Verisimilitude vs. legisimilitude (Q762133)

The recent definitions of truthlikeness - by R. Hilpinen, P. Tichý, D. Miller, I. Niiniluoto, and G. Oddie - have not paid attention to the distinction between lawlike and accidental generalizations. \textit{L. J. Cohen} has expressed this by saying that science should be viewed as aiming at ''legisimilutude'' rather than ''verisimilitude'' [Synthese 45, 489-510 (1980)]. \textit{G. Oddie} has given a reply to Cohen by defining the concepts of law and legisimilitude within higher-order logic [Synthese 51, 355-379 (1982)]. This paper gives a different reply to Cohen by treating laws as physically necessary generalizations and by defining the notion of legisimilitude for a statement h as the distance of h from a suitably chosen lawlike sentence. This idea allows the author to explicate the concept of legisimilitude by means of the same ''similarity approach'' that he has earlier used in the theory of verisimilitude for singular sentences, extensional generalizations, and first-order theories [Synthese 38, 281-329 (1978; Zbl 0388.03001)].