Pictorial Statistics
探讨高尔顿通过照片合成进行归纳推理的方法，发现其虽追求机械客观性，但主观判断仍不可或缺，且需非专家进行数据熟悉化。
This essay discusses Francis Galton’s method of inductive inference where the data are photographs of human faces. His aim of induction was to determine the typical characteristics of the natural class to which the individuals belong by composing the relevant photographs in a specific photographic way. The three populations studied by Galton were people suffering tuberculosis, Jews, and criminals. This essay argues that despite the fact that Galton aimed at mechanical objectivity, subjective judgements nevertheless appear to be a necessary part of this kind of inductive inference. At first sight, this seems very much in the line of Lorraine Das-ton and Peter Galison’s account of objectivity. They argue that in the twentieth century the awareness arose that mechanical-objective pictures still could contain errors that should be erased by trained judgement. Gal-ton’s case of inductive reasoning, however, departs from this account by showing that the correct composites were achieved by a combination of mechanical procedures and untrained judgements. To arrive at the typical characterizations one first has to familiarize oneself with the data, but the familiarization should be done by someone who is not an expert on the cases under study.