The Yudofsky scale is considered to be one of the best scales for measuring aggressive behavior developed to date. One of the chief shortcomings of the scale, however, is appropriate methods for scoring it in ways that make the resulting scores well-suited for data analyses. The basic scoring problem with the Yudofsky scale is that the scale is inherently discontinuous and non-linear; more specifically, it is a logical rather than classical psychometric scale. A variety of alternative scoring procedures are explored and compared empirically for the Yudofsky scale through an assault vignettes validation experiment conducted using a modified version of the scale. A simple approximation to C. E. Zeeman's swallowtail catastrophe theory model proved to be the best of all the procedures examined. The swallowtail scoring model produced Yudofsky scale scores that discriminated out control (verbal aggression only) from mild and severe assault vignettes with ratio level mean differences between the three vignette types as one would predict for these vignettes. The swallowtail scores were also logically coherent, directly interpretable, and psychometrically excellent. Catastrophe theory--a theory and mathematics of discontinuous and non-linear phenomena--was used effectively to score discontinuous and non-linear logical scales. Two tables are included, 13 references are attached, and the overt aggression scale is appended. (Author/TJH)