Tuesday 14 May 2013

Colonial Literacy, sentence complexity, and present day reading ability


Sanjoy Mahajan at freakonomics.com has an interesting piece on Colonial literacy, looking at the success of Tom Paine's 1776 "Common Sense" which in one year was read, or at the very least bought, by 20% of the American colonial population. Translated into today's US population, that is the equivalent of of a book selling 60 million copies. Few books attain that level. It took 8 years for The Davinci Code (a book not about common sense) to reach that level.

Although the words are short, the language and reasoning of Common Sense is rich and complex. Here is a very small example:

Society in every state is a blessing, but government even in its best state is but a necessary evil; in its worst state an intolerable one; for when we suffer, or are exposed to the same miseries By a Government, which we might expect in a country Without Government, our calamity is heightened by reflecting that we furnish the means by which we suffer.

You may be a different case, but I did a double-take or two, to make sure I was carrying the sense well enough to go on the the concluding punch line. Such complexity is at the heart of reading for meaning. The current National Assessment of Adult Literacy classifies the above text as being at the highest level, Proficient Reading, only attained nowadays by 13% of the American population. 

So, despite the apparent rise in intelligence suggested by the Flynn Effect, proficient reading has fallen from an implied 20% to a measured 13%, using comparable measures of complexity.

Mind you, 1776 was something of a special year for Americans, one which encouraged serious reading. They may have made a particular effort to understand the text. Either that or they read proficiently, inwardly digested, and came to their own proficient conclusions. Paine's view was clear: "To have had a part in two revolutions is to have lived to some purpose."

Regard him as one of the early bloggers.


  1. Perhaps this textual analysis can be used to provide a rigorous estimate of pre-1880 IQs (as opposed to Hart's ballpark estimates).

  2. Jan te Nijenhuis15 May 2013 at 07:58

    Dear Dr. Thompson,
    this is indeed a fascinating finding. As the previous poster already mentioned, Woodley, te Nijenhuis, & Murphy showed that the genetic intelligence of the Victorians is estimated to be 14 IQ points higher, so this finding of the excellent sales of a difficult to read book in the 18th century fits in with that.
    I guess there are linguists out there with computer programs that can give you a rating of the cognitive complexity of books. Would the complexity be higher in the Victorian time than it is now?

  3. If I may answer both of these very interesting questions at the same time (as opposed to simultaneously) I think that you jointly raise a very important point: we do not yet have a universal measure of item difficulty. Textual analysis, computer programs designed to measure the cognitive complexity of books, analysis of the differences in difficulty of maths problems and so on are the way forwards in understanding the contribution of intelligence to success in different walks of life. We have to answer the fundamental question: What makes problems difficult? I will try to post about this, but one suggestion is that we use Kolmogorov complexity to specify the difficulty level of each problem (by measuring the length of the shortest possible description in some universal language). This might well be the basis for a new field in intelligence research, if we could find the right team to do the job.