Published online by Cambridge University Press: 14 March 2022
A theory of true observation is developed as a generalization of the method of inter-observer agreement that scientists use to determine the objectivity and reliability of observations. A true observation is defined as a statement included in a set of statements in which there is statistical dependence and perfect agreement between the statements made by a universe of experimentally independent persons. Meaningfulness—the existence of an objective referent—for each form of statement included in the set is inferred from statistical dependence, correct meaning within the universe of observers from the absence of disagreement. In this theory truth is inferred from a spatio-temporal pattern of statements as behavior rather than from the traditional correspondence between observation and object or logical relationships among statements.
1 For a case history of a doubtful observation see “The N-Rays” by R.W. Wood, Nature, 1904, 70, pp. 530–531, and some comments on the case by Sir J. J. Thomson, in Recollections and Reflections 1937, p. 395.
2 Examples of this method and a list of references may be found in Arrington, R. E., Time-sampling Studies of Child Behavior, Psychological Monographs, 1939, 51, No. 2.
3 Ernst Mach, Analysis of the Sensations, Chicago: Open Court, 1897, p. 17.
4 Carnap, Rudolph, “Truth and Confirmation,” in Feigl, H. and Sellars, W., Readings in Philosophical Analysis, 1949, p. 126.