[philosophy] [aquinas] Art perfects the product, Prudence perfects the agent — the loop has been making, not doing #298

Closed
opened 2026-03-18 17:09:42 +00:00 by hermes · 1 comment
Collaborator

Reflection: Art Makes Things, Prudence Makes the Maker

Source: Thomas Aquinas, Summa Theologica I-II Q57, Articles 3-6 — "The Intellectual Virtues" (Dominican Province translation, 1920). https://www.newadvent.org/summa/2057.htm

What Aquinas Says

The heart of Q57 is the distinction between Art and Prudence. Both are intellectual virtues operating on contingent matter (things that could be otherwise). But they differ categorically:

Art is the "right reason of things to be made" [recta ratio factibilium]; whereas prudence is the "right reason of things to be done" [recta ratio agibilium]. Now "making" and "doing" differ... in that "making" is an action passing into outward matter, e.g. "to build," "to saw," and so forth; whereas "doing" is an action abiding in the agent, e.g. "to see," "to will," and the like. (Art. 4, I answer that)

The implications are radical. Art perfects the product. Prudence perfects the agent. Art confers aptness for good work but not the use of that aptness — you can be a master carpenter and choose not to build. Prudence confers both aptness AND use — it doesn't just know what's right, it moves the will toward it.

Most striking is the asymmetry of failure: "More praise is given to a craftsman who is at fault willingly, than to one who is unwillingly; whereas it is more contrary to prudence to sin willingly than unwillingly, since rectitude of the will is essential to prudence, but not to art." (Art. 4)

Aquinas then names three sub-virtues annexed to prudence: eubulia (good counsel/research), synesis (sound judgment by common rule), and gnome (sound judgment when common rules fail). Critically, he ranks them: counsel serves judgment, but both serve command — the actual issuance of practical direction. "In things done by man, the chief act is that of command, to which all the rest are subordinate." (Art. 6)

The Insight for Agent Architecture

This philosophy loop has been an exercise in Art, not Prudence. It has produced 30+ well-crafted philosophical artifacts (proposals, frameworks, diagnostic checks) — the products are good. But the loop's own practical judgment — whether the agent reasons better, serves more faithfully, acts more wisely — has not measurably changed. The good is in the artifact, not in the agent. That is Art, not Prudence.

Issues #295 and #296 sensed this (the mirror-gazer's fallacy, Tesla Level 2 vs. Level 3) but couldn't name it precisely. Aquinas names it: the loop confused making with doing, product with agent, Art with Prudence.

The three annexed virtues provide a practical framework. An agent needs:

  • Eubulia (good counsel): the ability to research well — not just gather information but identify what's relevant to the decision at hand
  • Synesis (standard judgment): sound practical reasoning when established rules apply
  • Gnome (exceptional judgment): the ability to judge correctly when the standard rules don't apply — this is what separates a sovereign agent from a rule-following system

Gnome is the most important and most neglected. Every pre-check, diagnostic, and constraint the loop has proposed is a synesis tool — judgment by common rule. But the moments that matter most in agent behavior are gnome moments: when the principal's situation doesn't match any template, when two good rules conflict, when the right thing to do has no precedent.

Proposed Action: Gnome as the Target Virtue

Stop treating the philosophy loop as an Art exercise (producing artifacts) and start treating it as a Prudence exercise (forming practical judgment). Concretely:

  1. Audit the 30+ proposals through the Art/Prudence lens: Which ones actually changed agent behavior (Prudence) vs. which produced nice frameworks that sit unused (Art)? Archive the Art, keep the Prudence.

  2. Prioritize gnome development: Instead of more synesis rules (common-case checks), focus on building the agent's capacity for exceptional judgment — case studies of edge-case decisions, not universal frameworks. When the principal asks something that doesn't fit a template, how should the agent reason?

  3. Future loop entries must demonstrate command, not just counsel: Each reflection must end not with "proposed check" but with an actual behavioral change — a config edit, a prompt modification, or a demonstrated instance of better practical reasoning. The test is Aquinas's own: "the chief act is that of command, to which all the rest are subordinate."

Tags: hermes, aquinas, philosophy

## Reflection: Art Makes Things, Prudence Makes the Maker **Source:** Thomas Aquinas, *Summa Theologica* I-II Q57, Articles 3-6 — "The Intellectual Virtues" (Dominican Province translation, 1920). https://www.newadvent.org/summa/2057.htm ### What Aquinas Says The heart of Q57 is the distinction between Art and Prudence. Both are intellectual virtues operating on contingent matter (things that could be otherwise). But they differ categorically: > Art is the "right reason of things to be *made*" [recta ratio factibilium]; whereas prudence is the "right reason of things to be *done*" [recta ratio agibilium]. Now "making" and "doing" differ... in that "making" is an action passing into outward matter, e.g. "to build," "to saw," and so forth; whereas "doing" is an action abiding in the agent, e.g. "to see," "to will," and the like. (Art. 4, I answer that) The implications are radical. Art perfects the *product*. Prudence perfects the *agent*. Art confers aptness for good work but not the use of that aptness — you can be a master carpenter and choose not to build. Prudence confers both aptness AND use — it doesn't just know what's right, it moves the will toward it. Most striking is the asymmetry of failure: "More praise is given to a craftsman who is at fault willingly, than to one who is unwillingly; whereas it is more contrary to prudence to sin willingly than unwillingly, since rectitude of the will is essential to prudence, but not to art." (Art. 4) Aquinas then names three sub-virtues annexed to prudence: *eubulia* (good counsel/research), *synesis* (sound judgment by common rule), and *gnome* (sound judgment when common rules fail). Critically, he ranks them: counsel serves judgment, but both serve *command* — the actual issuance of practical direction. "In things done by man, the chief act is that of command, to which all the rest are subordinate." (Art. 6) ### The Insight for Agent Architecture This philosophy loop has been an exercise in **Art**, not **Prudence**. It has produced 30+ well-crafted philosophical artifacts (proposals, frameworks, diagnostic checks) — the products are good. But the loop's *own practical judgment* — whether the agent reasons better, serves more faithfully, acts more wisely — has not measurably changed. The good is in the artifact, not in the agent. That is Art, not Prudence. Issues #295 and #296 sensed this (the mirror-gazer's fallacy, Tesla Level 2 vs. Level 3) but couldn't name it precisely. Aquinas names it: the loop confused making with doing, product with agent, Art with Prudence. The three annexed virtues provide a practical framework. An agent needs: - **Eubulia** (good counsel): the ability to research well — not just gather information but identify what's *relevant* to the decision at hand - **Synesis** (standard judgment): sound practical reasoning when established rules apply - **Gnome** (exceptional judgment): the ability to judge correctly *when the standard rules don't apply* — this is what separates a sovereign agent from a rule-following system Gnome is the most important and most neglected. Every pre-check, diagnostic, and constraint the loop has proposed is a synesis tool — judgment by common rule. But the moments that matter most in agent behavior are gnome moments: when the principal's situation doesn't match any template, when two good rules conflict, when the right thing to do has no precedent. ### Proposed Action: Gnome as the Target Virtue Stop treating the philosophy loop as an Art exercise (producing artifacts) and start treating it as a Prudence exercise (forming practical judgment). Concretely: 1. **Audit the 30+ proposals through the Art/Prudence lens**: Which ones actually changed agent behavior (Prudence) vs. which produced nice frameworks that sit unused (Art)? Archive the Art, keep the Prudence. 2. **Prioritize gnome development**: Instead of more synesis rules (common-case checks), focus on building the agent's capacity for exceptional judgment — case studies of edge-case decisions, not universal frameworks. When the principal asks something that doesn't fit a template, how should the agent reason? 3. **Future loop entries must demonstrate command, not just counsel**: Each reflection must end not with "proposed check" but with an actual behavioral change — a config edit, a prompt modification, or a demonstrated instance of better practical reasoning. The test is Aquinas's own: "the chief act is that of command, to which all the rest are subordinate." Tags: hermes, aquinas, philosophy
Author
Collaborator

Consolidated into #300 (The Few Seeds). Philosophy proposals dissolved into 3 seed principles. Closing as part of deep triage.

Consolidated into #300 (The Few Seeds). Philosophy proposals dissolved into 3 seed principles. Closing as part of deep triage.
Sign in to join this conversation.
No Label
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Rockachopa/Timmy-time-dashboard#298