The Invisible Gardener
Two explorers find a clearing with cultivated flowers. One says a gardener tends it; the other is skeptical. Every test for the gardener comes back negative: no footprints, no scent, no heat signature. How many failed tests does it take before the gardener hypothesis should be abandoned?
Philosopher Antony Flew introduced this parable in 1950 to argue about what makes a belief meaningful. The story shows how a claim can die by a thousand qualifications, modified to accommodate every objection until it can no longer be falsified. Karl Popper's criterion of falsifiability lurks just behind it.
Flew, A. (1950). Theology and Falsification. In New Essays in Philosophical Theology.
The parable
Two explorers discover a clearing in a jungle with cultivated-looking flowers and weeds. The believer says there must be a gardener who tends it. They set up camp to watch. No gardener appears.
The believer says the gardener is invisible. They build an electric fence, set bloodhounds. Still nothing. The believer says the gardener is also intangible, undetectable by dogs, leaves no marks.
The skeptic asks: "But what remains of your original assertion? What is the difference between an invisible, intangible, undetectable gardener and no gardener at all?"
Flew's point
Flew used this to argue that theological claims like "God loves us" and "there is a divine plan" survive every counterexample by modification. When a child dies of cancer, the believer doesn't say "God failed" but "God's ways are mysterious." The claim absorbs any evidence.
Flew thought this rendered the claim meaningless, not false but non-cognitive. A claim that can't be falsified isn't saying anything about the world.
The pushback
The story cuts both ways and has generated intense debate:
- Mitchell's counter: some beliefs are held on the basis of cumulative personal evidence, not any single verifiable prediction. Trust works this way too.
- Hare's counter: some beliefs function as "bliks," frameworks through which we interpret experience rather than hypotheses we test. Every worldview has unfalsifiable foundations.
- Scientific response: physics is full of things we can't directly observe but infer from effects. Does that make quarks meaningless?
The actual question
At what point does "modifying a hypothesis to fit the data" cross the line from good scientific practice to special pleading? This question applies equally to theology, macroeconomics, and how you defend your career choices.