Thursday, January 30, 2014

True Stories

One of the criteria for the use of stories in social science that Andrew Gelman and I are trying to promote is "immutability". What we mean by this is that the facts that a story conveys should be well-established to resist the pressures of whatever theory the story is trying to either challenge or illustrate. We don't mean that the story must be true. The facts of Hamlet's life are well-established despite being almost certainly fictional. You can get them wrong, revealing your ignorance of Shakespeare's play. (It's no good to defend yourself by invoking Saxo, by the way. Here the character's name is Amleth.) And that's central to our suggestion: if you can't get the details of the story wrong, then you can't use it to support an argument.

Some stories are only interesting if they are true, however. Their "basis in reality" is part of the story. If it didn't actually happen then the events in the story aren't really worth paying attention to. Other stories—either because they are deeply entrenched in the larger cultural narrative or because they are the products of exceptional minds, geniuses—are interesting even though they're "made up". Some stories work either way, but have different effects depending on whether they are taken to be true or false. Consider the difference between telling a story that begins "I'm the kind of person who would…" and one that begins "Guess what I did today." In the first case you will let the story illustrate a character trait, but in the second you come into possession of perhaps valuable information about how I live my life. If you use this story at my wedding, friend, you'd better get that important difference right!

Even when a story illustrates an important truth, that is, it is important to be truthful about the nature of the events it describes. I.e., whether or not they are actually fictions. And, for the sake of your own credibility, even when the facts are fictional you have to get them right. After all, your listener or reader might already have heard the story in another version, with different facts, or might know, contrary to your assurances that this is a "true story", or that this "actually happened", that this story has long ago been revealed to have been a hoax or a canard.

As my regular readers know, one of my favorite examples is Albert Szent-Gyorgyi's story about a group of soldiers in the Alps that found their way back to camp using a map of the Pyrenees. Among management researchers and consultants it's known as Karl Weick's story, mainly because he invariably fails to adequately credit his source, a poem by Miroslav Holub (that also attributes it to Szent-Gyorgyi). Weick's version is a plagiary of Holub's poem. Here's Weick's version:

Definitions not withstanding, I can best show what I think strategy is by describing an incident that happened during military maneuvers in Switzerland. The young lieutenant of a small Hungarian detachment in the Alps sent a reconnaissance unit out into the icy wilderness. It began to snow immediately, snowed for two days, and the unit did not return. The lieutenant suffered, fearing that he had dispatched his own people to death. But the third day the unit came back. Where had they been? How had they made their way? Yes, they said, we considered ourselves lost and waited for the end. And then one of us found a map in his pocket. That calmed us down. We pitched camp, lasted out the snowstorm, and then with the map we discovered our bearings. And here we are. The lieutenant borrowed this remarkable map and had a good look at it. He discovered to his astonishment that it was not a map of the Alps but of the Pyrenees. ("Substitutes for Strategy", p. 223)

While Weick reports the details of the story verbatim from Holub's poem (which is the same text, except with line breaks), he has subtly changed the story by the way he frames it. What Weick calls "an incident that happened", Holub calls a "story from the war". And while Holub doesn't specify exactly where it took place, except to say the soldiers were Hungarian and the mountains were "Alps", Weick sets the story explicitly in Switzerland. That is, Weick both presents the story as much more "true" than Holub, and introduces the entirely implausible idea that a group of Hungarian soldiers were wandering around in the Swiss Alps. Some people, when I show them Weick's version actually get stuck on this detail: "What were they doing in Switzerland?"

Interestingly, both the idea of presenting the story as "true" and the freedom to embellish it with details not provided in one's already slightly retouched source seems to be part of the tradition of telling it, at least, like I say, in the management community. I recently ran into a great example of this on the blog of StoryCare, a consulting product developed by Richard Stone at the health-care focused consultancy Synensis. Here's how he tells the story:

[Update: the quotation that follows is from the undated post on StoryCare's blog as it appeared before January 31, 2014. It has since been updated in light of my critique. Crucially, Stone has dropped the word "true" from the first sentence. I would argue, however, that the story is still being presented as true, which, as I say below is understandable given the way Weick characterizes the "incident".]

In his seminal book Sense Making in Organizations, Karl Weick tells a fascinating true story about a lieutenant in World War I who sends out a patrol into the French Alps to scout out the positions of the German troops. The small patrol took no provisions, because this was intended to be just a short search and they planned on returning to camp by nightfall. But about two hours into their trek it began to snow—so hard that it was soon a white out and the soldiers could barely see their hands in front of their faces. They were in trouble. Their leader led them to a small overhang in the side of a mountain where they settled in, hoping that the snowfall would break by late afternoon. But it continued to snow through the day, into the night, and for the remainder of the next day. It was one of those blizzards that comes around only every 500 years or so. By the end of the second day the team had gone through all their provisions and were growing hungry. Huddled together under that cliff to share their dissipating warmth, their hope for survival was growing bleaker by the moment. On the morning of the third day the snow was beginning to let up, but without any clear landmarks that hadn’t been obliterated by the 50 inch snowfall, they were lost and mentally preparing to die in the wilderness. One of the soldiers decided to rummage through his pack hoping to find some morsel of food that he might have overseen. There, folded at the bottom was an old map of the Alps. When he announced that he had found a map everyone’s spirits were buoyed. They made a decision to head out in the hopes of reconnecting with their regiment. Continually referring to the map for clues as to where they were, they slowly made their way through the drifts. Finally at nightfall one of them saw the glow of a light in the distance. They had found their way home. After the reunion and as his men warmed themselves by a fire and ate like they had never eaten before, their commander was curious about how they had escaped a frozen fate. He asked to see the map they had used. Examining it by his lantern, he looked closer to discover that this was in fact not a map of the Alps, but a map of the Pyrenees!

Notice all those extra details. When I asked Stone about it, he was entirely open about having embellished the story to bring it to life. As an aside, I disagree with him about the effect of the embellishment, which weighs the story down with unnecessary details. There's a reason this story has had such an impact on management theorists and practitioners: its imagery was crafted by a skilled poet. But the important thing to notice is that Stone presents this as (a) Weick's story and (b) a true story. He can be forgiven for thinking it is true, i.e., for taking the word of a professional scholar about the veracity of story that the scholar describes as "an incident that happened during military maneuvers in Switzerland". But there can be no question that Stone here misrepresents his source on a number of matters of fact. Some things he just gets wrong, others he simply makes up. First all, the original story certainly did not take place in the "French" Alps, and the enemy was not likely to be German (the Hungarians were on the side of the Germans in World War I). A whole list of details are pure speculation and window dressing:

no provisions
this was intended to be just a short search
they planned on returning to camp by nightfall.
two hours into their trek it began to snow
it was soon a white out
the soldiers could barely see their hands in front of their faces.
their leader led them to a small overhang in the side of a mountain
they settled in,
they hoped that the snowfall would break by late afternoon.
It was one of those blizzards that comes around only every 500 years or so. (!)
they went through all their [nonexistent] provisions and were growing hungry.
They huddled together under that cliff to share their dissipating warmth,
There had been a 50 inch snowfall, (!)

Then this little narrative misstep:

Folded at the bottom was an old map of the Alps.

Notice that this is a misleading statement. The trick to telling this story, as Holub understood, is to leave out the detail about what it was a map of. "What they thought was a map of the Alps," would have given away the ending. But this statement is, even within the wide berth of a fictional frame, a lie. It is simply contradicted by the surprise ending. Not very good storytelling here. Now, notice what happens next:

Continually referring to the map for clues as to where they were, they slowly made their way through the drifts.

This removes a crucial ambiguity in the story as it is normally told. There is a real question about how useful a wrong map would be, but especially if you "continually refer" to it. It might give you the hope you need, and make you pay extra close attention to clues in the environment (this is how it is normally explained), but surely the wrongness of the map will quickly become apparent if you keep going back to it to update your position. If it doesn't, and you still let it lead the way, you are very likely to wind up at the bottom of a ravine.

Richard Stone's readers are here being told a story, and even being provided with a source for it. And they are being told it is true. But, not only is the story not likely true in the first place, the version of the story they are getting misrepresents the source, which means that if they go on to tell it, French Alps and German positions and all, and happen to do so to someone who has actually read Weick, they will reveal that they don't know what they are talking about.

It's this kind of thing that Andrew and I think that especially social scientists, but also the consultants who invoke social science, should be better at. Storytelling is fine. But you have to get the story right. I guess it's sort of like that "amusing anecdote about a drug deal" in Reservoir Dogs.


Gabriela Rosschou said...

There are no true stories. Every narrative is produced as a communicated perception of a certain individual. Perception of situation, feeling or text.
We all have different filters rooted in experience (both personal and collective), environment (physical and social), patterns of behaviour, culture, expectations.
It might be useful in many ways to remember that every story is just a specific interpretation of certain involvement of a certain person at certain time (even our own interpretation of the same situation or text would change in time due to newly acquired experience).

Another thing is, of course, the art of telling a story for purpose, i.e. when we are using a story to support our precious activity as teachers, consultants, and, yes, parents. Being dedicated anarchist, I must admit that in this case sticking to the rules is important, if the story should work.
Thinking in terms of Bruno Bettelheim’s The Uses of Enchantment, The Meaning and Importance of Fairy Tales.

Thomas said...

My view is that there are "true stories" in whatever sense we need to differentiate them from fictions, frauds, and errors. It is possible to make up a story to entertain. It is possible to tell a story we know is untrue with the intention of misleading someone. And it is possible to get part of a story wrong.

I'm glad you agree about the rules as they apply in this case. Here, after all, the question is mainly one of getting the source right. But, as I point out, one of the things Weick gets wrong is the "truth" status of the story: his source says it's a "story from the war", Weick says it's an "incident". Stone then runs with this and calls it simply "true", which, I'd argue is as untrue as saying the troops were French.