The Ping Pong Method: A Human-in-the-Loop Framework for Iterative Knowledge Extraction with Large Language Models (LLMs)

This paper introduces the Ping Pong Method (PPM), a human-in-the-loop framework for co-creating high-precision outputs with large language models (LLMs). Developed through historical research on the 1915 Gastón Gadín case in Paraguay, the method combines adversarial prompting, iterative validation, and logic-based inference to enable non-experts to produce academically rigorous results. The PPM provides an accessible alternative to traditional prompt engineering, especially for independent researchers and digital humanities.