Finding the Truth in a Story

Heroes often begin as simple people who pass a test like pulling a sword from a stone and then become the giants we know. Scientists have to be careful that they don't tell that kind of origins story themselves.

Stephen Jay Gould used the sneering phrase, "just-so stories," that reduced evolutionary narratives to the level of amusing children's tales penned by Rudyard Kipling. I'll ignore the fact that Kipling's tales are true gifts to literature and keep my eyes focused on the big point: narrative is suspect in a scientific context. They enable "facile verbal arguments" to sound plausible without including evidence or data.

The point was brought home to me again while working on last week's post about John Shea's work. He mentioned a 1984 paper I had never heard of by Misia Landau titled, "Human Evolution as Narrative" (later turned into a book, "Narratives of Human Evolution"). Landau concludes, however, by doubling down on narrative. She asks:

…whether there is any way to present an evolutionary or historical account that does not involve storytelling. Trying to subvert narrative procedure, as do some contemporary fiction writers, is a possibility, though one requiring great literary ability. Given the difficulty of such an approach and the irreducibly diachronic aspect of history, a more realistic solution may be to treat narratives even more seriously than before. Rather than avoid them, scientists might use them as they are used in literature, as a means of discovery and experimentation.

Comments like these hit a nerve with me as (a) the author of a forthcoming narrative about speech origins and (b) a former literature major who, even if I was the only English scholar in school who also subscribed to Scientific American, has always taken stories very seriously.

It is probably impossible to write a scientific narrative of evolution without getting a little post-moderish, in the sense of raising questions within the narrative about the reliability of the story being told. Did it happen one way or the other? Storytellers are almost always forced to choose. It happened this way, they say and then go on to the next event. Classic narrative style, using an omniscient narrator, makes a choice silently and proceeds. Anybody telling an evolutionary story who does it in the classic manner is almost certainly going to start Gould rattling about in his grave. The storyteller has to say what the evidence is and admit that it could have happened some other way.

At the same time, too much digression in the manner of Tristram Shandy or uncertainty in the style of John Barth, is a disservice too because science amounts to something only when it sets forth a hypothesis. Others can then respond by wondering if the hypothesis is true. If there is no hypothesis, only a series of ponderings, the reader never gets anything to test and try to falsify. Landau justifies scientific narratives in exactly this way:

…by specifying a temporal sequence of events, narratives can be used to predict what future investigations will disclose about those events, and are thus open to falsification. The greater the number of events in the temporal sequence, the more testable the narrative.

Good narrative is usually very specific, very visual, and these things can be checked.

I'm reading Richard Panek's The 4% Universe just now and it reminds me that story points the way to theory. The Big Bang theory began as a speculative story that eventually became testable. If the story was true, there should be background radiation from the event. The radiation was found and Fred Hoyle's rival story of a steady state universe faded from view. Stories are the entryway to theory.

Particularly crucial to a story of human origins is the order in which things happen. Much early narrative has been falsified various orders already. The famous Piltdown forgery put a modern brain on top of an ape's jaw, confirming the expectations of the narrative of the time (1912) that the first step to be ing human was to get smarter. The Australopithecus discovery in 1924, put a small brain on a skull with many humanish features wasn't taken very seriously until after the Piltdown fossil was discredited in 1953, but even without that complicating hoax, the expectation that the big brain came first would have been enough to slow the embrace of Australophithecus as part of the human origins story. This kind of change in the narrative chronology has forced a reconsideration of what it is to be human.

Landau lists a number of events that must be sorted out by the narrator:

• Appearance of morals.
• Development of bipedalism.
• Growth of the brain.
• Introduction of technology.
• Onset of language.
• Shift from the trees to the ground.

I've listed these statements in alphabetical order, to give them a random structure. The task of every narrator is to organize them into a coherent chain of events. If science is to contribute to the story, it must provide the facts and evidence supporting the organization of the events.

A second task is also apparent from this list. These events break into subevents. "Onset of language," for example, is too brief and requires a number of other events. If I were telling a Chomskyan version of the onset of language for example, I would require at least a mutation producing recursion and, as a second sub-event, the externalization of thinking. As Landau says, the more events the narrative introduces, the more testable it becomes.

But there is a danger to narration. It is very old and has established many conventions that scientific storytellers must be careful to avoid unless there is strong evidence to embrace the convention, and there almost never is. Stories and myths of origins have a long history and a structure that Landau breaks down into "functions."

1. The initial situation: conditions of safety prevail. In accounts of human origins this setting is typically a forested one. Touché. But let's not forget that the most famous human origins account of all puts the initial situation in a paradise, a rich oasis of fruits and pleasures.
2. Introduction of the hero: evolutionary tales of human origins have no literal hero for they follow a lineage, sometimes of one species, sometimes of many. Heroes are often distinctive, outsiders who have some special, as yet, unappreciated gift. There is, however, no evidence or reason to believe there was anything special about the line that became us.
3. Change of situation: something happens that changes the initial situation, forcing the hero to respond. The change may be in the environment or in the hero, but whatever the change it accounts for the changes that the hero will face. Evolutionary stories usually begin with an environmental change, so this function will appear in almost any scientific narrative of an origin. 
4. The hero's departure: this reaction to function 3, rather than the change itself is the story's first turning point. This setting off on the adventure has no counterpart in a proper evolutionary narrative, for it marks a decision and evolution doesn't make decisions. A hero's story is a tale of heroic will; evolutionary stories are not. But there are a number of stories of human origins that try to base everything on a single turning point. Bipedalism gets much attention these days, I think largely because the fossil evidence puts it first. The big brain used to get much credit, until it turned out that the brain comes so late in the story. Even so, searches for events like the "Human Revolution" should inspire caution just because they fit so well into this function 4. Storytelling rather than data is building expectations.
5. Series of tests: the hero makes himself by following his destiny. The tests bring out the nature of the man. The human evolution story, we saw, has a number of events which eventually turn apes into humans. In a hero story, the tests specifically look for heroic virtues—courage, honor, fidelity. The events in an evolutionary story are not so focused on virtues, but the result of an event is a new trait. Again, the Biblical tale of Adam and Eve surprises by taking a different task. The hero has one test (don't eat the apple) and fails it!
6. Appearance of "the donor": the heroic tale has a wizard, wise guru—somebody—who aupports the hero. There is no counterpart in an evolutionary account, but peoplel often do appeal to the appearance of a mutation.  In the Adam and Eve story we get an anti-donor, Satan in the guise of a snake.
7. The donor presents a gift: in heroic stories the gift might be a ring or a magic cloak. In evolutionary stories. In evolutionary accounts the gift is more typically a trait—intelligence or maybe bipedalism. In the Human Revolution account of human origins archaic Homo sapiens gets ahold of symbols and is transformed into the modern species. The biblical account remains contrary; the gift is death. Adam has cursed the human race and is no hero. Evolutionary accounts can never have this srprise because evolution assumes death iis part of the story from the beginning. But we have to be careful. In congratulating ourselves for avoiding the biblical tale we can easily trip into a more ordinary heroic tale of King Arthur, or Hecules, or some such.
8. Another test: having become a hero, the hero is tested again to prove his humanity. In human origins accounts this test develops civilization. Thank heaven I stopped my account of speech origins 164 thousand years ago and don't have to worry about this part of the tale.

Science needs data, but it must be careful about interpreting that data. Be on guard for reading into the data what you want to find. And science tells stories, but it must be careful that the story doesn't tell the science. Evolutionary scientists should be as skilled critics of stories as they are skeptics of data. They need to be sure they are following the evidence and not a tradition.

But for all the risks in getting bum steers, when it comes to contemplating a series of events and causes, humans think in stories first. They are not going to go away, so evolutionary scientists should learn how to use them an criticize them.

Comments

[Raymond Weitzman]

You might be interested in watching the 10th lecture of the Stanford University series Darwin's Legacy on YouTube. The lecturer, George Lewis Levine gives a stunning literary analysis of Darwin's work. In fact, the whole lecture series is well-worth watching.
-----------------------------------------------------
BLOGGER: Here's the link: http://www.youtube.com/watch?v=qf4AEMVB46g

[Karthik Durvasula]

very nicely put!

[Dfried]

Blair,

A brilliant post, and I entirely agree with you about the necessity of narrative. But I do think someone should say a word for Stephen Jay Gould and the "just-so story." As I understand it, Gould's meaning was something much more specific than the unreliability of evolutionary narrative. Gould was referring to the popular misunderstanding of evolution, in which every discrete feature has to be separately accounted for by a specific adaptive advantage.

An example would be the distribution of our remaining hair, where the conservation of head hair supposedly has to do with keeping the brain cool, whereas axillary hair is there to trap pheromones and pubic hair is a visual exclamation point, a dry lubricant, etc. Different but equally plausible explanations were deployed a number of years ago in support of the theory that homo sapiens evolved by the seaside. We're essentially hairless because it helps us swim, but we have hair on our heads to protect our scalp while we wade. Of course, we could equally have evolved coats like otters!

Gould was not sneering at Kipling, but comparing the evolutionary naivete in such stories to "How the Elephant Got his Trunk" and similar origin fables. Gould's point was that features like axillary hair may be accidental byproducts of changes in mechanisms of growth and/or control driven by some other adaptive advantage altogether. Gould famously called such byproducts "spandrels."

Gould's acerbity on this point was of course partly driven by his liberal politics. Vulgar biological determinism constantly deploys such stories in the service of fatuous Panglossianism at best and social Darwinism and eugenics at the worst.

A good example is the bizarre and irritating idea that men are promiscuous and women monogamous because they are pursuing different reproductive strategies, as if they belonged to different species. The problem, of course, is that the proliferation of one's genes requires one's children to survive and reproduce, so that fathers should be no less invested in the nurturing and support of their offspring than mothers. If they're not, there must be some other explanation, of which the simplest and most parsimonious is that natural selection and therefore evolution is still going on, in this area as in every other.

As Richard Dawkins has been heard to complain, most Americans first learned their evolutionary theory from Gould. I'm one of them, and it wasn't a bad place to start.