Gish Gallop

You came to talk about one claim. You leave with twelve. Half of them are familiar errors, two are half-true, one would take a week to check, and the clock is gone. A reply that catches the worst three looks, to a tired audience, like a reply that ignored nine. The Gish gallop is that tactic. Quantity outruns the time available to answer, and silence on the remainder gets scored as consent.

The name comes from debate, not from a lab. Eugenie Scott, then at the National Center for Science Education, described it in 1994 and named it after Duane Gish, a creationist debater who could pour out more claims than a scientist could unwind before the moderator called time. Gish did not invent the move. Anyone in a comment thread, a cross-examination, a political interview, or a family argument can use it. The structure is the point. Speed plus volume plus a format that looks like a fair fight.

You can gallop by accident. Anxiety, enthusiasm, and a notes app full of grievances all produce a list longer than the other person's turn. The bug, when you are on the receiving end, is treating the list as a set of open cases you have now lost. The bug, when you are producing it, is mistaking coverage for a case. A claim is not stronger because its siblings did not get a reply.

• "You didn't answer points four through eleven, so you don't have an answer."
• "I'll just put the whole list here. They can rebut it if it's so wrong."
• "There are so many problems with this study" is followed by a paragraph of unrelated problems.
• "Address every sentence or you're dodging."
• "I brought receipts," and the receipts are thirty links, none of which you can open before the vote.
• "Even if that one was wrong, what about the other nine?"
• "I won because they got flustered."

The interview: a guest lists eight scandals in forty seconds. The host picks one. The guest says the host is afraid of the other seven. Viewers who wanted a verdict now have a mood. The mood is that something must be there, because nobody cleared the pile. Clearing the pile was not possible at that speed. That was the design.

The performance review: a manager arrives with a page of faults, mixing a real missed deadline with old tone complaints, a disputed metric, and a rumor. You answer the deadline. The writeup says you became defensive and left the rest unaddressed. The page was not a review. It was a gallop with a personnel form.

The thread you write: you are right about the main point and you add every side injury you have ever stored. The other person answers the weakest addition, because it is easy, and the main point dies in the pile. You galloped your own argument off the table.

Audiences confuse "unrefuted in the time allowed" with "unrefuted." Careful people look slow, which reads as losing, because care is slower than a list. Bad claims survive by traveling in a pack. Good claims get buried by their owner's extra claims. Institutions that reward the gallop, debate clubs, comment rankings, hearings with a two-minute clock, train speakers to bring more points than the format can test. The tested part shrinks. The implied part does the persuading.

Refutation is slower than assertion. A sentence can allege a study, a quote, a number, or a motive. Checking it means finding the source, reading the context, and saying what is wrong, which is three steps to their one. In a timed or social setting, the steps do not fit. Listeners also use a rough count: many particulars feel like a case, the way many anecdotes feel like a rate. Fluency and speed add a second illusion, that the speaker must have this at their fingertips because it is organized. A list can be organized and still be a pile of leftovers.

There is no experiment called the Gish gallop, and it would be a mistake to dress the tactic up as one. Scott's 1994 name stuck because it described a real debate problem: a format that looks symmetrical while the work of checking is not symmetrical. The underlying psychology is borrowed from better-studied bugs. A list of particulars recruits the availability heuristic, one vivid item standing in for a pattern. Unanswered items recruit a false feel of concession. Repetition of the pile, in quote-tweets and recap videos, can add an illusory-truth gloss: the claims have been "in the debate," which is not the same as having survived it.

The practical literature on public argument says the same thing Scott said. Debate is a sport with a clock. It is a poor instrument for settling empirical questions, precisely because the clock rewards the gallop. That does not mean you never debate. It means you do not treat a timed pile as a literature review. If the question matters, the answers have to live somewhere the clock cannot shout them down: a page, a shared document, a rule that each claim gets a turn of equal length.

• You are adding points because the last one might be answerable.
• You would be annoyed if they answered only your strongest point and ignored the rest, and that annoyance is the strategy.
• You cannot rank the list. Everything is supposedly essential, which means nothing has to carry the conclusion.
• On the receiving end, you feel you have lost because you ran out of minutes, not because a point landed.
• The audience's verdict arrived before anyone opened a source.

Say this, even if it sounds fussy: "There are twelve claims. I'll take the one your conclusion needs. If that one fails, the other eleven don't rescue it." Then take that one. If you are the person with the list, send the point and the source. The rest can live in a document nobody will pretend to have answered in forty seconds.

What if all twelve points are real?

Then they can survive a slower format. A true point does not need to hide in a pile. Offer to take them in order, in writing, with sources. Someone who needs the pile more than the points will resist that offer. The resistance tells you what the pile was for.

Isn't choosing one point a way of dodging?

Dodging is refusing the claim that carries the conclusion. Choosing that claim, and saying why the others can wait, is the opposite. The gallop calls this a dodge because the tactic dies if the audience accepts a queue. State the queue. Then answer.

A point you did not have time to open is not a point that won.

The interview

The review

Your own thread

Practice this pattern in the Reframing App - capture the trigger, label it (like Gish Gallop), check evidence, and write a more balanced thought.