Logical Fallacies

[Preamble | Manifesto | Agenda]

Logic and Logical Fallacies


From the Harvard Handbook for English, 4th edition. Copyright 1947, the President and Fellows of Harvard College. Copyright 1975 by Robert Gorham Davis. [permission to post it to IndiaPolicy has not been obtained but it is hoped that this violation will be condoned in the interest of wider dissemination of knowledge.]

 Expression does not exist apart from thought, and cannot be analyzed or profitably discussed apart from thought. Just as clear and effective organization is essential to good writing, so consistent thinking and coherence of mind underlie consistent writing and coherence of style. The faults and errors which [fall] under the headings of style and structure are closely bound up with orderly thought, as the student can hardly fail to notice. But some direct suggestions on the modes of consistent thinking and of analyzing and criticizing arguments and assertions ought also to prove useful. The following pages accordingly present some notes on logic and common logical fallacies.


 The first requirement for logical discourse is knowing what the words you use actually mean. Words are not like paper money or counters in a game. Except for technical terms in some of the sciences, they do not have a fixed face value. Their meanings are fluid and changing, influenced by many considerations of context and reference, circumstance and association. This is just as true of common words, such as fast as it is of literary terms such as romantic. Moreover, if there is to be communication, words must have approximately the same meaning for the reader that they have for the writer. A speech in an unknown language means nothing to the hearer. When an adult speaks to a small child or an expert to a layman, communication ma y be seriously limited by lack of a mature vocabulary or ignorance of technical terms. Many arguments are meaningless because the speakers are using important words in quite different senses.

Because we learn most words - or guess at them - from contexts in which we first encounter them, our sense of them is of incomplete or wrong. Readers sometimes visualize the Assyrian who comes down like the wolf on the fold as an enormous man dressed in cohorts (some kind of fancy armor, possibly) gleaming in purple gold. "A rift in the lute" suggests vaguely a cracked mandolin. Failure to ascertain the literal meaning of figurative language is a frequent reason for mixed metaphors. We are surprised to find that the "devil" in "the devil to pay" and "the devil and the deep blue sea" is not Old Nick, but part of a ship. Unless terms mean the same thing to both writer and reader, proper understanding is impossible.


The most serious logical difficulties occur with abstract terms. An abstraction is a word which stands for a quality found in a number of different objects or events from which it has been "abstracted" or taken away. We may, for instance, talk of the "whiteness'' of paper or cotton or snow without considering qualities of cold or inflammability or usefulness which these materials happen also to possess. Usually, however, our minds carry over other qualities by association. See, for instance, the chapter called "The Whiteness of the Whale" in Moby Dick. 

In much theoretic discussion the process of abstraction is carried so far that although vague associations and connotations persist, the original objects or events from which the qualities have been abstracted are lost sight of completely. Instead of thinking of words  like sincerity and Americanism as symbols standing for qualities that  have to be abstracted with great care from examples and test cases, we come to think of them as real things in themselves. We assume that Americanism is Americanism just as a bicycle is a bicycle, and that everyone knows what it means. We forget that before the question, "Is Father Coughlin sincere?" can mean anything, we have to agree on the criteria of sincerity.

When we try to define such words and find examples, we discover that almost no one agrees on their meaning. The word church may refer to anything from a building on the corner of Spring Street to the whole tradition of institutionalized Christianity. Germany may mean a geographical section of Europe, a people, a governing group, a cultural tradition, or a military power. Abstractions such as freedom, courage, race, beauty, truth, justice, nature, honor, humanism, democracy, should never be used in a theme unless their meaning is defined or indicated clearly by the context. Freedom for whom? To do what? Under what circumstances? Abstract terms have merely emotional value unless they are strictly defined by asking questions of this kind. The study of a word such as nature in a good unabridged dictionary will show that even the dictionary, indispensable though it is, cannot determine for us the sense in which a word is being used in any given instance. Once the student understands the importance of definition, he will no longer be betrayed into fruitless arguments over such questions as whether free verse is "poetry" or whether you can change ''human nature."


It is a common unfairness in controversy to place what the writer dislikes or opposes in a generally odious category. The humanist dismisses what he dislikes by calling it romantic; the liberal, by calling it fascist; the conservative , by calling it communistic. These terms tell the reader nothing. What is piety to some will be bigotry to others. Non-Catholics would rather be called Protestants then heretics. What is right-thinking except a designation for those who agree with the writer? Labor leaders become outside agitators; industrial organizations, forces of reaction; the Child Labor Amendment, the youth control bill; prison reform, coddling; progressive education, fads and frills. Such terms are intended to block thought by an appeal to prejudice and associative habits. Three steps are necessary before such epithets have real meaning. First, they must be defined; second, it must be shown that the object to which they are applied actually possesses these qualities; third, it must be shown that the possession of such qualities in this particular situation is necessarily undesirable. Unless a person is alert and critical both in choosing and in interpretin g words, he may be alienated from ideas with which he would be in sympathy if he had not been frightened by a mere name.


 Similar to the abuse of abstract terms and epithets is the habit of presenting personal opinions in the guise of universal laws. The student often seems to feel that the broader the terms in which he states an opinion, the more effective he will be. Ordinarily the reverse is true. An enthusiasm for Thomas Wolfe should lead to a specific critical analysis of Wolfe's novels that will enable the writer to explain his enthusiasm to others; it should not be turned into the argument that Wolfe is "the greatest American novelist," particularly if the writer's knowledge of American novelists is somewhat limited. The same questions of who and when and why and under what circumstances which are used to check abstract terms should be applied to generalizations. Consider how contradictory proverbial wisdom is when detached from particular circumstances. "Look before you leap," but "he who hesitates is lost."

 Superlatives and the words right and wrong, true and untrue, never and always must be used with caution in matters of opinion. When a student says flatly that X is true, he often is really saying that he or his family or the author of a book he has just been reading, persons of certain tastes and background and experience, think that X is true. Unless these people are identified and their reasons for thinking so explained, the assertion is worthless. Because many freshmen are taking survey courses in which they read a single work by an author or see an historical event through the eyes of a single historian whose bias they may not be able to measure, they must guard agains t this error.


Assertions of a general nature are frequently open to question because they are based on insufficient evidence. Some persons are quite ready, after meeting one Armenian or reading one medieval romance, to generalize about Armenians and medieval romances. One ought, of course, to examine objectively as many examples as possible before making a generalization, but the number is less important than the representativeness of the examples chosen. The Literary Digest Presidential Poll, sent to hundreds of thousands of people selected from telephone directories, was far less accurate than the Gallup Poll which questioned far fewer voters, but selected them carefully and proportionately from all different social groups. The "typical" college student, as portrayed by moving pictures and cartoons, is very different from the "representative" college student as determined statistically. We cannot let uncontrolled experience do our sampling for us; instances and examples which impress themselves upon our minds do so usually because they are exceptional. In propaganda and arguments extreme cases are customarily treated as if they were characteristic.

If one is permitted arbitrarily to select some examples and ignore others, it is possible to find convincing evidence for almost any theory, no matter how fantastic. The fact that the mind tends naturally to remember those instances which confirm its opinions imposes a duty upon the writer, unless he wishes to encourage prejudice and superstition, to look carefully for exceptions to all generalizations which he is tempted to make. We forget the premonitions which are not followed by disaster and the times when our hunches failed to select the winner in a race. Patent medicine advertisements print the letters of those who survived their cure, and not of those who died during it. All Americans did not gamble on the stock exchange in the twenties, and all Vermonters are not thin-lipped and shrewd. Of course the search for negative examples can be carried too far. Outside of mathematics or the laboratory, few generalizations can be made airtight, and most are not intended to be. But quibbling is so easy that resort to it is very common, and the knowledge that people can and will quibble over generalizations is another reason for making assertions as limited and explicitly conditional as possible.


Illustration, comparison, analogy are most valuable in making an essay clear and interesting. It must not be supposed, however, that they prove anything or have much argumentative weight. The rule that what is true of one thing in one set of circumstances is not necessarily true of another thing in another set of circumstances seems almost too obvious to need stating. Yet constantly nations and businesses are discussed as if they were human beings with human habits and feelings; human bodies are discussed as if they were machines; the universe, as if it were a clock. It is assumed that what held true for seventeenth century New England or the thirteen Atlantic colonies also holds true for an industrial nation of 130,000,000 people. Carlyle dismissed the arguments for representative democracy by saying that if a captain had to take a vote among his crew every time he wanted to do something, he would never get around Cape Horn. This analogy calmly ignores the distinction between the lawmaking and the executive branches of constitutional democracies. Moreover, voters may be considered much more like the stockholders of a merchant line than its hired sailors. Such arguments introduce assumptions in a metaphorical guise in which they are not readily detected or easily criticized. In place of analysis they attempt to identify their position with some familiar symbol which will evoke a predictable, emotional response in the reader. The revival during the 1932 presidential campaign of Lincoln's remark, "Don't swap horses in the middle of the stream," was not merely a picturesque way of saying keep Hoover in the White House. It made a number of assumptions about the nature of depressions and the function of government. This propagandist technique can be seen most clearly in political cartoons.


Often differences in degree are more important than differences in kind. By legal and social standards there is more difference between an habitual drunkard and a man who drinks temperately, than between a temperate drinker and a total abstainer. In fact differences of degree produce what are regarded as differences of kind. At known temperatures ice turns to water and water boils. At an indeterminate point affection becomes love and a man who needs a shave becomes a man with a beard. The fact that no men or systems are perfect makes rejoinders and counter-accusations very easy if differences in degree are ignored. Newspapers in totalitarian states, answering American accusations of brutality and suppression, refer to lynchings and gangsterism here. Before a disinterested judge could evaluate these mutual accusations, he would have to settle the question of the degree to which violent suppression and lynching are respectively prevalent in the countries under consideration. On the other hand, differences in degree may be merely apparent. Lincoln Steffens pointed out that newspapers can create a "crime wave" any time they wish, simply by emphasizing all the minor assaults and thefts commonly ignored or given an inch or two on a back page. The great reported increases in insanity may be due to the fact that in a more urban and institutionalized society cases of insanity more frequently come to the attention of authorities and hence are recorded in statistics.


The most common way of deciding that one thing causes another thing is the simple principle: post hoc, ergo propter hoc, "After this, therefore because of this." Rome fell after the introduction of Christianity; therefore Christianity was responsible for the fall of Rome. Such reasoning illustrates another kind of faulty generalization. But even if one could find ten cases in which a nation "fell" after the introduction of Christianity, it still would not be at all certain that Christianity caused the fall. Day, it has frequently been pointed out, follows night in every observable instance, and yet night cannot be called the cause of day. Usually a combination of causes produces a result. Sitting in a draught may cause a cold, but only given a certain physical condition in the person sitting there. In such instances one may distinguish between necessary and sufficient conditions. Air is a necessary condition for the maintenance of plant life, but air alone is not sufficient to produce plant life. And often different causes at different times may produce the same result. This relation is known as plurality of causes. If, after sitting in a stuffy theatre on Monday, and then again after eating in a stuffy restaurant on Thursday, a man suffered from headaches, he might say, generalizing, that bad air gave him headaches. But actually the headache on Monday may have been caused by eye-strain and on Thursday by indigestion. To isolate the causative factor it is necessary that all other conditions be precisely the same. Such isolation is possible, except in very simple instances, only in the laboratory or with scientific methods. If a picture falls from the wall every time a truck passes, we can quite certainly say that the truck's passing is the c ause. But with anything as complex and conditional as a nation's economy or human character, the determination of cause is not easy or certain. A psychiatrist often sees a patient for an hour daily for a year or more before he feels that he understands his psychosis.


Ordinarily when we speak of cause we mean the proximate or immediate cause. The plants were killed by frost; we had indigestion from eating lobster salad. But any single cause is one in an unbroken series. When a man is murdered, is his death caused by the loss of blood from the wound, or by the firing of the pistol, or by the malice aforethought of the murderer? Was the World War "caused" by the assassination at Sarajevo? Were the Navigation Acts or the ideas of John Locke more important in "causing" the American Revolution? A complete statement of cause would comprise the sum total of the conditions which preceded an event, conditions stretching back indefinitely into the past. Historical events are so interrelated that the isolation of a causative s equence is dependent chiefly on the particular preoccupations of the historian. An economic determinist can "ex plain" history entirely in terms of economic developments; an idealist, entirely in terms of the development of ideas. 


The formal syllogism of the type,

All men are mortal

John is a man

Therefore John is mortal,

is not so highly regarded today as in some earlier periods. It merely fixes an individual as a member of a class, and then assumes that the individual has the given characteristics of the class. Once we have decided who John is, and what "ma n" and "mortal" mean, and have canvassed all men, including John, to make sure that they are mortal, the conclusion naturally follows. It can be seen that the chief difficulties arise in trying to establish acceptable premises. Faults in the premises are known as "material" fallacies, and are usually more serious than the "formal" fallacies, which are logical defects in drawing a conclusion from the premises. But although directly syllogistic reasoning is not much practiced, buried syllogisms can be found in all argument, and it is often a useful clarification to outline your own or another writer's essay in syllogistic form. The two most frequent defects in the syllogism itself are the undistributed and the ambiguous middle. The middle term is the one th at appears in each of the premises and not in the conclusion. In the syllogism,

All good citizens vote

John votes

Therefore John is a good citizen,

the middle term is not "good citizens,)' but "votes." Even though it were true that all good citizens vote, nothing prevents bad citizens from voting also, and John may be one of the bad citizens. To distribute the middle term "votes" one might say (but only if that is what one meant,

All voters are good citizens

John is a voter

Therefore John is a good citizen.

The ambiguous middle term is even more common. It represents a problem in definition, while the undistributed middle is a problem in generalization. All acts which benefit others are virtuous, losing money at poker benefits others, therefore losing at poker is a virtuous act. Here the middle term "act which benefits others" is obviously used very loosely and ambiguously.


This phrase, meaning "it does not follow," is used to characterize is the kind of humor found in pictures in which the Marx Brothers per form. It is an amusing illogicality because it usually expresses, beneath its apparent incongrui ty, an imaginative, associative, or personal truth. "My ancestors came over on the Mayflower; therefore I am naturally opposed to labor unions.'' It is not logically necessary that those whose ancestors came over on the Mayflower should be opposed to unions; but it may happen to be true as a personal fact in a given case. Contemporary psychologists have effectively shown us that there is often such a wide difference between the true and the purported reasons for an attitude t hat, in rationalizing our behavior, we are often quite unconscious of the motives that actually influence us. A fanatical antivivisectionist, for instance, may have temperamental impulses toward cruelty which he is suppressing and compensating for by a reasoned opposition to any kind of permitted suffering. We may expect, then, to come upon many conclusions which are psychologically interesting in themselves, but have nothing to do with the given premises.


This means, in idiomatic English, "arguing off the point," or ignoring the question at issue. A man trying to show that monarchy is the best form of government for the British Empire may devote most of his attention to the character of George V and the affection his people felt for him. In ordinary conversational argument it is almost impossible for disputants to keep to the point. Constantly turning up are tempting side-issues through which one can discomfit an opponent or force him to irrelevant admissions that seem to weaken his case.


The first of these terms means to assume in the premises what you are pretending to prove in the course of your argument. The function of logic is to demonstrate that because one thing or group of things is true, another must be true as a consequence. But in begging the  question you simply say in varying language that what is assumed to be true is assumed to be true. An argument which asserts that we shall enjoy immortality because we have souls which are immaterial and  indestructible establishes nothing, because the idea of immortality is already contained in the assumption about the soul. It is the premise which needs to be demonstrated, not the conclusion. Arguing in a circle is another form of th is fallacy. It proves the premise by the conclusion and the conclusion by the premise. The conscience forbids an act because it is wrong; the act is wrong because the conscience forbids it. 


It is very difficult for men to be persuaded by reason when their interest or prestige is at stake. If one wishes to preach the significance of physiognomy, it is well to choose a hearer with a high forehead and a determined jaw. The arguments in favor of repealing the protective tariff on corn or wheat in England were more readily entertained by manufacturers than by landowners. The cotton manufacturers in New England who were doing a profitable trade with the South were the last to be moved b y descriptions of the evils of slavery. Because interest and desire are so deeply seated in human nature, arguments are frequently mingled with attempts to appeal to emotion, arouse fear, play upon pride, attack the characters of proponents of an opposite view, show that their practice is inconsistent with their principles; all matters which have, strictly speaking, nothing to do with the truth or falsity, the general desirability or undesirability, of some particular measure. If men are desperate enough they will listen to arguments proper only to an insane asylum but which seem to promise them relief.

After reading these suggestions, which are largely negative, the student may feel that any original assertion he can make will probably contain one or several logical faults. This assumption is not true. Even if it were, we know from reading newspapers and magazines that worldly fame is not dimmed by the constant and, one suspects, conscious practice of illogicality. But generalizations are not made only by charlatans and sophists. Intelligent and scrupulous writers also have a great many fresh and pro vocative observations and conclusions to express and are expressing them influentially. What is intelligence but the ability to see the connection between things, to discern causes, to relate the particular to the general, to define and discriminate and c ompare? Any man who thinks and feels and observes closely will not want for something to express.

And in his expression a proponent will find that a due regard for logic does not limit but rather increases the force of his argument. When statements are not trite, they are usually controversial. Men arrive at truth dialectically; error is weeded out in the course of discussion, argument, attack, and counterattack. Not only can a writer who understands logic show the weaknesses of arguments he disagrees with, but also, by anticipating the kind of attack likely to be made on his own ideas, he can so a rrange them, properly modified with qualifications and exceptions, that the anticipated attack is made much less effective. Thus, fortunately, we do not have to depend on the spirit of fairness and love of truth to lead men to logic; it has the strong sup port of argumentative necessity and of the universal desire to make ideas prevail.

Return to the main page