Tuesday, 24 February 2009

Free will with every packet of cornflakes

Yesterday I spent all day writing. What I produced was rubbish, all 2000 words of it. The words were flat and lifeless with no energy at all. In contrast, last week I was working on short story and the words zinged across the screen. I didn’t know where those words came from, I hadn’t planned them in advance. They just appeared.

Every fiction writer knows the excitement of setting up a situation and then being surprised at what happens next. Your characters can behave in ways that surprise you. Are they exhibiting free will? Even when you consciously plot out your characters’ stories, you have to be sensitive to what feels right and what doesn’t. If you force your characters to do something for the sake of the overall story, they may turn into puppets. They lose their ‘divine spark’ and the story can feel over-engineered. (Interestingly, this may only be true in literature. In Kleist’s famous essay on marionettes, he pointed out that they can be more graceful and more likely to achieve ‘perfection’ in their movements than humans, precisely because they lack human’s self-consciousness which, according to him, inhibits a complete understanding of the universe.)

Nowadays, free will is elusive. It suffered a fatal blow in the seventeenth century from Newton’s discovery of universal physical laws. The corresponding vision of a wholly predictable and mechanistic universe which only needs to be set going before it simply carries on for ever, doesn’t seem to need free will.

A universe that allows time travel (ours doesn’t appear to rule it out) must also impinge upon free will in some way, if only to prevent logical paradoxes (think of ‘Back to the Future’).

Is there any space for free will in quantum physics? After all, it (re)introduces uncertainty into physical systems, albeit on atomic scales. But this uncertainty (in the ability to know simultaneously an object’s position and momentum) is random. And when we exhibit free will, we don’t think we behave randomly.

Geneticists don’t have much time for free will, either. We’re creatures ruled by genes and all our behaviour can be explained by those genes’ desire to replicate themselves. Even apparent altruism to other people only seems to exist because we share practically all of our genetic material (99.9%).

So perhaps we enjoy reading about characters in novels who appear to exercise choice because that choice is a mirage in our real lives. Do our fictional characters have more free will than we do?

Thursday, 12 February 2009

In thee (author) I trust?

I’m currently reading a historical novel, Quicksilver by Neal Stephenson, which tells the story of the machinations of the Royal Society in the late seventeenth century and the row between Newton and Leibniz over who first invented the calculus. The book is a mixture of real and imaginary characters and it’s narrated by the latter. I think it’s these imaginary characters that move this book firmly into the realm of fiction, without them the reader would be more likely to wonder if what we were being told about Newton, Hooke, Leibniz, Charles II et al. were ‘real’. The addition of the fictional makes us feel secure that to a large extent all the interactions, even those between the real characters, are made up. (To that extent they’re not real characters at all).

The relationship between the reader and the author needs trust on several levels, overtly in the case of scientific narratives, less so for fiction. The most general trust requirement that the reader has, is that what they are about to read is interesting and worth spending their time (and perhaps money) on. In fiction, the reader needs to trust the author to invent a coherent universe, one that operates according to some specific laws (even if the author is the only one that knows those laws). If something happens in the story that seems to violate its fabric, then the reader stops trusting or caring. (See John Gardner’s excellent book ‘The Art of Fiction for a more detailed discussion of this point).

Of course, you don’t need to trust the narrator who may be unreliable and whose vision of events may be completely skewed. For example, Pip in ‘Great Expectations’ assumes that his mysterious benefactor is Miss Havisham and initially there is no reason to doubt this. The interest in the novel partly lies in exploring how this error affects Pip’s development. Barbara in ‘Notes on a Scandal’ tells us she only has Sheba’s best interests at heart. By the end of the novel, this is patently untrue.

Part of the fun in reading novels with unreliable narrators is working out the reasons for this unreliability. And even here, I think there needs to be some initial trust in the narrator. Someone, who from the very first page is clearly lying, is going to have to be incredibly engaging to keep the reader involved, because it will be so much more hard work just to figure out how this universe works.

Novels with unreliable narrators tend to be written in the first person, so are all first person narrators inherently unreliable due to their necessarily restricted and partial outlook?

Trust in scientific papers operates at various levels, some of them more explicit than others. Most obviously, the reader has to trust the authors’ integrity and believe that what is reported in the work actually happened. If this is violated it can have a huge far-reaching impact on the science. For example, one infamous case of scientific fraud is that of Paul Kammerer and the midwife toads in the twenties. Kammerer said he’d shown that the Lamarckian version of evolution was correct, in that an organism’s environment can directly affect how it develops and the attributes that it passes onto offspring (in violation of Darwinian natural selection which holds that genes mutate randomly and those organisms which are better adapted to their environment – by chance – are more likely to survive and mate and produce offspring, thereby ensuring that their genes are passed on.)

Male toads, from toad species which mate in water, have nuptial pads on their feet to enable them to hold onto their females. Toad species which mate on land don’t need, and therefore don’t have these pads. Kammerer claimed that by forcing the landlubbers to mate in water, he’d got them to develop nuptial pads in only a couple of generations. This caused a furore until it was shown that the nuptial pads on one of his specimens had been faked, and were caused by injections of ink. Kammerer committed suicide shortly afterwards.

This discredited Lamarckian ideas. And yet, how can a fake result be used to disprove a theory? Just because it doesn’t disprove Darwinian ideas, it doesn’t mean that the alternative is wrong. But this is what happened – bad science has been used to discredit Lamarckianism. (Alongside good science which favours Darwin, of course).

Fraud is (presumably) rarer than inadvertent mistakes. Science is difficult to do, mistakes happen. So how does the reader know whether to trust that the author hasn’t messed up?
Popper argued that science should be falsifiable i.e. you should be able to refute a theory if you get an experimental result that disagrees with it. In practice because of the possibility of mistakes there is more caution than this implies, and scientists are unlikely to chuck away an entire framework on the basis of a single result, simply because that result could be wrong.

More insidiously, scientists can be skewed towards a specific reading of their results based on their prejudices. So, trust is required to assume that the author has been open-minded and considered all options.

Of course another factor that influences the reading of a scientific paper is the authors’ reputations. If you know the authors, and accept their previous work, you are more likely to believe their current work. Is this true in fiction? Are you more likely to read a book based on the author’s reputation? Can the book ever stand alone?

You also have to trust what you are not being told. In fiction, much of the art lies not in writing but in editing, the cutting out of extraneous words to leave only the essentials. You have to trust that what has been left out is inessential. In science this is more problematic. Many experiments go unreported, because they don’t give interesting results. This is a particular problem in medicine with the testing of new drugs. If the results are inconclusive or unfavourable, they are more likely to go unreported than if they appear to support the hypothesis that the drug ‘works’.

Tuesday, 3 February 2009

From A to B

Writing fiction presents the writer with a near-infinite list of choices, not only on the subject matter, but also on the style and the process. For example, most stories are written in the past tense, but not all. Sometimes, I find that when I write in the present tense, my writing becomes more immediate and more fluid. I can ‘see’ my characters better as I can understand what they are doing right now. I don’t have to interpret their past actions for my readers.

However, this sort of writing can go badly wrong, and you can tell when it does because the characters in the story get stuck in a sort of realisation of Zeno’s paradox, in which an infinite number of steps are needed to get anywhere. The characters stop being able to do anything, because every single action occupies all of the present. When you’re concentrating on the present and not the future or the past, the present can seem like eternity. The writer chops the present down into finer and finer slices until time stops completely. This is probably a good thing if you’re meditating. It’s not a good thing if you want to write interesting fiction.

Of course, the antidote is to know which of the present moments are the important ones and leap from one to the next, leaving out all the others. A story never really flows continuously and smoothly from one moment to the next, and consequently the space-time in the story becomes rather lumpy and quantised. If the writer is good, the reader doesn’t notice these invisible joins.
The use of the past tense in literature is a method of telling the reader that the narrator has seen these events happen in the past and is relaying them to the reader. Using the present tense does away with this artifice, and replaces it with another. The reader is now ‘watching’ the events unfold before his or her eyes; this is the literary equivalent of watching a film. Was this method of writing influenced by the rise of film in the early twentieth century?

In scientific narratives (i.e. papers published in scientific journals), there is also a subtle editing of time and space. In these sorts of writings, you never read about the nights spent at the telescope waiting for the cloud to go away and the cabin fever brought on by spending two weeks alone at 10,000 feet observing the same star night after night. You never usually read about the things that go wrong, the wrong star observed, the wrong chemicals mixed, the endless debugging of the computer program. These things are not supposed to be relevant to the experiment.

The actual work reported in scientific papers is summarised in the past tense, although the results are reported in the present. So the work occupies a specific moment in time, but its outcome should stand for all eternity, even after the invisible narrator is long gone. And the passive voice is always used, to indicate that the presence of the author did not impact on the results. In fact, the scientist is the ghost in the machine. Events unfold with a certain inevitability, the stars were observed, the gases were mixed, the temperature was taken, the theory was developed, but who did all this?