Chapter Seven: The Fallacies


I.     Introduction

An efficient critical thinker is capable of acting and reacting quickly and effortlessly in circumstances where criticism is required. In an argumentative discourse, one may have the luxury of time, but more often than not, one must make do in a moment or two. One way to cope with resource limitations and increase efficiency is to know some bad arguments on sight, eliminating the need for sustained step-by-step analysis. Recall that good form does not a good argument make, nor do good form and content; a good argument is the whole package of form, content, and context. However, bad form does make a bad argument, as do bad content and bad contextual fit. While you cannot approve a good argument without going through each stage of the process described in the preceding chapter, you can suspect a bad argument at any stage where a flaw becomes apparent. Thus, if we can enable students to spot such flaws, we can enhance their ability to think critically in a wide range of circumstances, which should be one of our goals as educators.

When one makes an argument that exhibits such a flaw, one can be accused of committing a fallacy. This is usually the doing of the arguer, but if one hears a fallacious argument and allows it to influence one's belief system, one is also guilty of committing a fallacy. In this chapter, I provide a partial catalog of these argumentative fallacies. I begin by discussing the role of bad reasoning in critical thinking, and then discuss the plethora of conceptual frameworks available for classifying the fallacies. For our purposes, it will be useful to classify them in a way that lines up with the stages of analysis discussed in Chapter Six. Thus, I supply a threefold classification of the fallacies: formal fallacies, subject matter fallacies, and contextual fallacies. I close by warning against too liberal an application of fallacy-based critical thinking. Once you've learned to identify these flaws, you might become inclined to reject any argument that bears a resemblance to them, and this would be a mistake. In critical thinking, as in the justice system, it is a serious mistake to convict the innocent, and an infatuation with the fallacies can cause such mistakes.

II.     Taking the Bad with the Good

Most of us are guilty from time to time of producing or falling prey to weak arguments. It may be that we just don't think--everyone has their moments of weakness, after all--or that we fail to observe something important. On the other hand, it may be that we are taken in by an argument that appears compelling, but whose beauty is only skin deep. Whatever the reason, it's a fact that we can't avoid. Nevertheless, we can work to minimize the damage done by us and to us by weak arguments. To this end, we must learn to spot the signs of weakness in an argument, signs such as disconnection between premises and conclusion, false premises, or a failure of fit between argument and context. Familiarity with these signs will enable us to identify and avoid arguments that have no business influencing our beliefs and actions, and do so in a relatively quick and easy manner. The signs indicate argument distress, and they are often grounded in characteristics of arguments that tend to vitiate arguments wherever they are found. These characteristics are what we call fallacies. Knowledge of these all too common characteristics can increase response time to weak arguments significantly, and if their detection and detour become second nature, then one need no longer fear the arguments they weaken.

Before going forward with our introduction to the fallacies, a couple of additional words about weak arguments are in order. First, we could call every argumentative weakness a fallacy, perhaps even giving them all fancy Latin names, but this would defeat the purpose of focusing on the fallacies in the first place. Knowledge of the fallacies enhances critical thinking ability in large part because these weaknesses are not uncommon. If they were numerous or rare, the knowledge would just increase the search effort required to respond critically when confronted by a weak argument--if there were many different fallacies, it could take extra time to determine which of the fallacies, if any, is instantiated in a given case, and if they were rare, knowledge of them might be used so infrequently as to lose value. As it is, there are a few noteworthy flaws that weaken arguments, flaws that are keyed to the elements of a discourse episode or the specific context of that episode. For instance, it may turn out that an argumentative response to a challenge uses a term that receives a somewhat idiosyncratic interpretation by the one who issued the challenge. In this case, the argument will be weak because of a failure of communication that may never be repeated again anywhere in the universe. Again, we could call this a fallacy and add it to our list, but what would be the point?

Second, weak arguments are weak from a particular perspective, viz., the perspective of rationality. Strong arguments compel one to adopt their conclusions on pain of irrationality, while weak arguments do not. It is the underlying desire to be rational in belief and action that motivates people to think critically; if you are disinclined to value rationality, your response to arguments will not be shaped by rational pressures and you will not be favorably disposed to stronger arguments just because they are rationally compelling. It is not uncommon, however, that arguments which are weak from this perspective are strong from another. For example, flip on the television to a network for 15 minutes and you will undoubtedly see an advertisement make an argument that is weak as weak can be but is effective nevertheless. (Consider: what was up with that Taco Bell chihuahua? Were we supposed to buy chalupas because of that funny little dog?) Ads can make compelling arguments that are strong from the perspective of rationality, but they often don't, catering to that in us which is non-rational, such as our emotions or perhaps our overall aesthetic. In doing so, however, they do their job, insidiously convincing us that we need to buy Nikes because otherwise we won't be as fast or as inestimably cool as Michael Johnson. From the perspective of reason these are weak, but from the perspective of marketing they are strong. Further, it is not the case that reason always leads us--we may respond to an add with our heart and not our head, choosing to embrace the pitch because it meshes with our conception of what is beautiful or desirable. (Note that in such a case, there would be a good argument lurking in the vicinity of your interaction with the ad--viz., that if you wish to enjoy what you desire and find beautiful, you need to buy product A, and since you do, you buy product A--even though the ad itself may not make this argument.) This may not be rational, and perhaps it may even be irrational to some extent, but perhaps a bit of irrationality, like hot peppers, can enhance life's flavor.

III.     Framing the Fallacies

Fallacious reasoning, like all turkeys, can be carved up in different ways. In this section, I briefly describe several different ways of classifying the fallacies before introducing the framework that will structure the remaining discussion in this chapter.

A fallacious argument is an argument rendered weak by some flaw. It would be impossible to supply an exhaustive catalogue of these flaws, but some of them infiltrate argumentative discourse regularly enough to earn their own name. We introduce and illustrate a sampling of these fallacies below. Since our goal is to sensitize people to these flaws so that they can spot and avoid them in the give and take of daily life, it is important to introduce them systematically, paying close attention to relationships that bind them together into useful groups. Theorists offer different ways to do this that are keyed to different relationships.

One popular way to organize the fallacies is with elements recovered from analysis of a typical contribution made by an arguer to an argument episode. Call this the Episodic Framework. Three central elements emerge from even the most cursory inspection of an episode in which an argument is advanced. These are (a) the arguer, (b) the form of the argument, and (c) the content of the argument. When you make an argument, you typically do so to establish a conclusion about some issue, and those who respond to you do so as well. Argumentative flaws can be grounded in each of these, and so we can use them to aid in the classification of the fallacies. What results is this tri-partite classification:

  • Psychological Fallacies: these have the arguer as their source.

  • Logical Fallacies: these flaws are grounded in the structure of the argument.

  • Material Fallacies: these are fallacies that concern the content, or subject matter, of an argument.

A second framework privileges argument form. There are countless many people who argue, and even more topics, but argument forms are relatively few by contrast and so supply a relatively stable foundation for the classification of fallacies. Call this the Structural Framework. On this approach, there are two broad categories into which fallacies fit:

  • Formal Fallacies: these are argumentative weaknesses attributable to formal characteristics of an argument, i.e., characteristics that relate to how truth is channeled from premises to conclusion.

  • Informal Fallacies: if it isn't a formal fallacy, then it is one of these.

A third framework arises from the observation that good arguments typically have certain virtues. Among other things, they tend to be clear, relevant, and substantive. Fallacious arguments often offend one of these, being either unclear, irrelevant, or vacuous. Call this the Virtue Framework. (See Fogelin & Sinnott-Armstrong for a discussion of fallacies framed in this way.) For any virtue we recognize, offense against it would qualify as a type of fallacy. Among the fallacy types included in this framework, we would find these:

  • Fallacies of Clarity: these are exemplified by arguments that do not make their form or their content clear, perhaps because of vagueness or ambiguity.

  • Fallacies of Relevance: arguments instantiate these when they supply reasons for their conclusions that are not relevant to the claims made by those conclusions.

  • Fallacies of Vacuity: when an argument purports to supply a reason for a conclusion but fails to do so, perhaps because it assumes what it seeks to prove, then it exhibits this sort of fallacy.

While each of these frameworks has value, we will use a fourth framework, what I will call the Analytic Framework. Weak arguments emerge as such in the course of analysis. Their flaws are revealed either when light is cast on their form, their subject matter, or the context in which they are advanced. Given this, we can classify the fallacies according to when they typically make their appearance in the course of analysis. This correlation between stages of analysis and the fallacies leaves us with three categories:

  • Fallacies of Form: these fallacies emerge in the course of formal analysis.

  • Fallacies of Subject Matter: these emerge in the course of subject matter analysis.

  • Fallacies of Context: these are identified when one assesses the fit between an argument and its surrounding context.

The primary reason for selecting this framework is the conceptual connection between it and the approach advocated in the previous chapter, but this is not the only reason. For one thing, it accommodates each of the preceding three frameworks. The categories of the Episodic Framework fit rather neatly, with the logical and material fallacies corresponding to the fallacies of form and content, respectively, and the psychological fallacies forming one part of the fallacies of context. The formal fallacies of the Structural Framework correspond to fallacies of form, naturally, and the informal fallacies comprise fallacies of content and context. There isn't as neat a mapping from the Virtue Framework to the Analytic Framework, but the argumentative vices that underwrite the classification in the former are one and all revealed during analysis and so will find a home in one of the three categories of the Analytic Framework. The second additional reason for preferring this framework is that it promises to be exhaustive in a relevant and substantive way. Of the three approaches considered above, only the Structural Framework is clearly exhaustive--there will be fallacies of context left out of the first, and fallacies of all types left out of the third, failing an exhaustive enumeration of argumentative virtues, an enumeration that would be very difficult to do in a principled fashion. While the comprehensive nature of the classification in the Structural Framework is principled, it isn't clear that it is substantive, since it isn't clear that the informal fallacies form a natural class; instead, it looks more like a grab bag category, a fact that would undermine the ability of this framework to support the goal of quick detection and detour. Thus, there is reason to be sanguine about the quality of this framework as we charge ahead into the dense and frightening jungle of bad reasoning.

Before getting into the three categories, it should be noted that classification of some of the fallacies is tricky. In particular, the method is the source of some difficulty. This framework advocates classifying fallacies on the basis of when they make their appearance in the process of analysis, but this assumes that the stages of analysis are distinct so that a determinate decision can be made. Unfortunately, they are not so distinct. This point can be made from two perspectives. First, it is often unclear where one is in the process of analysis. For example, it isn't always clear where form stops and content begins. This is especially true with non-deductive arguments, since assessment of their non-deductive strength often depends on an appreciation of the claims involved. In addition, it isn't always clear where the argument stops and the context begins, rendering it difficult to determine what should count as a fallacy of context. Second, even if one is clear on where they are in the process, there is often on-going interaction between the levels of analysis. For instance, there is often a feedback loop of the following sort between formal and subject matter analysis: form is determined provisionally on the basis of the content of the claims, only to be reassessed later after a more thorough subject matter analysis. Thus, parallel processing of this sort, combined with vagueness of the levels, can make it very difficult to know exactly where a given argumentative flaw makes its appearance. Nevertheless, we will charge ahead and apply the method, classifying a number of fallacies below. I will make every effort to annotate those whose classification is especially uncertain.

IV.     Fallacies of Form

Fallacies of form are argumentative flaws typically discovered during formal analysis. These are flaws in the inferential structure of an argument, i.e., the relationships among the constituent propositions that convey truth from the premises to the conclusion. Because of this, fallacies undermine the support offered by reasons for conclusions. Since these are flaws in the structure of arguments and argument structure is repeatable, examination of it requires abstraction from specific details of a given argument. At this abstract level, the focus is on the structure of constituent propositions and the relations between propositions, considered structurally. The fallacies I identify in this section are argumentative moves that generally vitiate support by breaking or weakening the connections between constituent propositions. (Compare: recalling the canal analogy, these fallacies would be analogous to breaking the channel so that water couldn't flow from point A to point B.)

In this section, I list some of the fallacies of form, describing their character and illustrating them. This is not a complete list, but these are certainly some of the more common types. I organize them according to whether they apply primarily to deductive arguments, non-deductive arguments, or both.

Specifically Deductive Fallacies:

  • Affirming the Consequent: As we observed in Chapter 6, modus ponens is a valid inference pattern. This pattern can be symbolized as follows: If A, then B; A; therefore, B. A superficially similar pattern results by switching the second occurrences of 'A' and 'B' around, yielding: If A, then B; B; therefore, A. This, however, is an invalid move known as affirming the consequent. The second assertion in this pattern affirms B, which is the consequent in the conditional contained in the first assertion. This fallacy is grounded in the conditional, and it qualifies as a fallacy because it does not respect the way in which the conditional channels truth. That is, true conditionals guarantee the truth of the consequent given the truth of the antecedent, but this guarantee does not go in the opposite direction. (Put another way: antecedents are sufficient conditions on consequents, but consequents are not sufficient conditions on antecedents.) Thus, given a true conditional and a true consequent, the antecedent might turn out to be false, a fact that establishes this pattern as invalid. Consider: "If I'm in my office, my lights are on." If this is true and you see my lights on, you cannot infer that I'm in there, since it could be the janitor who has flipped on the lights. (Modus ponens, by contrast, could be called affirming the antecedent.)

Example: If Herman is deathly ill, then he'll be at home. Hey, Herman is at home. Therefore, he must be deathly ill. (This said of a guy who likes to stay home.)

  • Denying the Antecedent: We also saw in Chapter 6 that modus tollens is a valid inference pattern. This pattern can be symbolized as follows: If A, then B; not B; therefore, not A. A superficially similar pattern results by switching the second occurrences of 'B' and 'A' around, yielding: If A, then B; not A; therefore, not B. This, however, is an invalid move known as denying the antecedent. The second assertion in this pattern denies A, which is the antecedent in the conditional contained in the first assertion. As with affirming the consequent, this fallacy is grounded in the fact that the pattern fails to respect the logic of the conditional. True conditionals guarantee the falsity of the antecedent given the falsity of the consequent, because the consequent is a necessary condition on the antecedent. However, the antecedent is not a necessary condition on the consequent, and so this pattern does not go in the opposite direction. Thus, given a true conditional and a false antecedent, the consequent might turn out to be false, a fact that establishes this pattern as invalid. Return once again to my well-lit office. If the conditional, "If I'm in my office, then my lights are on" is true and you see me out on the town, you cannot infer that my lights are off, since the janitor could be back at work in my office. (Modus tollens, by contrast, could be called denying the consequent.)

Example: If I get a "B" in this class, then there is justice in the universe after all. What?!?! I got a "C"? There is no justice!

  • Fallacy of Division: One important aspect of deductive reasoning is that it makes content explicit that is implicit in the premises. We can think of the premises as constituting a whole and the conclusion of the argument qualifies as a part of this whole. Deductive reasoning just brings out something that is already contained in the premises. While whole-to-part reasoning of this sort is ineluctable when conducted properly, it can be fallacious when it isn't. One example of a fallacious form of this type of reasoning is known as the fallacy of division. Commission of this fallacy occurs when a property of the whole is illicitly taken to belong to each of the parts. Failure to appreciate the fact that while the whole in question may be divisible into parts, all of its properties need not be so divisible, and even those that are divisible need not line up with the natural whole-to-part division. In fact, this sort of inference only goes through when the property in question is inherited by the whole from all of its parts. Therefore, concluding that a part of the whole has a given property of the whole may well be fallacious unless it is clear that the property in question is inherited by the whole from all of its parts.

The current edition of the United States Supreme Court is conservative, as its opinion record attests. Therefore, each of the judges must be conservative. (Being a conservative court is not a property inherited by the whole from all of its parts.)

Specifically Non-deductive Fallacies:

  • Post hoc, ergo proptor hoc: This is a fallacy associated with abductive and confirmation arguments, depending on whether it is offered by way of explanation or prediction. Literally, "After this, therefore because of this". This fallacy is committed by those who take temporal conjunction to confirm causal connection. Temporal conjunction can be a sign of causal connection, to be sure, but it isn't always or perhaps even often such a sign. It is often coincidence that two events are temporally conjoined and not a sign of underlying causal connection. Circumspect reasoning requires assessing whether there are additional reasons for asserting the existence of causal linkage beyond mere temporal sequence. Non-deductive argumentation intended to confirm a causal claim is generally quite weak when supported by only one consideration; the more considerations, the stronger the argument.

"I received a visit from a strange man the other day, and then everything started going wrong. The phone stopped working, the car broke down, the Post Office lost a letter headed my way. I wonder who that man was and why he's doing this to me!" (Conspiracy theorists frequently wield this fallacy.)

  • Hasty Generalization: This is an example of a fallacy associated with inductive arguments. As we have seen, the more varied data supporting a generalization, the stronger that generalization will be. One commits the fallacy of hasty generalization when one generalizes too quickly from a paucity of data. What counts as "too quickly" will vary considerably from context to context, depending on the type of data, the purpose behind the generalization, and nature of the generalization, among other things. Also, note that there are two dimensions along which to evaluate data, viz., quantity and variety. One might have gathered an enormous number of data points that represent a very small part of the range in which these points could be gathered, in which case generalization might still be hasty even though the points could fill a computer.

"I had a bad experience with an Ethics class once--I didn't really like the teacher all that much, it was pretty early in the morning, and the material was tough sledding. As a result, I have resolved never to take another philosophy class. They're bad news, man."

  • Faulty Causal Generalization: As this fallacy occurs in connection with the move from particular observations to a causal generalization, it vitiates abductive arguments. It is similar in character to Post hoc, ergo propter hoc, although Post hoc arguments begin with the observation of temporal conjunction whereas faulty causal generalization can begin in other ways as well, such as with contiguity in space. In fact, this fallacy could be seen as the genus under which Post hoc is one species. As with that species, this fallacy is typically committed by one who moves too quickly from the appearance of causal connection to assertion of it. Once again, circumspect reasoning requires gathering as much evidence as possible, given the circumstances, before moving to a causal generalization. As with hasty generalization, what is possible will vary from circumstance to circumstance--what is a fallacious generalization in one context might be acceptable in another where resources are more limited.

"I've got a headache, and there are linden trees all over the place around here. They must be causing it--I've got to get away from these trees."

  • Faulty Analogy: This fallacy undermines argument by analogy. As arguments by analogy are complex, the fallacy of faulty analogy can take a variety of different forms. The characteristic feature of all forms is that the connections asserted between story and issue (see discussion of this form of reasoning in Chapter 6) do not support inference to the desired conclusion. A common form of this fallacy involves assertion of an analogy when only the most tenuous of connections bind story to issue. It is not uncommon to become impressed by a connection or two and then see an analogy where there really isn't one. (After all, everything is like everything else, at least in some respect or other, so mere similarity along one or two dimensions does not a strong analogy make.) If this analogy is then made to do argumentative work, voila! You have a faulty analogy. This flaw is revealed by close inspection of the relevant features of the story and issue respectively.

"My brother used to have gerbils, and they bit me whenever they got the opportunity. Hamsters are a lot like gerbils, so they will bite me too. Get them away!"

  • Fallacy of Composition: A central feature of non-deductive reasoning is that it generates conclusions that go beyond the premises, in the sense that the information contained in the conclusion is not contained in the premises. Like the type of reasoning described above in connection with the fallacy of division, this fallacy arises in connection with part/whole relations. In particular, one form of non-deductive reasoning involves moving from reasons that make claims about parts to a conclusion that makes a claim about the whole comprising them. One example of a fallacious form of this type of reasoning is known as the fallacy of composition. This fallacy arises when a property of the parts is illicitly taken to belong to the whole. Failure to appreciate the fact that while the parts together constitute the whole, the properties of each part need not also be properties of the whole, even if all of the parts have them. In fact, this sort of inference only goes through when the property in question is inherited by the whole from all of the parts mentioned in the reasons. Therefore, concluding that a whole has a given property that is also had by its parts may be fallacious unless it is clear that the property in question is inherited by the whole from all of the relevant parts.

Each of the players that the basketball coach has assembled has unbelievable game, and so the team they constitute will be truly outstanding.

Equal Opportunity Fallacies:

  • Non sequitur: A non sequitur is an argument in which the conclusion does not follow from the premises. Thus, all the fallacies we have so far named could be regarded as types of non sequiturs, where this title is applicable to any fallacy of form that undermines the connection between premises and conclusions. Any invalid deductive argument is, strictly speaking, a non sequitur, given that in such an argument the premises do not force the conclusion to be true and so the conclusion does not follow in the relevant sense. Inductive arguments can be non sequiturs as well, although these are more difficult to spot since one cannot look only at the formal structure to detect the flaw as one can with deductive arguments. In an inductive argument, one must examine the structure in conjunction with the world, although there are moves that often signal problems, such as hasty generalization. Such an examination might yield, say, a conclusion that fails to follow because it is not relevantly related to the premises. For example, one might observe the long ears of a pair of donkeys and fallaciously conclude that there are horses with long ears.

Example: "You're from Smith Center? Then you must know Bill Blankenship."
  • Fallacy of the Irrelevant Conclusion: A good argument is one in which the reasons adduced support the truth of the conclusion by revealing why the conclusion is true in the context where they are true. In the cases that typically matter, this requires that the conclusion be about the same topic as the reasons. (This is not always true, however, since deductive arguments are formally strong if either the premises cannot all be true at once or the conclusion can never be false, in which case there is no connection required for validity. Such a deductive argument is not structurally weak, but it is substantively weak and is revealed as such under the next category of fallacies.) If you present a set of reasons that supply good support for one conclusion and then draw another, you have committed the fallacy of the irrelevant conclusion. This is also known as ignoratio elenchi. (Note: if one is given a conclusion to defend and supplies a set of reasons that provide support for another conclusion, one commits the closely related fallacy of the irrelevant reasons.) This is the fallacy committed by the prosecuting attorney who tries to convince a jury of the guilt of the defendant by truthfully describing the heinous nature of the crime--the description of the crime may be true enough, but that alone does not implicate any one person in most cases. With arguments such as this, the conclusion is irrelevant because it makes a claim that is not motivated by the reasons, and so one has no reason to believe in it given the truth of the premises. Sometimes the irrelevance is minor and the leap is to a conclusion that might be attained if the train of thought is allowed to chug on for a few more steps, but in other cases it is major and represents a leap to an entirely different and unconnected track. As with the other fallacies, commission of this fallacy might be intentional, in a situation where there isn't sufficient reason and the arguer knows it, but it might also be an unintentional expression of biases that incline the arguer to see a connection that is not motivated by the reasons. This type of fallacy is formal, given that the proper structural relationships among the constituent propositions do not obtain, but it is also a subject matter fallacy, since the necessary agreement in subject matter among the propositions does not obtain. Thus, this flaw is represented at both levels of analysis.

Examples: (1) "Look, the elementary school is dilapidated, lacks air conditioning, and is asked to accommodate more students than it can hold. It is obvious that we must pass this bond measure." (There may be a good argument that begins with these reasons and ends with this conclusion, but this one isn't it.) (2) "Many of those who originally settled this great country were driven from their homes by religious persecution, and religion became the social backbone in the new world. In fact, religious practice and religious freedom have been an important part of the American culture. It is for this reason that we must allow prayer in the public schools."
  • Equivocation: The structure of an argument is grounded in the propositions that constitute it. These propositions have an internal logical character that depends on their terms, predicates, connectives, etc. In addition, they should be related to one another, e.g., they concern the same topic. Such relations are crucial to the flow of the argument, i.e., to the flow of truth from reasons to conclusion. Relations of this sort are often anchored in the appearance of the same words or phrases in the sentences used to express the propositions. Thus, one might speak of a "bank" in several of the sentences that constitute the argument, and thereby guarantee that the topic be common to all parts of the argument. However, one restriction on this is that the term must be used in the same sense in all of its appearances if it is to secure such relationships. Thus, if in one sentence 'bank' is used to talk about a lending institution and in another it is used to talk about the sandy side of a river, the two sentences are not in fact related, despite appearances. In some arguments, different employments of the same word will be signaled, and so while the multiple use may not be felicitous, it does not undermine the structure of the argument. With other arguments, however, different uses will not be signaled; in fact, the arguer may have no idea that she is using the term differently. In these cases, the same word will apparently anchor relations between propositions that do not in fact exist. If the argument depends on those phantom relationships, then it has fallen prey to the fallacy of equivocation. This fallacy is committed whenever an argument depends on the univocal use of a word or phrase that is in fact used in more than one way in the sentences that constitute the argument. One equivocates when one toggles back and forth between different meanings of an ambiguous word or phrase, and if this is done in the context of an argument, the argument could suffer as a result. As with the previous fallacy, this is one that could also be classified as a subject matter fallacy, given that the equivocation uncouples the connection in subject matter that must obtain among the constituent propositions.

(1) "The freedom/determinism debate concerns whether we are free to choose and act, with determinists arguing that we aren't free. I really don't get this debate, though. After all, we Americans enjoy freedom, as are many people around the world. Thus, determinists are obviously wrong." (The freedom/determinism debate is a metaphysical debate that often boils down to questions of causality, and so unfolds at a different logical level than the topic of political freedom. Thus, 'freedom' is used equivocally in this argument.) (2) "Tim McVeigh said he was sorry that the people in the Murrah Building had to die, and if he said he was sorry, he apologized. Therefore, McVeigh apologized for his crime." (This is not true--he didn't apologize. The first occurrence of 'sorry' relates to McVeigh's feelings, and one can feel bad about something without apologizing for it. The second occurrence is related to the sense in which 'sorry' is used in apologies. Thus, this argument uses 'sorry' equivocally.)

V.     Fallacies of Subject Matter

Fallacies of subject matter emerge when we are engaged in subject matter analysis. We notice the presence of these flaws when we are attending to the relation between the topics of reasons and conclusion, as well as to the truth of the reasons. Arguments that attempt to pass off unrelated propositions as related may conceal these, as do arguments that rely on propositions that are quite likely false due to their nature. Formal analysis typically misses these because they are not grounded in the abstract nature of propositions and propositional relations; rather, they are grounded in the substantive connections between specific propositions and the world, or among propositions on the basis of these worldly connections. (This is not true of all fallacies of form, since as we saw with irrelevant conclusion and equivocation in the previous section, the flaws that underwrite them can also underwrite associated fallacies of subject matter.)

As before, I devote the rest of this section to a list of common fallacies of subject matter. I describe the profile of each fallacy and then illustrate it.

  • Fallacy of Emotive Language: An argument commands attention because it is an attempt to win you over to a certain way of thinking. Arguments that rationally compel are those that convey you from premises to conclusion as if by luxury sports car--fast and smooth. Adoption of the reasons literally forces one to adopt the conclusion on pain of irrationality. However, there are tricks of the sophistical trade that can be used to win an audience over to the conclusion in the absence of rational compulsion. One of these is to spice the argument with emotive words, i.e., words that inflame the passions. Such words can distract an audience or perhaps shame them into buying a premise that is otherwise unacceptable. Use of emotive language is not always a bad sign, of course--the argument might be persuasive on its own, even if it weren't stated in such a highly charged way. However, use of this sort of language can tilt the balance in favor of the conclusion by short circuiting one's rational capacity, making possible an unmotivated move to the conclusion.

Example: "Think about abortion for a moment. The procedures--all of them--are acts of violence done to innocent babies. Surely this practice must be stopped." (Here use of 'acts of violence' and 'innocent babies' cloud the issue, making it seem like one must accept the premise and therefore the conclusion; however, there are other ways of describing the case that do not motivate this, and so it would take a further argument to support use of such emotive language in this case.)

  • Ad Hominem -- Against the Argument: An argument ad hominem is an argument against the person. Often, this is intended to undermine another argument or position by challenging the arguer in some way. This can be done in two ways. First, the ad hominem might be aimed directly at the person and have nothing to do with the particular argument in question. In such a case, it could be intended to challenge the arguer's motive or right to speak. Second, the ad hominem might be aimed at the person in a way that purports to be relevant to the argument itself. For instance, a challenge to an arguer's credibility could be relevant to the truth of the claims advanced in their argument. The first type of ad hominem is against the arguer and not the argument; since the character of the arguer is part of the context in which an argument is evaluated, it belongs in a discussion of argument context and so will be treated in the next section. The second type concerns the argument, and more specifically, the truth of the claims that constitute the particular argument at issue, and so belongs in this section. As should be plain, this type of ad hominem need not be fallacious. In some cases, an ad hominem can undermine an argument by calling into question the truth of a premise. Consider the following argument, advanced by a jailhouse informant: "Mr. Smith confessed to me in our cell that he kidnapped that man, and he was sincere. He must have done it." If it turns out that the informant has perjured himself in the past, this would constitute part of an effective ad hominem challenge to the argument. After all, we only have his word for the premise, and his word isn't exactly golden. Nevertheless, many instances of this type of argument are fallacious, since it is rarely the case that the character of the person is relevant to the character of the argument. If a perjurer adduces an argument that contains premises whose truth do not depend on him at all, then the argument should be evaluated on its own merits, irrespective of the moral character of the arguer. Another popular fallacy, the so-called tu quoque fallacy, or the "you are another" fallacy, is an instance of this type. One commits this fallacy when one dismisses criticism by noting that the arguer is guilty of the same thing he is criticizing. In general, calling attention to the arguer as a way of attacking the argument is often a sophistical sleight of hand, intended to call attention away from the argument and paint the argument with the same derision that is smeared on the arguer.

(1) "You say that the Sixers will win because Larry Brown is a championship coach in what is perhaps his final year of coaching, Iverson is unstoppable, and Mutombo will at least slow Shaq down. But I just heard you say over there that you think Larry Brown is a lucky chump! I really doubt that you mean what you are saying here." (2) "Look at you--you're smiling! You can't help but smile when you're lying, so what you've said has gotta be false." (3) Tu quoque fallacy: "I understand that you don't think we should kill cattle for our own use, but look at you, standing there with leather boots and a leather belt. I'm sorry, but I will not put down my hamburger."

  • Appeals Fallacies: In many arguments, the premises introduce a subject area and in so doing reveal their conclusion. Arguments such as these depend for their force on the subject matter connections among constituent propositions--the premises stay focused on the topic and motivate belief in their conclusion by what they say about the topic. This is not how it always works, though. In some situations, one might need to avail oneself of additional corroboration from sources other than those figure directly into the topic at hand. At these times, one might appeal to an authority on the topic, or perhaps to tradition or popular opinion. As with ad hominems, these can be relevant and compelling or irrelevant and fallacious. When an appeal supplies information that could buttress the truth of a conclusion, it is relevant and worth serious consideration; however, appeals can also be aimed at deflecting criticism away from arguments by raising issues that are not relevantly related to the topic. Often appeals of this second type play on one's emotions, generating a passionate response that is intended to produce a commitment to the conclusion irrespective of the rational character of the argument. Guided by these observations, I distinguish between two types of appeals: the truth relevant and the truth irrelevant. These do not represent mutually exclusive and exhaustive categories--instead, they mark out two ends of a spectrum of appeals. In what follows, I indicate which specific appeals fall into these categories and which fall between them. I also describe these appeals and illustrate their fallacious use.

Truth Relevant Appeals:
these are made to considerations that purport to bear on the truth of the claims that constitute an argument.

  • Appeal to Authority. When someone appeals to an expert or a standard text to support their claim, they make an appeal to authority. These can be good if the issue is one on which there are authorities, and if the person or text cited really is a credible authority and is cited correctly. Often, though, authorities are cited where there are none, or are cited incorrectly.

"Einstein believed in God, so God must exist." (One might wonder whether there could be an authority on this topic, although there are many who would argue that the Bible is one; however, even if there are, it is unlikely that Einstein was one.)

  • Appeal to Ignorance. One appeals to ignorance when you claim that something must be true because we don't know that it isn't. Here, ignorance of any conditions that would defeat the conclusion is taken to be an argument for the conclusion. If one's conclusion is tentative and there is good reason to believe that a lack of information is a significant indication, then this sort of appeal can be persuasive. (Consider: "It's after the notification deadline and we haven't heard anything; we must not have gotten it. Nuts!") However, many appeals of this type are fallacious--ignorance may be bliss, but it is rarely evidence. This is especially true of arguments with strong conclusions.

"He must have ratted you out to the cops. Think, do you know that he didn't? Have you any evidence that he didn't? There--that proves it!"

Truth Irrelevant Appeals:
these are made to considerations that have nothing to do with the truth of the constituent claims. Often, these address the emotions of the audience in an attempt to win them over to the conclusion via non-rational means.

  • Appeal to Pity. This move involves appealing to the sympathy of an audience so as to convince them of a particular conclusion. This conclusion could be a proposition or a course of action, i.e., a claim that one should believe P or that one should do A. If the former, the move is almost always fallacious, since feelings of pity will only rarely settle questions of fact. If the latter, it may be solid insofar as it introduces moral considerations into the mix and calls upon the audience to act morally. However, if the appeal is intended to obfuscate the issue and win people over to a course of action without any implicit moral argument, it is fallacious.

(1) Belief: "The US should give more money to African famine relief--have you seen the victims? Oh my, they're emaciated, on the verge of death, and they have this sad look in their eye. This is a policy we should endorse." (2) Course of Action: "Ahh, Dad, look at that kitten. He's so sad. And I'm sure that he won't last if we don't buy him--he's so thin and homely. No one will buy him when there are so many other kittens at the shelter. Please?"

  • Appeal to Threat. This type includes the classic "schoolyard bully" fallacy, "Believe it or I'll pound ya!", but it also includes other appeals. In general, an appeal to force is made when the audience is threatened with the consequences of failure to accept the argument. This threat could be carried out by the arguer, as in the bully case, but it need not be. This is also known as an appeal to force or appeal to power. The bully variant is compelling (especially if he is big and mean), but fallacious nevertheless. Threats of physical violence may compel one to believe, but this belief will not be motivated by the truth of the conclusion. Violence is irrelevant to the subject matter that constitutes the content of the argument. Other variants typically present consequences that are similarly irrelevant to the soundness of the positions they are introduced to support. As with other appeals, though, these need not always be fallacious, especially if the threat is introduced to warn someone off of pursuing a dangerous course of action. (E.g., "Don't go down that road--the gravel is so loose that you will crash, believe me.") One must ascertain whether the threatened consequences are relevant to the truth of the conclusion.

(1) Belief: "If you don't believe what I say in my lectures, I'll flunk you!" (It may well be true, but this does not supply any reason to believe that what is said in the lectures is true.) (2) Course of Action: "If you don't contribute to the fund, you will not be promoted." (Bad news, if true, but hardly a reason to believe that contributing to the fund is the right thing to do in general.) [Consider Pascal's Wager in this connection--is it an appeal to threat?]

  • Appeal to Flattery. If you try to win someone over to your way of thinking by telling them how great they are instead of presenting them with evidence for your conclusion, you make this appeal. The beauty, brains, and all-around fine character of your audience will be relevant in some contexts, but it generally will not be relevant to the truth of an argument's conclusion. (Of course, if the argument is about how great your audience is, then the whole thing will drip of flattery anyway, making such an appeal an integral part of the effort.) Such an appeal is often intended to make the audience feel good about the arguer, who then exploits the "feel good" moment for their own purposes.

"Someone as smart as you can certainly see that a vacuum cleaner like this one has no rival. And it's lightweight, so people who aren't anywhere near as strong as you can use it just as well as you can!"

  • Guilt by Association. An argument designed to win an audience over to a particular view by pointing out that the opposite position is associated with someone or something held in low esteem is a guilt by association argument. Because rejection of the argument is associated with someone or something that is unacceptable for some reason, then rejection of the argument is unacceptable. Often this involves noting that people who reject the conclusion are undesirable from the perspective of the arguer, in which case it is relevantly similar to ad hominem arguments. (The savvy arguer will select a target who they believe is undesirable from the perspective of their particular audience.) But the source of unacceptability need not be people--it could be that rejection of the argument is associated with arguments or views that are given little credence. This association could be structural, by analogy, perhaps, or substantive. Guilt by association arguments of the latter type can be persuasive, if the connections asserted in fact obtain and the associates are disreputable for good reason. Guilt by association arguments of the former type should be regarded with suspicion. In general, an argument should be evaluated on its own merits. A guilt by association argument of the former type can be used to motivate more careful consideration of the support proffered for the view in question, but the view should stand or fall based on the strength of that support. Given that this sort of argument does not in general speak to the questions of fact raised by the premises, it should not be taken to settle them.

"I thought about buying one of his paintings, but the Smith's own one and they are notorious for their bad taste."

  • Appeal to Personal Circumstance. Appeals of this sort introduce personal circumstances into arguments as relevant to the truth of their conclusions. In certain cases, the audience's self-interest is a relevant consideration, especially when courses of action are being considered. For instance, the flexibility afforded by a second car could well be a significant factor in convincing one to buy a second car. However, when the conclusion is a matter of weighty importance involving more than just the audience, self-interest loses its relevance. In such cases, what is good for me or for us might not be a factor on which we should rely in judging the conclusion. We might not like the fact that the conclusion puts us out, but life isn't always kind. In such a case, an appeal to personal circumstance will typically be fallacious.

"If the state legislature pass the cigarette tax, you'll wind up paying as much as 50 -cents more than you do now for your pack of smokes! That's a huge amount. You should definitely not support the tax." (This consideration might make one unhappy with the tax, but whether the tax is reasonable or not depends on considerations other than what will happen to one person's bank account as a result of its adoption.)

In addition to these, there are appeals that can be truth relevant or irrelevant, depending on how they are used. Two that stand out are:

  • Appeal to Popular Opinion. "Everyone is doing it!" "Everyone is talking about it, I'm tellin' ya!" These are appeals to popular opinion, designed in the first instance to motivate action and in the second belief. Such appeals introduce purportedly widespread opinion as relevant to the soundness of an argument or the legitimacy of a position. In some cases, the appeal is truth irrelevant and intended to convince an audience that they are all alone if they do not embrace the conclusion. Often this is enough to carry the day, since many people are not rugged individualists in argumentative matters, but this fact doesn't make them any less fallacious. As we have seen, arguments should stand or fall on their own merits, and this type of appeal seems designed to short-circuit that. In other instances, these can be offered as relevant to the truth of an argument's premises or conclusion, such as when they are intended to provide inductive or abductive evidence. If one is arguing for the popularity of something, such an appeal would supply relevant inductive evidence. Also, one might adduce such an appeal in arguing that an election winner is the best candidate for her constituency. ("After all, most people seem to like her, and that is a good sign.") Here popular opinion bears on the truth of the constituent claims. As with most appeals, these can carry some conviction, but often they are asked to carry far more than they can support. In such cases, they are fallacious. To detect fallacious employments, one must determine if popular opinion is in fact relevant to the soundness of the argument or position in question. If so, then one must be careful not to endow the appeal with more argumentative weight than it can carry.

(1) "Most of the people who have come through here in the last few days have purchased Playstation 2. I'm not saying that there isn't something that can be said for the others, but this really convinces me that Playstation 2 is the deck for you." (2) "Most people in the world believe in some religion or other; therefore, you should be religious."

  • Appeal to Tradition. When popular opinion has gray hair, it's tradition. Thus, an appeal to an opinion that has been popular for a long time counts as an appeal to tradition. Of course, such an appeal must identify a community in which the opinion is customary, and it must also make some sort of commitment to the length of time necessary for tradition formation, but typically this is not going to be a subject of much dispute. That is, you may well question whether the purported tradition really is traditional; if it isn't the appeal fails, but that doesn't qualify it as a fallacy. It is a fallacy if the very appeal itself (true or false) is irrelevant to the truth of the conclusion or the soundness of the argument. If this appeal is intended to fan the flames of reverence into an inferno that consumes all rational criticism, then it is truth irrelevant and most probably fallacious. Whether it is customary to believe or behave in a certain way does not establish that this way is defensible or even rational. As we all know, there are some pretty ridiculous traditions out there, so being consistent with a "tradition" is hardly compelling by itself. If an argument can't stand on its own, tradition will not help break its fall. Additionally, though, tradition can be used to supply inductive or abductive evidence for propositions in an argument, and so an appeal can be truth relevant. In these cases, one must determine if tradition is relevant in the manner suggested; if so, then it should be evaluated along with other reasons, but if not, then it is fallacious.

(1) "The O'Rourkes have been ardent supporters of higher education for many generations. This is one thread that binds us together as a family. You will go to college and you will appreciate it. Understand?" (2) "Faculty at this university have always regarded it a breach of decorum for students to refer to them by their given names, so it is wrong to do that."

  • Wishful Thinking: This fallacy is committed by those who conclude that something is true because they want it to be true. Rarely does a strong desire that something be true bear on its truth, unless perhaps the truth concerns one's own desires. ("I've been so apathetic of late, not caring about anything at all; man, I really wish that I wanted stuff again!") In most cases, though, wanting it does not make it so. Arguments of this sort are very fishy and should be subjected to serious scrutiny.

(1) "I am most definitely going to get that guitar for Christmas. I've never wanted anything more in my life! I have to get it!" (Note that this would not be fallacious if there was an implicit premise that went something like, "If I want something badly, my parents always get it for me.") (2) "Just think about how many people drink coffee--how many people love coffee! It can't be that bad for us, it just can't be." (Ahh, would that it were so...)

  • Begging the Question: An argument should provide one with independent reason to believe its conclusion. While it may not always be the case that arguments are proffered to convince the incredulous [note], one should be able to see them as having that potential. A person who seeks an argument because they do not believe a particular claim will require that the argument supply reasons that they can accept and that compel them toward the conclusion; otherwise, they will remain steadfast in incredulity. Surely, then, the argument will not perform its function if one must believe the conclusion in order to accept the premises, since the incredulous audience requires premises they can accept and we have already established that they do not accept the conclusion. Arguments that fail in this way beg the question. This fallacy, also known as petitio principii, is committed when one assumes what it is that one wishes to prove. When this is done, the arguer is in effect begging that the question that prompted the argument be asked again. Commission of this fallacy is a sign that one's argumentative burden has not been discharged. In such a case, one could assume the conclusion, smuggling it in under a different guise. An argument of this sort is logically circular, since it leads one from the premises right back around to the premises, given that the conclusion is one of the premises. Note, though, that it is not structurally flawed; in fact, it is deductively valid and so in fine shape from a structural perspective. Alternatively, one might set the argument up in a context that will only be acceptable if one has already bought into the conclusion. This argument would not be circular, but because it requires that one already be on board to accept the premises, it does not provide any rational support for the conclusion. Either way, these arguments will not do what they are called upon to do, viz., provide one with independent reason to believe the conclusion. They are the argumentative equivalent of pulling oneself up by one's bootstraps, a trick that should be left to magicians and not critical thinkers.

(1) Circular: "Free trade is good for the country, because it brings all the advantages of an unimpeded flow of goods between countries, and the unimpeded flow of goods between countries is good for the country." (From Fogelin & Sinnott-Armstrong, p. 409.) (Problem is, free trade just is the unimpeded flow of goods between countries, so the conclusion merely restates the second premise, rendering this argument circular and question begging.) (2) Non-circular: "You'll grant me that murder is wrong, right? So since abortion just is a type of murder, you are irrational to think that it isn't wrong." (This is clearly question begging, since one who disagrees with the conclusion will not accept the premise in which abortion is termed "murder"; however, the conclusion is not logically equivalent to any of the premises, and so it is not circular.)

  • False Dilemma: "It's got to be either A or B, and it ain't A, so B it is!" Arguments of this type are all well and good, so long as it really has to be either A or B. Problem is, many times the issue in question is much more complicated than this. The fallacy of the false dilemma is a direct result of this sort of oversimplification. It is easier to deal with a problem that is discrete, with a small finite number of mutually exclusive solutions, since you need consider only a small number of alternatives and any consideration that counts against one counts in favor of the others. (Dilemmas have two potential solutions, trilemmas three, and so on.) However, with problems that are more spectral, having many different but sometimes overlapping solutions, the situation is much murkier. A consideration that counts against one alternative may not count in favor of any of the others, and that makes the gathering of evidence and the process of judgment much more difficult. Consider abortion, as an example. Some would argue that it is either permissible or impermissible, period; i.e., all abortion procedures stand or fall together. However, many would argue that this is an oversimplification, and that the cases in which abortions are performed have to be considered individually, in light of the circumstances that attend them. If this is true, then the former way of characterizing the issue commits the fallacy of the false dilemma. Focusing specifically on the false dilemma, we can talk in terms of contraries and contradictories. Two options are contradictories if they are negatives of one another, e.g., true/not true, here/not here, big/not big. They are contraries if they are opposites, e.g., tall/short, big/little. There is no middle ground between contradictories, but there is a great deal of middle ground between contraries. One commits a false dilemma if one presents contrary alternatives as if they were contradictories.

(1) "You are either for me or you're against me. You didn't give to my campaign, so you are against me!" (One can fail to be for someone without being against them in any serious sense of that word; one could, for instance, be entirely apathetic about the whole process and fail to be for or against anything.) (2) "Your epistemology must be either rationalist, empiricist, or fundamentalist. Rationalism and empiricism are false, so fundamentalism is the only option!" (There are other unconsidered options in this case, e.g., pragmatism; however, if one could supply a principled argument for taking the other options to collapse into those listed, then the charge of false trilemma would be countered.)

  • Strawman Fallacy: Arguments in the real world do not arise in a vacuum. They are produced in contexts by people with agendas. Typically, they arise in contexts where there is disagreement about a position or a conclusion. In these contexts, one is often constructing arguments that are aimed at undermining the positions or arguments of another. To succeed in supplying a forceful argument of this type, one must be sure to set up the target position/argument correctly, i.e., in a way that would secure the approval of its proponents. Failure to do this will undermine the primary argument, since even if it is valid, the audience will deny that it applies to the position that they actually endorse. "That's a nice argument," they might say, "but it doesn't apply to my position." One commits the strawman fallacy when one characterizes an opposing view inadequately. This could involve distortion, oversimplification, unjustified extension, or incomplete expression, to name four forms of inadequacy. This move takes a real, flesh and blood position and attempts to pass it off as a faux, straw position. It is much easier to knock over a straw man than a real one, after all. This need not be intentional, since it could arise from the influence of unappreciated biases at work in the arguer.

Example: "Bush's energy policy just doesn't make any sense. He's from Texas, you know, and he is steeped in the oil business. When you get down to it, he wants to meet our energy by just building domestic oil and natural gas production. That's lunacy!" (A Bush supporter would have no problem countering this charge, since the policy is much more complex than this. The straw position attributed to Bush is not his actual position, so the supporter need not even address the charge that it makes no sense.)

  • Fallacy of the Mean: Some of us take things to extremes, but often this is unwise. It can be unhealthy, unsafe, or just plain obnoxious. These reactions suggest that we would prefer that they make like Aristotle and embrace the mean, i.e., moderate their behavior. The moderate approach is often a good one, but to assume that it is always the best one is ill advised. (Indeed, this amounts to taking an appreciation for the mean to extremes!) In some cases it isn't best, and application of the assumption can get you into trouble. When you apply the assumption where it is not applicable, you commit the fallacy of the mean.

"Clearly 10 drinks is too many over lunch, and I think zero drinks is too puritanical for me. So I'll stop with 5 drinks--that will be my golden mean."

  • Vague Concept Fallacies: Arguments that commit these fallacies attempt to establish precise conclusions on vague premises. Premises in an argument are claims (i.e., propositions), and claims consist of concepts that stand in a particular relation to one another. For instance, the claim that climate change is a serious issue involves the concepts climate change and serious issue. The nature of the claims, and so the nature of the argument, is determined by the constituent concepts. Some concepts, like the geometric concept square, are precise, i.e., there are clear standards that determine whether or not they are applicable. Other concepts, like bald, are vague, which is to say that there are no clear standards of applicability. Vague concepts often depend for their application on variable aspects of context and interest. There is nothing wrong with vague concepts, mind you, but there is something wrong with an attempt to make vague concepts do the work of precise concepts. The flaw here is structurally similar to the flaw that underwrites the fallacy of the false dilemma. Vague concepts are spectral, with cases in which application is uncertain sandwiched between cases in which they clearly apply and cases in which they clearly do not apply. Elephants are heavy animals and mice are not heavy, but what of dalmations? Or holsteins? The answer is not determinant and often depends on the interests of the person asking the question. Thus, with spectral concepts, it isn't often clear whether a concept continues to apply or fails to apply as you move from case to case, where with determinant concepts there are precise principles that can support such a judgment. (They may be very difficult to apply, but they will be available.) A vague concept fallacy will turn on use of spectral concepts like these as if they were precise, determinant concepts. In particular, it will depend on exploiting the fuzzy middle area, moving around in it as if it were absolutely clear when the concept applies and when it does not. In general, the steps in this sort of argument are as follows: (a) set up a sequence of incrementally related things or events, (b) begin at one end of the sequence and make an judgment about it, using the vague concept (i.e., assert either that the concept applies or does not apply), (c) proceed along the sequence in the direction of the other endpoint, noting at each point that the concept still applies in the same way, (d) finish up at the other endpoint, step back, and them make a general, overall judgment about either the endpoint you just landed on or about the sequence as a whole, where these can be judgements of identity, possibility, morality, etc. Arguments generated by this process permit you to pass judgment only after very small increments, and if you start with something to which a vague concept applies and change it ever so slightly, you will still be able to apply the concept. That is, there is no threshold effect with vague concepts. But when we actually make judgments in the world, we aren't just restricted to incremental judgments--we can make general, overall judgments that comprise several steps. In cases like these, our incremental judgments do not add up to our more general judgments, a fact that serves as a characteristic feature of vague concepts. Precise concepts are additive in this way because there is a threshold effect associated with them. Thus, we get a sort of argumentative sleight of hand, where you know what you start with and you know what you have at the end, but you have no reasonable idea how you got from one to the other. In what remains of this section, I introduce and illustrate four prominent vague concept fallacies.

  • Argument from the Heap: This fallacy is closely associated with the Sorites Paradox. This fallacious argument assumes that we know what is up with one endpoint of the sequence described above, i.e., we know that it is either a case in which the concept in question clearly applies or clearly does not; in addition, we know that if we change the case in a small way, we still have a case of the same kind as the one we started with. By examination of repeated small adjustments of this sort, we establish that we can never get to the other endpoint. If you start with a grain of sand, you don't have a heap; adding one grain of sand never takes you from a non-heap to a heap, so if you add a billion grains of sand to the first, one by one, you will still not have a heap. Hence the name.

(1) Assume Tim is hairy; taking one hair off of Tim's head will still leave him hairy; thus, by slow increments, we can establish that you could take a million hairs off of Tim's head and he'd still be hairy, so it is not possible for Tim to become bald. (2) "Life must be caused by a divine spark. Consider that if you start with an atom and you add another atom to it, it won't be alive, right? And if you keep doing this, you'll just have a bunch of atoms and no life. You can't get a living thing that way!" (3) Start with 1. Add 1 to it. You get a finite number. Add 1 again. Still finite. Keep doing this forever. Always you get a finite number. Therefore, there is no such thing as infinity.

  • Conceptual Slippery Slope Arguments: This argument exploits the vague middle area in order to show that there is no significant difference between the endpoints of the spectrum. Thus, you start with one of the endpoints and demonstrate, via the incremental sleight of hand described above, that we can get to the other without having to note a relevant difference at any point along the way. By setting up a sequence of increments small enough to justify the identical application of the vague concept, we fallaciously infer an overall conceptual identity.

"A human egg one minute after fertilization is not very different from what it is one minute later, or one minute after that, and so on. Thus, there is really no difference between just-fertilized eggs and adult humans" (Fogelin and Sinnott-Armstrong, p. 360).

  • Fairness Slippery Slope Arguments: As we have remarked, the existence of a vague middle makes it difficult to draw a line that indicates where the concept can be applied and where it cannot, even if the two endpoints of a spectrum are very different. If the concept is one that is associated with a moral property like fairness, we could become concerned about whether it would be fair to treat one case differently from a case that is not significantly different from it. Thus, we can use the sequence and the application of the vague concept to motivate an overall moral judgment. This is similar to a conceptual slippery slope argument, since you affirm both endpoints and argue that they are the same in an important sense. However, it is also similar to an argument from the heap since you start with a secure moral evaluation of one of the endpoints and proceed to argue that the same evaluation should apply to the other endpoint because there is no principled, morally acceptable way to "draw a line" that separates the endpoints.

Example: "Think about the animal kingdom. If you start with the most complex forms, you find that they are very similar to other, less complex forms. But this creates a moral problem. Consider that it isn't morally acceptable to kill a human. But where does it become morally acceptable? Where do you draw the line? You can't, so it must be morally unacceptable to kill rats in your house."

  • Causal Slippery Slope Arguments: These are perhaps the most prevalent slippery slope arguments. Also known as "domino arguments," they attempt to establish that an event which is apparently not bad will lead to one which is bad by claiming that if the first event comes to pass, it will cause a second, and this will cause a third, and so on, until we eventually arrive at the bad event. Thus, the first event must be bad after all and should be avoided at all costs. In setting this argument up, you construct a sequence of events and then assert that each event causes the event that occupies the next spot in the sequence. Here the vague concept is causality, and more particularly, causality in general. Once the sequence is constructed, it motivates a general, overall judgment that is evaluative in character. (Specifically, it is consequentialist in character.)That is, the conclusion is that an evaluative concept (e.g., being socially hazardous) which obviously applies to one end of the causal spectrum also applies to the other. As such, these arguments have aspects in common with both heap arguments and conceptual slippery slope arguments, much in the same way that fairness slippery slopes do. Note that if the causal links that are asserted to obtain between these events are strong, then such an argument can be a good one. While vague, causality is less so than other concepts that serve in these arguments.

Examples: (1) "We have to take a stand against sex education in junior high schools. If we allow sex education in the eighth grade, then the seventh graders will want it, and then the sixth graders, and pretty soon we will be teaching sex education to our little kindergartners" (Fogelin and Sinnott-Armstrong, p. 359). (2) "If a man indulges himself in murder, very soon he comes to think little of robbing, and from robbing, he comes next to drinking, and Sabbath breaking, and from that to incivility and procrastination. Once begin upon this downward path, you never know where you are to stop. Many a man dated his ruin from some murder or other that perhaps he thought little of at the time." (de Quincey, from Fogelin and Sinnott-Armstrong, p. 360).


VI.     Fallacies of Context

Finally, we come to fallacies of context. If you find yourself in possession of an argument that bears up to formal analysis and subject matter analysis but still fails to advance the agenda of the arguer, then it quite likely conceals a fallacy of context. These are commonly inconsistencies of one form or another, such as an inconsistency between the argument, understood as a means, and the end for which the argument was produced. But they can also be flaws that undermine the argumentative context by dooming contributions from the start. Fallacies of the latter sort create a context within which no quality argumentative work gets done.

In what follows, I introduce and illustrate four of these fallacies before commenting on the range of additional flaws that could receive attention under this heading.

  • Ad Hominem -- Against the Arguer: As noted above, an argument ad hominem is an argument against the person that is typically intended to undermine an argument or a position that the person endorses. We commented in the previous section on ad hominems that are used to undermine particular arguments by challenging the truth of their constituent claims. However, they can also be used against the arguer, irrespective of any particular argument they happen to make. This type of fallacy has several names: poisoning the well, damning the origin, and the genetic fallacy, to name three. When used in this way, ad hominems modify the context of the argument by discrediting the arguer in some way, typically by casting suspicion on his motives or his general character. (If an attack on a person's credibility is focused so as to vitiate a particular premise or set of premises, then it is an ad hominem against the argument; however, if the attack on one's general credibility, then it qualifies under this heading.) Here, the considerations adduced undermine an argument by convincing an audience to discount anything (including the argument) that comes out of the speaker's mouth. The arguer is the well, and by poisoning it we spoil anything that might come from it. These differ from ad hominems against the argument since they are not designed to cut away at a specific argument--they are more slash and burn in their use, whereas the first type of ad hominem is a more targeted criticism. Before describing and illustrating two popular forms of this type of argument, it should be noted that while it is possible to motivate a theoretical distinction between two general types of ad hominems, in practice it is often very difficult to do so. In many cases, there is only one argument in play, so any ad hominem consideration will be interpreted in conjunction with that argument, whether it is actually directed at it specifically or at the arguer. To the extent that distinction in such a case is possible, one can ask whether the considerations adduced could be used to undermine an argument on a different topic; if so, that is good evidence that this is an ad hominem against the arguer and not the argument.

(1) "I'm sorry, but you're a democrat, and where I come from, they run democrats out of town. I won't run you out of town, but I won't believe anything you say, either. So you just take your "we should get the asbestos out of the building" tripe somewhere else." (2) "So you think that Microsoft is blameless? That they shouldn't be split up? But hey, don't you like to buy high tech stocks? I bet you're saying that to protect your bank account."

  • Fallacy of Complex Questions: A question can be used to shape a conversational context by shifting the burden of the conversation to the person on the receiving end. If in addition the question puts the recipient on notice that he is expected to defend a belief or choice with his response, then it creates an argument context. However, some questions can create contexts in which no argument by the recipient can possibly succeed in making him look good. If in such a case the recipient is not in any way deserving of this sort of disadvantage, the person asking the question is guilty of committing the fallacy of complex questions. The question in this case is complex in the sense that its character tilts the argumentative playing field so much that the recipient is always running up hill. It does this by presupposing the truth of a false claim that, were it true, would justifiably put the recipient in a disadvantageous position with respect to the topic.

(1) "Have you stopped beating your wife?", asked of someone who has never beaten his wife; here the question presupposes that the recipient has beaten his wife. (2) "You aren't still taking drugs, are you?", asked of someone who has never taken drugs; here the false presupposition is that the recipient has taken drugs.

  • Fallacy of Means/End Incoherence: Most arguments are adduced in support of conclusions that, when established, fit into larger agendas. Thus, for example, we argue for the virtues of a particular restaurant so that it can become the place where our hunger is sated. We don't typically argue just for the sake of arguing. Given this, if an argument is to be successful, it must not only be good in itself, but it must also cohere with the larger agenda into which it fits. This requires that it be coherent with the end of that agenda; that is, the argument must support a conclusion the truth of which advances the arguer toward the attainment of her ultimate end. If the argument is good in itself but is irrelevant to the ultimate end, either thematically or structurally, then the arguer has committed the fallacy of means/end incoherence. The flaw in such a case could be due to thematic confusion on the part of the arguer, or to a subtle equivocation that produces the appearance of connection where there is none. (This fallacy is the context version of the non sequitur.)

If I succeed in crafting a quality argument for the conclusion that Gale Sayers was the best tailback ever in an argument about fullbacks, then I commit this fallacy. (Here I may have become confused, which could be at the root of many of these mistakes; however, it may be the case that subtle shifting of the context can change the subject and thereby shift the balance of power to the guilty party.)

  • Fallacy of Means/End Incompleteness: Production of a quality argument is all well and good, but if it is only part of a larger argumentative plan and the arguer fails to produce the other parts, then they will have committed the fallacy of means/end incompleteness. In such a case, the argumentative plan requires multiple steps, where each step requires an argument. An arguer who promises the full trip but only delivers part of it has failed to achieve her argumentative goals. This could be due to confusion about where in the process one is, or it might be due to a more nefarious reason, such as the desire to conceal an argumentative deficit of some sort or another.

On the heels of a promise to argue for the virtues of each of the Republican candidates for national office listed on the Idaho ballot, one argues for Bush and Otter and calls it quits, perhaps by uttering something like, "and so on." In this case, the generalization is out of place, given the promise of specificity, and so represents a sign of means/end incompleteness.

This list, while brief, is representative. One can commit a fallacy of context by (a) creating a bad one, or (b) failing to negotiate effectively through a good one. The first two fallacies cited above illustrates the first mode, while (3) and (4) exemplify the second. There are other forms these modes could take--one might presume guilt in a context where innocence should be presumed, say, or change the context in the middle of an episode (a variation on (3) above)--but we'll rest with these specific examples. One thing to note before taking up the remaining business of this chapter is that we can represent the argumentative context as itself an argument, in which case we will likely find contextual analogues of most of the argumentative fallacies described above.

VII. Convicting the Innocent

It is said that a little knowledge can be a dangerous thing, and this is especially true of knowledge about the fallacies. As I've argued, awareness of fallacious patterns of reasoning can reduce the amount of time and energy one must devote to critical thinking, and in the process enhance the overall quality of that thinking. But because it can engender a lazy approach to information evaluation, it can also have a very different effect. It is easy for folks to think that if they know the fallacies, they know all they must to be good critical thinkers. After all, by spotting fallacies they can spot bad arguments, and this will support them as the construct and critique pieces of reasoning. Spotting fallacies, however, is a very tricky business, as I have tried to make clear. Many of the patterns discussed above can be used to positive effect in the right circumstances. For example, a causal slippery slope argument, an appeal, or an ad hominem might be reasonable and even persuasive. (The patterns under the heading "fallacies of form" tend to be pretty fishy in general, but even there one can affirm the consequent if one is attempting to supply confirmatory and so non-deductive evidence for the antecedent.) Given this, careless use of fallacy knowledge can result in the unwitting rejection of a good argument.

Consider this analogy. Our criminal justice system is constructed so as to maximize the certainty associated with criminal conviction. The epistemic standards for conviction, viz., beyond a reasonable doubt, are high, and this is intended to ensure that a defendant is acquitted only if he is guilty. In general, it is considered a very bad thing to convict people of crimes they did not commit. (One way to ensure that the guilty are convicted is to lower the standard and convict pretty much everyone brought up on charges. Sure, you incarcerate innocent people, but at least all the guilty folks are locked up!) The same view informs critical thinking, i.e., it is a bad thing to reject compelling arguments. We avoid this by taking care to ensure that we have compelling reasons to reject an argument. In particular, we must make sure that an argument which instantiates one of the fallacies is in fact fallacious. This amounts to determining that the context, subject matter, and form do not render the argument persuasive in that instance. A little knowledge can be a dangerous thing, but a lot of knowledge can be a good thing.