Bias check II — fallacies of relevance

Premises that are not reasons

A fallacy of relevance offers a premise that has nothing to do with whether the conclusion is true. It works because the premise is emotionally relevant — it concerns a person, a motive, a consequence — and so feels like a reason. Van Cleave's six, in the order this course teaches them:

1. Ad hominem — attacking the person instead of the argument. "The trade printer's numbers can't be trusted; they're a competitor." Their motive says nothing about whether the numbers are right; check the numbers.
2. Straw man — refuting a weaker version of the opposing position. "Buying means sinking $260,000 into a machine that will be obsolete in three years" — nobody argued for buying on those terms; the buy case is a five-year loan and a press with a ten-year life.
3. Tu quoque — "you too": dismissing an argument because the arguer does not follow it. "Wren Advisory leases nothing itself, so its advice to lease is worthless."
4. Genetic — judging a claim by its origin. "That figure came from the vendor, so it's wrong." Origin lowers the source's credibility score (L5.7) and is a reason to verify; it is not a reason to reject.
5. Appeal to consequences — a claim is true or false because of what would follow. "The press can't be failing; we couldn't afford to replace it." Smith's list calls a version of this the red herring; the definitions differ, the error is the same.
6. Appeal to (irrelevant) authority — a claim accepted because of who holds it, where the person's expertise is not on the point. "Priya prefers the lease" as a premise for leasing; "the CFO wants option B" as a reason B scores higher.

When an appeal to authority is legitimate

Not every appeal to authority is a fallacy. Van Cleave's test: the authority must have relevant expertise on the exact question, and there should be consensus among such experts — or the appeal is to one person's opinion. A press engineer's statement about the press's remaining life is a legitimate premise about reliability; the same engineer's view on whether Lakeshore should borrow is not. Priya's preference is a legitimate input to stage 4 — the recommendation revisits the client's goals and values — and an illegitimate premise in stage 3, where the scores must come from evidence.

Find three in the clinic case

"Tomas has been running that desk for eleven years and he says software won't help" — appeal to authority; relevant on how the desk works, not on what software does. "The vendor's 40 % figure is marketing, so reminders don't work" — genetic; the figure needs verifying, and the peer-reviewed review (L5.7) is where to look. "Going digital means turning the clinic into a call centre" — straw man; nobody proposed that.

List the six cold

Ad hominem; straw man; tu quoque; genetic; appeal to consequences; appeal to authority. And the test for the last: relevant expertise, and consensus.