider them at all) that among these unsolicited and unpromising submissions there may perhaps be the twenty-first-century equivalent of Sherlock Holmes, James Bond, or Harry Potter; or perhaps a Booker Prize winner. She therefore has to give at least some serious consideration to these manuscripts.
Perhaps our agent is super-conscientious, a mistress of the management of time, and can manage without sleep. In those circumstances she may even do the initial trawl through the manuscripts herself. But that is unlikely to happen. It is more than probable that the busy agent will employ a reader to do the job for her. The reader will discard the manuscripts which are judged to be hopeless and leave a relatively manageable number for a final decision by the boss.
Over the past fifty years or so much has been written about the role of the slush-pile reader; the experiences described are mainly those of individuals who worked in publishers’ offices in the days when big-time publishers were still willing to consider submissions from the public; but we shall be safe, I think, in assuming that the process is much the same wherever it occurs.
Since the job of sorting through the slush pile is generally reckoned to be soul-destroying, it is almost invariably given to the newest and most junior member of staff: the one who is in no position to refuse. Such people are seldom given any training. (Until recently no one got any training in publishing anyway, unless it was in the form of ‘Sit by Nelly and watch what she does.’)
The volume of work is such that the reader cannot possibly give more than a few minutes to any one manuscript, unless it proves to be unusually promising. Often, those who have done the job claim that to read one paragraph is sufficient to enable a rejection decision to be made.
Here is what one publishing professional, Andrew Taylor, had to say about the task, writing in The Bookseller in 1996: ‘In an average day’s work at a publisher’s office, I aim to assess 7 to 10 submissions and write reports on each of them which vary in length from 2 to 500 words.’
Mr Taylor is more generous with his time than some publishers’ readers. Giles Gordon once stated that when he was the slush-pile reader at Gollancz, he learnt how to tell whether a manuscript was any good within 15 seconds. ‘It’s just a matter of practice,’ he said airily.
Literary agent Pat Kavanagh takes much the same view. ‘Two pages will tell you if a book from the slush pile is worth pursuing.’
The results of the search through the slush pile
It is generally reckoned that, however carefully or otherwise the slush pile is read, it is rare to find anything in it which is worth even the most cursory consideration as a candidate for publication.
The agent Pat Kavanagh, mentioned above, was asked how often she had found a book in the slush pile that was worth pursuing. ‘Never,’ she said. ‘I don’t believe it has ever happened to me.’
Barry Turner, in The Writer’s Handbook, once mentioned an agent who fared a little better than that, but not much. In 14 years of reading 25-30 manuscripts a month, the agent found 5 good ones. Another agent, at Curtis Brown, personally received 1,200 manuscripts in one year, and took on 2 of the authors as clients. One agent at perhaps the largest UK agency remarked recently that she was having to read 3,000 manuscripts in order to find 1 client.
In 1989, The Times reported that the well-known British imprint Hutchinson was receiving about 1,000 manuscripts a year. One of these unsolicited manuscripts might be published every couple of years or so. Maybe.
At Chatto and Windus the Times reporter was told that about 10 manuscripts arrived every day. Were they all read? Long pause. ‘Yes.’ Were any ever taken on? Long pause. ‘No.’
The largest publisher of romantic novels in the UK is Mills & Boon, or Harlequin Mills & Boon, to give the firm its full name. The Mills & Boon editorial director has stated that the firm receives 6,000 manuscripts a year from hopeful and so-far-unpublished writers. Out of these submissions, the company takes on, in a good year, about 10 new writers.
In 1995, the owner of two small publishing firms in the USA reported in Publishers Weekly that he had received nearly 7,000 offers of books in the previous twelve months, and had decided to accept 12 of these submissions.
A much larger and more prestigious American firm, Viking, agreed to publish only one unsolicited manuscript in 26 years; that was Ordinary People, by Judith Guest. The book went on to become a bestseller as well as the basis for a successful film.
Finally, the publisher Anthony Blond, writing in The Spectator, maintained that the acceptance rate of unsolicited manuscripts was 1 in 2,000, in both London and New York.
And so on. Taleb rightly advises us against drawing general conclusions from insufficient data (the Baconian flaw), but it would be wearisome, and it is surely unnecessary, to go further.
We can safely conclude, I suggest, that very few manuscripts are picked out of the slush pile – anyone’s slush pile, whether agent or publisher – with a view to being taken further.
It follows therefore, as dogs follow a bitch in season, that a writer’s chances of achieving any kind of success are extraordinarily small. There is only the slimmest chance that a new and as yet unpublished writer will be taken on to an agent’s list of clients; even if taken on as a client, there is no guarantee of publication; and even if the writer is published, the chances of achieving any kind of critical or commercial success are also small.
Few manuscripts are selected from the slush pile; but we know for certain that some of those which are rejected are in fact worthy of publication – worthy by any standards, whether literary or commercial.
We have already had one example of a black swan which was unrecognised by everyone when it was still in manuscript: Harry Potter. The sole editor in London publishing who was interested in the first Harry Potter book, by a completely unknown author, was Barry Cunningham of Bloomsbury.
‘If it hadn’t been for Barry,’ said J.K. Rowling in 2000, ‘Harry Potter might still be languishing in his cupboard under the stairs.’ A long succession of editors had previously described the book as ‘too long’, ‘too complex’, and ‘too old-fashioned’.
This instance is almost enough, on its own, to prove that slush-pile readers’ judgements are fallible. However, we don’t need to limit ourselves to one example. Here are some others.
MASH, which became one of the most famous series in the history of television, was originally a novel. It was rejected by 21 publishers over a period of seven years before eventually finding a home. After publication, it was adapted as a successful cinema film before being developed for television.
In a more literary vein, the most famous case is that of the American novelist, John Kennedy Toole. In the early 1960s Toole was made emotionally unstable by the frequent rejection of his book A Confederacy of Dunces, and in 1969 he committed suicide.
Toole’s mother then took on the task of trying to find a publisher for the book on which her son had laboured so hard. She finally managed it, and in 1981 the novel was awarded the Pulitzer Prize. A Confederacy of Dunces was hailed by the New York Times as a ‘masterwork of comedy’ (though I’m afraid I never got so much as a smile out of it myself).
There are many other cases in publishing history of books becoming famous and successful only after a long struggle to achieve publication. James Joyce’s Dubliners was rejected by 22 publishers; and Baroness Orczy’s The Scarlet Pimpernel by 12. There are so many such ill-advised rejections, in fact, that a whole book has been written about them: André Bernard’s Rotten Rejections. This book includes, among other things, a letter from a publisher in reference to an agent’s submission of an early John Le Carré novel. ‘You’re welcome to Le Carré,’ said the publisher. ‘He hasn’t got any future.’
George Greenfield, an agent, told a story about Enid Blyton, who was published by Macmillan, and whose innumerable books for children sold so many copies that they paid the salaries of most of the staff. On one occasion Enid had heard that her usual editor had left the firm, but no one had told her the name of her new editor. So, when she had completed her next book, she simply addressed it to ‘Macmillan & Co’. The manuscript went into the slush pile, and in due course was rejected.
A similar instance of the rejection of a work by a famous author occurred early in the career of Giles Gordon.
Giles was then working for Hutchinson, which was run by Robert Lusty. At the time, Hutchinson’s most profitable author was Dennis Wheatley, who is now largely forgotten but was then a sort of English Stephen King. At his death, Wheatley had 50 books in print, and total sales were 41 million copies. As with Enid Blyton, the income from his novels was making a massive contribution to his publisher’s profits.
Giles Gordon was a young man with fastidious tastes, and by his own admission he felt nothing but contempt for Wheatley; and so, when Wheatley’s new novel arrived, Giles had it sent out for a slush-pile report as if it was from an unknown.
The report which came back was not favourable. ‘The book is terribly hackneyed,’ declared the reader. ‘Decline.’
Giles showed this report to his boss, Robert Lusty, who was not amused. Giles was told to publish the book in short order.
‘In spite of my best efforts,’ said Giles, ‘Dennis Wheatley’s career continued to prosper.’
This story demonstrates a number of points. One, that young men are often arrogant, ignorant and stupid; two, that mature publishers usually develop a degree of common sense; and three, that slush-pile readers… Well, what can one say about the Hutchinson reader in this case? He or she failed to recognise a new book by the firm’s principal asset. Does this generate any confidence in that reader in particular, or the slush-pile process in general?
Writers hit back
Every so often, repeated rejection of what the author believes is a good book leads to an attack of blind rage, out of which comes a determination to prove that publishers are complete fools. Sadly, this is not too difficult to achieve, at least in particular instances.
The usual procedure to obtain revenge is to type out a chapter or two of a current bestseller, give it a new title, and submit it to publishers as your own work. I know of at least three occasions when this experiment has been carried out, with results which will not, I think, surprise you. I will quote only one example here.
In the summer of 2000, the French publisher Plon issued a novel which had been written by a famous television presenter; the book was a great ‘success’, in that the author was interviewed widely, made many personal appearances, and the public was persuaded to buy a large number of copies.
The magazine Voici decided, however, that this novel was less than interesting, and that it would never have been published at all had it come from an unknown author. Voici typed out the first chapter of the book and offered it, under a pseudonym, to every leading publisher in France. None of them wanted to read the full manuscript, and none even recognised it as the season’s hit – including Plon, which had published the book in the first place.
We have surely assembled enough evidence about the slush pile. We now need to consider some of Taleb’s types of erroneous thinking, apply these to the world of publishing, and see what can be learnt which might be of value.
Flaws in the slush-pile methodology
We noted above that the experiment with rats was designed to identify the ‘strongest’ rats, and that it failed to do so because of flaws in the methodology.
There is sufficient evidence provided above (and elsewhere) to convince the author of this essay, if no one else, that the slush-pile procedure is also flawed. It is intended to identify the ‘strongest’ or the ‘best’ books (however defined), and it demonstrably fails to do so.
Taleb tells us that, in any experiment or procedure, there will be a difference between the desired outcome (in this case identification of the best books) and the actual outcome if there is either variance in the base cohort, or randomness in treatment.
In the slush-pile procedure both of these factors are present.
The base cohort contains a wide range of variance. The books submitted will range from the sub-literate to the masterly. Some writers can spell and punctuate; some can’t. Some writers will reveal a lifetime of experience; some will display a youthful naivety.
Secondly, there is massive randomness in the treatment of the books submitted. A variety of readers are likely to be employed; they have their own preferences, their own likes and dislikes, and these will differ one from another. The source of a submission (author, unknown agent, high-powered agent) will itself colour the willingness of the reader to ‘give the book a chance.’
The slush-pile procedure, we will allow, does bring about the publication of books which reach a basic, but fairly modest, professional standard. But no more. It does not facilitate, much less guarantee, the identification of black swans. The procedure, as normally operated, is deeply unsatisfactory.