# Symbolic AI Symbolic AI — retrospectively nicknamed "GOFAI" (good old-fashioned AI) — is the original paradigm of [[Artificial Intelligence]]: the idea that intelligence consists of manipulating symbols according to explicit logical rules. On this view, to make a machine intelligent you encode human knowledge as symbols (facts, concepts) and rules (logic, if-then statements), and the machine reasons by applying them. It rests on the "physical symbol system hypothesis" of Newell and Simon — that symbol manipulation is sufficient for general intelligent action. This approach, grounded in [[Logic]], dominated AI from the Dartmouth Workshop (1956) through the 1980s, producing theorem provers, planners, and the Expert Systems that were its commercial peak. Its strength is transparency: a symbolic system's reasoning can be inspected and explained, rule by rule. Its fatal weakness is brittleness — it cannot easily handle the ambiguity, noise, and sheer common-sense breadth of the real world, and every rule must be hand-coded. The "knowledge acquisition bottleneck" (you can't write down everything a human knows) is what ultimately stalled it, ceding dominance to data-driven [[Machine Learning]]. ## Why it matters here Symbolic AI is mostly of historical and conceptual importance rather than political — included as essential disciplinary literacy and to make the field's central paradigm war legible (see Connectionism). One thread connects to the vault's wider concerns: the symbolic tradition's prizing of **explainability** is exactly what today's opaque [[Deep Learning]] "black boxes" lack, and it is enjoying renewed interest precisely because of AI's accountability and transparency problems. The contrast — interpretable-but-brittle rules versus powerful-but-inscrutable learned models — frames real debates about whether high-stakes AI decisions (in courts, lending, medicine) should be auditable. [[Neurosymbolic AI]] approaches that try to combine the two are an active research direction. ## See Also - Connectionism — the rival paradigm and the war between them - Expert Systems — symbolic AI's commercial peak - [[Machine Learning]] — the data-driven approach that displaced it - [[Logic]] — the formal foundation of symbolic reasoning - [[Neurosymbolic AI]] — the modern attempt to fuse symbolic rigor with neural learning