credit assignment

The vault's #2 concept and one it deliberately reads on two registers that live in different source communities. In machine learning it is the distribution of error backward through a network (structural) or through time (temporal) — the problem backpropagation solves. In the sociology of science it is who gets credit for a discovery — the problem priority disputes are about. The vault's signature bridge is that these are the same problem shape; the term "structural credit assignment," often traced to Minsky 1961, is a running attribution puzzle in its own right. Recent captures push toward "arbitration" as the general missing ingredient across both readings.

References

• claim-credit-assignment-spans-sociology-and-backprop · claim-minsky-1961-named-credit-assignment · myth-structural-credit-assignment-from-minsky-1961 · claim-credit-detectors-are-themselves-misattributed
• moc-backpropagation-origins · Captures: 2026-07-09-who-first-used-the-term-structural-credit-assignment, 2026-07-13-is-arbitration-the-general-missing-ingredient-across-the