Fallaciousness without Invalidity
John A. Woods, Douglas Walton · 1976
Professor James E. Broyles points out in a recent issue of this journal1 that arguments sometimes cited in the logic texts as paradigms of the fallacies of composition and division can quite legitimately be interpreted as instances of correct inductive reasoning. This observation, he concludes, raises some serious questions about the significance of these traditional classifications. We would like to emphasize the underlying theoretical significance of this point by drawing out one generai consequence of it for the theory of argument2 appropriate to the informai fallacies. The issue is whether invalidity is a necessary condition in an argument of fallaciousness. There are at least three interesting arguments for rejecting this implication, (i) Often it is felt that simply because an argument is an instance of an informal fallacy, it does not follow immediately that it is incorrect or invalid. For example, many instances of the traditional ad verecundiam turn out to be legitimate appeals to genuine authority or expertise.3 (ii) As Broyles points out, some instances of composition, e.g., 'players/basketball team/good'4 may be inductively yet fallacious if considered as a deductive argument, (iii) Various deductively valid forms of argument such as *p, therefore p' hâve substitution instances that commit the fallacy oipetitio principii.5 So it seems that an argument may be both valid and fallacious.6 Yet there are equally strong reasons for rejecting this rejection of the invalidity condition of fallaciousness, for since it is virtually uncontroversial that its converse does not obtain,7 rejection of it cuts formai logic and the informai fallacies asunder, apparently making the study of fallacies purely a matter of psychology, or some other discipline outside logic. And a tradition not to be despised suggests that an informai fallacy is an invalid argument even if its invalidity is not reflected in first order logic, or even in other standard or wellknown formalisms.