Lessons in Logic 7: Argument Analysis
In this lesson, you'll put together what you learned in Lessons 4 and 6 and use it to analyze the structure of arguments. By the "structure" of arguments, I mean the logical relationships of premisses and conclusion within the argument, as opposed to the internal logical structure of statements, which is a more advanced topic.
In the logical study of actual arguments, you need to analyze the argument before evaluating it. If your analysis is incorrect―for instance, if you mistake the conclusion for a premiss―your evaluation is bound to go wrong. It's important not to jump to an evaluative conclusion: you must first understand what the argument says before criticizing it. Mistakes in argument evaluation can result from either skipping the analysis step, rushing through it, or allowing a premature evaluation to prejudice the analysis.
For this lesson, you will be analyzing passages that contain only a single argument and nothing else. This is unrealistic, since most arguments come embedded in longer passages that contain non-argumentative material, and arguments are often linked together. However, it is useful to start out analyzing simple examples, and you will analyze more realistic ones in the next lesson.
To analyze the structure of an argument, use argument indicators to determine what the premisses and conclusion are. If every premiss of an argument came labelled with a premiss indicator, and every conclusion with a conclusion indicator, or if the conclusion always came at the end of the argument, then argument structural analysis would be easy. What makes it difficult is the fact that some premisses or the conclusion will not be marked―occasionally none of them will be marked―and the conclusion may not come at the end.
It's helpful, at least at first, to follow these steps in analyzing an argument:
- Identify the statements. How many statements are there in the argument? You might underline and number them.
- Is there a conclusion indicator? If so, which statement does it indicate is the conclusion?
- Are there any premiss indicators? If so, which statements do they indicate are premisses?
- Are there any statements not marked by indicators? If so, you may be able to tell whether they are premisses or conclusions by a process of elimination.
- Finally, reconstruct the argument and check to see whether it makes sense.
…[S]ince you're a compulsive planner, you've built in a pending sense of shame, because you know that you won't have time to do any of these things.
Source: Susan Gregory Thomas, "'Scuse me while I boot up a lecture", U.S. News & World Report, 1/13/1997
- There are three statements:
- You're a compulsive planner.
- You've built in a pending sense of shame.
- You know that you won't have time to do any of these things.
- There's no conclusion indicator.
- There are two premiss indicators: "since" and "because". "Since" indicates that the first statement is a premiss, and "because" indicates that the third statement is also a premiss.
- By a process of elimination, the second statement must be the conclusion.
- Here's a reconstruction of the argument:
You're a compulsive planner.
You know that you won't have time to do any of these things.
Therefore, you've built in a pending sense of shame.
This makes sense, so the analysis is complete.
Argument analysis is just like any physical or intellectual skill: you will become skilled at it only through practice.
Exercise: Analyze the logical structure of the following argument.
…[A]n entity cannot possess the right to life unless it has the capacity to desire its continued existence. It follows that, since fetuses lack the conceptual capacity to desire to continue to live, they lack the right to life.
Source: Don Marquis, "Why Abortion is Immoral", in Robert J. Fogelin & Walter Sinnott-Armstrong's Understanding Arguments (Fifth Edition), p. 471.
- Conclusion Indicators
- Arguments and Explanations
- Premiss Indicators
Next Lesson: Complex Arguments
Fallacy Files Book Club: Unspeak, Chapter 1
This is the first installment of a new feature: I'll be reading and commenting on Steven Poole's book Unspeak, as an alternative to doing a full review. Instead of reading the book and reviewing it as a whole, I'll be reading it approximately a chapter at a time and asking questions and making comments about each chapter as they occur to me.
If you wish to follow along, there is a link below to an online version of the introductory chapter that I'll be reading today, which is not the complete first chapter of the book. For future installments, you'll need to get ahold of a hard copy. If you have comments of your own about the book, or metacomments about my comments about the book, I look forward to hearing them. This is an experiment and, depending on how it goes, I may either speed up or slow down the pace. The initial plan is to complete at least one chapter a month. So, let's get started!
In the Introduction, Poole gives three examples of "unspeak":
- Tax relief
- Friends of the Earth
He goes on to explain why these are examples:
Each of these terms…is a name for something, but not a neutral name. It is a name that smuggles in a political opinion. And this is done in a remarkably efficient way: a whole partisan argument is packed into a sound bite. These precision-engineered packages of language are launched by politicians and campaigners, and targeted at newspaper headlines and snazzy television graphics, where they land and dispense their payload of persuasion into the public consciousness.
Based on these examples and what Poole says about them, he seems to mean by "unspeak" a type of loaded language, or "question-begging epithets", as Jeremy Bentham called them. Specifically, they are names given by advocates to political programs―"tax relief"―political movements―the "pro-choice" movement―or political organizations―"Friends of the Earth".
Poole calls them "unspeak" rather than "doublespeak" because they don't say one thing while meaning another, but instead attempt to pack an argument into an epithet, which is what makes them question-begging. Also, they are "unspeak", because they put those on the other side of the issue in an untenable position, such as being "anti-choice" or an "enemy of the earth". So, it's not because such terms are empty of meaning that they are "unspeak", but that they silence the opposition by a kind of poisoning the well.
I'm not too pleased with the term "unspeak", but I do think that Poole has put his finger on a distinct phenomenon, and one which has become more common in recent years. However, it also seems to be a rather narrow one, and I wonder how he can get a whole book out of it. To do so, I suspect that he'll have to stretch the notion of "unspeak" beyond the definition and examples given in this introductory chapter. Indeed, in the course of explaining why euphemisms aren't unspeak, he gives the example of "ethnic cleansing" and promises a further discussion of it in Chapter 4; but if it's a euphemism rather than unspeak, why discuss it any further? Well, we'll see.
At the end of this online introduction, Poole gives three further examples of unspeak:
- Anti-social behaviour
- War on terror
Only the last of these seems to fit the definition of "unspeak" and to be similar to the earlier examples. "Anti-social behavior" is a term from the social sciences, and "tragedy" is surely not unspeak all by itself, though it might form part of a larger unspeak term. Presumably, Poole will explain how and why these terms count as unspeak in the remainder of the book, so that we'll just have to see.
Those are my thoughts on the main substance of the online introduction, but I also have a few nits to pick:
- Poole tells a "just so" story about the history of the term "pro-life":
Campaigners against abortion had from the early 1970s described their position as defending a ‘right to life’. The opposing camp, previously known as ‘pro-abortionists’, then renamed their position ‘pro-choice’, rhetorically softening what they favoured. Defending a woman’s ‘right to choose’ whether to have a baby or not, the slogan ‘pro-choice’ appealed to an apparently inviolable concept of individual responsibility. It sought to cast adversaries as ‘anti-choice’: as interfering, patriarchal dictators. … Indeed, anti-abortionists quickly trumped that linguistic strategy by beginning to call themselves ‘pro-life’, a term first recorded in 1976. The phrase ‘pro-life’ appeals to a sacred concept of ‘life’, and casts one’s opponents―those who think abortion should be legally available―necessarily as anti-life, in fact pro-death. In a conceptual battle of two moral ideals, ‘life’ easily wins out over ‘choice’.
I'm dubious about this account for the following reasons: For one thing, I don't think that very many people ever called themselves "pro-abortion"―rather, it's a pejorative term used by "pro-life" people about their opponents. More importantly, it's easy to show that the term "pro-life" was in use well before 1976. In a footnote (not in the online version of the chapter), Poole attributes the date to William Safire and the OED. However, the Willkes' Handbook on Abortion of 1971, which was a pre-"Roe v. Wade" anti-abortion tract, uses the term "pro-life" three times in its Foreword alone. Of course, neither Safire nor the OED are infallible. In general, it's almost impossible to verify such word histories, because to do so requires establishing when the word was first used. However, to establish a word's first use requires showing that there are no previous occurrences, and such universal negative generalizations are almost impossible to establish empirically―try showing that no crows are white. It's much easier to refute such a history by finding a counter-example, as I've just done.
- I don't see why Poole thinks that George Orwell took a "defeatist" attitude towards political language in his great essay "Politics and the English Language", unless Poole is committing the "all or nothing" version of the Black-or-White fallacy. Orwell does say that political language consists "largely of euphemism, question-begging and sheer cloudy vagueness". One might argue with the "largely", but I can't see how one can deny that at least some political language is like this, and even that much is. It doesn't follow from the fact that much political language is hot air that it all is, and Poole is right that political euphemisms are worth paying attention to. When politicians put up a smoke screen, they're hiding something. Moreover, when Orwell mentions "question-begging" he may have had in mind language that Poole would classify as "unspeak".
In the next installment, I'll finish reading the remainder of the introductory chapter.
Source: Steven Poole, "Extract: Introduction", Unspeak
Debate Fallacies, Part 3
In the most recent Democratic presidential debate, the following exchange occurred between Anderson Cooper and Barack Obama:
COOPER: Senator Obama, Mitt Romney has accused you this week of saying that 5-year-old children should be getting sex education. Was he right?
OBAMA: Ironically, this was actually a proposal that he himself said he supported when he was running for governor of Massachusetts. Apparently, he forgot.
According to Fact Check, Obama was wrong in accusing Romney of supporting the same proposal. However, even if he were right about the facts, a Logic Check shows that he committed a Tu Quoque.
- "Part I: CNN/YouTube Democratic Presidential Debate Transcript", CNN, 7/24/2007
- Viveca Novak, et al., "Dems Face YouTube Interrogators", Annenberg Political Fact Check, 7/25/2007
Reader Response (8/1/2007): Paul Farrington writes:
It is my opinion that despite having the form of a tu quoque move, Obama's response as quoted by you is not in fact a "tu quoque". It seems to me to be essential to a tu quoque that the action being discussed be something that the user of the move himself consider to be wrong under normal circumstances (something implicit in the type name "two wrongs make a right").
In the case in point, Obama went on to defend the policy of providing sex ed to kindergarten children in a way that shows he clearly does not believe this to be a bad thing to do and it is clear that his endorsement of such a policy is not dependent on Romney doing likewise. Thus it would appear he was not committing a tu quoque.
I'm not even sure it's an ad hominem, though this is a trickier area: certainly Obama is implying that his opponent is a hypocrite, but (if the accusation were true) this would be a legitimate issue to raise in the context (especially given the tone of Romney's "accusation" as quoted by the Fact Check gurus). It seems to me to be a debatable ad hominem, not a tu quoque.
I don't think that Two Wrongs Make a Right requires the arguer to believe that the "two wrongs" are actually wrong; rather, it is usually the audience to whom the argument is addressed who believe this, and the goal of the argument is to convince them that the second "wrong" is not really wrong.
Also, Tu Quoque is a type of ad hominem, since it attacks the arguer rather than the argument. The distinctive characteristic of tu quoques, which sets them apart from other ad hominems, is that the critic is attacked for doing the same thing that he criticizes. That's certainly the case here. So, if the example is indeed an ad hominem, then it's a tu quoque, since it has the distinctive characteristic.
Obama could have defended sex education for young children without engaging in a personal attack on Romney, and I don't see how such an attack at this point in the campaign is relevant. If it so happens that both Obama and Romney are nominated and end up running against one another, then it would certainly be relevant for each to point out the faults of the other. However, at the moment Obama is running for the nomination against other Democrats. The reason for attacking Romney for supposedly supporting the very thing that he criticized is to divert attention from the criticism onto the criticizer, which is the classic tu quoque move.
Q: I've been looking for the name of a certain fallacy which I call "No correlation implies no causation". It goes like this:
Instances of A are increasing. Instances of B are decreasing. Therefore A does not cause B.
A occurs many times while B does not occur. Therefore A does not cause B.
The following are examples of these arguments:
Poverty has lessened in the last few years while crime has gone up. Poverty does not cause crime.
I've been smoking all my life and I'm not dead yet/don't have cancer. Cigarettes don't cause death/cancer.
C02 emissions increased over the decades 1940 to 1970 while the global temperature went down. CO2 doesn't cause global warming.
Is this a named fallacy? Am I correct in saying that cause does not necessarily imply correlation in any given case or period of time?―Justin Barker
A: The word "cause" is ambiguous in that we talk about the causation of both individual events and types of events. For instance, when we say that cigarette smoking causes cancer, we are talking about types of event. It's usually impossible to say of an individual case of lung cancer that it was caused by smoking, even if the sufferer smoked, since there are cases of lung cancer in non-smokers. Types of event may have many causes, for instance, lung cancer can also be caused by exposure to asbestos.
It is truly said that correlation is not causation, because two types of event may be correlated without either causing the other. So, correlation does not imply causation. However, it is also true that causation does not imply correlation, because of the complexity of causation for types of event. For example, suppose that an event of type E has two types of cause, say, C1 and C2. Then it is possible that neither C1 nor C2 is correlated with E. For instance, the rate of E may remain steady although the rate of C1 falls, if at the same time the rate of C2 rises concomitantly. Thus, neither C1 nor C2 is correlated with E, yet both are causes of it.
So, just as one should not jump to a causal conclusion based on a mere correlation, one should also not jump to a conclusion of non-causation based on non-correlation. The former type of error is common and usually goes by the name "false cause", but the latter as far as I know is nameless. Also, I don't know how common the "non-correlation therefore non-cause" error is―though there's a real-life example in the Resource below―so I hesitate to call it a logical fallacy, since a fallacy is a common type of error.
Nonetheless, it's worth keeping in mind that the relation between correlation and causation is never more than suggestive. A correlation between two types of event may suggest that one causes the other, but more evidence is required to establish it. Similarly, non-correlation may suggest no causation, but doesn't prove it.
Resource: Non-Correlation Isn't Non-Causation, 7/27/2004.
Use of jets to fight fires up in air
Source: Patrick O'Driscoll, "Use of jets to fight fires up in air", USA Today
Blurb Watch: Transformers
In the following ad blurb for the new movie "Transformers", the words of the critic are what's transformed:
"SHEER FUN AND EXCITEMENT…
'TRANSFORMERS' IS PRIME SUMMER ENTERTAINMENT."
Claudia Puig, USA TODAY
In the review, Puig calls the movie "sheer mindless fun and excitement". The blurb omits the word "mindless" without any notice―such as using an ellipsis to mark the deletion―which makes her comment seem more favorable than it is. The second sentence is actually the title of the review, which was probably supplied by an editor rather than the critic. Moreover, in order to turn the title into a sentence, the ad writer supplied the verb "is". The next critic undergoes a similar transformation:
"MICHAEL BAY SCORES."
Roger Ebert, CHICAGO SUN-TIMES
This blurb is actually taken from the caption to an image from the movie, which was probably written by an editor rather than Ebert. Both reviews are largely favorable, so the transformations these reviews have undergone are only slightly misleading.
- Ad for "Transformers", New York Times, 7/13/2007, p. B14
- Roger Ebert, "Transformers", Chicago Sun-Times, 7/5/2007
- Claudia Puig, "'Transformers': Prime summer entertainment", USA Today
Fifty Million Website Visitors Can't Be Wrong
"During an interview with a small French television channel in November, [French Cabinet member Christine] Boutin was asked if the Bush Administration might have been involved in the destruction of the World Trade Centre in New York, and the attack on the Pentagon. 'I think it is possible,' she said. Ms Boutin…said in the interview that she had been impressed by the strength of opinion on the internet that favoured the notion of a US conspiracy in the 9/11 attacks. 'I know that the websites that speak of this problem are websites that have the highest number of visits…. And I tell myself that this expression of the masses and of the people cannot be without any truth,' she said."
There are three fallacies lurking here:
- Boutin supports the claim that President Bush might have been involved in 9/11 with the notion that website traffic is some kind of measure of truth.
- Boutin made these remarks during an interview that was conducted before she joined the cabinet, and they were only recently publicized by conspiracy theorists. Apparently, the conspiracists think that it means something that the French Minister for Housing and Town Planning said it. This sort of appeal to dubious authority appears to be a standard conspiracist ploy, as we saw earlier with Charlie Sheen and Rosie O'Donnell.
- The conspiracist website that publicized the comments put up a video clip from the interview which was carefully edited:
Ms Boutin’s office noted yesterday that the internet video had dropped the next sentence, in which she said: "I’m not telling you that I adhere to that position." Christian Dupont, chief of staff for Ms Boutin, said that her remarks had been distorted to create a row.
This kind of video contextomy is another standard technique of conspiracist propaganda, as we saw earlier with the "pull it" quote.
- Charles Bremner, "Minister has to deny linking Bush to 9/11 attack", Times Online, 7/9/2007
- Oriane Raffin & Benoît Vitkine, "Christine Boutin rattrapée par ses propos controversés sur le 11-Septembre", Le Monde, 7/5/2007
Resource: Pull Quote, 5/12/2007
Check 'Em Out
- (7/5/2007) Ben Goldacre's latest "Bad Science" column gives a nice example of why playing the Hitler card is, in fact, fallacious. However, the larger point that Goldacre is trying to make with the example is that scientific studies should be judged on their own merits, and not on the basis of ad hominem arguments directed at their authors.
In a perfect world, I would entirely agree with Goldacre, but most people lack the knowledge, time, and inclination to be able to examine the data for themselves. So, most of us have to turn to experts―including Goldacre himself―to examine the data for us. Since there are many competing "experts" who say contradictory things, we have to decide who is trustworthy. If we could examine the data ourselves, we wouldn't need experts; but since we often cannot examine the data ourselves, we have no alternative but to examine the experts.
Of course, not every failing is relevant to trustworthiness. Goldacre asks:
…[H]ow bad would someone have to be for you to completely disregard the findings from their research, simply on the grounds of who they were? An adulterer? A recipient of private consulting fees? How about a cold-blooded racist, homophobic mass murderer?
The only one of these facts that's likely to be relevant is the morally trivial one. It's not obvious how being an adulterer, or even a mass murderer, could lead to misleading scientific studies; but it is obvious how a financial interest can lead someone to fudge data. For this reason, to criticize scientific research on the grounds that the scientist is an adulterer, a racist, a mass murderer, or a Nazi, is to commit an ad hominem fallacy. However, to criticize the scientist for receiving consulting fees is prima facie relevant, and thus not fallacious.
I share Goldacre's distaste for the ad hominem accusations that make up much of public debate of scientific issues. In an ideal world, we wouldn't have to resort to such distasteful shortcuts, but then an ideal world would have no need for a "Bad Science" column―or the Fallacy Files for that matter.
Source: Ben Goldacre, "Examine the data, not the author", Bad Science, 6/30/2007
- Slate has an article on how some parents have become convinced that a vaccine was responsible for their child's autism, despite a lack of scientific evidence for such an effect. One culprit, according to the article, is confirmation bias, which is the tendency to look only for confirmatory evidence of a favored hypothesis and to ignore or discount disconfirmatory evidence. Here's a quote from one of the parents:
"I know what happened to my son after he got his MMR [Measles-Mumps-Rubella vaccine] shot," [Mary Wildman] told me. "I have no doubt. There's no way they'll convince me that all these kids were not damaged by vaccines."
That's confirmation bias on steroids, since she insists that no evidence could convince her. However, it's also clearly post hoc reasoning, because it's the fact that the autism happened to her son after the shot that appears to have convinced her that the shot caused the autism.
Source: Arthur Allen, "True Believers: Why there's no dispelling the myth that vaccines cause autism", Slate, 6/29/2007
- Three statements:
- An entity cannot possess the right to life unless it has the capacity to desire its continued existence.
- Fetuses lack the conceptual capacity to desire to continue to live.
- Fetuses lack the right to life.
- Conclusion indicator: "it follows that". Conclusion: The third statement. This is tricky. The conclusion indicator is followed by a clause set off by commas which contains the second statement, but that statement is not the conclusion. The question to ask is: what does "that" in the conclusion indicator refer to? It may help to rearrange the sentence to make the logical relationships clearer without changing its meaning:
…[A]n entity cannot possess the right to life unless it has the capacity to desire its continued existence. It follows that fetuses lack the right to life, since fetuses lack the conceptual capacity to desire to continue to live.
- Premiss indicator: "since". Premiss: "Fetuses lack the conceptual capacity to desire to continue to live". This reinforces the fact that the second statement is not the conclusion since it is a premiss.
- The first statement is not marked by an argument indicator. However, since the third statement is the conclusion, the first statement must be a premiss.
- The following reconstruction makes sense, thus confirming the correctness of the analysis:
An entity cannot possess the right to life unless it has the capacity to desire its continued existence.
Fetuses lack the conceptual capacity to desire to continue to live.
Therefore, fetuses lack the right to life.