Trial and error is a fundamental, experience‑based method for solving problems and acquiring skills. It consists of making successive attempts, observing outcomes, and retaining or discarding responses until a satisfactory result appears. The approach is generally unsystematic: it does not presuppose insight, theory, or an organised analytical plan. Instead, progress depends on variation, feedback, and selection of what works.

Key characteristics

  • Repetition: multiple attempts are made, often with small changes between tries.
  • Variation: responses are varied, either by chance or by deliberate modification.
  • Feedback-based selection: outcomes guide which attempts are repeated.
  • Non‑insightful origin: successes can arise without a single moment of sudden understanding.

These features make trial and error robust in situations with little prior knowledge, but also inefficient where systematic strategies or prior models are available. It scales from simple motor tasks to complex exploratory searches, and can be blind (random) or guided by heuristics.

Historical background

Early experimental psychologists formalised trial and error as a topic of study. The phrase itself is attributed in historical commentary to C. Lloyd Morgan, as reported by W. H. Thorpe; Morgan experimented with alternate phrasings such as "trial and failure" and "trial and practice" before settling on the now‑familiar term. Morgan also proposed a methodological rule, often called Morgan's canon, which advises explaining animal behaviour in the simplest terms consistent with the facts. He illustrated this with observations of his terrier, Tony, whose gate‑opening behaviour appeared clever but, when closely recorded, resolved into a sequence of successful approximations rather than a single insightful act; discussion of this episode appears in accounts of C. Lloyd Morgan.

In laboratory psychology, Edward Thorndike formalised trial‑and‑error experiments using puzzle boxes and recorded learning curves that tracked improvement across trials. Thorndike's work emphasised that responses followed by satisfying consequences are more likely to recur, an observation later developed into more elaborate theories of learning and strengthened by behaviourist analyses such as operant conditioning.

Applications and examples

Trial and error appears across many domains. In everyday life it governs how people learn simple chores, mechanical fixes, or navigation of new environments. In animal behaviour studies it explains many learned sequences that do not require positing insight. In engineering and computing, unstructured search, certain heuristic optimization methods, and early stages of prototyping rely on iterative trialing and adjustment.

As a problem‑solving stance, trial and error is powerful when no reliable model exists and when feedback is immediate and informative. Its limits include inefficiency, potential for getting stuck at suboptimal solutions (local optima), and poor transfer when successful actions are situation‑specific. It contrasts with insightful problem solving, analytical planning, and model‑based methods, and overlaps conceptually with exploratory search, hill‑climbing heuristics, and reinforcement learning paradigms.

For further reading, historical and technical discussions may be found under the names cited here: C. Lloyd Morgan, Edward Thorndike, accounts of Morgan's canon, the anecdote about Tony, the notion of animal insight, modern summaries of learning theory, and treatments of operant conditioning.