Correctness in language: A formal theory
Johannes Gordesch, Burkhard Dretzke · Journal of Quantitative Linguistics · 1998
The article deals with the problem of correctness in language. Modern linguistic research shows that the dichotomy correct‐incorrect is insufficient to describe acceptability in language. One can establish five degrees of correctness ranging from acceptable, divided, ill‐established, dubious, to unacceptable usage. A mathematical model is developed for the five categories, while a generalised concept of probability, the real interval [0, 1] is replaced by a subset of a finite field, serves as the cardinal tool. Apart from standard statistical packages, computer algebra systems and systems for programming in logic provide appropriate computational means. This model is applied to Mittins’ set of data. It can be proved that in matters of correctness a tendency towards the statistical norm in language is inevitable.