R.F. Delderfield, To Serve Them All My Days (1972)

This post includes almighty spoilers about this superb novel, which I read recently. Normally I wouldn’t be bothered about that sort of thing myself: I consider myself too sophisticated to rely on such feeble devices as surprise. However, in this case, I was knocked flat by something, memorably, and I don’t want to ruin it. After all, the chance to let out a strangled sob-slash-whimper on a train is quite a rare thing for most people.
      This is a follow up to my discussion of the Predictive Processing conference in Cambridge, which you can read here. As you see in that post, I was very taken by a discussion of how this way of thinking about cognition (in which we are always making predictions, and then modifying them in the light of sensory inputs, and this is how our brains get so much done) might relate to psychotic behaviour. I will quote myself maladroitly paraphrasing Katharina Schmack, I will: ‘Is this a matter of faulty predictions, too strong, or too weak? Is it about a failure to deal with noise in the system, evidence from the world that shouldn’t affect predictions? Is it about problems in the re-evaluation, the feedback from input to prediction? Is it about how you deal with surprises, over-rating them or misjudging them?’.

*

So here’s the point I wanted to work with. We have a model there of why those suffering from psychosis might interact dangerously or unhappily with the world. Something is awry in the prediction-input-feedback loop, and what exactly that is might be unclear, or might vary from person to person, situation to situation. Now, it’s pretty obvious that a prediction-input-feedback loop is part of literary experience. We expect certain features of story or form to occur; those expectations are modified, sometimes confounded. We wouldn’t generally say that we are aiming to minimise surprise, which is a core idea in the Predictive Processing model; indeed, most of us would say that surprise can be pleasurable, but a lot of our experience is at least in dialogue with surprise.
      I wonder if there are cases in literature where the expectation-surprise dynamic is distorted such that we find ourselves apprehending the fictional world with a strange energy that shares something with the psychosis described above, giving us a disorienting excursion into a different way of predicting and responding to predictions. I was particularly taken with the idea that what could go wrong is the attention to noise: perhaps some people are unable to tune out those stimuli that should not modify their predictions. Some brief examples…

* How about the literature of jealousy? Do some works draw us into the mindset of a character’s psychosis by distorting our experience of how the world validates, or does not, predictions about it? In Shakespeare’s The Winter’s Tale, for example, I think we’re well outside the mind of Leontes when he says ‘there’s no truth in the oracle’, conspicuously failing to modify his predictions in response to evidence. But in Othello, I’d say, it’s not so clear, and perhaps that’s because we have been drawn into his particular distortion of expectation and corroboration — not that we suspect Desdemona of infidelity, but that we are no longer comfortable with understanding what’s just noise, and what isn’t?
* Briefly — Macbeth too? Having supped full with horrors, are we at times stuck in his gruelling mindset, sharing something of his logic as he maps his predictions onto the world?
* How about verse forms too? Surely there are some cases where some nuancing or violation of the rhythm or rhyme unsettles our ability to exist comfortably within a pattern of expectation and modification? I wonder about my current obsession John Skelton’s famous metrical innovation (known as Skeltonics): might they induce a kind of prosodic psychosis, leaving us responding in a spirallingly strange way?

*

But here is the point about To Serve Them All My Days. This is a novel about a schoolteacher, David Powell-Jones. Approximately 27% of the way into the novel (can you tell I read it on Kindle?) readers find themselves involved in the episode of the Winterbourne Divorce. This centres on a boy whose parents are divorcing, a big deal in the 1920s. His mother’s scandalous behaviour is in the papers, everyone knows, and he runs away. A search is ongoing, and Powell-Jones has just found, and admired, young Winterbourne’s watercolour paintings. Then a new section starts like this:

Venn’s lorry-driver, emerging from the quarry two-thirds of the way up Quarry Hill, was aware that something was amiss the moment he levelled out on the gradient, about one in six here but steeper beyond the ash coppices that grew on each side lower down the road. He braked as hard as he dared but the speed of the heavily-laden lorry increased so that he made a wild grab at the handbrake, throwing all his weight on it as the vehicle weaved the full width of the road, its speed increasing with every yard it covered.

As I read this, I was full of forboding, because it seemed like this agent of horrendous destruction was out of proportion to the Winterbourne plot; and I honestly think I had already slightly wondered whether the divorce episode was going to be enough to sustain the chapter. (There’s something in the punctuation. Delderfield is very good, I think, at creating paragraphs that are full of momentum; he does it here partly by avoiding commas.) Sure enough, this lorry crashes into Powell-Jones’s wife Beth’s car, killing her and one of his twin daughters, seriously injuring the other. The rest of the chapter describes him finding out, fleeing onto the moors in despair, and being found, ironically, by Winterbourne. The chapter completes one trajectory with the poor boy’s desperate and poignant query as to whether Beth was out searching for him when the accident happened.
      I think the structure of the chapter is superb, but (or rather, and) it also left me disoriented. In some ways it was a matter of simple surprise, a moment in fiction where prediction error slaps you in the face, but without real-life consequences. But perhaps, in the light of the Predictive Processing field, I think that for a while the novel inculcated a new configuration of prediction, sensory input, noise, feedback, and so on. The usual rules that fiction depends on, which help us judge the significance of the things we encounter, were suspended for a while.
      It would be consistent with my probings above to see this as a replication of something like psychosis, a syndrome in which the meaning of things, and dealing with those meanings, were all out of joint. That’s probably pushing it a bit too far. But I still think that literature can change our angles, training and then disrupting our predictions and our prediction errors, giving them a wider framework and a chance to do better next time, in books and in life. This might be a rather large thing for literature to be doing, over and over again.