This paper explores the historical evolution of views of score validity beginning with J. Guilford (1946) and ending with the writings of L. Cronbach (1989), S. Messick (1989), M. Kane (1992), P. Moss (1992), and L. Shepard (1993). Current treatments of validity have emphasized that it is the inferences made from scores, and not tests, that are valid (L. Wilkinson and the American Psychological Association Task Force on Statistical Inference, 1999). Recent treatments have increasingly emphasized the importance of considering falsification in evaluating validity. (Contains 15 references.) (Author/SLD)