Logic and Scientific Thinking

Misinformation

Bogdan G. Popescu

bgpopescu@tec.mx

Tecnologico de Monterrey

Welcome

Learning Outcomes

Locate the failure: the premises — false, or stripped of scope — vs. the link.

Diagnose the gallop as a tactic, not a fallacy — four criteria before the name.

Apply the defense: pick one claim, demand its evidence — and inoculate.

Recap

Question	Answer from the previous sessions
The four families?	Faulty reasoning · misused sources · distorted claims · emotional appeals
The three-step diagnosis?	Name the move · quote the words · say what would decide it
What did every fallacy attack?	The link between premises and conclusion — mostly

Today’s question: what if the premises are the problem?

Misinformation

What Is Misinformation

False or misleading information, presented as fact.
- A fact: a checkable claim that survives the check. Misinformation only looks like one.
It can sit inside a perfectly valid argument.
- Valid: if the premises are true, the conclusion must follow.
The new question: are the premises true?

From earlier sessions: valid form + true premises = sound. Misinformation attacks the premises.

Definition first, then immediately anchor it in the course’s ladder: validity checks the link; SOUNDNESS needs true premises — and misinformation is precisely an attack on the premises. “FACT” in course terms: a CHECKABLE claim that survives the check (the same “checkable claims” the debate wall counts later). THE OPINION GUARD, when a student asks whether a wrong opinion is misinformation: opinions (“the cafeteria’s food is boring”) aren’t checkable, so they can’t be misinformation — only claims in checkable territory can be true, false, or misinformed. Don’t open the fact→truth→belief chain; the slide’s job is only to point the checker at the premises. Say the consequence out loud: every tool from the formal session still works AND none of them is enough — you can run modus ponens flawlessly on a false premise and march, validly, to garbage. That is why this session exists. Terminology teaser (resolved next session): whether it’s “misinformation” or “disinformation” depends on INTENT — park the question visibly, students always ask early.

Where Arguments Fail

An argument has only two points of attack (Schopenhauer, The Art of Controversy, 1896):

The target	Its name	The toolkit
The link — “the conclusion doesn’t follow”	fallacy — formal or informal	validity · the twins · the families
The premises — his “grounds”	misinformation	fact-checking · sources

Everything else in his manual is conduct: tricks on the checker, not defects in the argument.

Today’s example of conduct: the Gish gallop — a tactic that hides both failures at once.

The session’s map, now with its 1830 pedigree: Schopenhauer’s “Basis of All Dialectic” (Saunders 1896, pp. 13–15) opens the cynic’s manual with the HONEST map — a thesis can be attacked only in its grounds (a premise is false) or its consequence (the premises don’t yield the conclusion) — and only then catalogues the 38 stratagems, which are all CONDUCT: tricks on the disputant and the audience, not defects in any argument. That structure IS this slide: two targets, then tactics. THE GALLOP sits on the tactics side — like the classic “red herring” (a diversion that changes the subject — filed under distorted claims in the catalogue): the herring diverts the checker’s attention, the gallop exhausts it; neither breaks an inference, which is why calling a gallop (coming) takes four criteria about conduct and intent, not a form-check. TWO HONESTY GUARDS, say them when the caveat fragment appears: (1) the premise/link split is a DIAGNOSTIC first question, not a partition — several familiar fallacies also poison premises (the loaded question smuggles one in, the false dilemma shrinks the menu, cherry-picking curates it), so ask “where is the MAIN failure?” and expect some arguments to fail in both places. (2) Strictly, a false premise is a property of an ARGUMENT; misinformation is false content in CIRCULATION — the two meet the moment someone reasons from the false content, which is why the table works, but they are not synonyms. Row 1 is everything so far: modus ponens/tollens and the twins (formal), the four families (informal) — all LINK failures. Row 2 is new: a false premise leaves the link intact, so link-tools can’t see it; you must check the INPUT (the data). FOR STRONG ROOMS, Schopenhauer’s second axis (same pages): ad rem — the claim clashes with reality — vs. ad hominem / ex concessis — it clashes with the opponent’s own admissions; a consistency-win against the person is not a truth-win about the claim (the persuasion session sharpens this). Worked micro-example if asked: “If a politician takes bribes, they should resign; X takes bribes; so X should resign” — perfect modus ponens; if the bribe claim is invented, the argument is valid, unsound, and pure misinformation.

Bad Premises

(An invented scenario — a classroom specimen.)

P1.	If a cafeteria scores below 60 on its health inspection, the university must close it.
P2.	“Our cafeteria scored 42.”
C.	The university must close our cafeteria.

Check the link: modus ponens
Every tool from the fallacy sessions comes back clean.
So where is the failure?

Bad Premises

P1.	If a cafeteria scores below 60 on its health inspection, the university must close it.
P2.	“Our cafeteria scored 42.” → the actual inspection report: 92.
C.	The university must close our cafeteria.

Misinformation is the wrong number — not the wrong connection.
A perfect form, fed false input.
The premise is just false.

True Words, Misleading Scope

The study: “Among 120 adults with mild symptoms, the treatment reduced self-reported pain by 8% over twelve weeks.”

The headline: “Scientists prove the treatment works.”

No quoted number is false.
The qualifiers vanished: which people · which outcome · how much · how long.

Misinformation can be manufactured by subtraction.

The misleading half of the definition, and the oldest trick in the catalogue (Aristotle, Sophistical Refutations): a claim true IN A QUALIFIED RESPECT presented as true ABSOLUTELY. Walk the deletion: the headline drops the population (120, mild symptoms), the outcome (self-reported pain, not cure), the size (8%), the window (twelve weeks), and the unstated comparison. Nothing asserted is false — the misinformation was manufactured by SUBTRACTION, so fact-checking each word returns “true” and the claim still deceives. The check is mechanical: restore the qualifiers, then judge. Invented example, as always — say so. THE 1830s SPECIMEN, from the assigned Schopenhauer (Stratagem I, Saunders 1896, p. 16): A claims the Peace of 1814 restored independence to “all the German towns of the Hanseatic League”; B cites Danzig, which lost its independence by that Peace; A escapes — “I said ‘all German towns,’ and Dantzig was in Poland.” The scope moved after the fact; restoring the qualifiers is the two-century-old defense. Connect forward, one sentence: the next session shows the industrial version — persuasion that runs entirely on true statements, selectively arranged. Bridge to the flood: one dubious premise — false or de-scoped — is checkable in a minute. What if there were four, arriving at speed?

Hiding the Link

The tactic: don’t fix the argument — flood the checker.

P1.	“The cafeteria scored 42 on its inspection”
P2.	“…and three students got food poisoning last week”
P3.	“…and the kitchen staff aren’t even certified”
P4.	“…and prices went up twice this year”
C.	The university must close the cafeteria.

In pairs (2 min): attack it. What do you check first?

The Gallop, Checked

P1.	“scored 42” → the number is false (the report says 92).
P2.	“food poisoning” → asserted, not shown — was it even the cafeteria?
P3.	“staff not certified” → what’s the link to the closure rule? Never stated.
P4.	“prices went up” → the connection is missing entirely.

One false premise + three claims that never connect — the flood hides both failure types.
Each check is cheap alone. Four at once, at speed, is the tactic.

The Debate, Again

This time, don’t follow the argument — count the claims.

The Flood

So — how many checkable claims did you just hear?

1917 · 313 ships · 285 ships · “the low 200s” · sequestration · the Air Force since 1947 · FDR · two conflicts at once · a trillion dollars · Congress proposed it · “it will not happen” · maintaining spending · 1916 · horses · bayonets · aircraft carriers · nuclear submarines · the Joint Chiefs · veterans’ support · “your budget doesn’t work”

In those keywords: a claim about every 7 seconds.

Is that a fallacy?

Not necessarily — a dense answer can be honest.

But volume used deliberately, with claims too weak to survive checking — that tactic has a name.

The claim-wall from the earlier debate session, back on screen BEFORE the name — felt first, named next slide, the bat-and-ball rhythm. Let the wall sit in silence a few seconds before the count appears; the visual does the arguing. Say aloud: back then, feeling lost in this wall seemed like a language problem — sequestration, FDR, the Joint Chiefs — but American viewers couldn’t check these at speed either; the lostness IS the flood working. You needed a pause button and fifteen minutes to do what the audience had two minutes to do live. THE NUMBER SURVIVES CHECKING: 20 checkable items over the 2:11 cut ≈ one per 6.5 seconds. If a student counts and challenges you, that IS the teaching moment — congratulate them for checking a claim about unchecked claims. THE CAUTION lives on the slide’s own fragments (not a fallacy, necessarily); the full convicts-neither-debater version is in the next slide’s notes — deliver it there.

Calling a Gallop

Many claims alone is not a gallop. Before using the name, show:

Who is doing it — name the speaker, not the exchange.
Speed: claims arrive faster than the format allows rebuttal.
Strategy: the quantity itself, not the content, is meant to persuade.
Weakness: several claims would not survive checking.

So what? Volume works on the checker — next: why it works, psychologically.

The criteria slide — the honest core of the session. The debate wall demonstrated the CHECKER’S experience (overload is real, and it worked on this room twice now), but it convicts neither debater: a dense policy answer can be many legitimate claims, and in that debate volume served both sides. Named for creationist Duane Gish, who used the tactic in evolution debates — coined by scientist Eugenie Scott, who campaigned AGAINST debating creationists at all. ARISTOTLE’S version of the strategy (Sophistical Refutations): one goal of contentious argument is forcing the opponent into repetition and “babbling” — the galloper’s intended conclusion may not be “my claims are true” but “look how incapable my opponent is of answering me.” The audience scores the LOOK of the exchange, not the ledger. AND THE ASYMMETRY, DOCUMENTED IN 1830 (Schopenhauer, Stratagem XXVIII, Saunders 1896, pp. 33–34): a geologist who is RIGHT gets laughed out of the room by a plausible objection (a molten early Earth would have boiled the sea away) because the true rebuttal — atmospheric pressure — “would require a treatise to demonstrate the matter to those who had no acquaintance with physics.” The cheap wrong claim beats the expensive right answer in front of any audience that can’t check — the gallop’s engine, two centuries before the name. The four checks turn a vibe (“so many claims!”) into a diagnosis; without them, “Gish gallop” becomes just another label to throw — the label-machine failure mode from the fallacies sessions.

Why It Works

Why Flooding Works

Why can a flood work? Three documented mechanisms:

We accept before we reject.
- You can’t understand “the ball is red” without briefly picturing it — believing is automatic; doubt is a second step. (Gilbert et al. 1993)
Quantity can look like quality.
- When people are not examining arguments closely, more reasons can seem more persuasive. (Petty & Cacioppo 1984)
Confidence sounds like competence.
- A confident speaker may seem more competent, even when that confidence is misplaced. (Price & Stone 2004)

Why Flooding Works

The flood exploits all three:

too fast to reject
too much to evaluate
too confident to dismiss.

NARRATION: This is not just my intuition — each of those three bullets is a documented psychological finding. Gilbert and colleagues showed that believing is automatic: understanding a claim means briefly accepting it, and doubting is a second, slower step the flood never gives you time for. Petty and Cacioppo showed that when audiences are not deeply engaged, more reasons feel like better reasons — people count instead of weighing. And Price and Stone showed that when you cannot check the claims, you trust the most confident voice. The flood is engineered for all three. And knowing this is itself the defense — you have just been inoculated. FRAMING CAUTION: these are general findings about persuasion under cognitive load — plausible MECHANISMS explaining why flooding can work — not studies documenting the Gish gallop itself. Present them as explanation, not as proof that any given flood is a gallop. One line per study if asked. Gilbert et al. 1993, “You Can’t Not Believe Everything You Read”: comprehension entails initial acceptance; under cognitive load or time pressure, the second “tag as false” step never runs, and false statements stay believed — a claim every 7 seconds IS a load manipulation. Kahneman’s retelling (2011, ch. 7) gives the classroom detail: participants holding digits in memory while reading nonsense claims (“a dinca is a flame”) labeled FALSE later remembered them as TRUE — keep the mind busy and the false sticks. Thirty-second demo if wanted (the Moses illusion, Kahneman ch. 6): “How many animals of each kind did Moses take into the ark?” — almost nobody catches that it was Noah; fluency slips errors past checking even when the fact is KNOWN. Petty & Cacioppo 1984 (Elaboration Likelihood Model): under low involvement, the NUMBER of arguments raises agreement regardless of quality — argument count works as a peripheral cue. Price & Stone 2004: the confidence heuristic — without a way to verify, listeners take speaker confidence as a proxy for competence. THE FOURTH MECHANISM, delivered verbally (keep the slide at three): the ILLUSORY TRUTH effect — mere repetition breeds believability (Hasher, Goldstein & Toppino 1977), and prior knowledge does NOT protect against it (Fazio et al. 2015). Repetition across channels is the flood’s industrial engine — plant it here; the persuasion session harvests it (familiar feels true). LIVE DEMO if wanted, 30 seconds: ask who has heard that “radio listeners thought Nixon won the 1960 debate, while TV viewers picked Kennedy.” Hands go up. The evidence is one non-representative poll, and scholars have called the story a myth for decades (Vancil & Pendell 1987) — yet everyone has heard it, because it is endlessly repeated. The room just experienced the illusory truth effect on itself: about a debate, about appearances, in a lecture on misinformation.

The Cost of Checking

Not stupidity — economics:

Accepting a claim is free. Checking it is work: understand → reason → decide.
Under a flood of claims, the cheap option wins — rationally.
You felt it during the debate clip — twice now.

Your capacity to check claims is finite — and floods are designed to exhaust it.

The Defense

Beating the Flood

%%{init: {"theme":"neutral","themeVariables":{"fontSize":"16px"},"flowchart":{"useMaxWidth":true,"nodeSpacing":50,"rankSpacing":50}}}%%
flowchart LR
  C["false claim<br/>cheap to make"] --> U["unprepared:<br/>believes it"]
  U --> R["repetition<br/>feels true"]
  R --> RB["rebuttal<br/>(costly, late)"]
  RB --> P["belief<br/>persists"]
  C --> I["inoculated:<br/>knows the trick"]
  I --> D["met with<br/>suspicion"]
  style P fill:#f3ded9,stroke:#b44527,color:#b44527
  style I fill:#cfe0d8,stroke:#4a7c6f
  style D fill:#cfe0d8,stroke:#4a7c6f

The defense:

name the tactic aloud
pick one representative claim
demand its evidence
refuse to chase the other twenty.

Close the loop: don’t answer all twenty — name the tactic, pick ONE claim, make them defend it, refuse the chase. The deeper defense is the session itself: inoculation. Students now know the criteria (so they won’t cry “gallop!” at density) and the mechanisms (so they recognize the pull when it works on them). Preview, one sentence, unresolved: the flood wins by overwhelming your checking — WHY that checking is so overwhelmable is the Kahneman reading’s question. ZALLER’S CONFIRMATION, one line (1992, p. 267): resistance to persuasion depends heavily on the availability of COUNTERVALENT communications — the flood wins by removing the counter-flow, and the one-claim challenge plus inoculation ARE the countervalent message. (The fuller Zaller treatment lives in the persuasion session.)

Inoculation, Tested

A false claim is about to reach your friend: “the cafeteria scored 42.”

No preparation → your friend hears the claim without knowing how to evaluate it.
Supportive defense → “The cafeteria has a good safety record.”
- This strengthens the existing belief but does not answer the coming claim.
Refutational inoculation →
- “You may hear that it scored 42.
- Ask to see the inspection report: the official score is 92.”

Move 3 is inoculation — sixty years of studies favor it; promising, not established (McGuire 1961; Banas & Rains 2010).

The three bullets ARE McGuire’s three experimental conditions, dressed in the session’s own cafeteria case — let the intuition land first, then name them. Condition 1, no treatment: McGuire (1961) used never-attacked beliefs (“brushing your teeth is good”); they collapsed at the FIRST attack. Condition 2, supportive defense: extra reasons the belief is right — little added resistance. Condition 3, refutational pretreatment: a WEAKENED version of the attack plus its refutation, delivered in advance — the later full-strength attack largely failed (McGuire’s own explicit analogy was to immunization — attribute the analogy to him rather than using it). Banas & Rains 2010 (Communication Monographs) meta-analyzed four decades: refutational pretreatment reliably beats both no-treatment and supportive-only messages. Roozenbeek & van der Linden 2019 (Palgrave Communications): the Bad News game puts players in the fake-news PRODUCER’S chair — impersonation, emotion, polarization, conspiracy, discredit, trolling — and after ~15 minutes players rate manipulative headlines as less reliable (pre-post, N = 15,000); the most susceptible improved most. Free online (getbadnews.com), translated into 20+ languages — a ready homework. THE HONESTY CAVEAT is now on the slide — deliver it, don’t skip it: replications are mixed on how far the effect generalizes (some outside Western samples find no improvement in true/fake discrimination), so present inoculation as promising, not established. CONNECT TO THE COURSE: the fallacies catalogue IS refutational pretreatment — every named move is a weakened dose plus its refutation; this session’s whole design (name, criteria, mechanisms, defense) is the inoculation the research recommends. And the generative finding justifies the exercise next: building the tactic builds the detection.

Exercise — Build a Gallop

In pairs (6 min):

Build (3 min): flood a harmless claim — “our cafeteria is the best in the city” — with 8+ quick claims: numbers, names, dates, comparisons. Weak is fine; fast is the point.
Swap & defend (3 min): pick one of their claims, demand its evidence — and refuse to chase the rest.

Debrief: how did flooding feel from each side? — producing the tactic trains detecting it.

The generative-inoculation exercise, licensed by the previous slide: producing manipulation improves detection. RUN IT TIGHT: assign the claim (cafeteria, a phone brand, a local team — harmless by design; say aloud this is a flight simulator, not misinformation practice). Builders discover the asymmetry (refuting costs far more than producing) from the producer’s chair — inventing ten claims takes seconds, and confidence does most of the work. Defenders practice the discipline: ONE claim, its evidence, refuse the chase — expect them to feel the pull to answer everything; naming that pull is the debrief’s payoff. Debrief questions: builders — “how easy was volume?”; defenders — “what did it cost to ignore nine claims?” TIME SAVER: run whole-class — you build the gallop live at the podium, the room defends. HOMEWORK OPTION: the Bad News game (getbadnews.com), ~15 minutes, the researched version of what they just did.

Conclusion

Where is the failure?

Schopenhauer’s map: two targets — premises and link; everything else is conduct.
Misinformation — false premises: a wrong number, or true words minus their scope.
The gallop — the conduct example: volume that hides both failures at once.
Calling it takes the four checks — overload felt is not strategy shown.
Checking capacity is finite — floods are designed to exhaust it.
The defense: one claim, its evidence — and inoculation before exposure.

Homework: find one real flood — and run the four checks on it.

The one-slide recap of the spine — read it as the session’s architecture: the map (two targets, then conduct) → the premise attacks → the tactic → why it works → the defense. Line to close on, verbatim: “Today the flood was anonymous — WHO floods on purpose, and at what scale, is the next session’s question.” (That is the hand-off to the persuasion session; do not resolve it.) DEFINITIONAL BRIDGE, one sentence, verbatim — it prevents a silent definition-shift between sessions: “Today, ‘misinformation’ named the defective information itself. Next time we split the SENDERS: the one who believes it, the one who knows better, and the one who never cared.” Homework options: the four checks on a real flood (a thread, a press conference, a launch), or the Bad News game (getbadnews.com, ~15 minutes) — the researched version of the exercise they just did.

References

Banas, J. A., & Rains, S. A. (2010). A meta-analysis of research on inoculation theory. Communication Monographs, 77(3), 281–311.
Gilbert, D. T., Tafarodi, R. W., & Malone, P. S. (1993). You can’t not believe everything you read. Journal of Personality and Social Psychology, 65(2), 221–233.
McGuire, W. J. (1961). The effectiveness of supportive and refutational defenses in immunizing and restoring beliefs against persuasion. Sociometry, 24(2), 184–197.
Petty, R. E., & Cacioppo, J. T. (1984). The effects of involvement on responses to argument quantity and quality: Central and peripheral routes to persuasion. Journal of Personality and Social Psychology, 46(1), 69–81.
Price, P. C., & Stone, E. R. (2004). Intuitive evaluation of likelihood judgment producers: Evidence for a confidence heuristic. Journal of Behavioral Decision Making, 17(1), 39–57.
Schopenhauer, A. (c. 1831/1896). The Art of Controversy (T. B. Saunders, Trans.). Swan Sonnenschein.