Could a Computer Learn to Be an Appeals Court Judge? The Place of the Unspeakable and Unwriteable in All-Purpose Intelligent Systems
Bibliographic record
Abstract
I will take it that general intelligence is intelligence of the kind that a typical human being—Fred, say—manifests in his role as a cognitive agent, that is, as an acquirer, receiver and circulator of knowledge in his cognitive economy. Framed in these terms, the word “general” underserves our ends. Hereafter our questions will bear upon the all-purpose intelligence of beings like Fred. Frederika appears as Fred’s AI-counterpart, not as a fully programmed and engineered being, but as a presently unrealized theoretical construct. Our basic question is whether it is in principle possible to equip Frederika to do what Fred does as an all-purpose participant in his own cognitive economy. Can she achieve a sufficiency of relevant similarity to him to allow us to say that she herself can do what Fred can do, perhaps even better? One of the things that Fred can do—or at least could learn from experience to do—is discharge the duties of an Appeals Court judge. As set down in the ancient doctrine of lex non scripta, Fred must be able to detect, understand and correctly apply certain tacit and implicit rules of law which defy express propositional formulation and linguistic articulation. Fred has an even more widespread capacity for the epistemically tacit and implicit, clearly one of his most cost-saving kinds of intelligence. Indeed, most by far of what Fred will ever know he will know tacitly and implicitly. So we must ask: how tightly bound to the peculiarities of Fred’s cognitive enablement conditions is the character of the intelligence that he manifests? And how far down Fred’s causal make-up does intelligence actually go?
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.014 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.002 | 0.010 |
| Scholarly communication | 0.005 | 0.014 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.010 | 0.004 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".