Social Graphs: Network Science and Mises

After blathering about social graphs through my last few blog entries, I finally had an intelligent idea: check the local used book store for volumes about social graphs.  Some real research is far better than my stream of consciousness, right?

And lo and behold, I came across Duncan Watts’s book Six Degrees: the Science of a Connected Age.  After leafing through it, I decided that the book was deep enough to warrant purchasing, and I took it home.  Only later did I realize that Duncan is one of the (very few!) people actually creating the new field of network science.

Network science is the brand-new field of science that uses tremendous amounts of computational horsepower to study the structure and behavior of large-scale networks, down to the level of the individual node.  Network science is such a new field because without large, fast computers, it is impossible to reveal any worthwhile results about networks with millions of nodes.

For example, starting on page 56 of his book Watts discusses the efforts of a mathematician at the University of Chicago named Anatol Rapaport who studied epidemiology in the 1950s.  Rapaport was taking a social graph approach to the problem, and actually achieved some initial results using just pencil and paper.  Here is the paragraph that ends Watts’s section on Rapaport:

“Back in Rapaport’s day, this realization was pretty much the end of the road, and reading his original papers you can see that he knew it.  Perhaps if the University of Chicago group had had the same kind of computers that we have today, they might have cracked the problem wide open, and network theory might have taken a very different route.  But they didn’t.  Blinded by a lack of data and hobbled by a dearth of computational power, the theory of random-biased nets struggled as far as its few protagonists could take it with their mathematical intuition, and then effectively disappeared.  It really was an idea for a future age, and like many such ideas, it had to do its time in purgatory.”

When I read this paragraph, I was reminded immediately of Mises.  Especially by the use of the word “purgatory.”  We all know that Mises spent his life in academic purgatory, and we all believe that he was way before his time.  But maybe Mises was also way before his time because the tools that he needed weren’t invented yet.

Mises was right to reject statistical economics – use of statistics automatically reveals that the user has no idea what is actually going on.  A quote from Human Action, section 5 of his “Prices” chapter: “The first variety is represented by the statisticians who aim at discovering economic laws from the study of economic experience.  They aim to transform economics into a quantitative science.  Their program is condensed in the motto of the Econometric Society: Science is measurement.  The fundamental error implied in this reasoning has been shown above …”

But while mathematical, Watts’s work is not statistical, and he includes almost no equations in his book.  Watts is using logic and computational horsepower to find new structure in networks.  He is studying dynamic networks to discover new theories behind their behavior.  One wonders how much further Mises could have gone if he had had a supercomputer in his day.

Leave a Reply

Your email address will not be published. Required fields are marked *