Christopher Prendergast’s critique of Graphs, Maps, Trees in ‘Evolution and Literary History’ raises objections of an empirical, theoretical and political nature.footnote1 The main disagreement is this: for Prendergast, nature and culture function in such incomparable ways that evolutionary theory, which was devised to account for the one, cannot possibly work for the other. This conceptual misalignment makes evolutionary ‘explanations’ of literature incapable of mastering any actual historical evidence, and forces them to rely on circular reasoning and various petitiones principii instead. In this analytical void, the market acquires an exaggerated importance, that makes it appear as ‘a cognate of Nature’; and the final result is that Graphs, Maps, Trees’s ‘no-nonsense realism . . . deteriorates fast into the language of the winner-takes-all attitude’ that is typical of social Darwinism.footnote2

Predictably, I dissent from this diagnosis, and will explain why in the pages that follow. But in the course of writing this reply I have also become increasingly (and uneasily) aware of how few concrete results have emerged so far from the models discussed in Graphs, Maps, Trees, and from others of a similar cast. As they are all rather recent, this fact does not invalidate them: they have gaps, yes, but for me they are still better than the existing alternatives. Still, a good method should prove itself by producing interesting findings, and the title of this article expresses my impatience with the purely methodological reply which, unfortunately, is what I am able to offer at present. ‘La metodologia è la scienza dei nullatenenti’, wrote Lucio Colletti in Marxism and Hegel: methodology is the science of those who have nothing (but nullatenente is harsher, loaded with sarcasm). A bitter truth, about which I will say more at the end.

Prendergast’s opening comments concern the ‘trees’ of detective fiction discussed in the last chapter of Graphs, Maps, Trees. For him, those images embody ‘an implicit syllogism: Doyle was to prove the most popular of the thriller writers; Doyle comes to use clues in a uniquely special way; therefore the way he uses clues explains his enduring popularity’.footnote3 Had I foregrounded clues because Doyle ‘used them in a uniquely special way’, he would be absolutely right: when the choice of the evidence pre-determines the results of the investigation, the reasoning is circular indeed. But that was not why I focused on clues in Graphs, Maps, Trees (nor in the earlier article on which it was based): rather, I did so because all theories of detective fiction place clues at the very centre of the genre’s structure, thus singling them out as its crucial morphological variable.footnote4 Between Conan Doyle and clues, in other words, there was no inevitable a priori agreement—if anything, it is striking how erratically and inconsistently they are used, throughout the Sherlock Holmes cycle.footnote5 But although Doyle’s solution was far from perfect, it was still better than those of his rivals, and this fact seemed to offer a good explanation for their different destinies. Readers liked clues, and so they chose Conan Doyle’s stories over those of his rivals.

Readers liked clues . . . But Prendergast doesn’t like that ‘like’. Having rightly noticed that ‘what readers like serves as the equivalent—or analogue—of “environment” in evolutionary thinking’, he then dismisses ‘this characterization of reading practices and preferences [as] a curiously lightweight scaffolding to build a theoretical model’.footnote6 But why lightweight? What more powerful agent of selection can there be than the choices of contemporary readers? Sure, there are publishing, and distribution, and their various appendices (reviewing, advertising, etc.); but even in the film industry, where their role is clearly much greater than in the book market of a century ago, genuine hits don’t acquire their typical momentum when these external pressures are at their strongest (that is to say, right away), but only weeks later, when they have largely been replaced by a chain of informal exchanges—I really liked that film—among common movie-goers.footnote7

Which brings me to that disgraceful ‘like’. It was, obviously, shorthand for something more complicated, on which I have so often insisted, from Signs Taken for Wonders on, that I have become reluctant to repeat it (especially as it’s not an original idea: Freud, Lévi-Strauss, Althusser, Orlando, Jameson, Eagleton and several others have all offered their version of it). In a nutshell, the idea is that literary genres are problem-solving devices, which address a contradiction of their environment, offering an imaginary resolution by means of their formal organization. The pleasure provided by that formal organization is therefore more than just pleasure—it is the vehicle through which a larger symbolic statement is shaped and assimilated. When readers of detective fiction ‘like’ clues, in other words, it is because the structure provided by clues makes them feel that the world is fully understandable, and rationalization can be reconciled with adventure, and individuality is a great but dangerous thing . . .footnote8

Until proved wrong, then, I will stand by the idea that literary history is shaped by the fact that readers select a literary work, keeping it alive across generations, because they like some of its prominent traits. But where is the ‘demonstrable causal relation’footnote9—where is the evidence that readers liked clues?

That readers selected Doyle because of his use of clues, writes Prendergast, ‘cannot simply be affirmed. It may well be that Doyle’s success can be accounted for in this way, but, subject to further investigation, it may well also be that it was due to quite different factors (for example, a fascination with the figure of Sherlock Holmes, the gentleman from Baker Street)’.footnote10