My natural experiment

By Joel Thurtell

For years — decades, in fact — I’d been telling the same story.

I was trained as a historian and I ought to know.

The famous dictum of the great medieval historian Marc Bloch that “history is the science of change” simply is not true.

How can history be science?

Why, the train of historical events is not chemistry.

You cannot replicate the course of history the way you conduct a laboratory experiment.

“Dad,” my son Abe said. “You are mistaken. You CAN do scientific history.”

To prove it, he threw my own (unfinished) PhD dissertation at me.

He obliged me to face the methodology I’d chosen for my thesis.

Historical demography.

The conversation came about after one of our visits to Michigan’s greatest privately-owned intellectual resource, that four-story warehouse in Detroit known as John K. King Books.

While Abe browsed in the Computer Science stacks, I was hanging out in Historiography.

Suddenly, my eye caught a title that began reeling back my own history.

It was The Historian and the Computer: A Practical Guide, by Edward Shorter, published by Prentice-Hall, Englewood Cliffs, New Jersey, 1971.

Nineteen-seventy-one.

That year, I came back from Mexico with a hand-carved pine box loaded with the “data recovery forms” I’d had specially printed so I could fill them with information I gleaned from colonial Mexican church records.

The data were the core of a dissertation I was planning to write about the ways in which Tarascan Indian civilization declined upon contact with the dominant Spanish culture that took over in Middle America after Cortes conquered the Aztecs.

Only, my interest was not in Aztecs, but in the Tarascans, a large group of Indians independent from and at war with the Aztecs whose empire fell to Cortés.

Show of hands: How many have heard of the Tarascans?

Well, get ready.

I’m getting back into the Tarascan thing, folks, so anybody who comes near me will hear about my Tarascan project.

I wish I’d had The Historian and the Computer when I came back from Mexico in 1971.

I came  to the University of Michigan as a graduate student in history in fall 1967. I was recruited to UM by the late Charles Gibson, a UM prof whose specialty was colonial Mexico. I’d won a fellowship in comparative history of colonization of the New World. Although I thought I was more interested in early modern English history, I took Prof. Gibson’s survey class in colonial Latin American history whose text book was his marvelously insightful and concise Harper Torchbook, Spain in America. I also enrolled in his seminar on comparative colonial history in the New World.

Because of my fellowship, I also got to attend amazing workshops and lectures by a Harvard political scientist, Karl Deutsch. Those meetings were held in the Institute for Social Research. This was all about social history, but more than that, it was all about harnessing the power of the computer to analyze historical data.

It was that whole computer thing that drove me out of the UM PhD program in 1971. More about that later.

I was enticed — entranced, you might say — by the wonders that computers could work.

I read E.A. Wrigley’s Population and History.

I learned that there are forces more powerful than armies and emperors, kings, presidents and generals — the normal focus of much written history.

Plagues, for instance, tend to wreck the best-laid plans and destroy the ablest of armies. Armies of rats are more potent than legions of humans. A drop in the mean temperature of the world can wreak incredible havoc.

One day, I happened into a lecture at the Michigan League. A handsome Frenchman was reading from notes.

He was really hard to understand. Why were people hanging on his mispronounced words?

This was Emmanuel Le Roy Ladurie. His classic book, Paysans de Languedoc, propelled him to the head of the class of historical demographers.

In smaller classes and over beer at the Village Bell bar, I listened to Ladurie and began to grasp what could be learned by painstakingly gleaning information from records that were kept for a purpose wholly different from the historian’s.

Roman Catholic priests diligently recorded every baptism, marriage and burial that took place in their parishes. Unwittingly, they provided data for future historians. They routinely recorded baptisms — births — deaths and marriages. From those leather-bound volumes, you can tally births and deaths, converting the decades and centuries into graphs showing fertility and mortality.

The Spaniards were obsessed with race, so they duly recorded the ethnicity of everyone whose name they wrote down. Some of the people were slaves. Duly noted.

I learned that if a historian could find those registers of vital data, they were a potential gold mine in terms of what you could learn about the way people lived and how people adapted to new circumstances such as the coming of the Spaniards.

That was what I wanted to do.

By that time, I was working part-time at the William L. Clements Library, taking independent reading courses and planning to take my oral and written exams for the PhD sometime within the coming millenium.

Besides working with Prof. Gibson, I was reading and discussing books with Eric Wolf, a professor of anthropology.

It was a nice life. I had my carrel in the grad library where I could hoard books, read and plot what the future held for me once I got my doctorate. No doubt about it, I was going to be a history professor.

It was early 1970. I was not draftable, thanks first to the Chicago Police Department and then to a high lottery number.

I got a call from Prof. Gibson: If I took my prelims (for some reason, at UM the general oral and written examinations for would-be PhD candidates are called “prelims”) asap, there was a $3,000 Rackham Prize Fellowship that would pay for me to do research in Mexico starting in fall 1970.

Three grand. Doesn’t sound like much? Adjusted for inflation, $3,000 in 1970 would have been worth $17,392.19 in 2011. Even today, 17 grand will buy a lot of tacos.

Suddenly, there were lots more books in that carrel. Prelims are very stressful, and this was so sudden. I’d been planning to take the exams in, oh, you know, one of these days. Now, the deadline was on me. I resigned my job at the Clements and dug in. Once I passed the tests, I got a diploma declaring me to be a Candidate in Philosophy. The next question was: What will I do for a thesis?

No doubt in my mind — I was going to Mexico in search of those parish registers of baptisms, marriages and burials that would enable me to piece together the structure of family and village life in colonial times. It is called “family reconstitution.”

Did I know anything about computer programming?

Ah, no.

Had I ever taken a class in statistics?

Come on!

How was my Spanish? Well, I took a so-called graduate reading class at UM designed to help me comprehend academic articles about history written in Spanish. Could I ask the way to the men’s room? Ha-ha!

None of these impediments was seen as a problem. My profs were enthused. Prof. Ladurie was looking forward to a family reconstitution study from Mexico.

Memories chugged through my head as I stood in the dim aisle of John King Books perusing The Historian and the Computer.

I was reading names of people I knew.

The late Charles Tilly, brilliant social historian then at UM and later at Columbia and one of the chief participants in those ISR workshops.

Jerome Clubb, UM history prof, author of articles on using computers in history and facilitator of those workshops.

The book is a nuts and bolts how-to manual for humanities and social science people whose theme is that this quantitative history stuff is not difficult. You don’t need to be a programmer or a statistician.

Wish I’d had this book in ’71!

Over the years, I’d explained my dissertation to Abe, who’s a UM grad with a Computer Science major.

However, it never occurred to me that I might show Abe the contents of that carved pine box.

That evening, after we returned home from John King, I brought the box upstairs and set it on the dining room table.

I pulled out a form. I’d had my forms made by a printer in Morelia, Michoacán. I designed the forms to reflect all the bits of information that appear in the colonial church registers I was reading.

I showed Abe a baptism recovery form. It had the first name of the baby, surname, date of baptism, date of birth, gender implicit from given name, and the baby’s race. The priest recorded similar information about parents and godparents.

Abe picked the form up and looked it over.

“Dad,” he said, “This is ready to be processed.”

If my data were ready to be processed, why didn’t I do it in 1971?

The Historian and the Computer, perhaps unwittingly, explains why.

The book was written partly out of frustration with humanities people like me who have an aversion to working with numbers. Or THINK they have.

I had the additional problem that I needed to make punch cards so my data could be processed by UM’s mainframe IBM computer.

I had a lot of trouble working with the people who ran the computers.

Actually, people like me never met those computer gurus.

We worked through social science people who acted like interpreters.

There was another problem that I now see more clearly.

What questions did I want to ask of my data?

Well, I had a hypothesis — more than one, actually. But converting my ideas into queries for the computer?

I had not a clue how to go about that.

By fall 1971, back in Ann Arbor, my fellowship money was exhausted. I was working in the UM Law Library pasting labels on books.

And then there was the blowup with Ladurie and another French historian, Pierre Goubert, in the lobby of the Michigan Union.

When Prof. Ladurie heard that I was not going to do a family reconstitution as planned, he got quite excited.

His English was quite comprehensible, and he clearly was not pleased.

The conversation was very heated.

Funny how you remember things.

For a time, I was living in a Tarascan Indian village in the mountains southeast of Pátzcuaro in the state of Michoacán. That is where I found a mother lode of parish registers, and I lived with the priest for about a month as I transcribed data from the 17th and 18th century books onto my data recovery forms.

The town I was living in, Cuanajo, still seemed pretty Tarascan to me. People were still speaking Tarascan, a language related to no other in the world. But about three kilometers away there was another village, Tupataro.

Now, both towns were Indian communities when the Spaniards took over in the 1520s. How is it that in Cuanajo in 1971 people still identified themselves as Indians, while in nearby Tupataro, they did not?

Answering that question became the focus of my thesis.

Prof. Ladurie did not like that idea.

Why was I not doing a family reconstitution?

The answer, I now know, is simple.

For a family reconstitution, you need a concurrent run of baptism, marriage and hopefully also burial registers.

I do not have registers of marriages, nor do I have records of individual burials. I have the priests’ tallies of burials. That is pretty good, but not enough for the work Ladurie had in mind.

The core of what I have is several runs of baptismal register data.

I’m damned lucky to have that.

Believe me, I spent a lot of time riding around in pinchis camiones — second-class rattletrap buses — searching for churches with viable records.

The standard line I was given was that due to inclement political conditions, many colonial records have been destroyed.

You know, wars of liberation from Spain in the 19th century, followed by endless wars to see who would rule.

Invasions by the United States and France.

The Mexican Revolution in the 1900s, followed by the Cristero Revolution, a religious war, in the 1920s.

War wreaks havoc with records.

So I was told.

I finally found some records in Pátzcuaro, and later on in Cuanajo.

“What you have, dad,” my son told me, “is a natural experiment.” It is — potentially — scientific history.

No, no, no, I said. History CANNOT be scientific. Unlike a discipline like, say, chemistry, you cannot duplicate historical events.

Not so fast, says Abe. Take a look at Jared Diamond’s Guns, Germs and Steel.

Diamond describes a “natural experiment” in which the destinies of different subdivisions of one original population are compared. You can study what happened in the dispersion of a people who moved from Asia to the Pacific islands and then adapted variously to differing physical conditions.

That is not a chemistry experiment, and yet it is possible to  study the different fates of the descendants of a core population and draw some conclusions about how people behave and adapt to differing conditions.

And so it is, Abe told me, with my data recovery forms.

The priests unknowingly were compiling uniform records of vital information. Dates of baptism, marriage, burial.

Why, I have graphs I made by hand in 1971 with lines showing annual fluctuations in birth and death. You can see when the epidemics hit. You can see how the people responded through increased or decreased conceptions.

These are data that can be analyzed by computer.

And we can learn neat things.

But we have to think hard about our questions.

That is what I’m doing now as I dig back into the data and ponder what questions I want to ask and how I want to put them.

That’s all I have time to write, now.  I’m working on a computerized form that duplicates my paper forms, and pretty soon I’ll be entering the colonial priests’ data into my new laptop so I can ask my questions.

But watch out — I’ll be writing more about my natural experiment and the Tarascans.

To be continued