CONCEPT
Post Production Systems
Emil Post’s 1920s formalism—the insight that any rule-based system reduces to replacing one string of symbols with another—the direct ancestor of expert systems, formal grammars, and every large language model that has ever generated a word.
A Post production system is a set of rules, each of which says: if the current string of symbols matches this pattern, replace it with that string. Strip away the meaning, reduce any system of inference to its bare mechanism, and this—Emil Post showed in the 1920s—is what remains. He called the result a production system, proved his normal form theorem (that any such system can be reduced to a single axiom and rules that merely delete a prefix and append a suffix), and established that from this one operation, all of mathematics can in principle be generated. The concept became the taproot of symbolic AI: every expert system of the 1970s and 1980s was a Post production system with handwritten rules; Noam Chomsky’s hierarchy of formal grammars is built on Post-style productions; every compiler that parses a programming language is applying Post’s formalism. The concept also becomes, on inspection, the architecture of large language models: a transformer takes a
Keep reading with YOU ON AI
Unlock the full book, 10,000+ field-guide entries, and a 1000+ thinker library. If you have a book code, register now — it takes a minute.