Logic In LD

by Jason Baldwin

Part I: Elements of Theory !

I. Preliminary Notions! II. Three Valid Forms! III. Practice!

1
1 2 7

Part II: Casing Applications !

I. Burdens! II. Research! III. Contention Mapping! IV. Value Premise(s)! V. Contention Development!

9
10 12 15 22 26

Part III: Cross-Examination and Rebuttals !

I. Basic Principles! II. Cross-Examination! III. Rebuttals! IV. Final Issue Selection!

31
31 32 36 39

Part I: Elements of Theory

The purpose of this article is to introduce LD debaters to a few basic elements of formal logic; later articles will explore some applications of logic to the practice of debate. Some people say that high school debate should be conducted and judged on the basis of logic alone-sola ratione, so to speak. Anyone acquainted with logic as a deductive science will have a hard time guring out what these people mean. If they mean simply that debaters should avoid being illogical, well and good-but then who has ever suggested that debaters should be illogical? If they mean that the resources of logic should be sufcient to settle the truth of any debate resolution, they are betraying a basic confusion about the power of logic. Trying to generate substantive truths using only logic would be like trying to generate money using only arithmetic. Hopefully the readers of this article will come to understand why. I. Preliminary Notions Despite its limitations, logic is an indispensable tool in debate. Logic is the science of arguments. An argument is a set of sentences, one member of which is the conclusion and the others of which are premises offered in support of the conclusion. An inductive argument is one in which the premises are intended to make the conclusion likely or probable. For example, (1) Most Alabamians own guns.(premise) (2) Pat is an Alabamian.(premise) (3) Therefore, Pat owns a gun.(conclusion)

If both premises of this argument are true, they provide evidence supporting the conclusion, but they do not render the conclusion certain. Perhaps Pat is a member of the tiny Alabama chapter of Handgun Control and believes it is immoral to own guns. A deductive argument, by contrast, is one in which the premises are intended to make the conclusion certain. For example, (4) No wombats are mammals.(premise) (5) J.S. Mill is a wombat.(premise) (6) So J.S. Mill is not a mammal.(conclusion)

If both premises of this argument are true, they make the conclusion not merely probable, but absolutely certain. Obviously anyone who wants to argue for a given conclusion will prefer a deductive to an inductive argument when the former is available. Most constructive arguments in LD can be represented as deductive, and deductive arguments are the only sort of

arguments which we can assess with a formal precision akin to that of mathematics. Therefore, we will focus exclusively on deductive arguments. Astute readers will have noticed that the sample deductive argument above-about J.S. Mill-contains a glaring problem: Both of its premises and its conclusion are false. And yet it is as good logically as an argument can be. What gives? The argument is valid. That means that if the premises are true, the conclusion must be true. To determine that the argument is valid, we do not have to know whether the premises really are true or not. We can simply recognize that if the premises were true, the conclusion would also have to be true. So if no wombats are mammals and if J.S. Mill is a wombat, then J.S. Mill is not a mammal. But the argument is not sound. A sound argument is one which is valid and in which all the premises are true. This means that all sound arguments are valid (by denition), but not all valid arguments are sound (because some valid arguments have one or more false premises). Given the denition of validity, you can see that a sound argument must also have a true conclusion. Please note that only an argument can be valid or invalid, sound or unsound; sentences or propositions do not possess these properties. Likewise, only sentences or propositions can be true or false; arguments cannot be true or false. Obviously someone making an argument wants a sound argument, not merely a valid argument. But logic can deliver only validity. That is, an argument may be as logical as you please but still have a false conclusion. Logic tells us what propositions (conclusions) follow from the propositions (premises) we already believe, but it cannot tell us which propositions to accept as premises in the rst place. Garbage in, garbage out, as they say. This is not just your author's opinion, or the majority view among professional logicians; nor is it a prediction or a generalization. It is the sober and indisputable truth about logic. Already, you should begin to see the problem with the view that any interesting debate could be settled on the basis of logic alone. The only way a debate might be settled by logic alone is if both sides agreed to the same premises and the dispute were simply over the validity of the inference from those premises to competing conclusions. But debates are almost never like that-the debaters argue to different conclusions from different sets of premises. Often both sides present valid arguments, and the real question is, whose premises should be believed? For better or worse, logic by itself cannot answer this question. II. Three Valid Forms We said above that an argument has the property of validity when its conclusion must be true if its premises are true. But how can you ensure that your own arguments have this valuable property? Arguments are valid or invalid on the basis of their form, not on the basis of their content. In other words, it doesn't matter what the sentences in the argument are about or whether they are true or false. As long as the sentences are declarative and are properly arranged (more on this below), the argument they compose
2

will be valid. The bad news is that there are literally an innite number of valid argument forms. The good news is that we can represent most of the deductive arguments used in LD with only three simple forms. To focus your attention on these forms, it will help to introduce a few symbols. Let us use the italicized letters p, q, r, and s to stand for declarative sentences. We will use the tilde (~) written before a sentence to signify the negation of that sentence. So if I assert ~p, then I am claiming that the declarative sentence p is false. Finally, let us symbolize conditional (if-then) sentences using an arrow () with the appropriate component sentences written at either end of the arrow. Conditional sentences are complex sentences-that is, they are composed of two simpler sentences. The rst of these component sentences comes after the "if" and is called the antecedent. The second component sentence comes after the "then" and is called the consequent. Consider, for example, this conditional sentence: If it rains, then I will skip the game. The antecedent of this sentence is "it rains," and the consequent is "I will skip the game." If p stands for "it rains" and q stands for "I will skip the game," then we can represent the conditional sentence as: pq. Now we can introduce the three valid argument forms we need. The rst is called modus ponens (MP). A modus ponens argument is of the form: pq (premise) p (premise) So q (conclusion)

Here is an example of a valid MP argument: (7) If capital punishment deters crime, then capital punishment is moral.(premise) (8) Capital punishment deters crime.(premise) (9) Therefore, capital punishment is moral.(conclusion)

Even if you believe that (7) or (8) is in fact false, you can see that the argument is valid: if (7) and (8) are true, then (9) must also be true. It may or may not be obvious to you that the order of the premises makes no difference to the validity of the argument. We could have switched (7) and (8), and (9) would just as surely follow. However, the order in which you present the premises of an argument can be very important to your audience's grasp and acceptance of the argument, and we will return to this subject briey in a future article. The second crucial argument form is modus tollens (MT), which is: pq (premise) ~q (premise) So ~p (conclusion)

In the abstract, modus tollens is less obviously valid to most people than is modus ponens. But in fact, every MT argument is valid, and we use this pattern of reasoning all the time. For example, (10) If all human life were sacred, then it would be wrong to defend the innocent with lethal force.(premise) (11) But it's not wrong to defend the innocent with lethal force.(premise) (12) So not all human life is sacred.(conclusion)

Again, the order of the premises in an MT argument does not matter. So long as one premise is of the form pq and the other premise is of the form ~q, the conclusion ~p necessarily follows. The third and nal valid argument form we need is called hypothetical syllogism (HS): pq (premise) qr (premise) So pr (conclusion)

Unlike modus ponens and modus tollens, every sentence in a hypothetical syllogism is a complex conditional sentence-there are no simple p's or q's as premises or conclusion. Here is an example: (13) If pornography demeans women, then pornography communicates ideas. (premise) (14) If pornography communicates ideas, then pornography is a form of speech. (premise) (15) So if pornography demeans women, then pornography is a form of speech. (conclusion)

You should memorize the three argument forms we have learned so far. Your ability to apply these forms to more complex debate arguments will depend on your ability to recognize the simpler forms instantly. Before learning how to apply MP, MT, and HS to more complex arguments, we should note two common fallacies or invalid forms which resemble the valid forms above. What makes a form of reasoning fallacious rather than valid? Simply that its premises even if true do not guarantee the truth of its conclusion. One common fallacy is denying the antecedent: pq (premise) ~p (premise) So ~q (conclusion)

The best way to show that a given form of reasoning is fallacious is to produce a counter-example to it-an example of the form in question which has obviously true premises but an obviously false conclusion. For instance: (16) If pine trees are pink, then pine trees are colored. (premise) (17) Pine trees are not pink.(premise) (18) So pine trees are not colored.(conclusion)

Both premises are true, but the conclusion is false-pine trees are green, and therefore colored. The moral: Don't deny the antecedent! The second fallacy you should take special care to avoid is called afrming the consequent, and it looks like this: pq (premise) q (premise) So p (conclusion)

We can use the same rst premise about pine trees to demonstrate the invalidity of this form: (19) If pine trees are pink, then pine trees are colored.(premise) (20) Pine trees are colored.(premise) (21) So pine trees are pink.(conclusion)

Here we have true premises but a false conclusion, proving that the argument is fallacious. Don't afrm the consequent! With these warnings out of the way, we are ready to combine the simple argument forms into more complex patterns. First, instances of HS can be stacked one upon another, and validity will be preserved. If we have as premises pq qr rs

then we can validly conclude So ps.

Think about it. We can apply HS to the rst two premises to conclude p--r, and we can apply HS a second time to this conclusion and the third premise to get p--s. In fact, logically speaking, it doesn't matter how many conditionals we stack up. We can apply HS as many times as necessary to collapse a stack of properly related conditionals into a single conditional conclusion. (To be properly related, the conditionals must be such

that the consequent of one conditional is the antecedent of another, whose consequent is the antecedent of another, and so forth.) Further, we can combine instances of MP or MT with instances of HS. So, e.g., from the premises ~r pq qr

we can validly conclude ~p.

How does this work? Again, we are simply combining premises two-at-a-time using MP, MT, or HS to reach subconclusions, which we then combine with remaining premises using further applications of MP, MT, or HS. In this case, we applied MT to the rst and third premises, which yielded ~q as a subconclusion. Then we combined this subconclusion with the second premise in another application of MT to reach the nal conclusion ~p. Instead of applying MT twice, we could just as easily (and validly!) have applied HS to the second and third premises to yield pr and then applied MT once to this subconclusion and the rst premise to yield the same nal conclusion ~p. From what we have said so far, you may already be able to see that we can combine the two forms of complex arguments to produce valid arguments of any length. If we let p* stand for a sentence different from p, we can use the premises p pq qr rs sp* p*q* q*r* r*s*

to conclude s*.

We do this by applying MP repeatedly, beginning with the rst two premises and continuing down the chain, or (if you prefer) by applying HS repeatedly to the second through eighth premises and then applying MP once to the rst premise (p) and nal

subconclusion (ps*). III. Practice And that's it. If you have followed me this far and have learned the three argument patterns explained above, you know enough formal logic to be an excellent debater. Of course, knowing this (or any) amount of formal logic does not guarantee that you'll be an excellent debater. You need to know how to apply your theoretical understanding of argument forms to real arguments. Later articles will discuss the application of the logical theory outlined here to LD cases and rebuttals. For now, you need to begin to recognize valid (and invalid) logical forms when you encounter them in spoken and written arguments. Arguments, as you know, do not typically present themselves as numbered lists of premises and conclusions. Instead, they come cloaked in prose, and you must learn to probe for the logical skeleton beneath the esh of prose. Here is an example of a typical written argument, the kind you might encounter (albeit at greater length) in an LD case: American colleges ought not to practice race-based afrmative action because doing so will lead to massive human suffering. Afrmative action will make colleges more appealing to students from other countries because it will increase their chances for admission. When more foreign students graduate from American colleges, more foreigners will stay in the U.S. and compete for highpaying jobs. This will put Americans out of work, which will destroy our economy. The collapse of our economy would destabilize the world politically, triggering a nuclear war. And we must do everything we can to prevent a nuclear war.

You'll notice that there is not a single if-then sentence in this paragraph. But we can easily translate the paragraph into a form which more clearly displays its validity: (22) Afrmative action will increase the chances of foreign students for admission to U.S. colleges. (premise) (23) If afrmative action will increase the chances of foreign student for admission to U.S. colleges, then afrmative action will make U.S. colleges more appealing to foreign students.(premise) (24) If afrmative action will make U.S. colleges more appealing to foreign students, then afrmative action will cause more foreign students to enroll in and graduate from U.S. colleges.(premise) (25) If afrmative action will cause more foreign students to enroll in and graduate from U.S. colleges, then afrmative action will cause foreign students to stay in the U.S. after graduation and compete with U.S. citizens for jobs. (premise) (26) If afrmative action will cause more foreign students will stay in the U.S. after graduation and compete with U.S. citizens for jobs, then afrmative action will cause U.S. citizens to lose their jobs.(premise)

(27) If afrmative action will cause U.S. citizens to lose their jobs, then afrmative action will destroy the U.S. economy.(premise) (28) If afrmative action will destroy the U.S. economy, then afrmative action will destabilize the world politically.(premise) (29) If afrmative action will destabilize the world politically, then afrmative action will cause nuclear war.(premise) (30) If afrmative action will cause nuclear war, then afrmative action will cause massive human suffering.(premise) (31) If afrmative action will cause massive human suffering, then American colleges ought not to practice afrmative action. (32) So American colleges ought not to practice afrmative action.(conclusion)

If you symbolize this argument using p's and q's, you'll see that is it almost identical in form to the long argument which concluded with s* above. If you compare my premise-conclusion restatement of the argument with the original paragraph version, you'll see that the conclusion and the last premise of the argument are both expressed by the rst sentence of the paragraph version. The next-to-last premise, that nuclear war causes massive human suffering, is never stated outright in the paragraph, but we added it to our formalization in order to ensure the argument's validity. Sometimes there are good rhetorical reasons to leave premises unstated, but for the purposes of mastering the logic of arguments, you should work to make every assumption explicit. Much of the best philosophy consists in identifying and challenging premises which are usually taken for granted. So how good is the afrmative action argument? Well, if we accept the formalization as suggested, the argument is logically impeccable. It is valid-the conclusion follows necessarily from the premises-and that is all we can demand logically from any argument. But logical or not, the argument stinks. Although the argument's moral (or normative) premise (31) is plausible, the argument also relies on a string of eight empirical (or descriptive) premises (22-29), all of which must be true for the argument to reach its announced conclusion. These empirical premises are not at all plausible. Perhaps there is good evidence supporting a few of them, but most look wildly exaggerated, and taken together, they are surely indefensible. Because of HS, premises (22)-(29) together entail that if afrmative action increases college admission chances for foreign students, then it will cause nuclear war. No one who understood all the words in that conditional could believe it, and no smart debater would want to have to defend such a far-fetched premise in order to attack afrmative action. But for now, the point is that despite its obvious weaknesses, the argument is logically valid, and any substantive debate about the argument will be decided not by logic alone, but rather by non-logical empirical or normative disputes about the truth of the argument's premises. You can nd arguments to analyze anywhere people are making arguments. Try looking at the editorial page or watching the talking heads on the news channels. Rarely will you nd an argument presented as explicit instances of MP, MT, or HS. Your challenge is to translate the natural language of the arguer into the somewhat stilted if-then premises
8

you have learned to check for validity. You will almost always have to add one or more premises which are only implicit in the argument's original presentation. If you hear what you know is an argument but have no idea where to begin in mapping it, rst nd the conclusion. If the conclusion is best represented as a conditional, you know all the premises will be conditionals. If the conclusion is not a conditional, then you should expect at least one simple premise-represented by a non-conditional such as p or ~q-to make up part of an MP or MT. Work backwards from the conclusion through the chain of premises until you arrive at the starting point. Finally, keep in mind that the three argument forms presented above, while adequate to represent the vast majority of, and maybe all, LD case arguments, are only a small fraction of the standard argument forms available in modern formal logic. Many real-life arguments, and some LD arguments, will have to be shoe-horned quite awkwardly to t into these patterns, and some arguments will not t at all. Fortunately, there are logical tools available to represent almost any deductive argument. Unfortunately, the Rostrum is not the place (and I am not the author) to acquaint you with that full set of tools. You may know other valid argument forms which allow you to represent this or that argument more gracefully than you could using only MP, MT, and HS. You should use all the logical tools at your disposal when you approach an argument. If you nd the notion of a logical calculus scintillating, you should take a formal logic course in college, or even pick up a logic textbook and teach yourself.1 The sequels to this article will presuppose only a familiarity with MP, MT, and HS, but everything they say will apply to all valid arguments, regardless of form or complexity.

Part II: Casing Applications

This article illustrates some applications of formal logic to the development of LD cases. It presupposes familiarity with the elements of formal logic introduced last month. By way of review, an argument is a set of sentences some of which (the premises) are offered in support of another of which (the conclusion). A valid argument is an argument of a form such that if its premises are true, its conclusion must be true, and a sound argument is a valid argument with true premises (and hence a true conclusion). We introduced three valid argument forms which can be combined with each other in a variety of ways: (1) MP-Modus Ponen (pq, p, so q); (2) MT-Modus Tollens (pq, ~q, so ~p); and (3) HS-Hypothetical Syllogism (pq, qr, so pr). If any of the foregoing is unfamiliar, you should review the earlier article before proceeding with this one. I will discuss the applications of logic to LD casing under ve headings: burdens, research, contention mapping, value structure, and contention development. My underlying thesis is that an LD case is best conceived as the presentation of one or more deductively valid arguments. Among my more interesting conclusions is that debaters should drastically revise the traditional value premise (or core value) and

criterion model of argument. I. Burdens There is much confusion about the burdens of LD speakers. What is each side's basic responsibility? What, if anything, must each side prove? This section is the most theoretical in this article. If you don't care about debate theory and simply want to know my answer to the above questions, you can skip to the last paragraph of this section and keep reading from there. Some LD students and judges are eager to embrace a picture of burdens borrowed from policy debate. In policy debate, there is a standing burden of proof on the afrmative and a corresponding presumption for the negative. That means that in the case of a tie (including the sort of tie that would result if neither side made any speeches), the judge should vote for the negative. The afrmative must prove the policy resolution is true, and the negative must destroy whatever arguments the afrmative makes. However, the negative does not have to prove that the resolution is false. You might wonder how a policy negative could destroy the afrmative's arguments without proving the resolution false in the process. The answer is, in part, that you can destroy an argument without claiming anything about the truth or falsehood of the argument's conclusion. A good argument is a sound argument. And a sound argument, you will recall, is an argument which is deductively valid and which has only true premises. But an argument which is invalid or which has false premises may still have a true conclusion. Nonetheless, it's a bad argument because we can't know the conclusion is true based on the premises offered. So if we represent the policy resolution as p, the negative team may undermine the afrmative's argument that p without proving (or even attempting to prove) that ~p. There are further reasons why a policy negative can refute the afrmative's argument without actually disproving the resolution, but those reason are related to conventions of policy debate and do not concern us here. What does concern us is that LD does not have prescribed burdens and presumption comparable to those in policy debate. The NFL's ofcial LD ballot instructs judges that "There are no prescribed burdens in L-D as there are in policy debate; no 'burden of proof' and no 'presumption.'" This means that if there were an argumentative tie in an LD round, the negative would not win by default. No one would win by default. More importantly, the lack of prescribed burdens means that it is not sufcient in LD, as it is in policy, for the negative merely to attack the afrmative's arguments. The negative must offer constructive arguments of its own, arguments which conclude that the resolution is false. For suppose the afrmative offers two arguments, each of which concludes that p (the resolution) is true. And suppose the negative follows the policy strategy of merely attacking those two arguments-showing that they are insufcient to prove that p. But the negative (as we are supposing) offers no independent arguments to prove that ~p, i.e., that the resolution is false. Then the negative has effectively
10

destroyed the afrmative's reasons for believing the resolution without offering any reasons to disbelieve it. And since there is no presumption in LD, if there are no reasons to believe the resolution and no reasons to disbelieve it, there is no reason for the judge to vote one way or the other. It's as if the debate begins with a tie (since there is no presumption) and the afrmative case attempts to break the tie by proving the resolution true. If the negative merely disarms the afrmative's attempted tie-breaker, we are back to a tie again. To win, the negative must do more than simply disarm the afrmative's tie-breaker. The negative must attempt to break the tie in the other direction, by proving the resolution false. So we might think of LD afrmatives as having a burden to prove the resolution true, but only in a sense of "having a burden" such that LD negatives equally have a burden to prove the resolution false. This means that there are no burden strategies which are open to the negative but not to the afrmative just in virtue of the rules of LD. In other words, the standing burdens of each side are symmetrical. Therefore, refuting afrmative arguments is not sufcient to negate an LD resolution. However, there may still be a range of legitimate negative strategies on this or that particular LD resolution. We can say generally that the negative must prove the resolution false, but there are different ways different kinds of propositions can be falsied. Consider the resolution that "The possession of nuclear weapons is immoral." It's clear enough what the afrmative is responsible to prove. But what must the negative prove? That the possession of nuclear weapons is morally good? If the negative proved that, the negative would certainly be proving the resolution false. But there are other ways the resolution could be false as well. The negative could prove, less ambitiously, that the possession of nuclear weapons is neither morally good nor bad; it is morally neutral. This conclusion logically implies the falsehood of the resolution. But the resolution is also false (it seems) if morality is a myth-if nothing is morally good, bad, or even neutral. For if no actions have moral properties, it is obviously false that any particular action (such as the possession of nuclear weapons) has the property of being immoral. So even on this simple resolution, there are several routes the negative could take to fulll its burden of proving the resolution false. Now consider a more complicated resolution, that "As a general principle, individuals have an obligation to value the common good above their own interests." People often refer to resolutions of this sort as "comparative," and it is comparative resolutions that generate the most controversy about negative burdens. Let us again use p to represent the resolution. What counts as ~p? Many people believe that the negative's burden on such a comparative resolution is to prove the converse of the resolution-in this case, to prove that individuals ought to value their own interests above the common good. And no doubt, if individuals ought to prefer their own interests to the common good, then it's false that individuals ought to prefer the common good to their own interests. So proving the converse is one way for the negative to satisfy its burden to show that ~p.
11

But there are other ways to show that ~p. Suppose the negative could prove that what an individual should prefer varies from situation to situation and from individual to individual. Then it would be false that as a general principle individuals ought to prefer the common good to their own interests. There is nothing incoherent about this position, and it has been defended by many philosophers. Or suppose the negative could prove that there is a general principle, but that the principle is that individuals ought to value their own interests and the common good equally. I think this principle would be difcult to defend, but if it's true, then the resolution is false, and the negative has yet another way to satisfy its burden. The position on negative burdens which I am now defending is a reversal of an earlier position. I used to believe that the negative had to defend the converse of a comparative resolution. I still believe that defending the converse often makes for the strongest negative positions, but I no longer believe that the rules of LD or the ethics of argument require the negative to defend the converse. My present position is likely to please no one. Some people will think it allows too many negative options, including "balance negatives," "situational negatives," or "critical arguments" (I put these terms in quotation marks because I am often unsure what people mean by them). Other people will think my view is too restrictive by requiring the negative to defend a proposition rather than simply attacking the afrmative's arguments. Of course, I think my view is just right. More specically, I think it best accounts for two relevant facts: The rst fact is that there is no prescribed burden of proof and no presumption in LD. This fact requires that the negative (if it wants to secure a win rather than just a tie) must prove a proposition, namely, the proposition ~p which is the denial of the resolution p. The second fact is that depending on what p says, there may be many ways to prove ~p-including more ways than simply proving the converse of a comparative resolution. I am not equally enthusiastic about all of these ways, but I think anyone who understands the logical signicance of denial must at least be open to them. So in sum we can say that the afrmative's basic burden is to prove that the resolution is true and the negative's basic burden is to prove that the resolution is false. Depending on the resolution, there may be many ways to prove the resolution false. But if we use p to represent the resolution and ~p to represent "it is false that p," then the afrmative case should make arguments the conclusion of which is p, and the negative case should make arguments the conclusion of which is ~p. II. Research There are at least two sorts of research relevant to constructing a debate case. The rst sort is general research conducted for the purpose of nding arguments to prove a resolution true or false. The second sort is focused research to support this or that
12

premise of an argument which you have already constructed. Here I am concerned only with research of the rst sort. It is best to begin research on every topic by reading several general survey chapters and articles before hunting for specic arguments. These surveys will often be entries in reference books, literature reviews in journals, chapters in textbooks, or introductions in monographs (books by a single author) and edited anthologies. In addition to suggesting further sources to pursue, this reading will acquaint you with the history of the controversy, the main areas of dispute, and the relevant terminology. Once you move from such background sources into sources which are actually defending a position on your subject, you should begin reading for valid arguments. First identify the conclusion the author of the source is defending. If there are several conclusions, identify the one which is most relevant to the resolution you're debating. Rarely will useful sources on a given topic be defending or denying exactly your resolution. Instead, they may be addressing a closely related area. For example, if you are searching for arguments on the resolution that in the U.S., the use of race as a deciding factor in college admissions is just, you may run across Matt Cavanagh's recent book Against Equality of Opportunity. Cavanagh is writing about the just distribution of jobs in Great Britain, but many of his arguments could be applied to the debate over U.S. college admissions. You should study any argument you think might be relevant to your resolution, even if its conclusion does not match the resolution perfectly. When you nd a relevant argument in prose, you should apply your understanding of formal logic to analyze the argument into numbered premise-conclusion form. This process is difcult to master, but it will greatly enhance your ability to understand (and, eventually, to attack and defend) arguments. In the course of explaining valid argument forms, we have used italicized letters (p, q, and so forth) to symbolize propositions. When you are formalizing real arguments, do not use letters to stand for propositions. The purpose of formalizing an argument is to grasp its substance-exactly what premises support a particular conclusion-so you should write out the relevant premises and conclusion in normal English. Most authors will not express their premises in conditional (if-then) sentences, so you will need to paraphrase in order to display the argument's logical structure. And since most authors do not state premises they assume will be obvious to their audiences, you will also have to make these suppressed premises explicit. That's OK as long as you are careful to make your reconstruction of the argument faithful to the author's original ideas. Your goal is simply to understand exactly how the author's argument works. As suggested above, it is easiest to begin this analysis by identifying the argument's conclusion. It is especially important at this stage to record the author's precise conclusion rather than your debate resolution (or its negation), if the two differ. If the argument's conclusion doesn't quite match your resolution, you can think about how
13

they might be reconciled at a later stage. But at this stage you will distort the author's argument if you try to shoehorn it to t your immediate debate needs. Once you have identied the conclusion, work backward to nd the other premises. For example, if the conclusion is the simple proposition r, then look next for a premise which you can express as the conditional qr. Does the author also straightforwardly assert q? If so, you have a complete MP argument. But perhaps the argument is more complex, and the author derives q only from the further premises pq and p. As we saw in the rst article, a valid argument may contain any nite number of iterations of MP, MT, or HS. Fortunately, the formal structure of most real arguments is relatively short; but rarely will an argument for an interesting conclusion be a single instance of MP, MT, or HS, so you should be prepared to encounter a bit of complexity. Often you will nd that the premises an author offers for a conclusion fail to support that conclusion validly. For instance, you may wind up with an argument of this form: (1) pq (premise) (2) qr (premise) (3) rs (premise) (4) So ~p (conclusion) Since (1)-(3) are conditionals, all you can derive validly from them is another conditional: ps. But the author has apparently jumped from this conditional to the conclusion ~p. When you nd what appears to be a logically awed argument, your rst impulse should be to ask whether you have accurately and completely represented the argument. Check your source very carefully to make sure you haven't overlooked or misinterpreted anything the author has said. If you remain convinced that you have formalized the argument fairly and completely, then ask yourself what, if anything, could be added to the argument to make it valid. In our current example, an additional premise ~s, when combined with the conditional ps, would license us (by MT) to infer the author's original conclusion ~p. Sometimes authors will deliberately take premises for granted if they believe them to be obviously true. This is more common in popular and journalistic writing than in scholarly writing, but it happens everywhere. For instance, a person might argue that gun control is ineffective from the premise that if gun control were effective, cities such as Chicago and New York which have strict gun control laws would have less gun violence than areas with no gun control laws. The defender of this argument may just take it for granted that cities such as Chicago and New York do not have less gun violence. If you nd such an implicit (or "suppressed") premise which is needed to make an argument valid, go ahead and write it into your list of premises, noting in parentheses that it is assumed or implicit rather than expressly stated by the author. Occasionally you will nd a published argument which is by itself invalid and which cannot be patched up by adding an obviously suppressed premise or two. For instance, you may discover that an author is denying the antecedent, inferring ~q from the
14

premises pq and ~p. This discovery is obviously well worth making. It allows you to eliminate a possible case argument from consideration, and it also alerts you to the fault in an argument your opponents may try to use. You should not, however, suppose that because a particular argument for a conclusion is invalid, the conclusion is false. All you have discovered is that one way of trying to reach the conclusion-the way taken by the invalid argument-does not succeed. There may well be (in fact, certainly are) other, valid ways of reaching the same conclusion. The more you research, the more arguments you can nd and formalize. Formalizing will help you to understand the arguments much better than you would if you just read over them in their original wording. It will often be harder than you initially expect, but you will probably feel a sense of accomplishment and mastery when you have captured an argument and laid bare its structure. You will be in a much better position to compare various argument possibilities and construct the arguments you will defend in your case. With enough practice, you can learn to formalize arguments in your head as you read or hear them, vastly increasing your power as a debater. III. Contention Mapping When you have read and analyzed enough arguments to give you a good feel for the main options on each side of the resolution, you are ready to begin mapping case contentions. This process can be very difcult (although it will be much easier if you have carefully analyzed the arguments from your research), but the guiding principle here is simple: Each contention should be a single valid argument which concludes in the resolution or its negation. If your research has gone well, or if you have a strong imagination, you should have more argument possibilities on each side than you can use in a single case. For most debaters, novice and varsity, it is best to map out all of the possible constructive arguments for each side. (An argument is constructive just in case its conclusion is the resolution or its negation.) You can make the most informed choices between possible arguments when you can compare them in the kind of detail a formal premiseconclusion outline reveals. Here, too, it is safest and easiest to start with the conclusion and work backward. The conclusion of every constructive argument on a given side will be the same: the resolution (if you are afrmative) or the denial of the resolution (if you are negative). What comes next, "after" the conclusion (if you work backward)? A conditional (if-then) premise which has the conclusion as its consequent. I.e., if your conclusion is the proposition q, the rst premise you construct should be of the form pgq for some proposition p. The trick is to gure out what proposition to use for the antecedent of this conditional. Suppose, for example, that you are mapping afrmative arguments for the resolution that "The United States is morally obligated to mitigate international conicts." You have come across several authors who argue, in various ways, that the U.S. should mitigate
15

conicts because the U.S. has promised to do so. How can you map the argument? You know the conclusion will be simply the resolution. And you know that the rst premise you map, working backward, will be a conditional statement which says: If p, then [resolution]. What should you ll in for p? Given your initial argument idea, the obvious candidate for p is "the U.S. has promised to mitigate international conicts." So your complete premise is, "If the U.S. has promised to mitigate international conicts, then the U.S. is morally obligated to mitigate international conicts." So far, so good. What more is needed for a valid argument? You cannot conclude q from pq alone. You need at least one further premise, p, added to your conditional to give you a valid instance of modus ponens. So you need to add as a premise, "The U.S. has promised to mitigate international conicts." If you stop at this point, you have mapped a single valid argument: (1) The U.S. has promised to mitigate international conicts. (premise) (2) If the U.S. has promised to mitigate international conicts, then the U.S. is morally obligated to mitigate international conicts. (premise) (3) So the U.S. is morally obligated to mitigate international conicts. (conclusion) There is nothing logically wrong with this argument as mapped. However, you may nd it useful to supplement the argument by adding more premises which clarify the details of the argument as you plan to defend it. The easiest way to supplement an argument is to ask yourself what you would say to an opponent who asked, about any of your premises, why that premise is true. You can usually answer this question by adding two or more new premises to support your original premise. What if someone asked why the rst premise of the argument above is true? Presumably the research that inspired the argument has given you a detailed idea of how exactly the U.S. promised it would mitigate international conicts. Perhaps the U.S. signed a treaty (we'll call it "Treaty T") in which it promised to "use all reasonable means, diplomatic and military, to restore peace to war-torn regions of the world." Assuming you have good evidence (in this case, a copy of Treaty T) to support your claim, you could represent your original (1) as the conclusion of further premises, like so: (1') The U.S. signed Treaty T.(premise) (1'') If the U.S. signed Treaty T, then the U.S. promised to "use all reasonable means, diplomatic and military, to restore peace to war-torn regions of the world." (premise) (1''') If the U.S. promised to "use all reasonable means, diplomatic and military, to restore peace to war-torn regions of the world," then the U.S. promised to mitigate international conicts. (premise) (1) The U.S. promised to mitigate international conicts. (premise, from 1', 1, 1') (2) If the U.S. promised to mitigate international conicts, then the U.S. is morally obligated to mitigate international conicts. (premise) (3) So the U.S. is morally obligated to mitigate international conicts. (conclusion)

16

Now you can see how you might supplement a simple premise p or q, such as premise (1): by making it the conclusion of further applications of modus ponens (or modus tollens). When you do this, the original premise remains part of the argument, but you justify it by making explicit the premises supporting it. What about a conditional premise like (2)? How might you support that premise? Again, think about what you would say if someone asked you why (2) is true. To supplement a conditional premise, you will need to add two new conditional premises to the argument. Here is the recipe: Write two new conditionals. Make the antecedent of the original conditional the antecedent of the rst new conditional. Make the consequent of the original conditional the consequent of the second new conditional. Finally, nd a new proposition to serve as both the consequent of the rst new conditional and the antecedent of the second new conditional. So you are supplementing a statement of the form pq with two statements of the form pr and rq. These two new statements yield the original pq by an application of hypothetical syllogism. Let's look at how this recipe works in practice. Think about the question, "Why believe that if the U.S. promised to mitigate international conicts, then the U.S. is morally obligated to mitigate international conicts?" This question really amounts to something like, "Why is there a moral obligation to keep promises?" To answer the question, you are probably going to have to appeal to a moral theory. Fans of strict promise-keeping tend to like Kant's theory, so we will suppose that's your theory of choice. According to Kant, what is wrong with deliberately breaking a promise is that doing so cannot consistently be willed as a universal law-i.e., breaking promises violates Kant's so-called categorical imperative. Given Kant's reason for thinking that we are morally obligated to keep promises, we can supplement premise (2) by adding: (2') If the U.S. has promised to mitigate international conicts, then the U.S. must mitigate international conicts if it is to act in accord with Kant's categorical imperative. (2'') If the U.S. must mitigate international conicts if it is to act in accord with Kant's categorical imperative, then the U.S. is morally obligated to mitigate international conicts. The two new premises (2') and (2'') justify (2) in our argument, making more explicit the thought behind that original premise. In an actual contention, as opposed to this argument map, you would want to explain in more detail just what Kant's categorical imperative is, why anyone should care about it, and why it requires promise-keeping. Expanded in the way we have now suggested, what began as a two-premise argument now has seven premises: (1'), (1), (1'), (1), (2'), (2''), and (2). Both the short and longer versions are equally valid. But the longer version will be more useful when you come to writing the actual contention, because it makes the content of the argument more clear
17

and detailed. We could go on asking the same "Why?" questions of the new premises in our longer argument and supplement the argument even further. We won't do that here, assuming you see roughly how the expansion process works. Keep expanding your map of an argument until the thought behind the argument would be clear to someone who carefully reads through the premises and conclusion. At this stage, you have found (or created) an idea for a constructive argument and have mapped that idea into valid premise-conclusion form. Your next task is to classify the premises of your mapped argument to get some sense of what kinds of support you would need if you were to adopt the argument as part of your case position. We can think of premises as falling into three broad categories: empirical, normative, and denitional. Empirical premises are claims which could, in principle, be tested scientically or with the physical senses. That January is colder than June, that Elvis is alive in Arizona, and that free trade hurts American wages are examples of empirical claims. (The rst is true; the second, false; and the third, contested.) Debaters should especially note that all causal claims (claims of the form "X causes Y" or "if X occurs, then Y subsequently occurs") are empirical claims. Normative (or evaluative) premises make claims about what is good or bad, moral or immoral, valuable or valueless. They are not typically subject to scientic or sensory validation. Examples include that cheating is wrong, that justice requires economic equality, and that citizens ought to vote. Almost all LD resolutions are themselves normative claims. Denitional (or analytic) claims are true in virtue of the meanings of their words; they are not substantive descriptions of any chunk of physical or moral reality. That all aunts are female, that every whole number is either odd or even, and that a treaty is a promise are all analytic claims. At most, they inform someone of the structure of our concepts and language. You might think that denitional claims are always uncontroversial, but this is not so. Is it true that human genetic engineering is, by denition, the curing of genetic diseases? This is a very controversial claim, and a debater lost the nal round of the Glenbrooks tournament several years ago when he built an argument on this premise but could not defend it with an actual denition from a legitimate source. Many philosophers now believe such distinctions among types of propositions are misleading. However, I believe they are useful for thinking about the premises of debate arguments, and I hope you have a rough feel for what kinds of claims fall under each category. Sometimes it will be hard to tell into which category a given premise falls. But even if a premise straddles two of the categories, it can be helpful to know at least that it denitely does not belong to the third. As an exercise, let's run through the seven premises of our last argument, classifying them by type. If you were doing this on paper or on a blackboard, you could simply write an "E," "N," or "D" next to each premise when you knew its type.
18

(1') The U.S. signed Treaty T. This premise is clearly empirical. It is a simple assertion of historical fact and should be conrmable by checking the relevant documents. (1'') If the U.S. signed Treaty T, then the U.S. promised to "use all reasonable means, diplomatic and military, to restore peace to war-torn regions of the world." This premise straddles the fence between empirical and denitional. On the one hand, we would probably need to nd documentary evidence of just what Treaty T said in order to assess the truth of the premise. On the other hand, the premise could be taken as simply dening (in part) the content of Treaty T-signing Treaty T includes that promise because that's just what Treaty T is. (1''') If the U.S. promised to "use all reasonable means, diplomatic and military, to restore peace to war-torn regions of the world," then the U.S. promised to mitigate international conicts. This premise is clearly denitional. It simply translates the language of the treaty into the language of the resolution. (1) The U.S. promised to mitigate international conicts. This premise does not fall under any of our three headings, because it is not, strictly speaking, a premise. It is a subconclusion of premises (1), (1), and (1). It does not aim to supply new information but simply helps the audience grasp the logical/ conceptual relationships among other premises. But if (1) were not a subconclusion and instead functioned as a free-standing premise, it would clearly be empirical, because it is a claim about a veriable action the U.S. has taken in history. (2') If the U.S. has promised to mitigate international conicts, then the U.S. must mitigate international conicts if it is to act in accord with Kant's categorical imperative. Classifying this premise is tricky because of the intricacies of Kant's philosophy. If Kant had asserted (2'), he probably would have meant it as a denitional truth: Once a person understands the concept of a promise, he will be able to see that it would be logically impossible to will as a universal law that promises be broken. (2'') If the U.S. must mitigate international conicts if it is to act in accord with Kant's categorical imperative, then the U.S. is morally obligated to mitigate international conicts. This premise is clearly normative. It makes a substantive claim about the content of moral obligations, namely, that a nation is morally obligated to act in accord with the categorical imperative.

19

(2) If the U.S. promised to mitigate international conicts, then the U.S. is morally obligated to mitigate international conicts. Like (1), (2) is no longer really a premise, because it is now a subconclusion from (2') and (2''). But if (2) were a free-standing premise, it would clearly be normative. (3) So the U.S. is morally obligated to mitigate international conicts. (conclusion) The conclusion is, obviously, a normative claim. It's the resolution, and LD resolutions (as we've said) are normative claims. You should classify the premises of all the possible case arguments you map. When you think about which one or two (never more than three) arguments to build into your case, you should compare the varying empirical, normative, and denitional burdens you'd be assuming by adopting each argument. It will not be enough in the case merely to state your premises and explain how the conclusion logically follows. You will also have to persuade your audience that each premise is true, and some premises are easier to sell than others. Any empirical premise is going to need empirical evidence to support it. That means if you don't have good evidence to back a premise at the time you rst map an argument, you will have to nd good evidence if you decide to include that argument in your case. LD students are especially prone to think that causal claims can be supported "logically" rather than empirically. This is false. Logic does not have the resources to prove causal statements about events in the world. (If you do not believe this, read Section IV of David Hume's rst Enquiry.) When someone says that "it's just obvious" that one type of action or event will cause another type of action or event, all that person really means is that it seems obvious to him. There is no way to know what actions will cause what effects without empirical investigation. Particularly when large-scale social phenomena are at play (as they often are in LD resolutions), armchair predictions about causes and effects have a very bad track record. You must have expert evidence that the world really works the way you claim it does if you expect your empirical premises to be credible to your audience. Good empirical evidence is hard to nd, and the more empirical premises an argument has, the more evidence you will need. Because consequentialist arguments depend on the actual empirical results of an action to determine its morality, such arguments will typically require more empirical evidence than deontological or other nonconsequentialist arguments require. Many debaters think that consequentialist arguments are somehow more objective and therefore easier to defend than are deontological arguments. But there is nothing in the structure of consequentialist arguments that should lead you to believe that. Any valid argument for a normative conclusion is going to contain at least one normative premise, and this is just as true of valid consequentialist arguments as of others. The only difference between consequentialist and non-consequentialist arguments on this score is that the former add controversial empirical premises to their controversial normative premises.
20

Debaters who wish to minimize their reliance on empirical evidence may want to avoid consequentialist arguments, especially consequentialist arguments with multiple causal premises. Notice that our sample argument above has an empirical premise (1') even though it is not a consequentialist argument. However, you should be able to see immediately that not all empirical premises are created equal. Premise (1') requires evidence, but the evidence will be an uncontroversial report of a public event, the signing of a treaty. This kind of empirical evidence is much easier to nd and to defend than is evidence on (say) the climatic effects of carbon dioxide or the deterrent effects of capital punishment. Normative premises will not require empirical evidence, but they will typically require persuasive explanation and perhaps a vivid illustration. Supporting a normative premise is likely to require more hard thought than any other element of an argument. So understand, when you compare argument options, that you will have to defend your normative premises in the case. Denitional premises vary greatly in the kind and amount of support they require. Sometimes appeal to a standard dictionary (or even to the linguistic intuitions of your audience) is enough support for simple denitional claims. At other times, especially when a claim concerns technical or otherwise unfamiliar concepts, you will need expert evidence. Finally, if a "premise" is really just a subconclusion of previous premises, you might want to explain how it follows so that your audience can fully grasp the logical connections you are claiming. Since this is not an article on casing per se, I will not go into detail on how the various types of premises should be supported (although I offer a few tips in Part V below). But it is important, before selecting your case arguments, to understand what kinds of premises each argument includes and to have a rough sense about what kinds of support each type of premise will require. If you end up choosing more than one argument for your case, check to be sure the premises of all your arguments are compatible. You do not want to defend a premise in one argument which you deny a few paragraphs later in a different argument. Be especially certain that the normative premises of all your arguments more or less tally with each other. It is not necessary to have the same normative premises in each argument (although this is often possible, and a good idea). But make sure you are not endorsing incompatible moral theories at different points in your case, and that you can explain, if asked, how a person could hold all your normative premises as part of a coherent moral outlook.

21

IV. Value Premise(s) In many regions, LD cases are expected to begin with a statement of a "value premise" and "criterion." Each of these items is expressed by a single word or phrase; together, they are supposed to provide standards by which judges should evaluate subsequent arguments. Sometimes the value premise is drawn straight from the resolution, as when the resolution mentions morality or justice. At other times, the value premise is chosen by the debaters as the single most relevant axis of evaluation for a generic "ought" statement. The criterion is typically thought of as a more detailed specication of the content of the value premise. For example, if a debater proposes a value premise of morality, she might go on to propose a criterion of maximizing the balance of pleasure over pain. The point would then be that the resolution is true if the course of action it prescribes is moral, and that that course of action is moral if it maximizes the balance of pleasure over pain. This is only a rough sketch of a very confused and confusing area of LD theory. I am going to challenge the traditional value premise/criterion model, but I want rst to acknowledge its merits. It is the model I was taught and used as a debater. It has served a valuable function by imposing order on cases that might otherwise have been unshapely piles of unrelated points. It can give judges a helpful focus for assessing the relevance of various claims at the end of a round. I have myself written a sympathetic exposition of the model (see the November 1997 Rostrum). However, the model has many shortcomings. First, it is notoriously hard to say precisely how the value premise is related to the resolution and how the criterion is related to the value premise. Debaters often speak of the value premise as "supporting" the resolution or of the criterion as "fullling" the value premise, but it is hard to know just what these descriptions mean. Second, because value premises and criteria are nouns or noun phrases, it is often unclear how they connect with the claims unique to either side of the resolution. Suppose the resolution is that "The possession of nuclear weapons is immoral." If the afrmative says, "My value premise is morality," what should this lead us to expect from his arguments? Will he be claiming that someone or some action is moral? The resolution would lead us to think the afrmative will be claiming that something is immoral. If so, why is the afrmative's value premise morality? If the afrmative succeeds in proving the possession of nuclear weapons immoral, he won't have proven that any person or action is moral. Instead, he will have shown that a certain action is not moral. So in what sense will anyone have "fullled" or "upheld" morality? And who would have done it-the afrmative debater, or the (non)possessors of nuclear weapons? A third problem with the traditional model is that it does not distinguish between necessary and sufcient conditions for satisfying a value. Suppose the value premise is morality and the criterion is respecting innocent life. Does that mean that if an action is moral, then it respects innocent life? Or does it instead mean that if an action respects innocent life, then it is moral? In the former case, respecting life would be a necessary condition of morality; in the latter, a sufcient condition. The simple terms that compose
22

a traditional value premise and criterion obscure this important logical distinction. A fourth problem, and perhaps the central problem, with the traditional model is that it separates the value premise and criterion from the contentions, where the substantive arguments are thought to be. Since many debaters are confused about how the value premise and criterion are supposed to function and they have what seem to be selfcontained arguments in their contentions, they end up ignoring the value premise and criterion and focusing only on contention arguments. Defenders of the traditional model will say (rightly) that this results from a misunderstanding of the model. However, that misunderstanding is widespread enough that the model is falling out of use among people who nd it too unwieldy to apply. These disadvantages give us reason to search for a better way of organizing LD arguments. To anyone wedded to the traditional model I have been criticizing, it should be a striking fact that moral and political philosophers who write about some of the same issues debated in LD never appeal a value premise or criterion (in the LD sense). Instead, their arguments resemble the formally valid sets of premises and conclusions we have examined above. The key point to notice is that there are no missing steps in a formally valid argumentthere is nothing left that a value premise or criterion could add. The premises of a valid argument are sufcient to prove its conclusion, and if that conclusion is the resolution or its denial, the argument proves enough for LD purposes. I suggest that LD cases be composed of valid deductive arguments with no logically independent value premises or criteria. So the model I am proposing really amounts to the model of deductive argument that philosophers have been using for centuries. This model avoids the four problems with the traditional value premise/criterion model I outlined above. I will not explain in detail how, but interested readers can examine a valid argument in the light of each problem and see that the problem does not arise for the argument. Do I suggest abandoning all talk of value premises? No, because many judges are expecting to hear debaters offer value premises and may penalize debaters who do not. (It may also be worth noting that the NFL's ofcial "Lincoln-Douglas Debate Guidelines" state that LDers should establish value premises and values criteria. They go on to dene values criteria as "a system upon which to judge values," but they do not say anything very helpful about what that means.) Fortunately, the valid argument model gives us a ready candidate for a revised version of the value premise that is not logically irrelevant to the constructive arguments. Where r is the conclusion of a constructive argument, the value premise will generally be the conditional premise in which r appears as consequent (if the argument employs modus ponens) or in which ~r appears as antecedent (if the argument employs modus tollens). Because our sample argument above employed modus ponens, the value premise would be (2), since this is the premise whose consequent is the conclusion (the resolution) which we are proving.

23

A major difference between this model and the traditional model is that the value premise is now a complete sentence rather than a noun or noun phrase. This is all to the good, since a premise is, by denition, a sentence or proposition. In fact, it is mysterious how a word or phrase could ever have come to be called a value premise in the rst place. While the new value premise avoids the problems with the old model, it can still serve a similar function if it appears near the beginning of a case or argument. It still provides the audience with a standard or goal to use in evaluating the success of the argument. It tells listeners when they will have a sufcient reason to conclude that the resolution is true-namely, when the debater has shown the antecedent of the value premise to be true (or, in the case of a modus tollens argument, the consequent of the value premise to be false). The fact that this value premise is pulled straight out of a valid argument also calls attention to the fact that it must be explained and defended just like any other premise if the argument is to succeed. Many debaters using the traditional model simply announce a value premise and criterion without really justifying their selection. It is easy to get by with this if the value premise and criterion play no essential role in the structure of one's arguments. But in the new model, constructive arguments will not be convincing unless the value premise is also convincing. This premise will always be a normative premise, and it will typically be the most important and controversial normative premise in an argument. Therefore, the value premise will often merit extensive explanation and support in its own right. Depending on the structure of the argument, it may be appropriate to spend more time arguing in support of one's value premise than in support of any other single premise of one's argument. Establishing this premise will often be the most philosophically challenging task of an LD case. This model makes better sense of the importance many judges place on "the standards debate." These judges want to see debate over conicting normative premises, and the new model is well suited to promote just such debate. A case that contains more than one independent argument may have more than one value premise. If separate case arguments require separate value premises, each premise should be stated and defended separately. Many traditionalists will view the possibility of multiple value premises in a single case with suspicion, and I share their concern. A case with multiple value premises may end up sounding incoherent or even inconsistent. I am inclined to think that for most students most of the time, developing arguments under a single value premise is the best strategy. I have found that LD students invariably regard themselves as exceptions to such generalizations; to any students reading this, I can only say, I mean you-you should probably try to construct cases with a single value premise. Recognizing the rare situations when multiple value premises might be appropriate requires a skillful sensitivity to the particulars of combinations of arguments which I cannot impart through this article. Here is the extent of what I can say in general terms: Multiple value premises are likely to work best in cases which seek to prove that some practice is immoral or not justied. There are many ways for actions to go wrong, so distinct value premises might identify distinct sufcient conditions for an action's
24

immorality or unjustiability; each independent contention could then prove the action immoral or unjustied by showing it failed to satisfy the requirements stated in one of the various value premises. But, as I say, it is usually safer and clearer to choose contention arguments that work together under a single value premise. I have proposed a revised understanding of the value premise, but is there any place for a revised version of the criterion in my model? Yes and no. Yes, in the sense that value premises of the kind I am recommending often function as criteria: they give judges standards or goals by which to measure an argument's success. In fact, often the newstyle value premise will contain a phrase that, by itself, could function as an old-style criterion. For example, in (2) above, the notion of keeping promises, which occurs in the antecedent, might serve as a criterion for morality, which occurs in the conclusion. When the resolution itself mentions a value term (such as "morally obligated" in our example), the complete new-style value premise will contain both the old-style value premise and the old-style criterion. But when the resolution does not mention a value term, there may be two normative premises, one of which is naturally treated as the value premise and the other of which is naturally treated as the criterion. Suppose, for example, that instead of directly mentioning moral obligation, our resolution said simply that the U.S. ought to mitigate international conicts. Now what had been a single premise (2) may be split into two separate premises: one which says that if the U.S. is morally obligated to mitigate conicts, then the U.S. ought to mitigate conicts, and one which says that if the U.S. promised to mitigate international conicts, then the U.S. is morally obligated to mitigate international conicts. The rst of these premises might be called the value premise since it proposes moral obligation as the relevant ought-making relationship. And the second might be called the criterion since it proposes promise keeping as a necessary condition of satisfying one's moral obligations. In situations of the rst kind I discuss, where a value term appears in the resolution, I believe proposing a criterion distinct from the value premise would be unnecessary and confusing. The value premise already contains the term which would have been treated as the criterion in the old paradigm, so nothing of substance is lost by simply dropping talk of the criterion. In these cases, the criterion is simply part of the value premise. However, in situations of the second kind I discuss, where the resolution does not mention a value term (typically using an open-ended "ought" instead), it may be wise to retain a distinct criterion. In these cases, the value premise and criterion will each be a complete conditional premise of the argument. Where r is the conclusion (the resolution or its negation), qr will be the value premise, and pq will be the criterion. If this is not immediately clear to you, study the example from the paragraph before last to see how such new-style value premises and criteria might be related. Like so much else in debate, deciding whether to retain a distinct criterion requires careful judgments about particular resolutions and arguments. Finally, is there a way to present genuine premises as value premises and/or criteria
25

without alienating traditional judges who expect words and phrases rather than sentences? I believe there is. A debater may simply pick the most value-laden word or phrase from her value premise or criterion and state this term as the (old-fashioned) value premise or criterion. She may then follow this word or phrase with a complete statement of the full premise. For example, a debater might say, "My value premise is respecting gender equality. In other words, if pornography disrespects gender equality, then government censorship of sexually explicit materials is justied." This statement should then be followed by an explanation and defense of the premise. V. Contention Development The toughest thought-work of case construction is now complete: extensive research has uncovered numerous argument ideas, and the best of these have been formally mapped and analyzed to show exactly which premises are required to sustain the conclusion and also what kinds of support will be needed to sustain the premises. One, two, or at most three of these mapped arguments have been selected for inclusion in the constructive speech itself, and authentic value premises have also been identied. The work that remains to transform these skeletal arguments into a presentable LD case falls largely into the domains of rhetoric and composition-subjects we will not take up here. But the logic-centered model of argument does have a few implications for this nal stage of casing which we should notice. The rst implication is one at which I have already hinted: Each contention should be a distinct valid argument concluding in the resolution or its negation. This means that if only one really strong constructive argument emerges from your research and brainstorming, you should expect to have a one-contention case. Some LD traditionalists have a fetish for two- or three-contention cases and will split a single argument into more than one contention if necessary. I cannot see what is gained by this practice, and I have often seen something lost. Students instinctively think of their contentions as free-standing argument units, and when pressed for time, they will sometimes abandon one contention (usually the last of two or three) in order to defend another. In the case of truly independent arguments, this strategy, while not ideal, need not be disastrous. But in the case of a single argument spread out over several contentions, the maneuver is fatal. No one contention in such a case provides a selfcontained reason to believe or disbelieve the resolution. If one contention falls, the others are worthless. If that same single argument had been presented as one contention instead of two or three, the student would have been less likely to think of any part of the argument as dispensable. Of course, the ideal is to present two or three truly independent arguments to afrm or negate the resolution. It is worth spending a lot of time in research and planning to try to discover two or three such arguments. But sometimes there really is only one excellent constructive argument, and to include another one or two mediocre arguments in the case would dilute the power of the strong argument. Since different arguments will appeal to different people, you should run your contention candidates by a variety of peers and adults (including plenty from outside debate) before you narrow your case to
26

a single argument. It will often be worth including a second or third argument which you personally nd uninspiring if other people nd it persuasive. The second implication of logic-centered argumentation for case writing is that each premise of a valid argument should be explicitly stated and defended. The key normative premises may be stated and defended as value premises before the contentions proper begin. But each premise should nd its way into the case somewhere. A premise need not be stated exactly as it appears in the argument map. To ensure formal validity, the argument map forced you to state most premises as if-then conditionals. Most such conditionals can be phrased more naturally in your case as non-conditionals. For example, consider this premise of our specimen argument above: (1'') If the U.S. signed Treaty T, then the U.S. promised to "use all reasonable means, diplomatic and military, to restore peace to war-torn regions of the world." No one would really talk like this in a persuasive speech. You could communicate the same point of the conditional (1'') with this non-conditional sentence: By signing Treaty T, our nation promised to abide by all of its provisions, including the requirement that we, quote, "use all reasonable means, diplomatic and military, to restore peace to war-torn regions of the world." Anytime you rephrase a premise from your argument map, you should be very careful that your rephrasing communicates exactly the same thought as the old premise. You will have preserved that equivalence of meaning if, roughly, the old premise is true only if the new premise is true, and the new premise is true only if the old premise is true. Note that premises should be stated and defended. Occasionally an argument will contain an especially obvious premise that needs no defense, other than one or two restatements in different words to make the point clear for listeners. But most important premises, including denitional premises, are controversial to one degree or another and should be defended. A crude rule of thumb is that each premise should receive a paragraph's worth of explanation and defense. This rule is crude because a few premises (as we have noted) will require little or no defense, whereas some premises may require many paragraphs to establish. Sometimes defending a premise requires constructing a mini-argument of which the premise is the conclusion. These further arguments can normally be incorporated into expansions of the argument map, as illustrated above. At this point, clever readers may wonder if it's really possible to defend every premise. If a premise is defended with a further argument, won't that argument itself contain premises which must be defended, and so on, to innity? This reaction is correct-it is impossible to deductively prove every premise on which an argument depends. At some point, the formal arguments must end, and the base-level premises must be treated as
27

postulates (OK if they are obviously true) or motivated by non-argumentative appeals (e.g., appeals to moral intuitions or emotions). Some people think LD should be about "pure logic," but that is not possible as long as LD is supposed to yield substantive conclusions. To get off the ground, every argument must begin with some premises that are simply asserted. No substantive conclusions follow from logic alone; this is just a fact about logic, not a shortcoming of LD. Empirical premises will require empirical evidence, although you will keep a better grip on your audience's attention if you explain most of the empirical background in your own words and quote outside authors only briey to conrm crucial details or conclusions. Reading long quotations is never impressive. It suggests the speaker has simply lifted his arguments wholesale from other sources and may not really understand the research. Whenever you use empirical evidence, you should sandwich it between your own explanation of its source, content, and signicance for the argument. Evidence is not self-interpreting. You must be scrupulously accurate in the ways you characterize the claims of your cited authors. It is tempting to exaggerate a source's limited claim so that it better supports the sweeping claim of your premise, but doing so is dishonest, and responsible judges will not reward the practice. Normative premises will require non-empirical support. Sometimes quoting a particularly clear or persuasive source (such as a philosopher) can bolster such a premise, but often you will have to resort to your own persuasive powers. One of the classic ways in which philosophers argue for their own normative premises is by the use of examples or analogies. Most of the normative principles that might apply to a given resolution are broad enough to apply to other issues as well. One of the traditional purposes of moral and political philosophy is to furnish a relatively small number of principles that will apply to many different situations. So chances are, the normative principle to which you appeal has a wider application. Pick some obvious non-resolutional situation to which the principle might apply. Describe a possible moral decision in that situation about which you and your audience will share moral intuitions and about which your princple gives the proper outcome-the outcome that matches everyone's intuitions. This nonresolutional example may then count as some evidence that the normative principle is correct. For denitional or analytic premises, you will need, at a minimum, to explain in full detail just why the premise is necessarily true. When appropriate, you should also quote authoritative denitions to support your claim. Consider a premise as trivial as, "If John is a bachelor, then John is a man." You might defend this premise by saying, "A bachelor is, by denition, an unmarried man. All bachelors, therefore, are men, so if John is a bachelor, he's a man, too." This explanation did not cite authoritative sources, because it assumes that every adult English speaker knows the denition of "bachelor." But if the claim were about a more technical or obscure concept, such as genetic engineering, the Second Amendment, or globalization, then quoting an expert source that described the concept would probably be necessary. Ultimately, there is no formula for adequately explaining and defending each premise of
28

an argument. Distinguishing the three types of premises as we have can be some help, but each premise will require a unique treatment. The best way to see how different sorts of premises can be supported is to study how they actually are supported by scholars in the relevant disciplines or by high-quality public intellectuals. By thinking of contentions as arguments and premises as the basic units of contentions, you can avoid the confusion that debaters often inject into rounds by their attempts to break contentions into different "subpoints." It is often unclear just how the various subpoints are distinguished from one another and how they are functioning in a constructive argument. Anyone who wants to retain talk about "subpoints" in his or her cases should probably think of each premise as a distinct subpoint. I don't nd most explicit subpoint talk helpful or attractive, but people who do should conform such talk to the logic of their arguments. A third implication of the logic-centered approach for contention building is that premises must be presented in an order the audience will nd persuasive and easy to follow. You might think that the best order of premises for any argument will be the order you have used in your argument map, but this is not always so. The argument map itself represents only one valid formulation of the argument; the same premises could have been ordered differently and still have yielded the same conclusion. So the map does not settle much about the order of presentation. We have already seen that where q is the conclusion of your argument, the conditional premise pgq in which q occurs as the consequent will typically be the value premise that is appropriately presented before the other premises of the argument. Apart from LD convention, there is a good rhetorical reason to present this premise rst. By doing so, you are setting a xed target for yourself. Consider this simple modus ponens argument: (4) Capital punishment deters crime. (5) If capital punishment deters crime, then capital punishment is morally justied. (6) So capital punishment is morally justied.

Here, (4) is an empirical premise, while (5) is clearly normative. If you presented the argument in just the order it is mapped above, some listeners might have the following reaction: "He just researched what capital punishment does, discovered it deters crime, and then cooked up an ad hoc moral principle to conclude that if capital punishment does whatever it does, it must be moral." In other words, it may look as if you have set a moving target for yourself-whatever capital punishment does, that's what you claim makes it morally justied. If instead you present (5) before (4), your approach looks more principled. It looks as if you have rst identied what quality makes a punishment, any punishment, morally justied, and then gone out into the world (so to speak) to see if capital punishment does, in fact, have that quality. The substance of your claims has not changed, but the argument may be more effective for some listeners.

29

As far as the other, non-value premises go, you will have to use your best judgment about the most effective order of presentation. Generally, you should be moving more or less sequentially through your argument map, but sometimes you may want to work from start to nish, other times from nish to start. Before you begin actually writing the contention, you may want to write each premise on a separate slip of paper and then experiment with different orderings to see which will make the argument most accessible to your audience. A fourth and nal implication of the logic-centered approach for case writing is that you must make the logical structure of your arguments clear to your audience. You cannot just assume that by presenting and defending a set of premises which do, in fact, entail an announced conclusion, you will have persuaded your audience that those premises really do entail that conclusion. Because you have spent much time researching and formulating the argument and have studied it in skeletal form, it is easy for you to see the logical relationships among its various parts. But your listeners are meeting the argument for the rst time when they hear your case, and they may not be accustomed to tracing the logical implications of different premises. The best way to ensure that your audience grasps the logical character of your argument is to call their attention to it at various points throughout your case. You do not need to use the jargon of premises, validity, modus ponens, and the like, nor do you need to retrace in explicit if-then form the full shape of the argument. But you do need to preview arguments before you make them and review them after you have made them. You can preview arguments near the beginning of your case or at the start of each contention. A suitable preview for the sample argument we mapped in Part III might be, "In my rst contention, I will show that the United States is obligated to mitigate international conicts because it has promised to do so in certain treaties." The equally brief review might be of all contentions together at the very end of the case. In addition to previewing and reviewing, you need to ag distinct premises and logical connections throughout each contention. You can do so by smaller-scale previews and reviews within the contention ("First, I will argue that the U.S. is morally obligated to keep its promises") and also by the use of logical conjunctions like "since," "because," and "therefore." The basic point is that as a speaker, you are responsible not merely to present logically valid arguments, but to make the validity of those arguments clear to your audience. The logical relations between your premises will be no more clear to many listeners than the truth or falsehood of individual premises. Just as you must persuade your audience to believe each premise of your argument, so you must persuade them that the premises, taken together, form a logical whole. We have covered a lot of ground in this article. A logic-centered approach to arguments has implications for interpreting burdens, conducting research, outlining contentions, framing value premises, and organizing contentions. Next month we will explore its implications for cross-examination and rebuttals.
30

Part III: Cross-Examination and Rebuttals

In earlier articles we have explored some basic elements of formal logic and the application of those elements to LD case development. Now we can consider the application of logic to cross-examination and rebuttals. These are the parts of a debate round where a working knowledge of logic can do most to transform a students performance from the ordinary to the outstanding. Students who have followed me this far may have the feeling that my advice here about rebuttals is less detailed than my earlier advice about (say) mapping case arguments. As the discussion moves deeper into the debate round, it becomes more difcult to offer helpful examples, because each example would require a description of an entire round, something space precludes. But I am condent that students who take the trouble to learn and implement the more detailed advice on formal logic and casing in the earlier articles will nd it fairly easy to apply the general advice I give here about cross-examination and rebuttals. In my rst Rostrum article (April 1994), I lamented the rise of what I called refutory debatedebate guided be nothing more than a reexive attempt to refute (usually several times over) anything and everything ones opponent says. I believe reexive refutation is still the norm in LD. The problem with such an approach, then and now, is that there is no real need to refute most of what ones opponent says, and the attempt to do so ruins the force and clarity of each individual response. It is all well and good to counsel students in the abstract to use a more selective strategy, hitting only the most important points. But the average student, confronted with such vague advice, may be unwilling to gamble on her own interpretations of what the most important points and most effective responses are. Instead, she will likely retreat to the mediocre comfort of peppering each of her opponents statements with as many objections as she can think up. Logic gives us the resources to be more helpfully articulate about strategies for selective, high-impact rebuttals. I. Basic Principles The key to making logic work for you in cross-examination and rebuttals is to interpret your opponents arguments (and, of course, your own) as formally valid premiseconclusion chains. This may sound obvious enough, but it can be surprisingly hard to do when you confront opponents who do not themselves have much understanding of the logical structure of their ideas and who therefore present them either in an undifferentiated stream of consciousness or in a logically confused series of subpoints. The jargon of points and subpoints is among the most deceptive at work in LD today. Many students and judges imagine they are clearing things up by labeling sentences as distinct points or subpoints, but these terms usually obscure the logical structure of a case. What is needed, of course, is an analysis of a case into distinct arguments, and of each distinct argument into its conclusion and supporting premises. Only in light of such analysis can a debater make condent, informed decisions about what to attack and
31

what to ignore. Debaters and judges who do not understand the logical structure of arguments tend to treat each distinct assertion (or subpoint) of a case or contention as a separate bit of reason to accept or reject a resolution. This is why debaters feel compelled to attack everything their opponents say and why certain progressive judges make decisions on the basis of dropped subpoints, whatever the points may be. Logically astute debaters know that individual claims by themselves may not lend any support to a conclusion. Just as a healthy kidney will not by itself keep you alive, so a single dropped subpoint will not typically lend any free-standing support to a resolution. A kidney contributes to the life of an animal only in concert with other organs, and a distinct premise supports a conclusion only in combination with an appropriate set of other premises. Arguments function as organic wholes; take away any necessary part (i.e., premise) and the whole argument dies. Your goal in cross-examination and rebuttals is to make your judges and opponents see that this is so. Analyzing an opponents argument into premise-conclusion form may reveal three types of problem. First, an opponent may present a valid argument with true premises for an irrelevant (i.e., non-resolutional) conclusion. Second, an opponent may present true premises in support of a relevant conclusion, but in a logically invalid form. Third, an opponent may present a valid argument for a relevant conclusion (the resolution or its negation), but with one or more objectionable premises. I will give an example of each sort of error and the corresponding response strategy below. Of course, it is possible that an opponents argument will display more than one of these three errors. The basic strategy of the logical debater is: (1) nd the weakness, and (2) attack the weakness. The logically savvy debaters task is further complicated by the fact that not only his opponents, but also most of his judges, will be unaccustomed to interpreting arguments and responses according to their logical structure. The truth is that many experienced LD judges think that argument means simply claim-warrant; they are not acquainted with the formal inferential patterns that distinguish logically adequate warrants from imitations. So debaters who intend to use logic to their advantage must be prepared not only to make the correct strategic decisions but also to explain those decisions and their strategic force to judges who otherwise may not understand whats happening. We might, then, say that there are three stages to the use of logic in cross-examination and rebuttals. In the rst stage, the debater must understand for himself the logical structure of an opponents position and decide how best to attack it. In the second stage, the debater must actually make the attack. And in the third stage, the debater must explain the impact of the attack to ensure that its power is not lost on judge or opponent. II. Cross-Examination There is a traditional story about cross-examination that goes something like this: Cross-examination has two main purposes. The rst is to clarify an opponents case
32

where it is unclear; this is the priority, but it is useful mostly for novices who havent learned to ow well. The second purpose, which is the main purpose for skilled debaters, is to discredit an opponents case by asking embarrassing questions about it. This story is misleading. It is true that clarication should be the priority in crossexamination; there is no point in trying to discredit an argument before it is clear just what the argument is. But it is not true that clarication is primarily a task for novices. Most cross-examination time for most debaters (including skilled and experienced debaters) is best spent in clarication of one form or another. Debaters would do better to leave the discrediting for rebuttals and use cross-examination exclusively to clarify for themselves and for their judges the logical structure and commitments of an opponents position. This is crucial since, as noted above, most opponents and judges will not have a ready grasp of logical structure and implication. And mapping the logical structure of an unfamiliar case can be a challenge even for logically competent debaters. Crossexamination is the time in the debate when logical structure can be laid bare. As you ow your opponents case, you should be trying to follow whatever logical structure there is and to impose whatever structure is missing. You are trying to identify each separate constructive argument, the conclusion of which is the resolution or its negation. Ideally, each contention will correspond to a separate constructive argument. But sometimes one contention will contain more than one argument, or one argument may be spread out over two or more contentions. Likewise, value premises and criteria will usually function as essential premises in constructive arguments, and denitions, observations, and other extra-contention elements may also contain premises on which constructive arguments depend. Even when you have identied a distinct constructive argument, its premises may be in an unnatural order which is hard to follow. Your primary goal in cross-examination is to review the logical structure of each constructive argument to prepare for your rebuttals. There need be nothing sneaky about this cross-examination strategy. You can be entirely upfront with your opponent: Id like to make sure I understand each of your arguments. You do not want to offer your opponent an open-ended invitation to restate each of her arguments. If you do, she is likely to (a) turn your cross-examination time into an extended constructive speech, and (b) restate her arguments in just as logically confusing a way as she presented them in the case. Instead, you should attempt to reconstruct each argument in a clear order, frequently asking your opponent if you have stated each step correctly. The goal is to translate each of your opponents lengthy, prosy constructive arguments into a short, clear set of premises which logically yield their conclusion. If an argument relies on unstated assumptions for its validity, you should probably make those assumptions explicit. If an argument validly draws an irrelevant (i.e., nonresolutional) conclusion, you can call attention to this fact by asking, So your second contention proves that [contention conclusion], right? It is best to wait until your rebuttal to point out that this conclusion is different from the resolution and therefore that your opponents argument is irrelevant.
33

When you have laid out the logic of an argument, you can narrow your rebuttal burden considerably by getting your opponent to agree with you about what you must do to kill the argument. You should call attention to the premise or inferential step you plan to challenge in rebuttals and get your opponent to agree that that premise or step is crucial for the success of the argument as a whole. You should not come right out and say, I plan to attack your value premise that if capital punishment deters crime, then capital punishment is moral. Instead, say something like, So it looks like this entire argument relies on the claim that if capital punishment deters crime, then it is moral. Is that right? Suppose your opponent has provided statistical evidence that capital punishment does, in fact, deter crime, and suppose that you do not want to get into a statistical battle. You can neutralize any future accusation of dropping your opponents statistical subpoint by now asking, So if I could prove that deterring crime does not automatically make capital punishment moral, then your evidence showing that it does deter crime would not matter, right? If there are sections of the case or contention which do not seem to be related logically to the main arguments, you can ask your opponent to explain briey how each of those sections supports the main argument(s) at issue. If her answer makes clear that a particular section really doesnt play an essential role in supporting (or defeating) the resolution, you can bracket this section as irrelevant for your rebuttals by asking, So by itself, this overview does not prove the resolution, right? Determining the logical structure of each constructive argument and reaching agreement about what would be sufcient to kill it will take most or all of your crossexamination time. As you become more procient, you may nish this task early enough to leave time for other lines of questioning. That is, you may have time to begin to discredit arguments during cross-examination. If you understand the logical structure of each argument and have made that structure clear to the judge, you can begin to chip away at just the weak spots you plan to emphasize in your rebuttals. The relevant weaknesses are the three mentioned in the last section: an irrelevant (i.e., non-resolutional) conclusion, an invalid form, or a false premise. Because this is not an article on cross-examination technique, I will not detail how to proceed on each of these three fronts. However, I will say that the best way to expose either the invalidity of an argument or the falsehood of a normative premise is to construct a counter-example. A validity counter-example is an argument with the same logical form as the argument under consideration but which has obviously true premises and an obviously false conclusion. For example, suppose a debater argues that: (1) If pornography is proven to cause rapes, then censorship of pornography is justied. (2) But pornography is not proven to cause rapes. (3) So censorship of pornography is not justied.

34

Readers familiar with the rst article in this series should immediately recognize this as the fallacy of denying the antecedent. To avoid the appearance of begging the question, a good counter-example should be on a completely different subject. Here is one: (1) If speeding is proven to cause mass starvation, then speed limits are justied. (2) But speeding is not proven to cause mass starvation. (3) So speed limits are not justied. Everyone recognizes that the premises of this counter-example are true but its conclusion is false. This is because there are many harms that, if caused by speeding, would justify speed limits (many true statements of the form If speeding is proven to cause X, then speed limits are justied). The fact that one such harm (mass starvation) is not in fact caused by speeding does not mean that there are not other signicant harms that are caused by speeding and that justify speed limits. A validity counterexample makes clear for an audience just what is wrong with invalid reasoning. To present a validity counter-example in cross-examination, you should simply ask your opponent if each of the counter-examples premises is true (Is it true that . . .?). Then ask if the counter-examples conclusion follows (Does it then follow that . . . ?). If your opponent agrees that the premises are true but the conclusion is false, move to the next point on your agenda; wait until your rebuttal to explain the logical parallel between your counter-example and your opponents original argument. A counter-example to a normative premise is really an application of modus tollens. It begins with a normative principle which you intend to attack. You show that if the principle is true, then certain consequences follow logically from it, but that these consequences are false or morally unacceptable. Therefore, the normative principle which implies them must also be false. For example, if your opponents normative principle is that whatever promotes Americas national interest is moral, your counterexample might consider slaughtering Yemens population and seizing their oil elds. This action might be in Americas national interest but would clearly not be moral. As an argument, the counter-example looks like this: (4) If whatever promotes Americas national interest is moral, then it could be moral for America to slaughter Yemens population and seize their oil elds. (5) But it could never be moral for America to slaughter Yemens population and seize their oil elds. (6) So its false that whatever promotes Americas national interest is moral. To present a normative counter-example in cross-examination, begin by asking your opponent to endorse the normative premise of his argument that you intent to challenge (Your argument depends on the premise that if an action is good for Americas national interest, the action is moral, right?). It is best to state the premise as a conditional, as in my example. Next, present the counter-example in such a way that it clearly satises the antecedent of the conditional (Would it be in Americas national interest to gain massive new petroleum reserves? [Yes] Suppose America could gain these reserves
35

by slaughtering the people of Yemen and taking over their land.). Now ask whether the counter-example satises the consequent of the conditional (Would slaughtering the people of Yemen to seize their oil elds be moral?). If you get the admission you want, move to your next point; wait until your rebuttal to explain how the admission undermines the normative premise on which your opponents argument depends. Finally, if you ask discrediting questions, it is unwise to draw out the full logical implication of your strategy in cross-examination. If your opponent agrees that the conclusion of a contention is somewhat different from the resolution, do not go on to ask, So this contention is really irrelevant to the debate, right? If you ask your opponent to admit outright that her argument is irrelevant, illogical, or unsound, she is going to deny the charge, and you may inspire her to do some quick thinking which will get her out of the bind youve helped to put her in. Lead your opponent right up to the edge of the precipice in cross-examination, but wait until your rebuttals to push her over the edge. III. Rebuttals Based on your cross-ex, try to nd the weakest two places in each constructive argument to focus your rebuttal attacks. If you attack more than two places in the argument, it is unlikely that any one of the attacks will be very persuasive or carry much weight, and you will likely fall back into a pattern of reexive refutation. The whole point of thinking explicitly about logic in rebuttals is to help you overcome the need to attack everything your opponent says, or even everything with which you disagree. Because arguments function as organisms, with each premise or inferential step playing an essential role, you can afford to ignore most of your opponents claims and focus your attention on only one or two vital organs. At the same time, it is true that different responses appeal to different judges. Therefore, it is wise to nd two ways to attack each argument, so that if you and your judge happen to disagree about which response is your strongest, you have not put all your eggs in one basket. Of course, stipulating two attacks per argument is somewhat arbitrary, and wise debaters (a vanishingly small minority) may recognize occasional exceptions. For example, if an afrmative case presents only one long contention which makes a single constructive argument, it might be worthwhile to kill the argument three times over instead of just twice. And there is one standing exception to my two-response rule: when your point about an argument is that its irrelevant, i.e., that it doesnt prove or disprove the resolution, you should make that your single response to the argument. If you claim that an argument is irrelevant and then go on to attack its premises in detail, you show that you really do consider it an important argument to refute after all. At the very least, you show that you are not condent in your claim that the argument is irrelevant. Here is a classic example of the irrelevance response: In an elimination round of Emorys prestigious Barkley Forum several years ago, a talented afrmative debater
36

who was favored to win presented two well-supported contentions which concluded that a journalists right to shield condential sources was a very good thing. The negative debater granted the truth of the afrmatives conclusion and the soundness of both her arguments. He did not attack a single statement made in the afrmative case. He simply pointed out that the resolution was that A journalists right to shield condential sources ought to be protected by the First Amendment, and that being a good thing was not equivalent to being properly protected by the First Amendment. His negative constructive had argued for a strict original-meaning interpretation of the First Amendment which excluded the right to shield sources. The negative advanced on a 3-0 decision. However, had the negative spent (say) 20 seconds arguing that the afrmatives contentions were irrelevant and then gone on to refute individual afrmative claims in detail, he might well have lost the round. The power of his single response depended on his willingness to stake the round on it. And since the afrmative was prepared to defend the soundness of her arguments (but not, unfortunately, their relevance), she would have come out looking much stronger had the negative engaged her on her own turf. There are three basic ways to attack an argument: its relevance, its validity, or the truth of its premsises. We have just discussed a relevance attack; now lets consider the other two options. Suppose you want to attack an arguments logical validity. If you have already set up a validity counter-example in cross-examination, you can refer back to it; if not, you will need to present a counter-example from scratch. When challenging validity, you must make it very clear from the beginning that your challenge, if successful, undercuts the entire argument. Before you present your counter-example, you should clearly identify the argument youre attacking and state explicitly that you are attacking the entire argument. You do not need to encumber your rebuttals with logical jargon, but you should preview your strategy in a more colloquial way. You can say simply that the argument youve identied purports to prove the resolution (or its negation), but that in fact, even if all your opponents claims (i.e., premises) are true, the conclusion does not follow. Then say you can use identical reasoning to prove a false conclusion, and present your counter-example. After presenting the counter-example, state again that your opponents argument has exactly the same awed logical form. If possible, point out in detail just where the inference goes wrong and explain further how your opponents premises are, in fact, compatible with your conclusion. The third basic refutation option, challenging the truth of a premise (i.e., challenging the soundness of the argument), will be more familiar to most debaters. Attacking individual claims is the bread and butter of the logically oblivious refutory debate which dominates many circuits. An understanding of logici.e., of arguments as complex inferential wholesenables one to refute premises much more selectively. If you have correctly mapped the logical structure of your opponents argument, you need only kill one of its premises to kill the entire argument. You do not have to refute each premise separately, and you do not have to provide multiple responses to a single premise. Indeed, if you plan to attack a premise, you are almost always better off explaining your best single
37

response in detail rather than tossing off six different half-sentence hints. Compare the martial arts master who downs her opponent with a single well-placed punch to the clumsy school girl who ails out of control slapping, scratching, and pulling whatever she can get her hands on. Choosing the weakest premise to attack and choosing the best single response to that premise are important skills, but they do not fall directly within the purview of this article. This is because logic, as you may recall from the rst article, cannot tell you what premises are true. It can only tell you whether conclusions follow validly from whatever premises you accept. When challenging a normative premise, you should use the counter-example strategy outlined above in the discussion of cross-examination. Point out your opponents normative premise and state that you can prove its false because it yields the wrong results in other situations. Then explain your counter-example using the modus tollens pattern illustrated in the last section. If your opponents original premise sounded promising before you attacked it, try to provide an alternative true normative premise which accounts for the truthful ring of your opponents premise but avoids the counter-example and also does not support your opponents conclusion. For example, if the original premise was that if a punishment deters crime, then it is justied, you might provide an example of an effective deterrent which is unjustly applied to an innocent person. Now it looks like deterrence alone is not a sufcient justication for punishment; but surely there is something right about the notion that punishments should make society safer. Perhaps you could propose an alternative normative principle that if several punishments t the crime equally well, then the criminal ought to suffer the one which is most likely to deter others from committing the same crime. This principle allows you to acknowledge that public safety matters while insisting on the priority of retributive justice. Although logically minded debaters will be especially skilled at spotting common fallacies of statistical reasoning, there is no general recipe for refuting empirical premises. The key, again, is to choose the premise and the response carefully and explain the response in detail for maximum impact. The primary contribution of logic to the refutation of individual claims is the understanding of how much damage an argument suffers from the death of a single premise. If the premise is necessary to the validity of the argument, the damage is fatal: death to the premise means death to the argument. As noted earlier, logical debaters must point this out to their judges at every opportunity. For judges accustomed to counting up dropped sub-points, the notion that refuting a single claim can kill a contention will be very surprising. You have already primed the judge to think of each argument as an interdependent whole in cross-examination. In rebuttals, you must preface and conclude each premise attack by pointing to the complete argument of which the premise is a part and explaining how the premise is crucial to the argument as a whole. You should say explicitly that even if every other claim in the argument is
38

true, the argument fails to sustain its conclusion if this single crucial premise is undermined. If the premise you are attacking is a value premise (or some other premise shared by all your opponents constructive arguments), you may be able to wipe out a whole case with a single response. But you cannot trust your judge to notice that the case or contention is dead. You must point this out and explain why it is so. Debaters are often admonished to impact their responses. I believe that pointing out the logical ramications of a single response for an argument or case as a whole is one sensible interpretation of what it might mean to impact that response. As much as possible, such impacting should be saturated with the language of the resolution so as to remind your judge that it is ultimately the resolution which is to be proved or disproved, and it is the resolution which your opponent has, in light of your strategic attack, failed to prove or disprove. IV. Final Issue Selection Many debaters and judges believe it is especially important to spend the nal minute or two of the negative rebuttal and the whole of the second afrmative rebuttal summarizing a few main issues in each round. Such a period of crystallization (as it is sometimes awkwardly called) may not be a bad idea, but it is probably less necessary for debaters who implement the model of selective, logical debate endorsed in the present series of articles. This is because the entire rebuttal speeches of such debaters will do exactly what the nal summary of less logical debaters is intended to accomplish: make clear why the resolution is true or false given the success of ones own constructive arguments and the given the failure of ones opponents constructive arguments. Logical debaters select their issues before they begin to speak, and they speak only about those issues. There is no shapeless mountain of reexively generated points and subpoints which must be knocked into some kind of order at the very end; the order has been obvious all along. Nevertheless, learning to seal the deal is a valuable rhetorical skill, and some judges may expect a distinct treatment of nal issues regardless of the clarity of your speeches as a whole. Conceiving of arguments as logical units will greatly simplify the construction of a nal summation. The advice given about nal issue selection has often been quite vague: pick the most important issues (but how do you know whats most important?), pick the issues youre winning (but what counts as an issue?), tell a persuasive story about the round (what does that mean?). If you have developed your own case arguments as logical wholes, you know what you must do to sustain them: sustain a set of premises which together entail the truth (or falsehood) of the resolution. Likewise, you know what you must do to defeat your opponent: show that for every distinct constructive argument she makes, her argument is either irrelevant, invalid, or dependent on a false premise. Each distinct constructive argument still in play at the end of the round deserves attention in your closing statementeither to show why it succeeds (your argument) or to show why it fails (your opponents argument).
39

In order to prove the resolution true (or false), you need sustain only one constructive argument. Therefore, if one of your original constructive arguments is badly mangled by the end of the round, you may wish simply to ignore it in favor of a more defensible argument. If you adopt this strategy, you should not call attention to the argument you are ignoring. Specically, you should not recite some awkward bit of debate jargon about kicking out of your original argument. Instead, call attenion to the argument you will defend by telling the judge that you believe it is your strongest and that it, by itself, is sufcient to prove the resolution true (or false). Of course, ideally you will not feel forced to choose between constructive arguments, since all your arguments will be ones you are well prepared to defend in the rst place. However many of your own arguments you are defending, your nal summation should remind the judge in a few sentences of the main premises of each argument in such a way that the resolution (or its negation) obviously follows. Then pick out the most damaging or prominent objections your opponent has made to the argument and state briey why the objections faili.e., why the argument survives. You will not normally have the option to choose which of your opponents constructive arguments to review at the end. Since any one constructive argument is enough to carry the day, you must show why all of your opponents constructive arguments fail. The only exceptions to this rule are if either (a) your opponent has narrowed her position for you by abandoning one or another of her constructive arguments or (b) you are attacking a premise (such as a value premise) which is shared by several arguments, such that separate treatment of each argument is unnecessary. If one of these exceptions occurs, you should remind your audience at the end that it has occurred. You do not want an abandoned argument to be rescued after it is too late for you to object, and you do not want a judge to mistake your very efcient attack on several arguments at once for a failure to attack any of the arguments. Whereas you may feel some pressure to respond to each major attack against your constructive arguments, your nal review of your opponents arguments can be briefer. This is because, again, you do not have to refute everything your opponent says in order to kill her argument. If you have attacked an argument on several fronts, you should choose the one or two deadliest objections to reiterate at the end. Rhetorically, it is usually unwise to spend a lot of time talking about your opponents arguments. You have already killed the arguments during the main body of your rebuttals, and you are now simply reminding your listeners of the results you have achieved. It is especially important to emphasize that your one or two objections, whatever they are, kill the entire constructive argument. A quick nal accounting of the fate of each constructive argument can be very helpful to a judge, suggesting in advance the process of deliberation you want him to use to reach a decision. As with every argument during the round, the focus at the end should remain squarely on the resolutionon why the resolution is true or false. Your devastating objection to a crucial premise may be fruitless if your audience does not understand
40

what bearing that premise has (via the argument of which it is a part) on the truth or falsehood of the resolution. Once again, the best way to ensure that you are explaining the full relevance of each distinct point is to force yourself to use the language of the resolution in making the point. There are other closing strategies besides the review of constructive arguments that may be effective and appropriate in particular rounds. Knowing which to use when requires a more sensitive judgment than can be prescribed in an article like this one. But the review of constructive arguments is a safe and useful default strategy. And any good closing strategy must talk explicitly about the truth or falsehood of the resolution, which is usually going to involve discussing constructive arguments in one form or another. These articles have obviously not provided complete instruction on how to debate. Indeed, it has been a recurring theme that logic alone will not sufce to establish the truth of any proposition (one must begin with true premises), and it should also be clear that logic in the strict sense of formal validity is not the decisive issue in most debate rounds. But logic is an indispensable tool of competent debaters. Similarly, no amount of anatomical and physiological knowledge would sufce to make a person a great surgeon; successful surgery also requires excellent judgment and physical technique. But an aspiring surgeon had better learn the anatomy and physiology of the human body. Random cutting and patching are at best pointless and at worst dangerous, even if they are technically graceful. Logic reveals the anatomy and physiology of arguments. Excellent debate requires trained judgment and rhetorical excellences beyond the scope of logic. But debate without a solid grasp of logic is just so much random cutting and patching.

41