Functorial Transfer Probe¶
Transfer experiment — instantiates Composable Relation Modeling
Transfers a relational pattern from one domain to another and tests whether its arrows, composition, and invariant survive the crossing.
A Functorial Transfer Probe takes a relational model that works in a source domain and asks whether it can be carried, structure intact, into a different target domain. Its defining commitment is that a transfer is legitimate only if it is functorial: every object maps to a target object, every arrow to a target arrow, and — the load-bearing part — composites map to the composites of the images, so that "do f then g, then transfer" equals "transfer, then do the images." The probe is exploratory and cross-domain by nature: it is not confirming that one already-chosen map is clean (that is a checklist's job) but searching for whether a mapping between two materially different domains can exist at all, and reporting what it costs. Because real transfers are rarely free, the probe also names the tradeoff or adjunction — the pair of directions between the domains and what each preserves or loses.
Example¶
A hospital's discharge process is chronically congested: patients medically ready to leave wait hours for a bed to be cleaned, a ride arranged, a pharmacy order filled. An operations lead who once worked in electronics manufacturing suspects the factory pull system — where a downstream station signals upstream to release the next unit, capping work-in-progress — might transfer. Rather than analogize loosely, they run a Functorial Transfer Probe. The source model's objects are workstations and its arrows are "release next unit"; the target objects are discharge stages (bed-ready, transport-ready, meds-ready) and the candidate arrows are "signal readiness upstream."
The probe's discipline is to check that composition transfers, not just objects. In the factory, chaining two pull signals composes into one longer pull with a predictable invariant: work-in-progress stays capped. Does the composite survive in the hospital? The invariant of interest they insist on is bounded patients-in-limbo. Testing the composite reveals a break: unlike factory units, a patient's "readiness" can regress (a new symptom re-admits them upstream), so a composed pull chain does not preserve the cap the way it does on the line. The probe does not throw the pattern out — it marks the tradeoff: pull transfers for the transport and pharmacy legs (irreversible, unit-like) but not for the clinical leg. The output is a partial, honest transfer rather than a seductive false analogy.
How it works¶
- State the transfer rule as an object-and-arrow mapping. Say explicitly which target object each source object becomes and which target arrow each source arrow becomes. A transfer with no arrow-level mapping is a metaphor, not a probe.
- Test that composition is preserved. Pick source composites, map them, and check that the image of the composite equals the composite of the images. This is the step that separates real structural transfer from surface resemblance.
- Fix the invariant that must survive. Name the property the source pattern buys (a cap, a guarantee, an equivalence) and require the transferred version to deliver the same property, not a look-alike.
- Mark the adjunction or tradeoff. Record the paired directions between domains and what each preserves versus loses, so the transfer's price is explicit rather than discovered later.
The probe consumes an existing source model; it does not inventory objects or draw the base diagram itself.
Tuning parameters¶
- Transfer fidelity bar — how strictly composition must be preserved to accept the mapping. A high bar rejects seductive analogies; a low bar admits patterns that will break under composition.
- Invariant scope — how much of the source's guarantee the target must reproduce. Demanding the full invariant yields trustworthy transfers but few of them; accepting a weaker invariant yields more transfers of lower assurance.
- Directionality — whether the probe seeks a one-way transfer or a two-way adjoint pair. Two-way is stronger evidence of shared structure but is rarely available across genuinely different domains.
- Partiality tolerance — whether a transfer that works on some arrows but not others is accepted as partial or rejected wholesale. Partial acceptance is usually the honest outcome.
When it helps, and when it misleads¶
Its strength is that it upgrades analogy from rhetoric to test: it is the mechanism that catches when a pattern "transfers poorly because surface analogies were preserved but relational structure was not." When it succeeds, it licenses reusing hard-won structure across domains; when it fails, the failure is specific and instructive.[1]
Its failure mode is the false functor — a mapping that matches objects and even individual arrows convincingly but silently breaks on composition, so the borrowed guarantee evaporates exactly when several steps chain. The classic misuse is declaring transfer on object-level resemblance alone ("workstations are like discharge stages, so pull will work") without ever testing a composite. A subtler misuse is quietly weakening the invariant until the transfer "works" by definition. The guarding discipline is to make the composition-preservation test mandatory and adversarial, to hold the invariant fixed rather than negotiating it down, and to prefer an honest partial transfer over a flattering total one.
How it implements the components¶
structure_preserving_transfer_rule— its core: the explicit object-to-object and arrow-to-arrow mapping between domains, accepted only if composites map to composites.invariant_of_interest— the property the source pattern guarantees, which the transferred version must reproduce for the transfer to count.adjunction_or_tradeoff_pair_marker— the recorded paired directions between domains and what each preserves or loses, making the transfer's price explicit.
It does not inventory the objects or draw the base arrows (object_role_inventory, typed_arrow_inventory — that is Object–Arrow Diagram), and it does not run the within-model, single-map preservation review (composition_rule, identity_arrow_baseline, associativity_invariant — that is Structure-Preservation Checklist, its nearest twin; the probe crosses domains and searches, whereas the checklist confirms one in-domain map).
Related¶
- Instantiates: Composable Relation Modeling — the probe is the archetype's cross-domain transfer experiment.
- Consumes: Object–Arrow Diagram supplies the source model being transferred; Structure-Preservation Checklist can verify the resulting target map in place.
- Sibling mechanisms: Structure-Preservation Checklist · Object–Arrow Diagram · Composition Table · Commutative Path-Equivalence Diagram · Identity and Associativity Test Suite · Source/Target Type Check · Interface-Contract Category Map · Categorical Refactoring Workflow
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: Functorial Transfer Probe operates as a bounded trial, probe, simulation, or rehearsal that generates evidence from performance because it transfers a relational pattern from one domain to another and tests whether its arrows, composition, and invariant survive the crossing.
Independent corroboration: The frozen evidence defines Functorial Transfer Probe as 'Transfers a relational pattern from one domain to another and tests whether its arrows, composition, and invariant survive the crossing', so its operative form is Experiment, Test & Rehearsal.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Mathematics
Origin pattern: Single lineage
Present-day reach: Universal
Rationale: Functorial preservation of identities and composition comes directly from category theory.
Review resolution: Both reviewers agree that mathematics is primary: Functorial preservation of identities and composition comes directly from category theory. I retain no alternate origin domains only where the reviewers identify formative lineage, not every later application. I resolve origin_mode as single_lineage because the mechanism is best explained as one continuous professional lineage. I resolve domain_reach as universal because it is broadly applicable across essentially all domains. Encyclopedia synthesis is true because the exact generalized packaging is an encyclopedia-authored combination or refinement.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
References¶
[1] Gentner, D. "Structure-Mapping: A Theoretical Framework for Analogy". Cognitive Science 7(2), 155–170 (1983). Models analogy as mapping structured knowledge from a base domain into a target domain. registry ↩