BOOK REVIEW
A Statistician in Embryo
John Aldrich
Karl Pearson: The Scientific Life in a Statistical Age.
Theodore M. Porter. x + 342 pp. Princeton University Press, 2004. $35.
Today Karl Pearson (1857–1936) is chiefly remembered as a
statistician. Practitioners associate his name with a correlation
formula, with the chi-squared test and with "Pearson
curves," whereas historians generally consider him to be the
founder of the 20th-century discipline of statistics. As the subject
for a biography, he is both promising and challenging: His life was
an extraordinary intellectual adventure, embracing physics,
philosophy of science, biometry, history, the “woman
question” and eugenics. How to make his story coherent is the problem.
Biographies of scientists are often written by other scientists, who
want to show how the subject's discoveries and contributions that
are now judged to be important grew out of his or her life. Karl
Pearson: The Scientific Life in a Statistical Age, by
Theodore M. Porter, is more about "the moral and intellectual
presumptions of the statistical life" than about discoveries.
Porter, a historian at the University of California, Los Angeles, is
interested in how quantification became institutionalized in the
modern world, and this is his third book on the subject. These
volumes look like a trilogy, although they were not planned as such.
The Rise of Statistical Thinking, 1820-1900, published
in 1986, examined how quantified evidence altered the practices of
business, research and record keeping. A decade later, Trust in
Numbers: The Pursuit of Objectivity in Science and Public
Life described how the ideal of expert opinion based on
quantified evidence became established in the course of the 19th
century. Now Karl Pearson: The Scientific Life in a Statistical
Age investigates the vocation of the quantifier from the
inside, through the biography of a major figure.
Porter chose Pearson in part because he finds him fascinating but
also because Pearson tried on alternative lives, in fact and in
imagination. Measurement did not come naturally to the "great
quantifier," but writing did. Porter uses the fiction, the
histories and the masses of letters Pearson wrote to reveal what he
thought was involved in becoming a scientist or a statistician. The
becoming is more interesting than the being, because Porter’s
engagement slackens after the 35-year-old Pearson has been
"seized" by the "statistical impulse." A
statistician-biographer would not have raced through the decades of
statistical achievement, as Porter does.
Porter was motivated to write this book by the fact that by 1900
Karl Pearson had made himself the world’s leading mathematical
statistician. In 1880, and even in 1890, no such position had
existed, and had such a position been imagined, Pearson would not
have seemed a likely candidate to fill it.
In 1880 Pearson had recently graduated in mathematics from the
University of Cambridge, coming in third in the Mathematical Tripos,
the country’s most prestigious intellectual competition, used
as much for picking future judges and bishops as for anointing
promising mathematicians and physicists. His father was a lawyer,
and it was expected that Karl would become one too. However, as he
prepared himself for the law, he pursued his real
enthusiasms—for philosophy, literature and history. He
published an epistolary novel (an autobiography from one of those
imagined lives) and a passion play, committed works but artistically
unsuccessful. However, his essays on history and sociology were
professional pieces, "deeply researched and startlingly
original." Porter guides us through Pearson's changing
interests—the poetry, the medieval history, the history of
customs, and the like—as he describes the young man's evolving ambitions.
By 1890 Pearson was established in an academic career, but not as a
historian or sociologist, rather as a physicist and applied
mathematician. His specialty was elasticity, but this was not
exciting enough, so he worked intermittently on fundamental physical
theory, although without success. The zoologist W. F. R. Weldon
became his colleague at University College in 1890. This led to
their partnership in biometry—work that proved to be a
rehearsal for the statistics that came later. Porter covers these
developments well.
The author finds a "deep tension" in Pearson's life,
between the objective, almost mechanical, quantifier and the thinker
whose work could not be fragmented without destroying it. To a large
degree, the tension has to be inferred, for although we are given an
inside view of the man who would become the statistician, we have no
inside view of the statistician. Apart from a very telling document
Pearson wrote within hours of his wife's death in 1927, in which he
questioned his whole life, we see only the public life. Porter is so
familiar with the sources that I assume they contain no record of
Pearson's inner life. The contrast between the first seven chapters,
which are about the future statistician, and the two chapters on the
actual statistician is striking, although not easy to interpret.
Porter's biography of the young Pearson, the statistician in embryo,
exceeds all expectations in recreating the intellectual worlds in
which Pearson tried to find a home. The breadth of the reading and
the depth of interpretation are impressive and seem to me to go far
beyond the requirements of investigating the vocation of
statistician. It is also refreshing to have a study emphasizing the
young Pearson. The older Pearson is a familiar, villainous figure in
stories from genetics, feminism, politics and statistics: He is
either on the wrong side of the argument or behaves like a monster.
For example, those telling the story of Pearson's relations with R.
A. Fisher usually cast Pearson as Goliath to Fisher's David. But
Porter shows us a young Pearson, clever and brave, who has a burning
passion to understand things.
This book adds much to our knowledge of Pearson and is a must for
the Pearson specialist. But what of the general reader? Anyone
encountering Pearson in statistics and curious about what he did and
why he did it might better go to The Rise of Statistical
Thinking, or to Stephen Stigler's The History of
Statistics, for orientation. Then the reader might tackle this
book, perhaps starting with the statistical chapters (8 and 9) and
then reading the book from the beginning!—John Aldrich,
Division of Economics, School of Social Sciences, University of
Southampton, England