Agatha Christie deconstructed:  Pierre Bayard's Who Killed Roger Ackroyd?

Back to main English 359 page

NB ~ Don't read the following until you've finished the novel in question.  Please note the suggestions at the end of this discussion for further analysis.

Pierre Bayard, in Who Killed Roger Ackroyd? (1998), takes as a point of departure the rules formulated (see link mentioned week 1- by the American detective writer S. S. Van Dine (the pen name of Willard Huntington Wright, whose novels of the 1920s introduced the scholarly and snobbish amateur detective Philo Vance).  The most important of Van Dine's rules, Bayard suggests, are the two rules of the 'disguise principle':

1.     'the truth must be hidden throughout the book.   The detective novel has meaning only if the truth is not revealed until the end of the book'
2.     'While being hidden, this truth must be accessible to the reader, even in plain view.'  This second rule, the 15th in Van Dine's list, means that if the reader should reread a detective novel he would 'see that the solution had, in a sense, been staring him in the face'

Christie, Bayard says, 'seems to have perfected' the 'disguise principle', and as a result her fiction allows us to explore in great detail the 'wealth of the modalities of blindness and of the production of meaning':

'Few writers have mined the eminently Freudian question of psychic blindness as systematically as Agatha Christie.  Why don't we see?  Agatha Christie poses this question in all her plots, but also relates the question to the reader, for beyond the detective story, every book tells the same tale about the blindness of those who read it.
'The type of blindness we are dealing with here is of interest to psychoanalysis, which naturally attends to things that escape consciousness.  In this respect, Agatha Christie's work displays the unconscious, not by calling for the decoding of hidden meanings, but by inventing different devices each time that aim to prevent the reader from perceiving the truth.  And while her work experiments with many forms of blindness, it stages the proof of that blindness by its opposite, the unblindingthe interpretation and action by which meaning is produced.'

Christie's novels manifest a 'rigorous and systematic application' of the two key rules, using three main techniques:

1.     Disguise: the murder or murderer disguised as something or someone else.  This is 'the most classic way' of misleading the reader, and can take the form of, say, disguising a murder as a natural death or suicide, or of disguising the nature or function of the murderer (his or her public image or profession - or 'more subtly understood' the literary function or narrative role of the murderer - may divert our suspicions).
2.     Distraction: the reader's attention is drawn to other, more suspicious clues or characters.  This is 'a kind of negative disguise' (not the truth made unrecognisable but the false 'dressed up to draw attention to itself'); it can involve characters, clues and other plot devices, and is, Bayard suggests, one of the things that separates detective fiction from most other literature, since its purpose is to 'prevent thought'  - 'to prevent an idea taking shape', rather than to stimulate ideas.
3.     Exhibition:  making the truth invisible by making it apparent everywhere - 'by inscribing it in every letter' ('The murderer is, shall we say, hidden behind the murderer.').

These modes of concealing the truth can often be employed simultaneously, or single modes can be 'doubled' (disguise, especially, Bayard suggests, 'profits from being reinforced' - as when, for example, a murderer is given a respected status and also made to seem a victim).  The application of these modes of concealment produces what Bayard sees as a central paradox in the genre:

      -  detective fiction works by complicating events, which requires the proliferation of signs (if not, there would be no plot, no point in reading it)
      -  but at the end it asks us to accept that there is in fact one truth, that all signs are determinate - yet the lengthy process of complicating the truth leaves a residue of doubt in the reader's  mind:  'The possibility of combining these three major modes of concealmentconsiderably augments the virtual number of solutions to a mystery by setting up a plural logic' - and what most interests Bayard here is the consequent difficulty of 'arresting meaning':  'In the end, Christie's work generates a model of polysemy (a multiplicity of meanings) in which each indefinitely reversible element becomes subject to caution.'  The text proposes a 'series of meanings' that 'overarch the whole text like a set of reserve hypotheses that are valid until the final resolution imposes one of them.The question is, then, to what extent this virtual multiplicity of meanings finally generates undecidability, therefore annulling the model of readability that Van Dine proposes.'

Turning specifically to The Murder of Roger Ackroyd, Bayard suggests:

The novel mainly uses the technique of disguise: Sheppard masquerades as someone unlikely to be a murderer.
-  But this is taken to higher level (the mode of concealment is doubled): the murderer is concealed in the narration itself (since the first-person narrator is conventionally the mouthpiece for the truth).
This effect is augmented by 'two other closely connected methods that are essential to the production of illusion':
1.              double-edged discourse: statements that can mean two completely contrary things. By means of such statements, - such 'ambiguous formulations' - Sheppard 'manages to tell us nothing but the truth'.  It requires a second reading for us to read them 'correctly' (and in fact, unlike the traditional detective novel, The Murder of Roger Ackroyd would seem to be 'explicitly composed for the sake of a rereading').
2.              the 'lie by omission': Sheppard may tell 'nothing but the truth', but he does not tell the whole truth.  Look out for ellipses - in the sense of gaps between one statement and another, into which we would retrospectively have to insert things that the narrator has suppressed (e.g., picking up the dagger when he is examining the curios).

As a result of these complexities, Bayard maintains, Roger Ackroyd might be taken, more widely, to support an argument that:

-  the narrators in detective fiction are always deceitful - artfully presenting the truth through suspense and artifice;
the logic of 'lie by omission' applies to all narrative: all narrative is selective; there are things that happen in the fictional world that we are not told about;
-  a text depends ultimately on the reader to fill in the gaps and this opens it up to deluded readings.

If we think about things in this way, Bayard suggests, we will see that there are 'massive improbabilities' in Christie's novel - improbabilities that 'shatter the idea that with all the facts given on the final page, the text will come to a neat conclusion without the reader's participation.  We see how this opens a door that will be difficult to close.  In effect, if ellipses can be added anywhere - including within sentences - then what we were reading until now as a finite and coherent whole is merely an incomplete portion drawn from a much vaster text, and we run the risk of setting no limit on reading and interpretationinfinitely multiplying the number of virtual solutions.'

(Pierre Bayard, Who Killed Roger Ackroyd?  The Murderer who Eluded Hercule Poirot and Deceived Agatha Christie (1998; London: Fourth Estate, 2000), pp. 19-40)

Suggested further analysis:  In the light of this summary of Bayard's critique, reread the last two chapters of The Murder of Roger Ackroyd, together with the key parts of chapters 4-5, describing the murder and its 'discovery' (say, pages 35-50).  In chapters 4-5, find examples of Christie's 'modes of disguising the truth' as summarised above, and then consider (i) Bayard's question of why the reader decides to believe the second account rather than the first; and (ii) his suggestion that if we once start to 'add ellipses' and to question the narrator's reliability, we have not closure but 'a multiplicity of meanings.' (in short, who else might have killed Roger Ackroyd?)