Experimental Abstraction: Francis Galton, John Venn, and Cambridge Anthropometry, 1887-1891

The story of statistics before 1900 is one of a logic common to every science that emerged from the interplay of two developments: the combination of observations and the use of probability mathematics. Both having separate beginnings, these two developments intersected in the first decades of the 19th century, only to spread as a single method horizontally – across scientific disciplines – and vertically – in terms of technical sophistication. This neat story comes at the prize of a loss of historical accuracy. The main reason is its focus on abstract concepts and lack of attention to the material and local aspects of the interplay between observations and mathematics needed to establish sufficiently abstract statistical knowledge. The present paper draws on original archival research to describe a remarkable, yet hitherto little-known episode in the history of statistics: Francis Galton’s collaboration with John Venn, between 1887 and 1889, in an unofficial psychometrical laboratory at Cambridge. Its focus is on the various difficulties Galton and Venn in their joint endeavor, which ranged from choosing a suitable room and weighing the reliability of instruments to aligning statistical techniques with measurement results. In doing so, Galton and Venn were forced to use their polymathic skills to come up with hands-on ways to find out what were relevant statistical associations. The paper concludes by placing the Galton-Venn laboratory into the context of the emergence of psychology at Cambridge and by considering its importance for the discipline of statistics in the 1880s-1890s.