We are fictional characters of our own creation

At the climax of Anna Karenina, the heroine throws herself under a train as it moves out of a station on the edge of Moscow. But did she want to die? Various interpretations of this crucial moment in Tolstoy’s great masterpiece are possible. Had the ennui of Russian aristocratic life and the fear of losing her lover Vronsky become so intolerable that death seemed the only escape? Or was her final act mere capriciousness, a theatrical gesture of despair, not seriously imagined even moments before the opportunity arose?

We ask such questions. But can they possibly have answers? If Tolstoy says that Anna has dark hair, then Anna has dark hair. But if Tolstoy doesn’t tell us why Anna jumped to her death, then Anna’s motives are surely a void. We can attempt to fill this void with our own interpretations, and we can debate their plausibility. But there is no hidden truth about what Anna really wanted, because, of course, Anna is a fictional character.

Suppose instead that Anna were a historical figure and Tolstoy’s masterpiece a journalistic reconstruction of real events. Now the question of Anna’s motivation becomes a matter of history, rather than a literary interpretation. Yet our method of inquiry remains the same: The very same text would now be viewed as providing (perhaps unreliable) clues about the mental state of a real person, not a fictional character. Lawyers, journalists, and historians, rather than critics and literary scholars, might put forward and debate competing interpretations.

Now imagine that we ask Anna herself. Suppose that Tolstoy’s novel was indeed an account of real events, but the great steam engine slammed on its brakes just in time. Anna, apparently mortally injured, is conveyed anonymously to a Moscow hospital. Against the odds, she pulls through and chooses to disappear to escape her past. We catch up with Anna convalescing in a Swiss sanatorium. As likely as not, Anna will be as unsure as anyone else about her true motivations. After all, she too has to engage in a process of interpretation: Considering her memories (rather than Tolstoy’s manuscript), she attempts to piece together an account of her behavior.

Even if Anna does venture a definitive account of her actions, we may be skeptical that her own interpretation is any more compelling than the interpretations of others. To be sure, she may have “data” unavailable to an outsider — she may, for example, remember the despairing words “Vronsky has left me forever” running through her mind as she approached the edge of the fateful platform. However, any such advantage may be more than outweighed by the distorting lens of self-perception. Our interpretations of our own actions seem, among other things, to assign to ourselves greater wisdom and nobility than might be evident to the dispassionate observer. Autobiography always deserves a measure of skepticism.

Are we all fictional characters?

But isn’t the same true of the stories we tell ourselves as our lives unfold? We have all heard the oft-quoted remarked that “journalism is the first rough draft of history” (attributed to Washington Post president and publisher Philip L. Graham, and many others). But we might also say that our moment-by-moment stream of consciousness is the first rough draft of autobiography. And if autobiography deserves a measure of skepticism, perhaps the first rough draft of autobiography deserves a double dose.

In my book, The Mind is Flat: The Remarkable Shallowness of the Improvising Brain, I argue that modern neuroscience, psychology, and AI push us even further: to the conclusion that the stories we tell ourselves about our motives, beliefs, and values are not merely unreliable in their specifics but are fictitious through and through. They are improvisations, created in retrospect by the astonishing story-spinner that is the human mind. When we imagine, query, or debate Anna’s motives, we know that there is no right answer about the true motives underlying Anna’s actions because Anna isn’t real. Yet the very same story-spinning machinery our brains use to create explanations for the actions of fictional characters are used when we interpret the actions of people around us, and indeed, ourselves. We are, in a very real sense, fictional characters of our own creation.

Consider three strands of evidence. First, the neuroscience. The linguistic explanations we create of our own behavior are generated by the language centers in our left cerebral cortex. In people whose brains have been surgically split in two, by severing the corpus collosum that links the left and right cortex, this means that language-generating machinery in the left cortex is necessarily completely oblivious to the machinations of the right cortex. As it happens, the right cortex sees the left half of the visual field and controls the left hand. So, when people with split brains are asked verbally to explain the actions of their left hand, you might expect them to be utterly mystified. But not at all! They are all too ready to confabulate a credible-sounding (though entirely baseless) explanation.

In one classic study by UC Santa Barbara’s Michael Gazzaniga, a person with a split brain is asked to match pictures on cards to an image shown on a computer screen . The trick is that the two halves of the brain are shown different images: The left (language) half of the brain is shown a chicken claw, and the right half of the brain sees a snowy scene. The person then chooses which of the picture cards best match the image. The right brain directs the left hand to choose a picture of a snow shovel — matching the snowy scene, of course. But the left, linguistic, brain knows nothing of this — it has seen only seen a chicken claw. Yet when asked to explain the actions of the right hand, the left brain is ready with a fluent, immediate, and apparently convincing answer: that the shovel was chosen because you need a shovel to clean out the chicken shed. This is a wonderfully creative answer: the left brain is doing its best to link the chicken claw with the shovel. It is also plainly wrong. But what is really striking is that it is generated at all, let alone fluently and with conviction. It makes one strongly suspect that what Gazzaniga calls our left brain “interpreter” is always a master of invention — it never has direct access to the actual causes of behavior.

Second, the psychology. Decades of experiments have found that we are story-spinners about our own motives, thoughts, and emotions. We imagine that we find people more attractive when we have just walked over a high and wobbly bridge (otherwise, why the adrenaline?). If you have had an injection of adrenaline, you rate annoying behavior to be more annoying (you interpret the adrenaline as a clue that you are really riled up). More recently, the amazing phenomenon of choice blindness shows that people can be tricked into thinking they preferred one face, flavor of jam, or even political view to another — and can fluently and compellingly justify a choice they never actually made at all.

Finally, the evidence from artificial intelligence. If we could reveal (not just invent stories about) the real causes of our behavior, then experts from every field should be able to tell us what they know and why. Imagine if we could just put that knowledge into a database and use it to re-create that expertise in a computer. If only it were so easy! In the 1970s, artificial intelligence researchers tried this strategy and it failed comprehensively. It turns out that experts have no idea how they diagnose diseases, forecast the weather, or play chess: Their explanations are both full of holes and hopelessly self-contradictory. In retrospect, perhaps, this should not have been a surprise — after all, two millennia of philosophy have surely demonstrated the baffling puzzles and contradictions that arise when we try to explain our everyday statements about good and evil, freedom and responsibility, or the nature of cause and effect.

The mind is a spectacularly inventive, if wildly inconsistent, storyteller, generating a continual stream of explanations, speculations, and interpretations, including of our own thoughts and actions. And these stories are so fluent and convincing that we often mistake them for reports from a shadowy inner world. But introspection is not some strange inner perception; it is the human imagination turned upon itself.

This article was partially excerpted from The Mind is Flat. It is republished with permission of the author.