# Symbolic AI and the "Golden Age" (1956–1970s)

The early decades of AI were dominated by symbolic, rule-based approaches to reasoning. This sub-article expands [Section 3 of History](../history.md).

## Search and Problem Solving
- **Logic Theorist** and **General Problem Solver** demonstrated that machines could manipulate symbols to prove theorems and plan actions.
- Search algorithms (A\*, heuristic search, constraint satisfaction) became core tools.

## Game Playing
- Arthur Samuel's checkers program (1959) used self-play and learning.
- Later, chess programs improved steadily toward Deep Blue (1997).

## Languages and Tools
- **Lisp** (McCarthy, 1958) became the dominant AI language, with garbage-collected list processing.
- **Prolog** (1970s) enabled logic programming and declarative knowledge representation.

## Neural Networks Enter and Retreat
- **Perceptron** (Rosenblatt, 1958) was a simple learning network.
- Minsky & Papert's *Perceptrons* (1969) showed limits of single-layer perceptrons, dampening neural research for years.

## Robotics and Planning
- **Shakey the robot** (Stanford, 1966–1972) autonomously planned, perceived, and navigated.
- STRIPS (1971) formalized planning in terms of preconditions and effects.

## Why It Matters
Symbolic AI established reasoning, representation, search, and planning as core AI problems. Its strengths (explainability, logic) and weaknesses (brittleness, knowledge engineering) shaped later cycles.

## Sources
- Newell, A. & Simon, H. (1972). *Human Problem Solving.*
- Minsky, M. & Papert, S. (1969). *Perceptrons.*
- Fikes, R. & Nilsson, N. (1971). *STRIPS: A New Approach to the Application of Theorem Proving to Problem Solving.*

*Back to [History](../history.md) · Next: [Expert Systems](expert-systems.md).*