We said that all arguments, deductive and inductive, are attempts at justifying (establishing the truth of) a conclusion on the basis of one or more premises. An argument is good if this attempt is successful, i.e., if the (one or more) premises really do justify the conclusion. (Justification simply is good argumentation, good reasoning.) In order for any argument to be a good one, it must satisfy three criteria (i.e., conditions).

For a deductive argument to be good, it must (1) have only true premises; (2) be (deductively) valid; i.e., the premises must indeed necessitate the conclusion (not merely be claimed to do so, as they are in any deductive argument, good or bad); and (3) be noncircular. For suppose the argument is valid: this means that if the (one or more) premises are all true then the conclusion must also be true. That, however, does not yet establish that the conclusion is true. For that, we must know in addition that the premises are in fact true; that's why a good deductive argument must satisfy also criterion (1). If criteria (1) and (2) are both met, then the conclusion must be true.

Yet even a valid deductive argument with only true premises may still be bad, i.e., may still fail to establish (justify) the conclusion. Here's one: 'Obviously Socrates is mortal. [How do I know?] Just because Socrates is mortal.' (Or 'because it's just so.') Now, the premise of the argument is surely true. Also, the argument is valid: if the premise is true, then so must the conclusion be (since the two say the same thing!). Yet, although criteria (1) and (2) have thus been met and hence the conclusion must indeed be true, it plainly has not been established: the argument infers the conclusion from the premise, yet it got the premise by having it restate the conclusion. The argument is circular; it begs the question.

Not all circularity is as blatant as this. A circular argument is one in which the conclusion or a sentence equivalent to it is also (or is also presupposed by) a premise. Note that any argument that is circular is therefore deductive and valid (although of course still bad because of its circularity); for if an argument has the very conclusion or its equivalent (or a sentence presupposing this equivalent) as a premise, then of course the premises necessitate the conclusion (i.e., if the premises are true then so must the conclusion be)!

For an inductive argument to be good, it must (1) have only true premises; (2) be (inductively) valid; i.e., the premises must indeed make the conclusion probable (not merely be claimed to do so, as they are in any inductive argument, good or bad); (3) be unbiased; i.e., the premises must not deliberately or carelessly omit relevant counterevidence, i.e., evidence that goes against the conclusion. Criteria (1) and (2) connect in the same way as they do in deduction. And although inductive arguments cannot be circular, bias resembles circularity inasmuch as deliberate omission of relevant counterevidence also keeps even valid arguments with only true premises from justifying the conclusion.

Back to the Start Page of Werner Pluhar's Experimental Pages