Skip to main content
note
This page was translated from the German original, partly by machine. Some passages may read awkwardly or contain inaccuracies. When in doubt, please read the original.

16. Informal Fallacies - Presumption Part 4

 

Summary​

This fourth session on fallacies of presumption narrows the focus to faulty assumptions in inductive reasoning. Induction is the process of observing evidence to arrive at a probable conclusion or explanation: we notice that objects fall and posit a law of gravity. Inductive fallacies arise when the supporting evidence is incomplete, distorted, misused, or misinterpreted. (Some philosophers hold that induction is always fallacious, the "problem of induction," which the course defers to a later video on the philosophy of science.) Di Donato organises six fallacies under two subcategories: generalization and reduction.

Generalization​

A generalization derives a characteristic from repeated experience of similar particular instances and asserts it universally across a target class. Seeing many furry cats, we conclude that all cats have fur.

1. Dicto simpliciter (sweeping generalization / fallacy of accident). From the Latin a dicto simpliciter ad dictum secundum quid, "from the statement unqualified to the statement qualified." The fallacy occurs when a general rule is cited even though it is false, or when a reliable rule is misapplied to a special case it does not cover. It rests on confusing an accidental trait (one a class member may have but that is not required by its nature, such as brown hair in humans) with an essential trait (one universal to the class, such as being warm-blooded). The error is treating an accidental trait as essential and applying the rule despite a qualifying exception. Examples: "Airplane pilots are usually male, so our pilot will be male"; "It should be illegal to cut people with knives, so surgery should be illegal too"; "Lying is universally wrong according to Kant, so Germans who lied to hide Jews during the Holocaust did wrong."

2. Hasty generalization (converse accident). This is forming a general rule from too small or unrepresentative a sample. Induction is not itself fallacious: we reasonably accept conclusions drawn from a sufficiently large body of data. The fallacy occurs when the sample is too small and the whole would yield a different conclusion. It is the mirror image of the fallacy of accident: accident assumes what is true of the many is true of the few, while hasty generalization assumes what is true of the few is true of the many. Examples: "Every squirrel I've seen has grey fur, so all squirrels are grey"; "That person from Charlotte was rude, so Southerners are rude"; "That computer from Best Buy broke, so Best Buy only sells bad products."

3. Cliché reasoning. Substituting popular, overused phrases ("conventional wisdom") for relevant evidence. Such clichés are usually over-generalizations, and competing clichés cancel out: "Nothing ventured, nothing gained" versus "Better safe than sorry." Neither offers a real reason for action. Another instance: "You can't trust that account of the Punic Wars because history is written by the victors."

Reduction​

Reduction makes a complex issue look simpler than it is by focusing on only a limited aspect of it.

4. Faulty analogy (weak or false analogy). An improper comparison that stresses superficial similarities while ignoring substantial differences relevant to the issue. Every analogy eventually breaks down (otherwise the two things would be identical); the analogy fails when it breaks down on the relevant point. Examples: "Water boils at 212°F and olive oil is also a liquid, so it boils at the same temperature" (ignoring different chemical compositions); "Dogs and turtles are both animals, so they live about the same length of time"; "Jumping out of a plane is no different from stepping off a sidewalk" (ignoring speed of impact).

5. "All answers or no answers." Assuming a position is false because its holder cannot answer every question surrounding the issue. It corners an opponent by demanding answers to loosely related questions, making them look foolish so the audience rejects all their positions. It carries a built-in emotional appeal, linking back to earlier relevance fallacies. Examples: "If you can't explain what the locusts with human heads in Revelation 9 represent, your whole interpretation is false"; "It's impossible to know everything about an infinite being, so I'm agnostic" (which slides from "can't know everything" to "can't know anything").

6. Post hoc ergo propter hoc (false cause). From the Latin "after this, therefore because of this," the fallacy reduces a causal relationship to a merely temporal one: because B follows A, A must have caused B. Chronology is never equivalent to causality. Di Donato adds four cautions:

  • Oversimplified cause: treating A as the sole cause when B has multiple contributing factors.
  • Necessary vs. sufficient conditions: a necessary condition must be present for the effect but does not alone produce it; a sufficient condition is enough to produce it. Temperatures at or below 32°F are necessary for snow, but precipitation is also required for sufficiency.
  • Confusing cause and effect: did the appliance shut off because the power failed, or did a short in the appliance blow a fuse and cut the power? Cause and effect can even be simultaneous.
  • Mistaking correlation for cause: events occurring together (the cum hoc ergo propter hoc, "with this, therefore because of this" variant), such as ice-cream sales rising with beach openings.

Examples: "We won the lottery because I played my lucky numbers"; "John sneezed after the phone rang, so he's allergic to ringtones"; "Becky lost weight after starting college, so higher education causes weight loss"; "Drownings rise with ice-cream sales, so ice cream causes drowning" (both driven by the common factor of warm weather).

Looking ahead​

The session adds six more fallacies to the running list. The fifth and final part on presumption will turn to inductive reasoning's category of missing evidence.