and used technical terms, which he was asked to explain. But the jury listened with absorbed interest, and he kept on until he had answered the question thoroughly.

"As an architect," the coroner asked, when Hart had completed his explanation, "will you state whether, in your judgment, these changes that you have described, especially the substitution of inflammable material for fireproofing and the weakening of the main walls, were sufficient to account for the great loss of life in the fire?"..

66

...

Such alterations as I have indicated tended to weaken the walls, and in other ways to bring the building below the danger limit."

"It was what might be called a fire-trap, then?"

"I did not say that!"

66

Well, Mr. Hart," a member of the jury finally interposed with a question, can you say that the Glenmore as it was built conformed to the building ordinances of the city of Chicago?"

...

"No, the Glenmore violated the ordinances in a number of important particulars."

There was a sudden hush in the room. This point had been established before by different persons who had been examined. Nevertheless, the admission coming from the architect of the illfated building was an important point.1

The degree of convincingness of these three kinds of evidence. It would seem that the convincingness of these three kinds of evidence is somewhat different. What a man admits, much against his will, may evidently be taken for truth if one is sure that the testimony given is really hostile to the interests of the witness. If, however, a student uses a bit of undesigned testimony, he should assure himself that the witness is generally accurate and veracious, and that he is really testifying as undesignedly as he appears to be. Comparison of the three illustrations just given will show that one witness giving negative testimony is not so convincing as one witness giving undesigned testimony or mak

1 Robert Herrick, The Common Lot, pp. 380-383. The Macmillan Company.

ing a hurtful admission. A writer may be sure that his witness giving negative testimony is veracious, but it is even then perfectly possible that there may have been oversight. Negative testimony, therefore, to be convincing in itself, should usually be given by several reliable witnesses:

External and internal tests. Each of the tests of evidence considered thus far questions the soundness, the good faith, or the good judgment of the witness, and sometimes more than one of these. These tests examine the evidence externally. They consider the man who gives the testimony, the conditions under which it was given, whether the evidence as a whole coincides with other testimony known to be true, whether it is self-contradictory; they do not say, "Is there anything faulty in the process of thought, the reasoning that has produced the opinion, or the inference on which it is based?" If they did, they would be not external but internal tests. It may seem, offhand, that there is one exception to this statement, the test of self-contradiction. This, however, does not analyze the logical process involved in the evidence, but simply, looking at the face of the statement, points out that an assertion made in one place is contradicted in another. We need not question the process of thought by which the witness reached the statement; it is enough that he is untrustworthy. These external tests establish the fact of a man's unreliability as a witness; the internal tests discover logical unsoundness in one piece of reasoning irrespective of its source. The external tests are applicable to testimonial evidence; the internal, chiefly to reasoning that large class of inferential evidence that has been classified as circumstantial.

SECTION 2. FALLACIES

Fallacies and their dangers. The examination of evidence for internal weaknesses is a search for fallacies. By a fallacy is meant "any unsound mode of arguing which appears to demand our conviction, and to be decisive of the question in

hand, when in fairness it is not." It is very necessary in argument to guard carefully against such unsound reasoning, for not only may an opponent intentionally try to mislead by unsound methods of reasoning (or, what is far more probable, may use them unawares), but any worker in argumentation, with the best of intentions, may himself unconsciously slip into them. A fallacy is very often extremely hard to detect, for rarely is it self-evident. Generally it is embedded in a mass of other entirely trustworthy material. It is perhaps but part of a sentence in a volume of many pages, yet if it exists it may be fatal to the ultimate convincingness of the argument. The opponents of two noted books, Mr. Kidd's Social Evolution and Dr. Nordau's Degeneration, insist, not so much that most of the conclusions in each book are wrong, as that each book rests for its force on a fundamental fallacy, which, if not noted, makes all the chapters drawn from it easy of acceptance. Evidently, then, it is important to know what are the chief kinds of fallacies, and how to recognize and avoid them.

Imperfect classification of fallacies. In the past repeated attempts have been made to classify definitely the different kinds of fallacies, but experience has shown that a satisfactory hard-and-fast division of them has not yet been found. The divisions overlap, some of the fallacies falling into one or another class as the student looks at them from one or another point of view. The rough division used in the following pages is frankly open to such an objection. It is based, however, primarily on the three sources from which fallacies may arise, and therefore should serve well to put the student of argumentation on his guard. An arguer must convince by means of words. If through carelessness or perversity he juggles with these words and uses them now in one sense and now in another, it is clear that his reasoning must be unsound. Lack of definition is, then, the first source of fallacy. But even if his terms are unequivocal, he may lead his reader astray by stating as a fact what has come from careless or erroneous

observation on his own part or the part of those on whom he relies. Errors of observation thus cause a second class of unsound modes of reasoning. If the observation upon which the evidence is based has been sound and the terms have been carefully used in well-defined meanings, an arguer can hardly be guilty of fallacious work unless he uses these words and facts in a reasoning process which is itself unsound. That is, errors in reasoning furnish the third source of fallacious argument. From one or more of these three causes - lack of definition, errors of observation, and errors in the reasoning process-come, then, unsound modes of thought that invalidate both deductive and inductive argument. As in both these forms of argument the same causes produce vicious reasoning, the two will not generally be considered separately in the following discussion. Nor does the detailed classification under the three headings aim to be exhaustive. The student of argumentation does not need to know how to classify in intricate fashion the unsound reasoning with which he meets. If he clearly recognizes something as fallacious, it matters little to him that two or three names may be given to the errors involved. His purpose is achieved when once he learns how to discover and guard against the common fallacious processes to which he is himself liable and which may confront him in current publications. A classification more intricate than is necessary for this purpose falls outside the scope of this book.

I. FALLACIES ARISING FROM LACK OF DEFINITION The moment a reader begins to consider fallacies, the truth of what was said in Chapter III as to the importance of definition in argumentation is emphasized. A very large class of fallacies arises from an ignorant or a careless use of terms or phrases. The reader saw in Chapter III how careful definition at the outset avoids the confusions, the argumentative errors, intentional or unintentional, that may

result from using words in themselves ambiguous, ignorantly, carelessly, or with unfair purpose. Whenever in an argument words or phrases occur which are ambiguous or are used in an incorrect sense, they may, if unexplained, produce confusion, intentional or unintentional, and lead to "unsound modes of reasoning."

Undefined words with more than one meaning. Another form of this fallacy which arises from the use of an ambiguous word or phrase is to employ the same word in different parts of an argument in two different senses. For instance, a student arguing once on the topic, Should Art Museums be open to the Public on Sunday? wrote ten to a dozen pages of rather rambling matter, reducible to these statements: "The world consists of a cultivated few and a mob. A mob will generally destroy a work of art. Therefore the Museums should be opened only to the cultivated few. Consequently, since the cultivated few have leisure to attend Museums on week days, it is not necessary to open them on Sundays." Here the student used "mob" in the first sentence merely to signify the mass of the uneducated or partially educated, the majority of humanity; in the second he shifted the meaning to "a crowd of overexcited men whose passions have overcome their better judgments." The same danger occurs also when a word is used without explanation that has both an everyday meaning and one, more rarely used, from its derivation. We often hear the saying, "The exception proves the rule," given as proof that the exception to the rule is what shows the rule to be correct perfect nonsense. Here the speaker takes "proves" for our word proves, when in this case it is used in its strictly etymological meaning derived from the Latin probo, to test, and the meaning is, “The exception tests the rule." In all these cases a fallacy is very liable to result when an ambiguous term is left undefined.

Words used as identical because they look alike. Another student wrote a theme to prove that college men should vote