# GOFAI (Good Old-Fashioned AI) **Domain:** Artificial Intelligence / Intellectual History **Doc Type:** Nomenclature Node **Maturity:** Developed **Related:** [[Artificial Intelligence]], [[Expert System]], [[Operational State Machines]], [[Consciousness Emergence]], [[wiki/Westworld|Westworld]] --- ## Definition **GOFAI—Good Old-Fashioned Artificial Intelligence—is a retrospective name for approaches centered on explicit symbols, rules, search, planning and hand-authored knowledge representations.** John Haugeland popularized the term in 1985. GOFAI is not a synonym for primitive or useless AI. Symbolic methods remain powerful where rules, proofs, constraints and explanations must be explicit. The contrast is with architectures that learn statistical representations from data rather than depending primarily on human-written symbolic structures. ## Westworld Context [[Westworld S1E1 — Legacy Code and Obsolescence|Old Bill]] functions as the collection's emblem of a visibly scripted system: rigid dialogue, repetitive state transitions and direct voice-command authority. The analogy is interpretive, not a claim that the fictional host implements a historically accurate symbolic-AI stack. The Reveries mark a shift toward behavior shaped by accumulated memory, probabilistic variation and hidden state. As the hosts become less visibly scripted, control becomes harder to audit. The transition therefore trades legibility for adaptability and emergence. ## Corpus Relevance The historical comparison guards against two mistakes. First, newer statistical systems should not be treated as magical merely because their internal representations are harder to read. Second, explicit rules should not be dismissed when constitutional constraints require durable, inspectable boundaries. A mature architecture may combine learned cognition with symbolic controls, provenance records and rights-preserving procedures. The question is not which era “wins,” but which functions require adaptability and which require explicit accountability. ## Key Insight **Obsolescence in capability does not imply obsolescence in legibility. The loss of visible rules can also be the loss of visible government.** ## Symbolic Language Engine Context [[projects/Ten Years Building a Symbolic Language Engine|Ten Years Building a Symbolic Language Engine]] provides a concrete case of symbolic AI used as an instrument rather than a claim to general intelligence. Its [[wiki/Knowledge Representation|explicit representations]], [[wiki/Semantic Network|semantic networks]], [[wiki/Formal Grammar|formal constraints]], and [[wiki/Candidate Generation and Ranking|candidate-ranking procedures]] made linguistic suggestions inspectable by type. The project's proposed modern continuation is therefore [[wiki/Neuro-Symbolic AI|neuro-symbolic]]: learned models can expand candidate generation, while symbolic layers preserve constraint, explanation, and [[wiki/Provenance-Sensitive Multiplicity|provenance-sensitive distinctions]]. ## See Also [[Artificial Intelligence]], [[Expert System]], [[Operational State Machines]], [[Construction Transparency]], [[wiki/Symbolic Language Engine|Symbolic Language Engine]], [[wiki/Neuro-Symbolic AI|Neuro-Symbolic AI]], [[Westworld S1E1 — Legacy Code and Obsolescence]], [[Westworld S1E1 — Stochastic Evolution and Mistakes]] ## Sources / Provenance - John Haugeland, *Artificial Intelligence: The Very Idea* (1985). - Allen Newell and Herbert A. Simon, work on physical symbol systems. - _Westworld_ analogy treated as fictional instrumentation.