Fallacy and Falsehood

From StoneHome

It is a common mistake to transpose the terms fallacy and falsehood. Much like translucency and transparency, it is common for people to believe the meanings of each word are swapped with what each word really means; alternately, many people believe fallacy is a fancy way to say falsehood, or that it's somehow the correct (effete) noun form for literati in the know. This page, along with its sister pages in my English category, exists to help clear up the confusion surrounding the words.

The words from above

First, a pair of simple definitions.

Falsehood

Falsehood is the simpler of the two ideas. In pseudoformal terms, falsehood is incorrect data within an argument. The conclusion reached from a falsehood is invalid because the decision was based on incorrect facts. Examples are trivial:

• "Two plus two is five."
• "California is a radioactive wasteland."
• "Tiger Woods is a white woman with a peg leg."
• "France doesn't suck."

Only data may be false, not arguments; data may only be false, not fallacious. Falsehood is the basis of lies, misinformation, and whatever religion I've decided to hate this month.

Fallacy

Fallacy is the more complex of the two ideas, and unfortunately comes in many forms, so it's easy to begin to believe it blindly applies to anything wrong in the early flood of examples. There are a bunch of categories of error which have been well known since Ancient Greek times, which unfortunately means that the best way to refer to them is in Greek (or in a few rare cases, Latin.) Confounding the issue, some fallacies weren't established until the modern day, and those fallacies are expected to be translated rather than to be named in their traditional language. However, these days a cosmopolitan individual can get away with localized names for everything, and since I'm an American I think of English as everything, so though I'm sticking to their proper names, I provide lookup tables and titular translations for each to English when available and appropriate.

Again in pseudoformal, a fallacy is a flawed reasoning within an argument, rather than flawed data. The conclusion reached by a fallacy is invalid because the reasoning used doesn't make sense, frequently for subtle reasons to which certain people won't listen. Examples are less simple without sucking:

• "Super Actionman smokes cigarettes, so I'm going to smoke cigarettes" (Appeal to Authority,)
• "Everybody knows you can get AIDS from drinking fountains" (Appeal to the Public,)
• "The Zarkonians control the banks, the schools, the government; they are responsible for our ruin; we must kill them all" (Straw Man,)
• "Well you better be happy there are UN peacekeepers in Iraq because otherwise we'd have to send in the tanks again" (Appeal to Force,)
• "We've been hand-tilling these fields for a hundred years, we don't need tractors" (Appeal to Tradition,)
• the Chewbacca Defense (Red Herring.)

Reasoning may only be fallacious, not false; only reasoning may be fallacious, not data. Fallacy is the basis of misdirection, doubletalk/twisting the truth, and prejudice.

It is worth noting that although a conclusion derived from fallacy is invalid, that doesn't actually make it wrong; just because the supporting arguments are nonsense doesn't mean that the conclusion is actually incorrect, simply that the path taken to the conclusion has no meaning. One of my favorite examples comes from third-or-so grade mathematics: ¹⁶/₆₄ = ¹/₄, at which one may arrive through simple division. However, there is a fallacious line of reasoning which suggests that in order to find out what ¹⁶/₆₄ boils down to, simply remove each six. Obviously that's wrong, and for sake of being complete, it fails on ²⁶/₆₄. Note that the conclusion is not actually wrong in the original case; it's certainly invalid, but coincidentally it also happens to be correct. This is a critically important understanding in logic - people who try to shoot down conclusions with fallacy, rather than the routes to those conclusions, are making a subtle but serious comprehension mistake, and should be watched carefully for others.

Examples

Fallacy and falsehood are not mutually excusive; something may be both false and fallacious. However, neither invokes the other. Some examples follow:

False

Simple

• A two foot sphere of lead weighs only one microgram, so it will float in air.
  • The reasoning is good: a two foot sphere which weighed one microgram would indeed float in the air. There is no fallacy here.
  • Lead in fact weighs more than that. There is a falsehood here.

Intermediate

• As a Libertarian, Ross Perot supported Ralph Nader
  • The reasoning is good: in typical, members of the same political party support one another. There is no fallacy here.
  • Ross Perot is not a Libertarian. There is a falsehood here.

Fallacious

Intermediate

• Blacks are an inferior race. Why, look at these intelligence tests from 1955 in South Africa. (Biased Sample)
  • Those tests do exist, and they're probably accurate. There is no falsehood here.
  • In the 1950s, blacks in the nation South Africa were a badly underpriveleged undereducated group subject to a wide variety of social pressures against intelligence including disbarment from libraries, higher schooling and access to state facilities. Using the status of blacks in South Africa in the 1950s is not representative of the norm for blacks, but rather the norm for an abused people.
  • The specific faulty reasoning here is called Biased Sample and is the mistaking of trends caused by Effect A as actually caused by Effect B. Though the example chosen is particularly caustic, mistaking socioeconomic advantage for racial caliber, less horrible examples of this mistake run all throughout science; placebos, double-blind tests and their kin and the scientist's intense statistical training are designed specifically to eliminate this fallacy.
  • Similar examples of the degeneration of opressed populace exist among whites (Victorian-era Irish, 1930s American-Italians and Roman-era Gothic peoples), orientals (the Ainu, the Chinese during Mongol rule) and the New World peoples (especially during the European colonization, many individual tribes dissolved inbetween white persecution, liquor and the destruction of the traditional ecosystem in various parts of both North and South america, the most prominent examples probably being the five nations of the Iroquois Council and the collapse of the Aztec empire at the hands of barely enough Spaniards to annoy a soccer stadium.)

Complex

• Dversky and Kahneman's associativity study example works well here.
• Suppose in the 1960s you have a woman named Linda, well educated, liberal, a philosophy major and an outspoken advocate regarding race relations. Is it more likely that: A) She is a bank teller, or B) She is a bank teller and a feminist? (Conjunction Fallacy)
• The study mentioned above found that slightly under 85% of people asked answered B, that she was likely to be a feminist.
  • The way I've phrased that carefully exposes the underlying error: the question tends to make people think "well, she's probably a feminist, so it's probably B." The problem is that bank teller is in both halves. That might be a reasonable judgement if B) was just that she was a feminist; the data may support the conclusion based on other behavioral patterns. Hell, it may even be true that (say) 87.3% of the female philosophy majors involved in race relations in the 1960s were also feminists. However, the set of feminist women bankers by definition cannot be larger than the set of women bankers; any secondary qualification you add is guaranteed to decrease the likelihood of B, if it has any effect at all. The fallacy attempts to put feminism against non-feminism by attaching the unnessecary banker qualification to hide that the other set is effectively everybody. If you remove the banker qualification, it becomes "is it more likely that A) she exists, or B) she is a feminist? ."
  • Admittedly I injected that 87.3% in the previous example for support here; let's also pretend that women had exactly thirteen jobs available to them in the sixties with equal probabilities. A more natural way to see the fallacy is to rename things for cards, which we are used to seeing in terms of probability. Say Joe has used a special cheating machine, which sorts diamonds towards the top of the deck such that they will be drawn as the next card with a probability of 87.3% . The chance that she's a banker is 1/13, the same as the chance that this card is a four. Therefore, The question above is equivalent to asking "Is it more likely that: A) the card drawn is a four, or B) the card drawn is a four of diamonds?" It doesn't matter how good the cheating machine is; if it's not perfect, the chance that it's a four is always higher than the chance that it's the four of diamonds. The bit about being a four card (being a banker) is to distract you from the real question: is 100% or 90% more likely? (Oooh, 90%. It's probably B.)
  • This is the Conjunction Fallacy, which asks you about the likelihood of a stricture in terms meant to encourage you to think of it instead as a categoric range.
  • This also provides an amusing anecdotal counterexample to appeal to the masses: apparently 85% of people asked can be wrong, even when the correct answer cannot be debated.

False and Fallacious

Simple

• "We have been attacked by the peoples of the Middle East. Saddam is a recurring threat which does horrible things; we must hold him accountable for the attacks of 9/11."
  • Like most Republican diatribe, this is relatively easy to take apart. The falsehood is that we were attacked by the peoples of the Middle East; in truth we were attacked by a single extremist splinter group with denounced religious affiliations.
  • The fallacy is in scrubbing Saddam together with 9/11. Saddam had nothing to do with 9/11, other than being the real villain's neighbor. This is the Bad Company Fallacy, also known as the Fallacy of Guilt by Association.
  • There are other faults with this argument, though I won't belabor them; suffice it to say that this argument is false at least twice and fallacious in at least three seperate ways.

Intermediate

• All Americans are xenophobes. John, an American, has travelled, and is not a xenophobe. Therefore, John isn't really all that American. (No True Scotsman)
  • Being a member of a group does not denote characteristics of that member, no matter how badly we want to stereotype someone. For example, not very many Huron were the cannibalistic super-warriors we all heard about, nor did all that many of the Incan cities gut their young. Even if Americans could be broadly characterized as xenophobic this would not characterize John. For example, ARIS 2001 suggests census data of a US 76.1% majority of assorted Christian denominations; it still could not be said that John is not really an American if he's not Christian (he could be Islamic, Hindu, Buddhist, Atheist, Agnostic, or whatever; pretty much anything but Jewish, who spell it without the H.)
  • Not all Americans are xenophobes. (In fact, we mostly lash out because we're bored; very few of us hate foreigners of any kind other than the French.) There is also therefore a falsehood here.
    • To wit, my jabs at the French can be broadly ascribed to quite a few fallacies: No True Scotsman, Appeal to Ridicule, Argument from Ignorance, Invalid Proof, Appeal to Pity, Appeal to Spite, Appeal to Fear, Appeal to the Populace, Appeal to Emotion, Correlation Implies Causation, Fallacy of Distribution, Fallacy of Composition, Dicto Simpliciter, Hasty Generalization, Genetic Fallacy; hell, even Trivial Objections at times, and god only knows how much other stuff (particularly Misleading Vividness - I don't describe their aroma in nearly sufficient detail.) It's fun, and they're one of the few cultures on Earth which can take it. Lighten up.
  • No True Scotsman is almost always accompanied by a falsehood; its nature almost demands a falsehood as the supporting criterion. Nonetheless, this fallacy/falsehood pair is quite common: no jock likes math, no skateboarder's afraid to smoke weed, no black man takes seriously a career in US politics. Those three criteria all happen to fail on a friend of mine from college named Lincoln, by whom I would not be surprised to see the first black entrance into the US Presidency (he's even got the old family money, traditional education and upper crust social background that we pretend someone doesn't need, the ridiculously high quality diction and pretty boy face that we pretend someone doesn't need, the military background, the sports, etc. 'Course, he'd also be the first Jewish president, and lord knows you could pile six different scotsmen fallacies on the three way surprise of his black British Jewish parents...)

Complex

Be Careful!

Just because it was reached by fallacy doesn't mean it's incorrect. With solid logic and false data, one reaches false conclusions. However, with solid data and fallacious logic, one frequently still comes to correct conclusions; one simply doesn't have an established proof as one believes. For example, one of the fallacies - enthymeme - doesn't actually comment on the quality of the output of the logic in any way; it simply quantifies the argument unsound because critical logic goes unspoken. The classic example of an enthymeme is "Socrates is a man, therefore socrates will someday die." This is not actually sound logic: unspoken has gone the rule that as mortals, all men someday die. (In fact, our grandchildren, the magical promise of nanotechnology, gene therapy, stem cells and whatever other buzzword can do absolutely anything this month may see this unspoken rule broken.) Though the conclusion is fairly obviously correct, the logic is not.

So, for an alternate example, consider the disadvantaged black test scores example from earlier. Were that test re-done for a society of men and apes, and were the apes subject to the same pressures as 1950s South African blacks, the test would still be invalid based on the unfair playing ground placed between man and ape; nonetheless, the conclusion that apes really are dumber than man would be true. Therefore, that the 1950s tests skewed against blacks are invalid doesn't actually mean that blacks aren't less able (which I can say in public because there's an assload of other evidence that does.) All the fallacious reasoning means is that you have not shown the correctness of the conclusion, not that the conclusion is incorrect.

It is therefore important to realize that whereas the presence of bad data in contact with good logic creates a false conclusion, the presence of a fallacy is not an argument against its conclusion, no matter how badly you may hate the conclusion; it is only argument against the conclusion's original supporting argument. This is also known as Argumentum ad Logicam, or the fallacy fallacy. Using fallacies to attack a conclusion is usually a warning sign that you're also submitting to Appeal to Emotion and frequently also Argumentum ad Hominem.

In Summation

If the information is bad, it's a falsehood. If the reasoning is bad, it's a fallacy. If both are bad, it's both. Falsehoods are simply incorrect data, and are very easy to spot; however, many fallacies are very difficult to spot, and a familiarity with the list of fallacies can often help one dissect what precisely it is about an argument which doesn't feel right, or display in concrete terms why the way someone blamed you for their own mess doesn't make sense, or whatever. In many ways, fallacy defines a set of ground rules for debate: no biting, no eye gouging, and no references to public or nonauthoritarian belief. Take the time to familiarize yourself; these can be surprisingly useful.