Chapter 16. Question-Begging Arguments

**Chapter 16. Question-Begging Arguments** **Part A. Assuming What Needs Proving** In ordinary English, to say "it begs the question" means "it raises the question," but the term has a different meaning in logic: a **question-begging argument** is one that uses its conclusion as a premise. By doin...

**Chapter 16. Question-Begging Arguments** **Part A. Assuming What Needs Proving** In ordinary English, to say "it begs the question" means "it raises the question," but the term has a different meaning in logic: a **question-begging argument** is one that uses its conclusion as a premise. By doing so, the argument essentially just *assumes* the very thing it's supposed to *prove*. Consider the three arguments below. [Argument 1] [Argument 2] Each of these arguments commits the **fallacy of begging the question** because each one uses its conclusion as a premise in some way or other. Argument (1) does this most blatantly. Argument (2) is only a slightly less blatant case: one of the conjuncts of the conjunctive premise is the argument's conclusion, and we could just as easily have written it on a line by itself. Argument (3) is not as obviously question-begging as (1) or (2), but since "all dolphins are nonfish" is a statement logically equivalent to the statement that no dolphins are fish, the conclusion really is being used as a premise to support itself; it's just been reworded a bit. If one of the premises of an argument (or one of the conjuncts of a conjunctive premise) is equivalent to the conclusion, then the argument assumes the very thing it's supposed to prove, and that's fallacious reasoning. Note something important: arguments (1)-(3) are all *sound* arguments. Since every statement validly entails itself (we learned this in Chapter 4), using the conclusion as a premise guarantees the validity of the argument, and in this case the premise in each argument happens to be true, which makes the arguments sound, as well. So, *question-begging arguments are always valid, and sometimes they're even sound*. The problem with such argumentsthe reason we consider them fallaciousis that they are *rhetorically flawed*. *Rhetoric* is the art and science of *persuasion*: it's about persuasive writing and speechmaking. When we say that an argument is rhetorically flawed, we mean that it won't persuade rational people to accept its conclusion. No rational person who did not *already* accept the conclusion of Arguments (1)-(3) would be persuaded by any of these arguments, since none of them provides any independent evidence to support the conclusion. Good arguments provide us with independent grounds for accepting the conclusion instead of just using the conclusion to "prove" itself. Contrast Arguments (1)-(3) with the rhetorically superior (and sound) argument below. [Argument 4] Neither premise (1) nor premise (2) merely restates the conclusion; (1) does not even mention dolphins, and (2) does not mention fish. In combination the premises entail the conclusion, but that's (obviously) unobjectionable. A rational person could be persuaded by Argument 4. **Part B. Circular Reasoning** Consider the reasoning below. I can trust Jones because Smith vouched for him, and I can trust Smith. I can trust Smith because Roberts vouched for him, and I can trust Roberts. I can trust Roberts because Jones vouched for him, and I can trust Jones. If we were to diagram these inferences, they would form a circle: **Circular reasoning** is what you're engaged in when you use a statement, P, to support (directly or indirectly) another statement, Q, and then you offer Q as your (direct or indirect) support for P. The support in each case can be indirect because, as in our example, there might be three or more statements involved. The "tightest" circle results when P is your direct support for Q, and Q is your direct support for P. Consider the example below. A: Jones's lies prove how corrupt he is. B: How do you know Jones has been lying? A: Because corrupt people are always lying. By this logic, A infers that Jones is corrupt from all the lying he does, but A's only stated reason for thinking Jones lies is his belief that Jones is corrupt. That's (tight) circular reasoning:  In order to break out of this circle, we need independent evidence for at least one of the two statements here. If we prove that Jones is corrupt by showing that he's a liar, and then we find independent support (say, multiple convictions for perjury) to show that Jones really is a liar, then there is no circle. The reasoning would then have a more "linear" structure, so to speak: Perjury convictions Jones is a liar Jones is corrupt When you reason in a circle, you end up using a statement to prove itself when the circle closes. This fact can make it seem as if circular reasoning and begging the question are the same thing, but they're not. A *single* inference can beg the question, but circular reasoning requires at least *two* inferences, as our examples show. You must infer one statement from a second, and you must infer that second one (directly or indirectly) from the first. Note also that a person's reasoning is circular when he *tries* to infer P from Q, and *tries* to infer Q from P. That could happen *even if neither of these inferences is question-begging*, *because* P *and* Q *are not the same statement*. Suppose a person tries to infer *all gems are rubies* from *all rubies are gems*, and viceversa. The arguer is reasoning in a circle here, because he's trying to support each statement solely by inferring it from the other statement. But neither of these inferences is question-begging, because the premise and the conclusion are not equivalent in this case, and neither validly entails the other; perhaps the arguer simply didn't realize this. Question-begging inferences, by contrast, are always valid, because an inference is question-begging only when the premises somehow contain the conclusion; that condition is not satisfied here. For this reason, as well, reasoning in a circle and begging the question are not the same thing, although each represents a fallacious way of reasoning. **Exercises** 1.What, very generally, does it mean (in the context of argument) to *beg the question*? 2.What does it mean to say that an argument commits the fallacy of begging the question? 3.Suppose I want to prove to you that your solution to a problem is wrong (you think it's right). My argument is below. Does it beg the question? 4.In Chapter 4 you learned that the inference below is valid. In the remaining exercises, assume each argument is being used for persuasion. 5.Indicate whether the argument below is question-begging. 6.Indicate whether the argument below is question-begging. 7.Indicate whether the argument below is question-begging. 8.Indicate whether the argument below is question-begging. 9.Indicate whether the argument below is question-begging. 10. Indicate whether the argument below is question-begging. 11. Indicate whether the argument below is question-begging. 12. Indicate whether the argument below is question-begging. 13. Some arguments, we've seen, are question-begging because one of the premises just is, or is equivalent to, the conclusion. An argument can also beg the question by having a premise that "contains" the conclusion. When we talk about a premise "containing" the conclusion, what do we mean? 14. Consider the argument below. If the argument were being used to persuade you that Jeremiah is a liar, would it beg the question? The answer is yes. Why? 15. There is another context in which some people use the term *question-begging*, although it doesn't technically amount to begging the question in the logician's strict sense, since it can happen even in the absence of any argumentsay, in just asking a question. It involves using a certain word or phrase *in a context where using it assumes what still needs to be proven*. When a prosecutor during a murder trial refers to the defendant as "the murderer," for example, she's assuming what hasn't yet been proven: that the defendant committed murder. She's not allowed to use that word in this context; the defense attorney would justifiably object to that characterization of his client. In each case below, indicate whether there is a "question-begging characterization" being used. allegation). The cop asks, "Who was the first person to stop you as you exited the store?" e. Richards has just been found not guilty of tax evasion, and he's being interviewed outside f. Richards has just been found not guilty of tax evasion, and he's being interviewed outside 16. Answer the questions below: 17. Is this reasoning circular? "I once had doubts about the accuracy of the reporting in the *New York Times*, but they published an article recently with the results of their own study that shows they practically never make mistakes. That cleared up my doubts." 18. Is A's reasoning below circular? A: Krauss is being honest. 19. Is A's reasoning, below, circular? 20. Is A's reasoning, below, circular? [Optional exercise 1]: Consider the argument below. [Optional exercise 2]: The French philosopher Rene Descartes famously entertained the possibility that everything he perceived might only be a dream. "Skeptics" sometimes use this possibility to put in doubt whether we really know anything about the world outside our own minds. If someone were to argue against the skeptic by saying that he's assured by the people around him that he's most definitely not dreaming, then would this argument beg the question? [Optional exercise 3]: The Scottish philosopher David Hume famously wondered whether we have any good reason to expect that the regularities we've observed in nature to this point will continue to hold in the futurewhether, for example, dropped objects will continue to fall. If someone argues that in every case when he's had these doubts the regularities in nature have continued to hold, and that therefore he can reasonably believe they'll keep on holding, then would this reasoning beg the question? **Chapter 17. Arguments from Ignorance** **Part A. Arguments from Ignorance** Some people use the term *argument from ignorance* to refer to the following (obviously fallacious) form of argument: 1. There is no evidence that P is true. 2. So, P is false. Alternatively, one might argue that because there's no evidence that P is *false*, P must be *true*. This inference, too, would be fallacious. Merely from the absence of evidence that P is true it does not follow that P is false. It might be that no one ever looked for any evidence that P is true, or that P is difficult to prove despite being true. Hence, an argument like the one below would represent fallacious reasoning: 1. There is no evidence that electrons, like protons are neutrons, are composed of stillsmaller particles. 2. So, electrons are not composed of still-smaller particles. Suppose, on the other hand, that we were to add another premise according to which *if* electrons *were* composed of still-smaller particles, *then we'd have evidence of that by now*, or probably we would. This would augment the argument in a way that would make it much stronger (whether it would be cogent is another issue). We'll consider this augmented form of argument from ignorance in the next section. **Part B. Augmented Arguments from Ignorance** A somewhat different and more plausible way of arguing for a proposition's falsity involves claiming that if P were true, then probably there would be good evidence of that by now. Consider the example below. 1. There is no good evidence that an intelligent civilization lives on the moon. 2. If there were an intelligent civilization living on the moon, then probably we'd have good evidence of that by now. 3. So, probably there is no intelligent civilization living on the moon. Premise (2) is, in this particular case, extremely plausible, and the argument provides good reason to accept the conclusion. If, improbably, (2) needs a subargument, would could support it by noting among other things the lunar landings and the fact that we've made close observations both of the side of the moon that always faces Earth and its dark side. (Strictly speaking, our conclusion should state that *unless it's antecedently very likely that an intelligent civilization lives on the moon*, there probably isn't one there. From the premises *if P, then probably Q, and not Q*, it doesn't follow that *probably not P*, since P might be something highly antecedently probable.[^1^](#fn1){#fnref1.footnote-ref} We'll let this be understood in our examples.) Contrast the preceding argument with the one below. 1. There is no good evidence that extraterrestrial life exists. 2. If there were extraterrestrial life, then probably we'd have good evidence of that by now. 3. So, probably there is no extraterrestrial life. In this case, premise (2) is not very plausible. The universe is vast, and the finite speed at which signals and spacecraft can travel means that even if intelligent (to say nothing of unintelligent) life exists elsewhere in the universe, making contact with it or gathering any indirect evidence of its existence could be extremely difficult. Indeed, for all we know, we're not even searching by the right methods; perhaps an alien civilization would use some as-yet unsuspected means of communication. If there is life elsewhere in the universe that we haven't found, that wouldn't be surprising.[^2^](#fn2){#fnref2.footnote-ref} We'll use the term **augmented** **argument from ignorance** to refer to the type of reasoning in the two sample arguments we just looked at. Even this name is apt to suggest that all such arguments, like the one about extraterrestrials, are fallacious, but not all of them are, as the first example shows. This is a plausible form of reasoning provided that we can reasonably expect to have evidence of P's truth, *if* P is true. What will determine whether that's a reasonable expectation? All of the following will enter into it: 1. Have we actually been looking for evidence? Would we know it when we see it? 2. How large is the space we've been searching for evidence? (In some cases, the space is a physical one. In other cases, the space is a conceptual space of possibilities, such as when we search for solutions to a problem.) 3. How long have we been searching for evidence, and how many people have been searching? If the space is small and uncluttered, a shorter search might suffice. 4. Have we been searching by the right methods? We raised this question in the extraterrestrials example. 5. Is there any special reason why evidence for this thing might be hard to find? Might extraterrestrials deliberately conceal their existence for some reason, for example? Might something we're looking for be camouflaged? Might a problem be a particularly difficult one to solve, even if a solution does exist? We have to consider factors like these and decide whether we can reasonably expect to have evidence for P by now, if P is the case. Another example is below, where P in this case is the existence of Bigfoot. 1. There is no good evidence that Bigfoot exists. 2. If Bigfoot did exist, then probably there would be good evidence of that by now. 3. So, probably Bigfoot does not exist. Taking the truth of the first premise for granted, is the second premise plausible? That depends on how large the space is where Bigfoot might live, how long people have been looking for Bigfoot, whether they're searching by the right methods (say, by searching during the daytime, if Bigfoot is supposed to be active during the day), whether Bigfoot would be camouflaged, and so on. In this particular case, the second premise and the overall argument seem pretty solid, even if we concede that some animals are fairly elusive in the wild. But other cases may be more difficult to assess. **Exercises** 1.What is the basic form an un-augmented argument from ignorance? Why are all such arguments fallacious? 2.What is the form of an "augmented" argument from ignorance? Are all such arguments fallacious? 3.Suppose a cop and his drug-sniffing dog spend ten minutes searching a car for marijuana. They don't find any; the dog never alerts to the presence of any drugs. Now evaluate the augmented argument from ignorance below. 1. The K-9 team has found no marijuana (or evidence of marijuana) in this car after a ten minute search. 2. If there were marijuana in this car, then probably the K-9 team would have found it by now. 3. So, probably there is no marijuana in this car. 4.Suppose a brutal dictator argues that the people under his rule must be happy, because he never hears any complaints (i.e., he has no evidence suggesting that they're *unhappy*). Evaluate this augmented argument from ignorance: 1. I've had no complaints about the conditions in this country. 2. If people had complaints, then probably I'd have heard them by now. 3. So, probably the people have no complaints. 5.Evaluate this augmented argument from ignorance. *People have been searching for the Loch Ness monster for years, and nobody has ever come up with any good evidence that it exists. So, I think we can reasonably conclude that it doesn't exist*. 6.Evaluate this augmented argument from ignorance: *My psychiatrist says I'm harboring some resentment toward my brother. I've done some soul-searching on this, and I don't find in myself any such resentment. I think I'd know if I resented someone in my family. So, it's fair to say my psychiatrist is just wrong.* 7.Consider the augmented argument from ignorance below. Why might someone claim that premise (2) is questionable in this case? 1. Romeo gives no signs of being in love with Juliet. 2. If Romeo were in love with Juliet, probably he'd give signs of it. 3. So, probably Romeo is not in love with Juliet. 8.The augmented argument from ignorance below against the existence of any carbon monoxide in my home is a good one. It would have been fallacious if the manner of searching had been...what? Also, does the argument below show that probably *no* CO is present, or should the conclusion be a somewhat weaker statement? 1. My carbon monoxide (CO) detector has not sounded. 2. If there were CO here, then my detector would have sounded by now. 3. So, probably there is no CO here. 9.If two ornithologists spend two hours searching for a rare species of woodpecker in a 10,000 square mile rainforest and come up with no evidence that it lives there, can they reasonably conclude that it doesn't live there? 10. Evaluate the following augmented argument from ignorance: *Twin primes are primes separated by two (e.g., 5 & 7, or 11 & 13). According to the Twin Prime Conjecture, there are infinitely many twin primes. Mathematicians have been trying for many years to prove this conjecture, without success. Surely, if the conjecture were true, someone would have found a proof by now. So, we can reasonably infer that the Twin Prime Conjecture is false*. 11. Consider the augmented argument from ignorance below, and indicate how it could turn out that premise (2) is implausible. Could the argument be strengthened by altering the conclusion? 12. Evaluate the augmented argument from ignorance below. 13. Evaluate the argument below. In this case, an explicit subargument is provided for the second premise. 14. Passenger pigeons are believed to have become extinct in the early 1900s, largely as a result of hunting and deforestation. Now evaluate the argument below. 15. Evaluate the pattern of reasoning below -- an extreme form of the type of reasoning we've been discussing. [Optional exercise 1]: Below is a philosophical application of augmented argument from ignorance (in the philosophy of religion, it's called the *Argument from Divine Hiddenness*). Consider for yourself whether the argument is compelling.[^3^](#fn3){#fnref3.footnote-ref} 1. There is no good evidence that God exists. 2. If God did exist, then probably by now he'd have provided us with good evidence of his existence. 3. So, probably God does not exist. [Optional exercise 2]: Consider again the argument in one of the exercises above to show that probably the Loch Ness monster does not exist. Suppose someone attempts to counter our argument this way: *For all we know, the Loch Ness monster has the power of invisibility. And maybe it can shield itself even from sonar, so that it's no surprise at all that we haven't found it.* *So, the fact that we haven't found it doesn't make it reasonable to infer that it doesn't exist*. How should we respond to this? (The Latin term *ad hoc* is useful here.) **Chapter 18. Analogical Arguments** **Part A. Analogies** An **analogy** is a comparison of two things, generally with the aim of making one of them easier to understand. If you explain computer viruses to someone by likening their behavior to that of biological viruses, then you're using an analogy to make computer viruses easier to understand for people already familiar with the biological kind. Two related terms are below. - To say that one thing is **analogous** to another simply means that they're alike in some way. The behavior of a computer virus is in some ways analogous to the behavior of diseasecausing pathogens. - To say that one thing is the **analogue** of another is to say that it's the counterpart of that thing in an analogy. If you compare the flow of blood through a body to the flow of traffic through a city, then blood cells are the biological analogues of cars, veins and arteries are the biological analogues of streets, and a clot is the biological analogue of a traffic jam. Analogies generally are not perfect: they break down where the things being compared are dissimilar in some important way, which is why you never want to stretch an analogy too far when it's used as a basis for decision-making. Computer viruses are in some ways similar to biological viruses, but obviously not in every respect. As long as we don't stretch the analogy beyond its usefulness, it can be heuristically useful despite the differences.[^4^](#fn4){#fnref4.footnote-ref} **Part B. Analogical Arguments** An **analogical argument** is a (usually inductive) argument in which we draw a conclusion about one thing based on the similarities it bears to something else. Here is a concrete example: Notice that the similarities mentioned in premise (1) are all relevant to a car's gas mileage. We would not mention that both cars are blue even if it were true, since that would not make it any more likely that your car gets the same gas mileage as mine. Since the number of similarities we find between the two things we're comparing can vary from one case to another, and since their relevance to the inferred similarity (gas mileage in this case) is a matter of degree, some analogical arguments are stronger than others. If we wish to attack an analogical argument, then we can question the *relevance* of the similarities mentioned, or we can suggest that there are not *enough* similarities to make a strong argument, or we can identify relevant *dissimilarities* between the objects that tend to weaken the analogy. If, using our previous example, one car has properly inflated tires while the other does not, then this will make it less likely that they get the same gas mileage and thus weaken the argument. Consider another example: Thus far, the courts have rejected arguments like this one against prostitution laws. There's at least one crucial difference, they claim, between consensual homosexual sex and commercial sex transactions: the former does not threaten to cause harm to others in any way, but prostitution does so in a variety of ways, such as by contributing to human sex trafficking and slavery. This difference spoils the attempted analogy between private consensual sex acts and the sex trade, such that laws against prostitution are not, in fact, unconstitutional invasions of privacy.[^6^](#fn6){#fnref6.footnote-ref} Lawyers deal in analogical reasoning often when comparing cases and the rationales behind court rulings. When an economist compares an ongoing recession to an earlier recession and claims that the government interventions that ended the previous recession probably will work to end this one, she also is reasoning analogically. **Exercises** 1.Define the following: 2.What is the basic form of an analogical argument? 3.True or False: Analogical arguments typically are inductive arguments. 4.What factors are most important in determining the strength of an analogical argument? 5.Consider the arguments below and select the correct option. 6.Evaluate the analogical argument below and answer the questions that follow. a. What else might we learn about Smith and Jones that would strengthen the argument? b. What might we learn about Smith and Jones that would weaken the argument? c. What might we find out about these two that would almost certainly have no relevance to the argument? 7.Evaluate the analogical argument below. 8.Evaluate the analogical argument below. 9.Consider the analogical argument below and answer the questions that follow. 10. Evaluate the analogical argument below. 11. Consider the analogical argument below and then indicate which of the statements below, if true, would be relevant to the argument's strength (i.e., it would tend to make the analogy stronger or weaker). 12. Consider the analogical argument below and then indicate which statements below, if true, would be relevant to the argument's strength. 13. Consider the analogical argument below and then indicate which statement below, if true, would strengthen the argument. 14. Evaluate this analogical argument: 15. Construct an analogical argument for the conclusion that animals, like people, can feel pain. [Optional exercise 1]: So-called "skeptics about other minds" are those who deny the possibility of having a justified belief in the existence of any mindeven a human oneother than your own. Is it possible to construct an analogical argument for the existence of other human minds? (Other people can tell us they have minds, of course, but how do we know this isn't merely the unthinking behavior of automatons?) Would the argument be much different from the one for animal pain? [Optional exercise 2]: A famous argument for the existence of a creator god is an analogical one offered by William Paley in his 1802 treatise *Natural Theology*. Consider whether you find it compelling. [Optional exercise \#3]: Consider again an argument used in one of the exercises above: **Chapter 19. Argument to the Best Explanation** **Part A. Explaining the Facts** Your argument for a statement P is an **argument to the best explanation** when you infer that P is true because it provides the best overall explanation for some set of facts.[^7^](#fn7){#fnref7.footnote-ref} Below are three examples. - Your car's engine doesn't crank and the instrument panel doesn't light up when you turn the key. There is more than one possible explanation for these facts, but *the battery is dead* seems like the most plausible hypothesis, and so you're likely to conclude---tentatively--- that the battery is in fact dead. - A prosecutor reasons to the best explanation in court when she puts forward what she takes to be the most plausible theory of the crime. That *Smith committed this murder* might be the best overall explanation for a range of evidence that includes DNA, remarks made by Smith to the police during interrogation, facts about the relationship Smith had with the victim, ballistics and fingerprint evidence, and so on. - Astrophysicists regard the theory that *the universe came into existence in an explosive event around 13 billion years ago* as the best overall explanation for a range of data that includes, among other things, a background of microwave radiation that pervades space, the abundance of light elements in the universe, and the way the galaxies are moving. It should be clear from these examples that argument to the best explanation is a very common form of argument used in a wide variety of contexts. Such arguments typically are inductive, and they always start from a set of facts that need explaining. Are there other ways to explain the facts mentioned in the examples above? Yes. Perhaps someone other than Smith committed the murder, but someone planted evidence to frame Smith for the crime and the police coerced a confession from him. A thousand other theories might be formulated to explain all the evidence we've gathered. Whenever you get a certain experimental result in a lab, there will be multiple ways to explain why you got that result. This point is crucially important: *for any finite set of facts, there is always more than one logically possible explanation for those facts*. In some cases there will be just one or a few explanations that seem to have any realistic chance of being right, but there will always be other ways to explain the facts that are at least abstractly possible. So, you're always going to be confronted with multiple possibilities, and you'll have to figure out which explanatory theory is relatively the best. You'll conclude that probably that theory is true, or at least more likely true than any other known theory.[^8^](#fn8){#fnref8.footnote-ref} **Part B. Criteria for Theory Choice** How do we decide which of several competing theories or hypotheses provides the best overall explanation of the facts? We ask certain questions about each one. Below are some of the main questions we ask.[^9^](#fn9){#fnref9.footnote-ref} 1. Does this explanatory theory *make sense* of the facts? We want the explanation not only to be consistent with the facts we've gathered, but to make us expect those facts, as a dead battery makes you expect your car's lights not to work. Merely being told that the car is old is consistent with the lights not working, but it doesn't particularly make you expect the lights not to work or help you understand why they're not working. Partly what this illustrates is that making sense of the facts involves providing some clue about *how* those facts came to be: how does the car's being old cause the lights not to work, for example? The account an explanatory theory provides can achieve varying levels of depth and specificity; generally, the more depth, the better.[^10^](#fn10){#fnref10.footnote-ref} In general, we want an explanation to make us think to ourselves, "Okay, now it makes sense why we find these facts." 2. Does the explanation make sense of *all* the facts, or just *some* of them? We want theories that have broad *scope* in that they help us understand all the facts we've gathered. This seems obvious, but it often happens in science that none of the competing theories is yet able to explain literally every fact gathered from observations and experiments. It's (partly) a question of which one explains the most. 3. Is the explanation *simple*, or at least simpler than the alternatives? "Occam's Razor" (named after the medieval philosopher William of Occam) tells us that, all else being equal, the simplest explanation for a set of facts is the one most likely to be true. (The trick here is to provide a coherent account of what explanatory simplicity involves. What makes one theory simpler than another, exactly? Going even deeper, we can ask questions about whether some types of simplicity are more "truth-indicative" than other types.) 4. Is the theory *antecedently plausible* in the sense that it coheres well with other beliefs we already have about how the world works? If I'm to believe (what the defense alleges) that the victim in our trial died when the gun Smith was holding "just spontaneously went off," then I'll have to abandon my belief that guns don't just go off by themselvesa belief strongly supported by a wealth of experience. That's a good reason to assign extremely low antecedent plausibility to this theory. There is no mechanical procedure for applying these criteria to all the theories we're considering. You think about the various theories in light of all these criteria and you make your best guess about which one, overall, best satisfies the criteria. You conclude that this is the one most likely to be the truth. Now, you can't really make a comparative judgment about which explanation is the best until you've had a look at the candidate explanations, and we've noted that there will, strictly speaking, be infinitely many candidates. Of course you can't compare them *all*. But in most cases, at least when the number of facts we have is fairly extensive (albeit finite), the number of candidates that have any *realistic* chance of being true will be manageably small, and we can confine our attention to the frontrunners. Sometimes rational people disagree about which explanatory theory is the best; this is especially common in the empirical sciences, where often there are competing theories, each with its proponents, to explain the observations and experimental results gathered so far. **Exercises** 1.What is the structure of an argument to the best explanation? 2.True or False: Arguments to the best explanation are inductive arguments. 3.True or False: For any finite set of facts or evidence we've gathered (i.e., in any case where we haven't gathered every fact that could possibly be relevant to the event or state of affairs we're trying to explain), there is always more than one logically possible explanation for those facts. 4.How do we determine which explanation for a set of facts is overall the best explanation? 5.Could an explanatory theory E1 satisfy some of the criteria for explanations better than another explanation E2, but satisfy other criteria less well than E2? 6.Can there be cases in which there is not enough evidence yet to make a strong argument for the truth of any particular explanatory theory? 7.Consider the facts below: a. What explanation seems most plausible here? b. Consider this alternative explanation and indicate which explanatory virtue it most obviously lacks: *Mrs. Jones took her jewelry to be cleaned*. c. Consider this alternative explanation and indicate which explanatory virtue it most obviously lacks: *the pastor at the Jones's church busted down the door and took the jewelry as a practical joke with every intention of returning it.* d. Consider this alternative explanation and indicate which explanatory virtue it most obviously lacks: *the jewelry was stolen by burglars who picked the lock to get in, but later and independently some vandals broke through the front door and made a mess of the house*. e. What piece of information that might later be added to the list of facts would enhance the probability that the Jones's house was burglarized by street criminals? 8.Consider the facts below and then evaluate the argument to the best explanation that follows. 1. Jones was a man in his mid-20s 2. Jones was found dead with a gunshot wound to the head 3. Jones had no history of mental illness 4. Jones had a job 5. Jones was not married and had no children 6. No suicide note was found 1. The facts are as stated in (1)-(6). 2. The overall most plausible explanation is that Jones committed suicide with a firearm. 3. So, probably that's what happened. 9.When you perform an experiment in a science class and don't get the result the text says you should get, you probably assume that you botched the experiment somehow, and you're almost certainly right. How is reasoning to the best explanation involved here? 10. Physicists would like to formulate a theory that could explain everything we know about the behavior of matter and energy by reference to just a few fundamental particles or other entities and (hopefully) without having to assume that there are (say) over twenty dimensions of space. What two criteria for theory choice are most obviously implicated here? 11. True or False: Arson investigators use reasoning to the best explanation. 12. True or False: Doctors never use reasoning to the best explanation. 13. True or False: Argument to the best explanation is used in the sciences and in crime investigations, but not in everyday life. 14. True or False: Two people can disagree not only about what the best explanation for a set of facts is, but about what are the facts that need explaining. 15. True or False: Whether a certain explanatory theory "makes sense" of a certain fact is not a matter of degree. 16. True or False: Although we prefer theories that fit with things we already believe about how the world works, some explanatory theories can lead us to abandon some of these background beliefs. 17. The chapter pointed out that lawyers at trial use argument to the best explanation. Lawyers, like everyone else, want their theories (of a crime) to be simple, explain all the evidence, and so on. But in the context of a criminal trial, there are two extra requirements placed on any such argument. What are they? 18. Geologists believe there once was a huge "super-continent" on Earth that later broke up, the pieces drifting apart. What most prominent facts about the continents seem best explained by this theory? 19. Some people claim to have had "near-death experiences" (NDEs). What are some of the relevant facts here, and what seems like the best explanation for these experiences? 20. Some people claim to have been abducted by aliens. What are the relevant facts here, and what best explains them? [Optional exercise 1]: Although we generally take it for granted that, all else being equal, the simplest explanation is mostly likely the correct one, some philosophers question the implicit assumption that the universe can be expected to do things in the simplest ways. Is there any good epistemic (as opposed to pragmatic) argument for favoring simpler theories, assuming we can formulate a reasonably precise definition of explanatory simplicity? [Optional exercise 2]: *Explanatory depth*, mentioned above in connection with "making sense" of the facts, is about how far an explanation can go in answering a persistent series of "why" questions. Suppose, in regard to the burglary exercise, that someone asks *why* the Jones residence was burglarized. My explanation may or may not provide an answer. If it doesif I say, for example, that it was the work of a street gangthen I might be asked again why that street gang chose to burglarize the Jones residence. And if I have an answer for that, I might be asked yet another, deeper why-question. Here's a question: must any plausible explanation have such depth that it can answer literally every why-question someone might raise? [Optional exercise 3]: Connected with the issue raised in the preceding question, is it possible to provide an explanation for literally *everything*? Suppose we have an indefinitely-long conjunction of facts describing the world and everything in it down to the last detail, including every event that ever happens. Can that conjunctive fact have an explanation? **Chapter 20. Testimonial Arguments** **Part A. Arguments Based on Testimony** Much of what we believe are things we accept on the testimony of others. The proportion of our beliefs that we're in a position to confirm for ourselves is small. Here are some examples: - Julius Caesar was assassinated. - Helium atoms have two protons and two neutrons in their nuclei. -  has an infinite, non-repeating decimal expansion. - Mt. Everest is around 29,000 feet high. - Your car needs a new alternator. If you believe any of these statements, then most likely you believe it because you've been told it's true by someone else. Examples could easily be multiplied. The point is that we very often reach conclusions based on the say-so of others. When you argue for a conclusion this way, you're using a **testimonial argument**. The argument below is a testimonial one. 1. Jones says that diamond is the hardest natural substance. 2. Jones's testimony is credible. 2a. Jones is a gemologist. 3. So, probably diamond is the hardest natural substance. What makes the argument a testimonial one is the fact that we're to accept the conclusion because Jones tells us it's true and because, according to (2), Jones's testimony carries weight. The argument below, by contrast, is *not* a testimonial argument. This argument reaches the same conclusion, but not on the basis of anyone's say-so. (That these tests were performed, and that diamond came out on top, is something we might accept on testimony, but that would be a different argument with a different conclusion.) Instead, non-testimonial evidence in the form of hardness tests are the basis for accepting (3) in this case. We have a testimonial argument for P when we're to accept the conclusion just because 1. Some person or group says P is true, and 2. This person's or group's testimony is credible enough to warrant our acceptance - The characteristics of the testifier/source - The intrinsic likelihood or unlikelihood of that for which the testimony is given - The extent to which the testimony is corroborated by other trustworthy sources We take up the first issue in the next section. Let us pause to note, however, that one sometimes hears it said that all arguments based on testimony or authority are fallacious. In order to make that claim, one must suggest either that every source lacks credibility, which is false, or that no source's credibility can ever be known, which also is false, or that even credible testimony cannot be trusted, which is incoherent. What is certainly true is that testimonial arguments based on *incredible* (i.e. not credible) testimony are fallacious. What is arguably true is that first-hand knowledge, *where it can be had*, is generally more desirable than knowledge based on testimony. **Part B. Source Characteristics** What factors are relevant in assessing the credibility of the testifier? Consider the following: - *The source's sincerity, confidence, and sobriety*. For obvious reasons, testimony only counts for something when it's offered sincerely. And because sometimes even sincere testimony is offered only with some trepidation or tentativeness, the source's confidence clearly is relevant, too. In referring to the source's "sobriety," we mean more than simply that he or she is not drunk; we mean, more generally, that the person is clear-headed when the testimony is given, as opposed (say) to being under extreme stress or fatigue. - *The source's qualifications to speak on the matter at hand*. When the topic is one that requires some particular expertise, we often refer to the argument that results as an **argument from authority**, but in its essence it's really just an argument from testimony. An "authority" does not have to be a white-coated lab experimenter with a Ph.D., after all. A chef is an expert on cooking, a cop is an expert on the crime patterns in his or her jurisdiction, and an auto mechanic is an expert on car engines. In any such case, we consider whether the person speaking has the requisite knowledge to be relied upon, or whether he's in some other way well-situated to speak on the matter at hand, such as (in the case of eyewitness testimony) having been present during some event with a good vantage and reasonably well-functioning sense faculties. - *The source's track record of reliability*. If this person or organization has proven reliable in the past, then it's natural (and reasonable, all else being equal) to put our trust in the source again. Here we need to be specific, however, since even generally reliable people might have a spotty record of reliability when it comes to the specific topic at hand. - *The source's bias, if any, on the subject at hand*. When someone does *not* ask you to take her word for something, but instead provides you with independent evidence to support what she's saying, her bias is no longer relevant; you have to examine the evidence and never mind where it came from in such cases (we'll expand on this point in the chapter on "ad hominem" arguments). But when the argument is a genuine argument from testimony, which means *all we have to go on is the testimony itself,* with no independent evidence to back it up, the testifier's bias, if any, is a relevant consideration, even if not a decisive one. It's one thing to consider in ascertaining the inductive strength of the testimonial argument. Putting all these together in an overall assessment of a source's credibility is not an exact science, of course. You'll have to use your best judgment here. But there are other factors to consider even when the source seems reliablefactors that will partially determine whether the testimony is credible *enough* to warrant our acceptance. **Part C. Antecedent Probability** Suppose I tell you with apparent sincerity that I saw the King of England shopping alone in a Walmart here in the U.S. last week. You're not likely to accept my testimony as accurate, even if I work at Walmart, I'm generally honest, and so on. The reason is obvious: what's being testified to in this case is something highly improbable, and this has to affect your evaluation of the conclusion's likelihood. If I tell you with equal sincerity that I saw our local mayor shopping at Walmart, then you're likely to take my testimony more seriously. So, along with the testifier's own credibility, the antecedent (or "intrinsic") probability of what's being claimed by the testifier is relevant to the credit we can give to the testimony. **Part D. Corroboration** Corroborated testimony is testimony backed up by other testifiers. Ideally, those would be independent sources of at least equal reliability. The more corroboration we have, the stronger a testimonial argument we can make, which is why being able to call on multiple independent witnesses to some fact at trial strengthens a legal case, and why getting a second opinion in medical matters is worthwhile, especially in serious cases. Enough corroboration might be enough to overcome the antecedent improbability of what's being claimed in many cases. It's not clear just how many people it would take to convince a rational person that the King of England really was in the local Walmart yesterday, but if enough independent sources (shoppers, store personnel, and local police) swear to it, you're likely to start believing it, and justifiably so, even before it's confirmed by the UK government (which, of course, is for you just another source of testimonyas would even be the King's own statements on the matter). **Part E: Summary** As should be clear at this point, evaluating testimonial evidence can be a complicated affair, although it's something we do practically every day in a more or less intuitive way. In addition to the main general factors we've mentioned, there is a grab-bag of other issues that might factor into our deliberations, such as the context (am I getting what seems like sincere testimony from multiple and usually reliable sources, but on April 1^st^?), or the remove at which we stand from original testimony passed through multiple persons before reaching us. Information has a way of getting distorted as it passes through intermediaries, especially when they feel the need to simplify or paraphrase it before passing it on. All this being said, to the extent that we consider the variety of factors listed above, our assessments will be more informed and, hopefully, more reasonable. **Exercises** 1.What is the form of a testimonial argument? 2.For each argument below, indicate whether it's a testimonial argument. 3.True or False: Testimonial arguments generally are inductive. 4.Consider the testimonial argument below and explain how we should assess the truth of the second premise. 5.What is the **fallacy of appeal to illegitimate authority?** Give an example. 6.Which of the three arguments below commits the fallacy mentioned in the preceding exercise? 7.Of the criteria mentioned in this chapter for evaluating testimony, which ones would be most relevant to evaluating the testimony of a used car salesperson about the condition of one of his or her offerings? (Incidentally, some would suggest that you need not rely on testimony here, since you can have the car independently inspected. Is this right?) 8.Why would the testimony given in Argument 1, below, have more credibility than the testimony in Argument 2, even if given by the same person under the same circumstances? 9.One specific type of testimonial argument is called (in Latin) ***argumentum ad populum*** 10\. Evaluate the *ad populum* arguments below (each one is formulated enthymematically, with the second premise left implied). Is one stronger than the other? 11\. Which of the two *ad populum* arguments below (if either) is the stronger? 12. What's wrong with the *argumentum ad populum* below? 13. In some casessometimes as a stalling tacticwe're told that we should delay taking action on something or other because the testimony of experts is mixed, meaning there is no consensus among them. What questions ought we to ask about this "mixed" testimony before deciding on a course of action? 14. Suppose we're told that experts disagree about the connection between a certain substance found in a local water supply and cancer. Without more specific information, it's difficult to be sure exactly what they disagree about. List three possibilities. 15. List three ways in which to evaluate the testimony you find online. [Optional exercise 1]: The *principle of credulity*, applied to testimony, basically tells us that we should believe what other people tell us unless we have some good reason not to, which basically means that all testimony comes with *prima facie* credibility. Do you agree? Why or why not? (In ordinary English, *credulous* is sometimes used interchangeably with *gullible*, which means being too easily fooled by others. In the present context, however, *credulous* only refers to a tendency to believe what you're hearing. \[Hence, when you're *incredulous*, you're inclined *not* to believe what you're hearing.\]) [Optional exercise 2]: The eighteenth century Scottish philosopher David Hume once argued against the rationality of believing in miracles based merely on testimony. Hume's argument is notoriously open to varying reconstructions, and the version below has a conclusion that might be stronger than the one Hume actually endorsed, so evaluate it on its own merits. (See the essay "Of Miracles" in Hume's *An Enquiry Concerning Human Understanding* for the full presentation of his argument.) [Optional exercise 3]: At trial, the testimony of a witness can be attacked in many ways, including many of those we've talked about. But partly because trials use various rules about the admissibility of evidence, there are some other ways in this context to attack witness testimony. Can you list any? ::: {.section.footnotes} ------------------------------------------------------------------------ 1. ::: {#fn1} Suppose you're performing an experiment in a science class and that the textbook tells you which result you should get. If the text is right, you'll probably get (this) result. But suppose you don't get (this) result. It doesn't follow that probably the text is wrong, since it's antecedently very likely that the text is correct.[↩](#fnref1){.footnote-back} ::: 2. ::: {#fn2} If there is intelligent life elsewhere in the Milky Way galaxy in particular, would it be surprising not to have any evidence of that? Some suggest that it is, especially given the number of extrasolar planets we've discovered. This is known as "Fermi's Paradox."[↩](#fnref2){.footnote-back} ::: 3. ::: {#fn3} For more on this argument, see J. L. Schellenberg, *Divine Hiddenness and Human Reason* (Cornell, 2006).[↩](#fnref3){.footnote-back} ::: 4. ::: {#fn4} Some standardized test problems are labeled "analogies," and they usually take the form of a problem such as "quiet is to loud as soft is to (?)" The idea is that the pairs are analogous in respect to the relation between the items in them.[↩](#fnref4){.footnote-back} ::: 5. ::: {#fn5} *Lawrence v. Texas* (U.S. Supreme Court, 2003).[↩](#fnref5){.footnote-back} ::: 6. ::: {#fn6} *Erotic Service Provider v. Gascon* (Ninth Federal Circuit, 2018).[↩](#fnref6){.footnote-back} ::: 7. ::: {#fn7} Argument to the best explanation is sometimes called "abductive" argument.[↩](#fnref7){.footnote-back} ::: 8. ::: {#fn8} Whether arriving at the objective truth is always (or ever) the goal in science is disputed by some philosophers of science, but we'll set this issue aside here.[↩](#fnref8){.footnote-back} ::: 9. ::: {#fn9} Any theory that could be true must be internally consistent, but this criterion is obvious enough that we'll let it be understood.[↩](#fnref9){.footnote-back} ::: 10. ::: {#fn10} See optional exercise \#2 for more on the issue of explanatory depth.[↩](#fnref10){.footnote-back} ::: 11. ::: {#fn11} Some authors call this style of argument "bandwagon" reasoning and treat it as if every instance of it were fallacious. But not all such arguments are fallacious; they must be considered case by case.[↩](#fnref11){.footnote-back} ::: 12. ::: {#fn12} That Hume actually defended only the weaker claim that no miracle *so far* recorded in history can be believed, although some miracle reports might in principle be rationally believed, is suggested by his "eight days of darkness" example in Part II of "Of Miracles." Whether it could rationally be believed even in this case that the event was caused by a deity is another issue.[↩](#fnref12){.footnote-back} ::: :::