2(3)7-11Responses.pdf text- Lves, md indue- !- 1er lms, thich :hat the ld at l a who not losophy . ~ect i, :>n- ttion :tere ce- an in- teluc- ion to much :>uld ts t? lman • s News- l see te- :>mmit :tive rea- ies. 1ve to eir il how of nclu- ive duc- :mfir- otheses this make her tive. ism, e sup- darn and e the ing uen- and· argu- To ,e .othing hlnk, :gories Deductive and Inductive: Types of Validity, Not Types of Argument David Hitchcock McMaster University Recent contributors to the Informal Lflic Newsletter (1, 2, 3, 5} have d1.scussed e legitimacy and nature of the common distinc- tion between deductive and inductive argu- ments. I shall argue that Weddle (.5) has abandoned the distinction for the wrong rea- sons and has rejected what is worth preserving in it. Fohr Cll argues convincingly against one way of construing the distinction, I shall maintain, but proposes an alternative which is equally unacceptable for several reasons, including one pointed out by Weddle (5) and one pointed out by Govier (.3) but not taken to its logical conclusion. I shall propose instead that we regard the distinction be- tween deductive and inductive as a broad and exhaustive distinction between types of val- idity. On my account Wellman's (6) distinc- tion between inductive and "conductive" is a distinction between two types of inductive validity. Finally, I shall note that my pro- posal is not novel. Weddle holds that all carefully drawn argu- ments, whether traditionally deductive or traditionally inductive, provide conclusive grounds for their conclusions, in the sense that it is (logically, mathematically or physically) not possible for the premises to be true and the conclusion false. In this sense the inference of every carefully drawn argument is deductively valid. The distinc- tion between deductive and inductive there- fore disappears, though we can still distin- guish syllogisms, analogical arguments, statistical generalizations, causal arguments, arguments from authority and good reasons arguments. The trouble with Weddle's position is that he is mistaken in his claim that "when an arguer properly hedges the conclusion of a traditionally inductive argument, the result assumes the role held to belong exclusively to deduction." (p . 3) Consider Weddle's own example: "When a low pressure ridge moves down from the Gulf of Alaska (etc.) we usually get rain the next day, and a low pres- sure ridge is moving down right now (.etc.). : hence it is likely to rain tomorrow." Weddle claims that with pr~~ises thus filled out and the conclusion hedged, it is impossible for the premises to be true and the conclusion false. But suppose some facts not mentioned in the premises (.such as a competing high 9 pressure ridge coming from a different direc- tionl make it highly unlikely that it will rain tomorrow. Then, despite the truth of the premises, the conclusion is false. In general, traditionally inductive arguments of this sort become deductively valid only if we stipulate that the premises describe a closed system or we add an open-ended premise to the effect that no other factors obtain which would make the predicted state of af- fairs improbable. But adding such premises changes the traditionally inductive argument into a traditionally deductive one. Hedging the conclusion of a traditionally inductive argument in fact involves a claim about the strength of the link between the evidence cited in the premises and the occur- rence predicted in the conclusion. Since there may be uncited counter-evidence, the predicted occurrence may be in reality highly improbable, even though the premises provide probable grounds for thinking it will occur. Although Weddle's elimination of the dis- tinction fails, there are good arguments against three traditional ways of defining it: in terms of logical form, in terms of the strength of the link between premise(s) and conclusion, and in terms of the claimed or intended strength of this link. Weddle (5) himself provides convincing counter-examples to attempts to make the distinction on the basis of logical form, such as the contention that inductive arguments draw universal conclusions from particular premises and deductive arguments draw partic- ular conclusions from universal premises. Skyrms (4, pp. 13-15) gives examples of all possible combinations of particular and universal statements in both deductively valid and inductively strong arguments. Fohr (1) rightly points out that a distinc- tion in terms of the strength of the link be- tween premises and conclusion runs foul cf the requirement that there can be bad, i.e. in- valid or weak, instances of each type of argument. Arguments which are neither deduc- tively valid nor inductively strong (nor conductively valid, etc.) will have no place in a supposedly exhaustive classification. One way of patching up this approach is to label "inductive" all arguments which are not deductively valid, but, as Weddle (5) notes, the practice of logic texts indicates that this is not how logicians make the distinc- tion. Most recent logic texts define deductive arguments as arguments which involve the claim that the conclusion follows necessarily from the premises, and inductive arguments as arguments which claim only that the premises provide probable grounds for the conclusion. Fohr's (.1) proposal to base the distinction on whether the person putting forward the argument intends his premises to provide conclusive or merely probable grounds for the conclusion is a variant of this approach, for the only way of detecting such intentions is to notice what·the arguer claims. Although this third type of proposal escapes the ob- jections to the two previous approaches, there are two strong arguments against it. In the first place, as Weddle (5) points out, the claimed or intended strength of the --- - ----- - - - - inferential link is partly a function of psychological temperament. Bold reasoners will claim more strength than the inference has, timid ones less. In assessing their arguments, however, we would discount the hyperbole of the one and the hesitation of the other, and apply the standards of correct- ·ness appropriate to the argument itself, per- haps noting parenthetically the rashness or timidity of the claim the arguer makes for his argument. We would, in other words, treat the bold reasoner's argument as induc- tive, despite his claim that the conclusion followed necessarily, and the timid reason- er's argument as deductive, despite his hesitation about drawing the conclusion. In the second place, arguers may in fact have no intentions at all about the strength of the link between premises and conclusion. This is not merely the point made by Govier (4) and conceded by Fohr (1) that, where arguers make no explicit claim about the link, we may not be able to determine what they intend, and may thus have to cover our- selves by assessing the argument both ways. It is the stronger point that the arguer may simply intend his premises to convince his hearers of the conclusion, without either intending that the conclusion follows neces- sarily or intending that it is made probable by the premises. Suppose, for example, I say to my wiie: "You should help me paint the kitchen this evening. You promised you would." My intention is to convince her to help me paint the kitchen, on the ground that she promised she would. But I make no claim, nor (let us suppose) do I have any intention, about the strength of the link between my premise and my conclusion. Though this imag- inary example is sufficient to establish my point, I suspect that in the real world a great many people have no such intentions when they put forward arguments. So the dis- tinction between deductive and inductive arguments fails to be exhaustive. If traditional attempts at defining the distinction break down, we have to remind ourselves of the point of making the distinc- tion in the first place. We can find a clue in Weddle's (5, p. 4) "exaggeration" that "what distinguishes deductive from inductive arguments is the sections of logic books in which they happen to be found." Precisely, although Weddle errs in characterizing the difference between these sections. In the sections of logic books which deal with so- called "deductive arguments", we develop the theory of the circumstances in which an argu- ment is deductively valid or deductively in- valid--that is, in which it is impossible or possible for its premise(s) to be true and its conclusion false. Within this general category, there is a variety of types of logic: the logic of truth-functional sen- tence connectives, the logic of first-order quantifiers, the logic of identity, the logic of the Aristotelian relations between non-empty sets, the logic of Lewis-55 modal- ities, and so forth. In the sections of logic books which deal with so-called "inductive arguments", we develop the theory of the circumstances in which an argument is induc- tively strong or inductively weak--that is, in which it is more or less probable that its conclusion is true given that its premise(s) 10 are true. Within this general category, there is a variety of types of logic: the logic of the confirmation and disconfirmation of hypotheses, the logic of analogical argu- ments, the logic of inferences from sample characteristics to population character- istics, the logic of controlled experiments to prove causal claims, the logic of conduc- tive or balance-of-considerations or good reasons arguments, and so forth. we ought to assess an argument on the basis of which of these specialized types of logic seems to provide the most appropriate frame- work--in other words, on the apparent logical form of the argument. In doing so, we may be guided by the claim or intention of the ar- guer about the strength of the link between premises and conclusion. But such a claim or intention is at best of heuristic value, and may have to be discounted. The main question to be asked in this connection about any argument is how strong the link is be- tween the arguer's premises and his conclu- sion, not whether the arguer's claim about their link is correct. Let me conclude by noting that others be- fore me have rejected the distinction be- tween deductive and inductive arguments in favour of a Qistinction between deductive validity and inductive strength. One such author is Skyrrns (4, p. 12). Since he has not succeeded in convincing everyone, there may nevertheless have been some point in bolstering his position with the preceding arguments. WORl