Blurb Watch: Paper Man
|"'PAPER MAN' IS AN INTELLIGENT, METICULOUSLY CONSTRUCTED, WELL-ACTED MOVIE…"―Stephen Holden, THE NEW YORK TIMES||"'Paper Man' is an intelligent, meticulously constructed, well-acted movie that has been written to death."|
- Ad for "Paper Man", The Dallas Morning News, 4/30/2010
- Stephen Holden, "Triangle: Man, Wife and Friend (Imaginary)", The New York Times, 4/23/2010
What Fodor and Piattelli-Palmarini Got Wrong
Kenan Malik has reviewed Jerry Fodor and Massimo Piattelli-Palmarini (F&P)'s new book What Darwin Got Wrong, which I discussed last month (see the Resource, below). I've started reading the book, and have already read the last few chapters, which appear to be the most relevant to the "intensional fallacy" issue. Though I think that I now understand what an intensional fallacy is, the book is not really clearer in explaining it than the reviews have been. Here's Malik's description:
At the heart of the argument is the critique of what Fodor and Piattelli-Palmarini call the 'selection-for' fallacy: the belief that natural selection chooses particular phenotypes because they provide the organism with an evolutionary advantage. This cannot be, they argue, because of the presence of 'co-existent' traits. Some bodily features that Darwinists claim have been selected for in the course of evolution possess more than one trait or property, only one of which actually increases fitness (the Darwinian term for the capacity to reproduce), but all of which may be correlated with increased fitness.
An intensional fallacy in the relevant sense is not the Masked Man Fallacy, but a similar fallacy involving the substitution of co-extensive predicates within an intensional context. For example, to use one of F&P's examples, "having an organ that pumps blood" is co-extensive with "having an organ that makes thumping sounds". "Selection-for", according to F&P, is an intensional context. The process of natural selection has selected animals that have hearts, but it would be fallacious to infer that it has selected animals that have a thumping organ. The heart's noises are a "free rider" on its pumping of blood, meaning that the noises were not selected-for. Why, exactly, this is a problem for the theory of natural selection is not as clear.
If I understand Malik correctly, his criticism of F&P's argument is similar to mine:
Ironically, though, it is Fodor and Piattelli-Palmarini who confuse artificial and natural selection. … The fact that natural selection is 'blind' to traits that do not impact on fitness should not lead to the conclusion that it is 'blind' to all traits.
Read the whole thing.
Source: Kenan Malik, "Pigs Won't Fly", Literary Review
Resource: Intensional Fallacy, 3/5/2010
Update (5/2/2010): I've found another review of the book, this one by philosopher David Papineau, that makes similar points. After discussing Fodor's "first objection" to the theory of natural selection, namely, that it doesn't explain everything about evolution, he writes:
Fodor and Piattelli-Palmarini's second objection is much more peculiar. Not only does natural selection fail to explain everything―it doesn't explain anything.
Here's what he has to say about the "intensional fallacy" argument:
Fodor and Pialtelli-Palmarini do have an argument for these strange claims. They say that natural selection is insensitive to the difference between good traits and bad ones. One of their central examples is the evolution of the heart. Hearts pump blood, but they also make thumping sounds. As a result, any selective mechanism that favours blood pumpers will willy-nilly favour noise makers too―and, while blood pumping may help survival, thumping noises are themselves unhelpful. So, they conclude, selection favours unhelpful traits as well as good ones. … This isn't a good argument because, at its centre, lies a simple confusion. It is true that when an adaptive and a nonadaptive trait are tightly yoked together, natural selection will be forced to take them both or not at all, and so in these specific cases will be "blind" to the difference. But it doesn't follow that there is no relevant difference at all between the two traits in question. Of course there is. One trait helps survival and the other doesn't. And in general natural selection certainly isn't blind to this kind of contrast. Organisms with helpful traits tend to leave more offspring than those without. This means that there is a general tendency for helpful traits to spread, even if unhelpful traits like thumping sounds sometimes come along for the ride.
Of course, the fact that Malik, Papineau, and I have independently arrived at similar objections to the book's argument doesn't show that we're right and F&P are wrong. However, it does strongly suggest that, if F&P are correct after all, then a lot of people are misunderstanding them in similar ways. Either a step in the argument is unclearly expressed or it's just plain missing.
Source: David Papineau, "Review of What Darwin Got Wrong", Prospect Magazine, 2/2010
Poll Watch: UK Edition
Dan Hungerford sends the following report from across the pond:
Back in 2008, you pointed out many of the fallacies (logical and rhetorical) that were made in the context of the US presidential election. One of the more common ones, of course, was journalists treating polls as more accurate than they really were, i.e., committing the Fake Precision fallacy. Now they're doing the same in Britain, ahead of the General Election! Even the Times isn't immune to such errors:David Cameron scored a narrow victory over Nick Clegg in the second leaders’ debate last night after an impassioned contest turned personal. The Tory leader scraped a win―on 37 per cent to Mr Clegg’s 36 per cent―according to a Populus poll for The Times. Some 27 per cent gave the verdict to Gordon Brown as the pressure of tightening polls saw the trio trading ill-tempered barbs.
The Guardian wasn't quite so egregious, as it estimated a 4% lead rather than just 1%, but still definitely fell for the booby trap.Last night's outcome suggests that the dynamics of the campaign―which saw the Lib Dems emerge as a powerful third force in the election―have not been reversed. An hour after the end of the debate, a Guardian/ICM poll of 504 voters who watched the broadcast gave Clegg a narrow win, with the Lib Dem leader on 33% and Cameron and Brown on 29% each.
The simple fact that the two polls had different outcomes indicates that they should be taken with a grain of salt. If anything can be said at all, it's that the election is fairly close, at least as far as proportional support for the different parties is concerned.
I don't follow British politics, so I was a little surprised to see how differently the British newspapers report polls as compared to American practice. There is a lack of basic information in the British reports about how the polls are conducted, such as the number of people polled, the confidence level used, and the resulting margin of error. American news outlets usually report this information, though often in fine print at the end of the article. Of course, they frequently ignore the margin of error in the body of the article, but at least they provide it for those who know what it means. A different Guardian article on the same poll reported above (see Source 1, below) mostly followed the American practice.
The Times, to its credit, does have a useful "Poll Watch blog" by Peter Riddell that provides some of this information, for instance, that Populus polled 1,067 people. However, it also says that this was an online poll, which raises a red flag: without further details about how these 1,067 people were selected, we can't tell whether this is likely to be a representative sample of British voters.
An additional point about the Guardian poll is that its sample was only 504 people, which is about half of what is standard in American national polling. Thus, the margin of error is approximately plus-or-minus four percentage points, so that the difference between the three candidates is within that margin. Unfortunately, neither Guardian article cites the margin of error.
Dan makes an excellent point about comparing polls, and Riddell helpfully compares five polls taken immediately after the debate. The resulting chart graphically shows why polls should not be treated as if they are perfectly precise, since no two of the polls agree on all of the results. If the results were precise to the percentage point, then only one of the polls could be correct; but which one?
Unfortunately, one thing that British and American newspapers seem to have in common is that―with the exception of Riddell's "Poll Watch"―they only report on their own polls. This makes it difficult for the reader to do what Dan did, that is, to compare the results of more than one poll, which usually gives a better sense of polling precision.
- Julian Glover, "Guardian/ICM poll: Nick Clegg scores narrow victory in second TV debate", The Guardian, 4/23/2010
- Celinda C. Lake & Pat Callbeck Harper, Public Opinion Polling: A Handbook for Public Interest and Citizen Advocacy Groups (1987), p. 74.
- Peter Riddell, "Poll Watch: Five polls, so won the second leaders' debate?" [sic], Times Online, 4/23/2010
- Roland Watson, "David Cameron edges second leaders’ debate, according to Times poll", Times Online, 4/23/2010
- Patrick Wintour & Polly Curtis, "Election debate: Nick Clegg survives the storm", The Guardian, 4/23/2010
Are you more logical than a legislator?
In 1974, during a congressional committee debate on impeaching President Nixon, a motion was made by Representative McClory. Here is the exact wording of the motion as read into the congressional record:
Mr. McClory moves to postpone for 10 days further consideration of whether sufficient grounds exist for the House of Representatives to exercise its constitutional power of impeachment unless by 12 noon, eastern daylight time, on Saturday, July 27, 1974, the President fails to give his unequivocal assurance to produce forthwith all taped conversations subpenaed by the committee which are to be made available to the district court pursuant to court order in United States v. Mitchell.
Much of the subsequent debate on the motion was concerned with its wording and how it should be interpreted. Some congressmen criticized the motion's wording, while others defended it and explained its meaning. In the course of the debate, the following characterizations of the motion were given:
- Mr. McClory: "I am offering this motion, Mr. Chairman, to defer the conclusion of our proceedings for a period of 10 days, providing that we get assurance from the President by tomorrow noon that 63 of the 64 tapes which the Supreme Court has ordered to be made available to the district court, to Judge Sirica, are also made available to this committee."
- Mr. Dennis: "Mr. Chairman, this appears to me to be an exceedingly moderate resolution, which really ought to draw the support of everyone here. All it says is that if the President gives us assurances by tomorrow noon, that he is going to produce forthwith the matter which is to be given to the Special Prosecutor, that then we will give him the short period of 10 days in which to do so."
- Mr. Latta: "…I just want to call his [Mr. McClory] attention before we vote, to the wording of his motion. You move to postpone for 10 days unless the President fails to give his assurance to produce the tapes. So, if he fails tomorrow, we get 10 days. If he complies, we do not. The way you have it drafted I would suggest that you correct your motion to say that you get 10 days providing the President gives his unequivocal assurance to produce the tapes by tomorrow noon."
- Mr. McClory: "If the gentleman will yield, the motion is correctly worded. It provides for a postponement for 10 days unless the President fails tomorrow to give his assurance, so there is no postponement for 10 days if the President fails to give the assurance, just 1 day. I think it is correctly drafted. I have had it drafted by counsel and I was misled originally, too, but it is correctly drafted. There is a 10-day postponement unless the President fails to give assurance. If he does give assurance, there is a 10-day delay. If he fails to give it, there is only a 24-hour or there is only a―23½ hour day."
Who was right? Which, if any, of the above characterizations of the motion is correct and which incorrect?
Source: Debate on Articles of Impeachment: Hearings of the Committee on the Judiciary, House of Representatives, Ninety-Third Congress, Second Session (1974), pp. 137-8, 141 & 145-7.
Philosopher Antony Flew has died. He was the author of the wonderful book How to Think Straight, which I highly recommend. It's probably the best book of its kind, that is, a short book for general readers on what would now be called "critical thinking". Flew also, if I'm not mistaken, in that book identified what he called "the no-true-Scotsman move", which I've long intended to do an entry for if I can only figure out where it should go in the Taxonomy.
Source: Piers Benn, "Antony Flew obituary", The Guardian, 4/14/2010
Update (4/22/2010): The current eSkeptic has an article on the controversy over Flew's final book in which he argued for the existence of a god after a lifetime of atheism. Apparently, the book was ghostwritten by his supposed co-author Roy Abraham Varghese, though it was published with Flew's name first and most prominent. See the Sources below for the whole story.
Even if he didn't write the book, it appears that Flew did indeed become a theist at some point during the last several years of his life, as he was subjected to an unseemly tug-of-war between christian apologists on one side and atheists on the other with him in the middle. I guess that some christian apologists think that well-known atheists are authority figures for other atheists; what's worse, some atheists act like they agree!
Should the fact that Flew changed his mind mean that every atheist should suddenly convert to theism? More specifically, Flew actually became a deist; so, should everybody, christian and atheist alike, now become deists? Of course not. At best, we should be curious as to why Flew changed his mind, and open to changing our own minds in the light of any new evidence or arguments he had to offer. Unfortunately, if the reviews of the book that he didn't write are to be believed, there aren't really any new arguments to be found in it. Given that Flew didn't himself write it, I doubt that I'll bother to read it―I still haven't read some of the books that he actually did write.
What did the apologists think they were accomplishing in exploiting a dying man in order to produce pro-religious propaganda? I guess that some religious people are inclined to project the authoritarianism that often characterizes religion onto philosophy. In religion, we are often told to believe something because a holy person, or a holy book, said it. In philosophy, there are no authorities: no infallible philosophers or inerrant books. A philosopher or a book is only as good as its arguments, so those who took advantage of Flew's decline committed a fallacious appeal to authority of the first kind (see the entry for Appeal to Misleading Authority). The apologists have a lot to apologize for.
- Anthony Gottlieb, "I’m a Believer", The New York Times Sunday Book Review, 12/23/2007
- Kenneth Grubbs, "Antony Flew, 1923-2010: Following the Argument Wherever it Leads", eSkeptic, 4/21/2010
- Mark Oppenheimer, "The Turning of an Atheist", The New York Times, 11/4/2007
Where's the harm?
I've previously pointed (see the Resource, below) to one of Ben Goldacre's Bad Science columns concerning the mindboggling case of a Dutch nurse named Lucia de Berk, who was serving a life sentence for murders that may well not have even happened. In his latest column, Goldacre explains one of the statistical fallacies that convicted her:
The case against Lucia was built on a suspicious pattern: there were 9 incidents on a ward where she worked, and Lucia was present for all of them. This could be suspicious, but it could be a random cluster, best illustrated by the "Texas Sharp Shooter" phenomenon: imagine I am stood in front of a wooden barn with a machine gun in each hand, maniacally firing off a thousand bullets into the wall. I remove my blindfold, walk up to the barn, find 3 bullets which are very close together, and carefully paint a target around them. Then I announce that I am an olympic standard rifleman.
Read the whole thing.
Source: Ben Goldacre, "Lucia de Berk―a Martyr to Stupidity", Bad Science, 4/9/2010
Resource: Check it Out, 4/8/2007
Obama's Approval Rating Hits New Low
That's the headline of a CBS news report of recent poll results. Here's the relevant part of the story:
The latest CBS News Poll, conducted between March 29 and April 1, found Americans unhappier than ever with Mr. Obama's handling of health care―and still worried about the state of the economy. President Obama's overall job approval rating has fallen to an all-time low of 44 percent, down five points from late March, just before the health bill's passage in the House of Representatives. It's down 24 points since his all-time high last April.
However, the helpful graph that accompanies the report tells a different story: it shows that Obama's approval rating has been essentially unchanged all of this year. Unfortunately, the graph doesn't show the exact numbers for the approval rating this year, but they can be found in the individual poll reports (see Source 1, below). 44% is only two points lower than the previous "all-time low" at the beginning of the year and, thus, within the margin of error (MOE) of ±3 percentage points. Of course, supporters of the President may be disappointed that his approval rating has not improved since the passage of the health care bill, but there's no evidence in this poll that his approval has actually declined, despite the headline. (4/8/2010)
China Offers High-Speed Rail to California
An underwater train?
Check it Out, Too
I've often mentioned the dangers of shifting definitions in statistical studies (see, for instance, the Exposure section in the Fallacy linked below), but I probably can't mention it often enough. It's one of the most common sources of misleading numbers in the news, and you don't have to know anything about statistics to spot it! John Allen Paulos' latest "Who's Counting?" column examines this phenomenon through four recent examples: maternal deaths, violent crime, autism, and human trafficking. Read the whole thing!
Source: John Allen Paulos, "Redefinition: Humpty Dumpty and the News", Who's Counting?, 4/4/2010
A "RIVETING" Contextomy
An ad for the new movie The Art of the Steal gives a one-word quote of The Washington Post's Ann Hornaday: "riveting". As a general rule, you should be skeptical of one-word blurbs. The shorter a quote, the more context has been removed, and a single word is about as short as it can get.
In this case, the word "riveting" does not even appear in Hornaday's review of the movie in the Post (see Source 1, below). In addition, she gives the movie only two-and-a-half stars out of a possible four. If Hornaday really found the movie "riveting", wouldn't she have given it more stars, together with an enthusiastic review?
So, where did the word "riveting" come from? Apparently, it comes from an earlier article in the Post by Hornaday which previews movies coming out later in the year. However, it's clear from the way the article is written that Hornaday had not at the time seen some of the movies mentioned. Here is all that she wrote about The Art of the Steal:
The story of Albert Barnes, his breathtaking art collection and the fight over where it will reside should lead viewers on an eccentric, riveting ride.
So, it should prove riveting, but apparently turned out not to be. This is not the first time I've seen blurbs quoting a critic who hasn't yet seen the movie, but it would be nice if the studio could be shamed out of ever doing it again.
- Ann Hornaday, "Controversy, in broad strokes", The Washington Post, 3/26/2010
- Ann Hornaday, "Movie critic Ann Hornaday's picks for spring 2010", The Washington Post, 1/31/2010
Check it Out
I'm usually skeptical about tests that claim to diagnose your personality―especially online ones―but I was amazed at how accurate the following free online test was. Try it!
Source: "What's your Spectranality?"
Answers to "Are you more logical than a legislator?": McClory's motion is a compound proposition made up of two simpler propositions:
A. The committee will postpone for 10 days further consideration of whether sufficient grounds exist for the House of Representatives to exercise its constitutional power of impeachment.
B. By 12 noon, eastern daylight time, on Saturday, July 27, 1974, the President fails to give his unequivocal assurance to produce forthwith all taped conversations subpenaed by the committee which are to be made available to the district court pursuant to court order in United States v. Mitchell.
Thus, as a whole, the motion has the form: A unless B. However, B contains a negation in the word "fail", that is, B negates the simpler proposition:
C. By 12 noon, eastern daylight time, on Saturday, July 27, 1974, the President gives his unequivocal assurance to produce forthwith all taped conversations subpenaed by the committee which are to be made available to the district court pursuant to court order in United States v. Mitchell.
So, more specifically, the motion has the form: A unless not C. "Unless" means the same as "if not", so that the motion is equivalent to: A if not not C. Of course, the double negations cancel out, giving: A if C. Or, in other words: if C then A.
Now, let's look at the individual claims:
- McClory's rephrasing of his motion is correct, since "providing that" means the same as "if", so that McClory is saying: "A if C".
- Dennis' restatement of the motion is also correct: "if C then A".
- Latta begins by correctly restating the motion in his second sentence. However, he then goes on to infer―the word "so" is an argument indicator―that "if he [the President] fails tomorrow, we get 10 days". That is: "if not C then A". This is incorrect, since the recess is conditional upon the President not failing, that is, giving his assurances. In the next sentence, Latta claims: "if not A then C", which is equivalent to his previous, incorrect claim by transposition. Finally, Latta's proposed rewording of the motion is: "A if C". As we've seen, this is equivalent to the original motion, and perhaps clearer because it lacks the negation that seems to have confused Latta.
- McClory begins his defense of the motion's wording by correctly restating it, but then fallaciously infers (again, the word "so" is an argument indicator): "if not C then not A", which is an Improper Transposition. McClory again restates the motion correctly, then correctly states it as a conditional proposition: "If he does give assurance, there is a 10-day delay", that is: "if C than A". Finally, however, McClory again improperly transposes his own motion.
While the component propositions that make up the motion are rather long and complex, its basic logical form is not. So, this is a good example of how a statement that is logically fairly simple can still confuse intelligent people―at least, intelligent enough to get elected to Congress!
Acknowledgment: I was first made aware of this example by a single printed page of excerpts from the debate that I discovered in some of my papers from graduate school. Unfortunately, the page has no identifying information on it, and I don't remember where I got it. Sally Thomason (see the Source below) apparently received a copy of the same excerpts from the logician Nuel Belnap in 1983. There's no date on my copy, so I don't know exactly when I got it, but I'm sure that it was after '83, though probably only a few years later. So, it's possible that I received the printed page as a handout at a talk given by Belnap, or in some indirect way that traces back to him.
Source: Sally Thomason, "On Not Avoiding Negatives", Language Log, 2/21/2004
Update (4/21/2010): Typo in section 3 above corrected.