Fallacies are counterfeit or sham arguments. They may fail to be arguments—(1) because their antecedents are false; (2) because the antecedents though true are not arranged dialectically, and do not suggest the right conclusion; (3) because the language is equivocal.
To take the last first. So many things are called by the same name, and so many different names may be applied to the same thing, that if we attempt to argue from words alone, without any personal knowledge of the things or judgments that are in question, we shall certainly make mistakes. The only security against this sort of fallacy is much experience, and the self-denial necessary to relinquish argument and the criticism of arguments, when we have no sufficient knowledge of the data.
The degree of imperfection in observation which should be considered to render the theorem fallacious, is no easy matter to determine. One class of logicians (the Formal) get over the difficulty by declaring that dialectic is not concerned with concrete knowledge at all,[17] but only with its general properties (as conceived by Aristotle), and they have set up as a standard of logical truth the capability of being imagined. A centaur is to them as true a fact as a horse, and they would accept as valid such a theorem as this: 'All centaurs object to be shod with iron; Gryneus is a centaur; therefore we may conclude that he would resist being shod with iron.' No amount of conceivability or formal coherence can make this other than nonsense.
J. S. Mill and his followers go to the opposite extreme. They study all the sciences and endeavour to master their methods of reasoning—which is well; but they do so with the prepossession that there exists some absolute standard of knowledge to fail in attaining which involves fallacy. They thus condemn as false all theorems based on superseded notions of nature and man. Only modern thinkers can argue rationally—the ancients were all and habitually victims of fallacy,—and of the moderns only the few are rational who have mastered the latest theories on every subject. This is the principle of Mill's doctrine on the fallacies of observation; we can see that he regarded all beliefs as fallacious which he had himself outgrown or did not feel a need of. 'Truth' was simply the facts and judgments that happened to suit Mill's mental constitution.
From the Substantial point of view this is an untenable position.
No degree of observation is intrinsically defective if it serve the purpose of intellect, which is to protect the mind. There is no intellectual truth as a thing in itself. The thoughts of a sparrow or a child are as perfect as those of a man, if they afford the necessary defence to the individual's sentiment. As we change our inner mental character, new intellectual ideas have to be acquired and the old are discarded, perhaps completely forgotten. They appear now to be ignorances and fallacies—mal-observation and bad reasoning. The new seem to be so much truer—perhaps infallibly true. All that is illusion. We make another advance, and the thoughts that a week before were as stable as rocks are now cast aside as absurd. Perhaps the belief in the certainty of present judgments is a condition of our making the best use of them; if so they should not be shaken until we are ready to enter on the next stage of knowledge.
It is quite true that one man may know more than another, but the ground on which the more intellectual is generally considered to be superior to the latter is not the right one. He is not better for his intellectual acquirements, but he is better if his mind, being of a finer sort, required a superior intellect to defend it. At bottom, then, the general cause of mal-observation—there are particular causes which interfere with the general rule—is inferiority of sentimental character. We do not see what we do not need to see, and we see imperfectly what is not essential to our well-being. That we should be ignorant or reason badly about what does not concern us is not in itself a defect.
It is inconsistent with these views on the function of intellect to admit that any sort of non-observation or mal-observation can be always and for all alike fallacious. If there are things which we habitually ignore, the presumption is that they do not concern us—that the knowledge they would confer is not essential to our welfare and would be intellectual lumber.
I should therefore abstain from condemning as fallacies theorems drawn in good faith from facts believed to be true, and which serve as motives of conduct. They are sophisms only when the reasoners have not taken ordinary pains to verify their data, or, knowing the antecedents to be false, pretend to believe them for some immoral purpose.