Argument from Example: When One Case Proves a Point — and When It Doesn't
"Japan has universal healthcare and some of the best health outcomes in the world — that proves the model works." "Look at Venezuela — that's what socialism leads to." "My grandfather smoked his whole life and lived to 97." "There's this woman in my town who got off welfare and built a million-dollar business." Examples are everywhere in argument — and they are doing more work than we usually notice. Sometimes that work is legitimate. Sometimes it is deeply misleading.
What the Argument from Example Does
The argument from example is the basic inductive move: using one or more specific cases to support a general claim. In philosopher Douglas Walton's taxonomy of argumentation schemes, it is among the most fundamental:
- Example premise: In case C₁ (and possibly C₂, C₃...), proposition P holds.
- Generality premise: Case(s) C are representative of the general class G.
- Conclusion: P holds generally for G (or at least presumptively, in the absence of countervailing evidence).
All inductive reasoning follows this pattern at some level. Science builds theories from observations; law develops doctrine from precedents; policy is shaped by case studies. Without the capacity to generalise from specific instances, we could not function as rational agents. The question is never whether to use examples, but how many, how well selected, and how honestly interpreted.
When Examples Constitute Good Evidence
Under the right conditions, a well-chosen example can be genuinely probative — that is, it can provide real evidence for a claim, not just rhetorical decoration.
Existence proofs. A single example can definitively establish that something is possible. If someone claims that no democratic society has ever transitioned to authoritarianism peacefully, a single counterexample refutes the claim outright. This is the logical structure of falsification: one contradicting instance destroys a universal claim. When the purpose of an example is to show that something can happen, one suffices.
Illustrating a mechanism. Examples that illuminate a causal mechanism are more powerful than pure enumeration. If I want to argue that social media platforms amplify emotionally provocative content, a detailed case study showing the engagement algorithm's effects on a particular media ecosystem can support the claim by showing how the mechanism works — making the general claim more plausible, even if the single case cannot prove the generality statistically.
Cumulative examples. When multiple independent cases from diverse contexts all point in the same direction, the convergence is genuine evidence. The strength of the argument depends on the independence and variety of the examples, not just their number. Ten examples from a single dataset, a single region, or a single type of context count for much less than three examples from genuinely independent, varied sources.
In law and precedent. Legal reasoning is explicitly built around the argument from example. The doctrine of precedent (stare decisis) holds that how similar cases were decided in the past is evidence for how they should be decided now. This is not mere appeal to tradition — it encodes values of consistency, predictability, and equality before the law. The relevant question is always whether a new case is genuinely analogous to the precedent invoked. See also: argument from precedent.
The Classic Failures
The argument from example goes wrong in several well-documented ways, each of which deserves scrutiny.
Hasty generalisation. Drawing a broad conclusion from a small, unrepresentative sample is the most basic inductive error. "My grandfather smoked and lived to 97" is not evidence that smoking is harmless — it is one data point from a distribution that contains millions of others. The plural of anecdote is not data. Single vivid examples are systematically overweighted in human cognition (see availability heuristic) relative to base rates and statistical patterns. We remember the grandmother who got off welfare and built a business; we do not hold equally present in mind the thousands who could not. See also: hasty generalisation.
Cherry-picking. Perhaps the most common and consequential abuse: selecting only the examples that support your conclusion while ignoring or actively concealing the ones that don't. Climate change sceptics citing cold winters while ignoring long-term temperature trends. Policy advocates highlighting success stories from their preferred programme while burying failure data. Literary critics quoting passages that support their interpretation while omitting passages that complicate it. Cherry-picking is the argument from example weaponised — it presents a misleadingly favourable sample as if it were the whole.
What makes cherry-picking so insidious is that the individual examples cited are often genuinely true. The fraud lies not in fabrication but in selection. A confirmed example of a welfare recipient becoming a millionaire does exist. The question is what that example establishes about the population of welfare recipients — and the honest answer is: very little, without knowing the base rates and the full distribution. See also: confirmation bias, which makes cherry-picking feel natural and invisible.
The unrepresentative example. Even without deliberate cherry-picking, a single case may be unrepresentative for structural reasons the arguer doesn't notice. Japan's healthcare outcomes may depend on factors — dietary patterns, social cohesion, working culture, demographic structure — that cannot be transplanted to another context. Venezuela's economic collapse may owe as much to oil dependency and political dysfunction as to economic ideology. Treating one case as transparently representative of a category is a form of false cause reasoning: assuming that what explains the example explains the general case.
The exception proving the rule — misunderstood. The phrase "the exception proves the rule" is often invoked to protect a generalisation from a counterexample. But it is almost always misused. The phrase originally meant "the exception tests the rule" (from the Latin probat in the sense of probe/test) — a genuine exception is evidence against the rule, or at least a signal to sharpen it. It is not a magic phrase that lets generalisations absorb contradictory data without modification.
The Representativeness Problem
The critical question in any argument from example is representativeness: is this example typical of the class it is claimed to illustrate? This is harder to assess than it sounds.
Human cognition is systematically biased toward memorable, vivid, concrete examples — the statistical concept of representativeness (in Kahneman and Tversky's sense) captures this well. We tend to treat examples that are salient, emotionally resonant, or narratively coherent as more representative than they are. News media selects for unusual events, which means the examples that reach most people are systematically unrepresentative of base rates. A plane crash is reported extensively; the ten thousand flights that landed safely that day are not. The result is a systematically distorted picture of aviation risk — driven entirely by the argument from example operating on a skewed sample.
The fix is not to stop using examples — it is to be explicit about the relationship between the example and the claimed generalisation. Ask: what is the reference class? How large is it? How was this example selected? What does the distribution look like? These questions force the move from anecdote to evidence.
Demanding Better Examples
When someone offers an example in support of a general claim, the productive challenge is not to dismiss the example but to ask what it actually establishes:
- How was this example found? Deliberately sought examples supporting a predetermined conclusion are weaker evidence than examples that emerged from systematic search.
- How representative is this case? What do we know about the population it comes from? Are there known selection effects?
- What is the base rate? If this example is exceptional, how exceptional? How often does the general pattern hold?
- Are there counterexamples, and how are they being handled? A claim that ignores inconvenient contrary cases is not well-supported, however compelling the positive examples are.
- What mechanism does the example illustrate? Examples that illuminate a causal mechanism are more valuable than examples that merely enumerate instances.
The Rhetorical Power of Examples
It would be wrong to end without acknowledging why examples are so central to human communication — and why their persuasive power sometimes exceeds their evidential weight.
Psychologist Paul Slovic has documented the "collapse of compassion" phenomenon: statistical presentations of large-scale suffering generate less concern and charitable giving than the presentation of a single identified victim. The narrative and emotional engagement that a well-chosen example creates is not irrational in itself — it is part of how we make abstract propositions concrete and humanly meaningful. Examples are not just evidence; they are tools of comprehension.
The task of critical thinking is not to expunge examples from reasoning — they are too valuable for that — but to ask, consistently and honestly, what the example actually establishes, what it might be concealing, and whether the move from the specific to the general is as smooth as it appears.
Sources & Further Reading
- Walton, Douglas N. Argumentation Schemes for Presumptive Reasoning. Lawrence Erlbaum, 1996.
- Walton, Douglas, Chris Reed, and Fabrizio Macagno. Argumentation Schemes. Cambridge University Press, 2008.
- Kahneman, Daniel. Thinking, Fast and Slow. Farrar, Straus and Giroux, 2011.
- Slovic, Paul. "The More Who Die, the Less We Care." Psychic Numbing and Genocide. 2010.
- Wikipedia: Faulty generalization
- Wikipedia: Cherry picking