
Researchers at Helsinki University have used artificial intelligence to analyze extensive unpublished notebooks written by Kurt Gödel, the influential Austrian logician whose incompleteness theorems transformed modern mathematics. The project, led by Finnish logician Jan von Plato since 2017, focused on 16 workbooks consisting of 1,009 pages written by Gödel in the early 1940s, much of which was composed in Gabelsberger, an obsolete German shorthand.
Initial progress on understanding the material proceeded slowly, with researchers advancing approximately one page per week. The situation changed significantly in September when von Plato incorporated Claude, an AI language model, into the research effort. Within days, Claude produced a comprehensive analysis of the entire collection, revealing that Gödel had proved hundreds of previously unknown theorems and developed numerous mathematical ideas that would not be formally recognized for decades. Von Plato characterized the discovery as remarkable, noting that the workbooks contained anticipations of results not known until the 1970s and 1980s, along with 14 unexplored themes in logic.
One significant finding involved Gödel’s work on the independence of the axiom of choice. In 1966, mathematician Paul Cohen received the Fields Medal for proving the independence of the continuum hypothesis and the axiom of choice, work that Gödel claimed to have completed in the 1940s. Using Claude’s analysis, researchers discovered that Gödel had indeed constructed nearly complete proofs of these results, missing only one definitional step that he had actually corrected in his notes two years prior. The reconstructed proof successfully passed verification through Coq, a formal proof-checking program.
Von Plato initiated the Gödel project upon facing mandatory retirement, securing substantial funding from the European Research Council. He became one of only six people worldwide capable of reading Gabelsberger shorthand, a skill he taught himself for the research. Claude’s ability to synthesize information across all 1,000 pages simultaneously proved invaluable, as the AI could identify connections and patterns across Gödel’s work in ways that would be prohibitively time-consuming for human researchers. Through continued collaboration with Claude, the team identified 663 theorems for which Gödel had either completed proofs, sketched proofs, or developed proof strategies.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI