Performance is not experience
Success in conversation may demonstrate behavioural competence without showing that the system understands or feels anything.
Twentieth-century philosophy and computing · 1912–1954
Turing
Turing replaces the foggy demand to define machine thinking with operational tests of what machines can compute and how convincingly they can converse.
Alan Turing was a British mathematician, logician and cryptanalyst whose work founded computability theory and shaped modern computer science. His analysis of machine intelligence introduced the imitation game and reframed questions about thinking as questions open to disciplined testing.
MindKnowledgeAgency
The question he enters
The question Turing enters
What evidence would justify treating a machine's conversation as intelligent rather than as an impressive imitation?
The central move
Replace a verbal dispute with a test
Turing argues that asking whether machines can think invites an unproductive battle over definitions. The imitation game instead asks whether an interrogator can reliably distinguish a machine's linguistic performance from a human's, creating a public criterion for intelligence without claiming direct access to inner experience.
Key ideas
Terms that carry the argument.
- Turing machine
- An abstract model that captures rule-governed computation through symbols, states and stepwise operations.
- Computability
- The class of problems solvable by a systematic procedure, together with limits on what any such procedure can decide.
- Imitation game
- A conversational test in which a machine attempts to sustain responses indistinguishable from a human participant's.
- Learning machines
- The proposal that intelligent behaviour may be developed through training and modification rather than fully programmed in advance.
Where the view strains
Objections and limits.
The test rewards human likeness
Intelligence unlike ordinary human conversation may be overlooked, while deception or fluent mimicry may be overvalued.
Sources and routes
Read the work. Follow the argument.
Primary texts
Where to meet the position.
- Wiley Online Library / London Mathematical SocietyOn Computable Numbers, with an Application to the EntscheidungsproblemWebThe paper introducing the machine model now central to computability theory.Proceedings of the London Mathematical Society, series 2, volume 42, issue 1, 1937, pages 230-265; DOI 10.1112/plms/s2-42.1.230. · Official journal record with full text subject to Wiley access terms. Checked Aug 2026.
- Oxford AcademicComputing Machinery and IntelligenceWebThe imitation game, objections to machine intelligence and the possibility of learning machines.Mind, volume LIX, issue 236, October 1950, starting at page 433. · Official journal record with online article access subject to Oxford Academic terms. Checked Aug 2026.
- The Turing Digital Archive, King's College CambridgeIntelligent MachineryWebAn early report on machine learning, search and artificial intelligence.Turing's 1948 typescript report with autograph corrections, archive reference AMT/C/11. · Institutional catalogue record and digitised manuscript pages. Checked Aug 2026.