Teleological fallacy

“They’ll just mutate to beat it”

“There’s no point developing new antibiotics. Bacteria mutate in order to become resistant to whatever we throw at them.”

Bacteria don’t mutate in order to do anything. Mutations happen whether or not they help, and resistance spreads when a resistant variant already exists and the drug kills its competitors. That difference is why the conclusion fails: if resistance needed only the need, nothing could slow it, but because it depends on the right variant turning up, it can be made less likely. Giving drugs in combination, so that a germ would need several resistance mutations at once, is a standard treatment for tuberculosis for exactly that reason.

“It must be there for something”

“Men have nipples, so they must do something. Evolution doesn’t keep useless parts around.”

This assumes every feature exists because of what it does. Male nipples are better explained as a by-product: male and female embryos share the same early body plan, and nipples form before the two develop differently. Selection only removes a feature when carrying it costs enough to matter, and these cost almost nothing. A feature can be present because of how the body is built, not because of anything it is for.

“Viruses evolve to become milder”

“This new virus will get milder over time. It’s in the virus’s interest to keep its hosts alive, since a dead host can’t spread it.”

This sounds like biology, and for a long time something close to it was the standard view, known as the avirulence hypothesis. It reads the virus’s “interest” as a goal that evolution pursues. But selection favors whichever variants spread most, and how harmful those turn out to be depends on how harm trades off against transmission. As Alizon and colleagues (2009) review, the field moved away from the idea that parasites always evolve toward harmlessness, and the trade-offs involved make sweeping predictions hard. The virus may well get milder; “it needs its hosts alive” isn’t a reason to expect that it will.

The fallacy runs straight from what a feature does to why it exists:

The philosopher Larry Wright (1973) argued that “F is there to do G” is a proper explanation when F exists because it does G, as when selection kept F for doing G. On this view the problem is not the words “in order to” but a missing middle row. The educational researcher Kostas Kampourakis (2020) makes the same point about biology teaching: purpose-talk backed by selection is scientifically sound, and the error is explaining a trait as present because it was needed, or as if it had been designed.

• Need as cause: the need for a trait is said to produce it. “Finches diversified in order to survive.”
• Purpose in the non-living: features of rocks, weather or planets explained by what they do for living things. “Rain falls so that crops can grow.”
• Everything has a function: assuming every trait was selected for something, so a purpose must be found. Stephen Jay Gould and Richard Lewontin (1979) criticized this habit in evolutionary biology as the “Panglossian paradigm”, after the tutor in Voltaire’s Candide who explains that noses were made to wear spectacles.
• Forecasting from a goal: predicting where evolution will go from what would be good for the organism. “Humans will evolve bigger brains because we need them.”

The same name is sometimes used in history for a different error: reading the past as if it had been heading toward the present. That isn’t covered here.

Purpose-talk is legitimate when a real mechanism links the purpose to the cause.

• Shorthand for a selection history. “The heart is for pumping blood” summarizes the fact that hearts were kept because they pumped blood.
• A real agent with a goal. People make tools for purposes, and many animals act to reach goals. A beaver’s dam is built to hold back water.
• A question, not an answer. Asking “what might this be for?” is a useful way to start an investigation, as long as the answer is tested and “nothing; it’s a by-product” is allowed.
• Plain description of what something does. “The ozone layer absorbs ultraviolet light” names an effect without claiming it explains why the layer exists.

The test: what links the benefit to the cause? If you can name the selection, the designer or the goal-directed agent, the explanation can stand. If the benefit is doing the work alone, it can’t.

The heart is there to pump blood

A biology teacher: “The heart is there to pump blood. That’s why it has valves: to stop blood flowing backward.”

This is purpose-talk, but it’s shorthand for a selection history. Hearts that moved blood, and valves that stopped backflow, made their owners more likely to survive and reproduce, and that is why they’re there now. Unpacked, every “in order to” becomes a “because” about the past.

Predicting pesticide resistance

“If we spray this field with the same insecticide every year, the beetles will become resistant to it.”

This predicts how a species will evolve, which sounds like forecasting from a goal. But it rests on a mechanism: in a large population, a few insects that happen to survive the spray are likely to exist, the spray kills the rest, and their offspring take over. The prediction doesn’t depend on the beetles needing anything, which is why it can also be prevented (by rotating products, for example).

Purpose-based explanation seems to come naturally. The psychologist Deborah Kelemen (1999) asked children and adults to choose between explanations for features of prehistoric rocks and animals. Children in the early school grades often preferred teleological answers even for non-living things, such as that the rocks were pointy “so that animals like Cryptoclidus could scratch on them when they got itchy”, while adults mostly reserved them for living things. Explaining beforehand how rocks form physically shifted that preference only in the oldest children tested.

Adults show the tendency too when they have to answer quickly. Kelemen and Evelyn Rosset (2009) had undergraduates judge explanations under time pressure or without it. Under time pressure, they accepted more scientifically unwarranted teleological explanations (such as “the sun radiates heat because warmth nurtures life”) but did no worse on other false explanations. Kelemen, Joshua Rottman and Rebecca Seston (2013) found the same pattern in physical scientists at top American universities: they accepted fewer such explanations overall than comparison groups, but more when rushed. The authors read this as a default way of explaining that education overrides without removing.

It’s also reinforced by legitimate language. Biologists rightly say that wings are “for” flight, and the shorthand is easy to extend to cases with no selection behind it. And the real mechanism is harder to picture. The dark form of the British peppered moth spread in the nineteenth century where soot-darkened bark made pale moths easier for birds to find. But it began as a single mutation, which van’t Hof and colleagues (2016) dated to around 1819, and it spread because it already existed when darker bark made it useful. “The moths turned dark to hide” is shorter, and leaves that out.

• Ask what happened first. Was there variation for selection to act on, a designer, an agent with a goal? If the answer is “the need came first”, that’s the gap.
• Ask whether it could be a by-product. Not every feature has a function of its own.
• Ask what the purpose predicts that the mechanism doesn’t. Usually the teleological version promises an outcome (resistance always arrives, viruses always weaken) that the mechanism makes conditional.
• Don’t treat purpose-talk as the error by itself. “Eyes are for seeing” is fine; it depends on whether the speaker could unpack it.