How to Win an Argument
Contents
I have always had a soft spot for strange subjects and for anything that claims to get at the truth. A few years ago I heard an interview episode on a Taiwanese podcast, Mindi Xuandu. The guest, Hsiung Jen-chien, was a Taiwanese kid who went alone to India in junior high and buried himself in a Tibetan Buddhist monastic college. What he described was the entire training system behind Buddhist debate, and I sat there riveted. The full curriculum runs close to twenty years, roughly these blocks:
- Collected Topics: the entry level, logic and the definition of terms
- Pramana: orthodox Buddhist logic and epistemology
- Prajnaparamita: the appearance of the universe and the map of practice
- Madhyamaka: the hardest philosophy Buddhism has, emptiness
- Abhidharmakosha: Buddhist physics of the cosmos and psychology of the mind
- Vinaya: the law and daily administration of the monastic community
The one that got me was Pramana. It has a structure of steel called the three-part proof, which is the Buddhist version of the syllogism, and it forces you to break any claim into three pieces: the thesis (what you assert), the reason (why), and the example (your evidence). Miss one of the three and the argument doesn’t stand
Inside that curriculum, the live part is taking the seat of the defender. Monks go to the debate courtyard every afternoon and every evening to do it for real, and somewhere in their thirties or forties, having finished all of the above, they become Geshe candidates (yes, only candidates). To get the Geshe degree itself, the highest honor, roughly a doctorate in Buddhist studies, they still have to pass the one exam that decides everything
That exam is the Great Debate of Lhasa, the final and largest version of taking the defender’s seat. With thousands of senior monks, elders and the entire monastic body of every major Lhasa monastery watching, the candidate sits alone in the middle of the floor. The chief examiners and the sharpest debaters in the hall take the challenger’s role, standing up in turn, pulling questions at random from the five great treatises, and hammering him with the fiercest logic they have at the fastest speech they can manage. He gets no preparation and no scriptures to check. All he has are the standard replies, I accept, the reason is not established, there is no pervasion, and he has to dismantle every attack with those. One logical contradiction, or one question he can’t answer, and points come off on the spot
That rhythm doesn’t survive on paper. There’s a recording of a debate held in Mandarin, nowhere near the level of the Great Debate of Lhasa, but the flavor is identical: a question, a clap, and you answer right now
If he survives it, the examiners weigh his performance in that debate, his record over the years, and the sheer volume of scripture he can recite, and that decides which rank of Geshe he gets. From the top down:
- Lharampa Geshe: the highest honor, the equivalent of top scholar in the country or the highest grade of state doctorate. Across all of Tibet only a few dozen places open up in a year. A Lharampa can go straight into the Upper and Lower Tantric Colleges for esoteric study, and is in the running for abbot of a major monastery, or even for Ganden Tripa, the head of the Gelug school
- Tsokrampa Geshe: an outstanding graduate who passed a hard debate in front of the whole monastic assembly, roughly a doctorate from a top university
- Lingse Geshe: a doctorate awarded by internal examination inside one’s own dratsang, the monastic college
- Dorampa Geshe: awarded after passing a basic debate test at the monastery gate or in front of the main hall
The Geshe isn’t the end of the hardcore life either. He can spend another year or two in a tantric college on esoteric practice, learning visualization, mantra recitation and mandala construction, and in the Gelug school nobody without a Geshe is allowed into the tantric colleges for the advanced material at all; or he can return to lay life and carry that brain, forged by two decades of merciless logic, back into the ordinary world to start a company, do academic work, or teach. From enrollment to degree is somewhere between twenty and twenty five years, and the moment he has it, he is recognized across Tibetan Buddhism as a master of logic and philosophy
That interview made me pay attention to whether a claim, its reasons, and its examples actually connect. Yet in the public eye, that often seems to matter little when deciding who has won an argument.
Later, a project required me to analyze public opinion, so I read many papers, online arguments, and exchanges in the news. Some arguments had clear problems in their reasoning yet still came out ahead. I gradually noticed the same kinds of errors recurring. That is why I wanted to write this article: to set out eight common logical fallacies.
Emotion and Reasoning
In Thinking, Fast and Slow, Kahneman describes two ways we think. System 1 responds quickly and intuitively; System 2 takes more effort to examine a claim. An accusation or vivid example can provoke a reaction before we have checked the reasoning behind it.
It often takes longer to explain a faulty inference than to make one. If I cannot answer immediately, that does not mean the other person’s argument holds. I can first identify the claim and ask what reason supports it.
Eight Common Logical Fallacies
The names are less useful than understanding where each argument goes wrong. These eight examples fall into three groups.
Shifting the Topic
Ad hominem

The most primitive move, plainly put: play the man, not the ball. You put a specific number on the table and the other person doesn’t even look at it, just comes back with “how much did they pay you” or “why would anyone believe someone with your politics.” Political talk shows run this scene nearly every night. He doesn’t care whether he can refute your point. What he wants is for everyone else in the room to decide you’re not worth listening to
Red herring

The name comes from hunters dragging a strong smelling smoked fish across a trail to pull the hounds off it. The method: throw in a new topic that has nothing to do with the original one but is very good at stirring feelings, and your attention goes with it. A politician pressed on his own corruption case reaches for “well, when the other side was in power, were they any less rotten.” You asked about A and end up fighting about B, and by the time you’ve chased it down, nobody remembers what the question was. It works just as well at home: you ask your partner “how could you treat me like that,” and back comes “and you stay up eating midnight snacks every night, have you ever thought about how I feel,” and in one move you’ve gone from plaintiff to defendant
Smoke screen

A cousin of the red herring, and harder to guard against. The red herring throws in something unrelated. The smoke screen takes a related issue and deliberately stirs it until it’s too complicated for you to finish answering. Any time power shortages come up here, what you want to ask is simple, whether there’s enough electricity and whether the supply holds, and very quickly someone drags you into the global energy transition, natural gas futures and international net zero timelines. All of it genuinely does relate to electricity, but the complexity itself becomes the smoke, and the simple question ends up buried underneath where nobody can see it
Distorting the Claim
Straw man

First they build a dumber, more extreme fake version of you, then turn around and beat the fake one up. When Taiwan was fighting over same sex marriage, what the supporting side actually said was “let two people who love each other marry.” What the opposing side shouted at the public was “so at this rate, will people be marrying animals next?” Nobody on the supporting side ever said that. But it was the absurd version, the one that got beaten up, that a lot of people remembered
False dilemma

Something with many possible shapes gets pressed down into two extremes so you have to pick a side. “If you won’t eat ractopamine pork, Taiwan can forget about joining the international community” is the classic form: swallow it, or be cut off from the world. The options in the middle, negotiate, inspect, label it clearly and let consumers choose, get talked about as if they don’t exist. Very few things in the real world come down to two roads, but once you’ve been forced to choose between two extremes, you have no room left to think of a third
When Evidence Does Not Support the Conclusion
Circular reasoning
Using the conclusion to prove the conclusion. It goes around in a circle and proves nothing. Every election season you hear this exchange: “there must be something wrong with this guy, look, he got indicted.” “And why was he indicted? Because there is something wrong with him.” From beginning to end, the one thing you wanted, evidence independent of the claim, never appears
Appeal to ignorance
Treating “no evidence” as “guilty.” The most famous product of this move is probably the long stretch during which nobody in the Chinese speaking world dared help an elderly person who had fallen. The line from the Peng Yu case in Nanjing that got quoted everywhere, “if you didn’t knock her down, why would you help her up,” translates into plain speech as: since you can’t prove you’re innocent, you’re the one who did it. But “can’t produce evidence” gets you as far as “this is unresolved” and no further. It doesn’t reach “you’re guilty.” This move puts an equals sign between the two
Hasty generalization
Grab a few cases and paint over everything. A while back Taiwan was debating whether to bring in migrant workers from India, and the internet went up fast: “India has so many sexual assault cases, isn’t this importing a public safety problem along with them?” A few lurid news stories plus a nationality label, and a few hundred million different people in one country get painted into a single shape. However real an individual case is, it can’t stand for the whole, and what hasty generalization is betting on is that you won’t stop and ask one question: is this sample actually representative
Knowing the names is a start; the next step is recognizing the gap in a particular argument. Each of the eight questions below illustrates one fallacy. As you answer, look for the point where the claim, reason, and evidence stop connecting:
Spot the Fallacy
Eight questions, eight movesThree Steps for Examining an Argument
In an actual argument, I do not need to recall every fallacy by name. I can pause, examine what was said, and return to the original question.
Pause. When someone uses an emotionally charged label, first ask how it relates to the question under discussion.
Examine. Are they addressing my claim or attacking me? Are they responding to something I actually said? Does their evidence support their conclusion?
Return. If the conversation has shifted, I can say: “That is a separate question. Let’s first check the number we were discussing.”
The point of recognizing a fallacy is to identify the step in the reasoning that fails.
Winning and Explaining
Someone using a straw man or a hasty generalization may not be trying to mislead anyone. They may simply have missed a step in their own reasoning. Knowing the name of a fallacy does not tell me what they intended.
Winning an argument does not mean the argument is sound. Even when I can identify a mistake in someone else’s reasoning, I still need to check whether my own claim has enough evidence behind it.