Ontological Parsimony
About this pattern
This is a generated FPF pattern page projected from the published FPF source. It is canonical FPF content for this ID; it is not a FPF Reference product feature page.
How to use this pattern
Read the ID, status, type, and normativity first. Use the content for exact wording, the relations for adjacent concepts, and citations to keep active work grounded without pasting the whole specification.
Type: Kernel parsimony and admission discipline pattern Status: Stable Normativity: Normative unless a section is explicitly informative
Use this pattern when FPF work proposes a new U-kind, core relation, dependent durable value, or public structural name and the current question is whether existing ontology can express the claim without the proposed durable ontology addition.
Relations
Content
Use This When
Use this pattern when FPF work proposes a new U-kind, core relation, dependent durable value, or public structural name and the current question is whether existing ontology can express the claim without the proposed durable ontology addition.
Typical moments:
- a new U-kind seems useful after
E.24.UK; - a proposed root kind may actually be a dependent value, slot, relation, record, publication form, lens, local frame, or C.3
U.Kind; - two candidates overlap strongly;
- a name is convenient but the ontology may already be expressible through existing patterns.
Primary EntityOfConcern. The EntityOfConcern is the parsimony claim for one candidate ontology addition.
First useful move. Recover the candidate with E.24.UK or the subject pattern, then find the best current FPF expression for the exact receiving claim or use.
In this pattern, an existing governed expression is an admitted kind or dependent value, slot, current relation, record, publication form, lens, local frame, or direct-pattern claim whose direct owner defines its use.
What goes wrong if missed. FPF grows duplicate kinds for claims already carried by governed expressions. Later patterns then argue over words instead of recovering the EntityOfConcern, exact relation or slot, and admissible claim.
What this buys. A small ontology can still express rich project situations: the pattern either admits a new durable value with a boundary, or identifies the existing governed expression that already carries the claim.
Not this pattern when. The current question is only a local display name, publication title, naming taste, or ordinary glossary cleanup. Use the relevant Part F naming pattern unless the name is being asked to carry durable ontology.
Problem Frame
FPF needs enough primitives to be useful, but every new primitive creates learning cost, bridge cost, and future repair cost. Ontological parsimony is not anti-growth. It is the rule that FPF adds a new kind only when composition, reuse, dependent-value settlement, and subject patterns cannot express the action-facing claim without material loss.
When source or draft wording proposes a candidate durable value in U.* form, treat that as an admission claim. Apply A.11 after E.24.UK recovers the governed object and before naming patterns choose a public label.
For a relation-kind candidate, first apply A.6.P. If the participants are exact but no current direct relation expresses the named receiving claim, apply A.6.RCD. Stop when it returns an existing exact predicate, a local compound claim, or subject-bounded or reusable predicate-definition content. A derived relation-kind candidate continues only when a named use needs stable occurrence semantics and supplies its proposed direct settlement; an irreducible primitive relation-kind candidate continues only under A.6.RCD disposition 4.
Problem
A useful project word, slot-position label, publication form, diagram element, mathematical lens, or repeated source term can start acting like a durable FPF kind before the governed object and subject pattern are recovered. The problem is to decide whether the candidate preserves an action-facing distinction that the best existing governed expression cannot carry, or should remain that expression or a local name.
Forces
Solution
Use four gates before admitting the new ontology addition. Apply every gate to the same exact candidate, receiving claim or use, and current facts. Record the best existing expression first; then state the exact loss, overlap discriminator, newly admissible claim or action, and nearest excluded case for that same use.
Use this compact record:
Possible dispositions:
- retain as a root U-kind;
- retain as a dependent durable value under a root settlement;
- retain as a local C.3 kind or typed claim;
- express through an existing governed expression;
- keep as source wording or a local name;
- for a relation-kind candidate, stop at the exact
[A.6.RCD](/generated/patterns/A.6.RCD)existing-predicate, local-compound, subject-bounded-law, or reusable-predicate-definition result; - retain a derived relation-kind candidate only with the required occurrence semantics and direct settlement; or
- retain an irreducible primitive relation-kind candidate only when
[A.6.RCD](/generated/patterns/A.6.RCD)disposition 4 passes.
Archetypal Grounding - Maintenance
A retained addition also needs a reopen condition. Reopen or lower the admission when usage collapses, overlap with an existing value is discovered, an existing governed expression becomes adequate, the boundary becomes fuzzy, or the name starts hiding that expression. This is maintenance discipline, not a fixed calendar ritual.
Bias-Annotation
A.11 corrects kind-inflation bias. A useful word, field name, record label, or diagram element can start behaving like a universal kind because it appears often, feels important, or has prestige in a source tradition. The repair is ontological: recover the governed object and try the best existing governed expression before admitting a new durable value.
It also corrects false-parsimony bias. A compact ontology is not achieved by refusing every new value. If composition hides a reviewable distinction or blocks an action-facing claim, parsimony admits the new value and states its boundary, overlap, and reopen condition.
Conformance Checklist
Common Anti-Patterns and How to Avoid Them
- Slot label becomes kind. A system-role designation, transformation-participant label, source-maintenance position, carrier position, or boundary slot is renamed as if the label created a new universal kind; recover the admitted System, exact local system-role kind or direct relation, any separately obtaining assignment, and Work only when each is current.
- Publication form becomes ontology. A card, record, view, dashboard, figure, or report title is treated as the governed object instead of the episteme, relation, or carrier it publishes.
- Mathematical lens becomes object. A graph, tuple, algebra, metric, coordinate, or threshold is admitted as an ontology object without naming the EntityOfConcern and lens-use claim.
- Local project name becomes kernel vocabulary. A useful project label is promoted to durable FPF vocabulary before composition and direct-pattern expression are tried.
- Overlap is ignored. A candidate is admitted even though an existing pattern already carries the same claim with clearer boundaries.
- Parsimony as refusal. A new value is rejected because "fewer kinds is better" even though existing composition loses a distinction users need to claim, compare, repair, stop, or rely on.
Consequences
Rationale
Ontological parsimony preserves FPF's ability to handle many domains without turning every local distinction into a root object. The pattern follows the same discipline used by E.24.UK: recover the governed object first, then decide whether a new durable value is needed. Using an existing governed expression is a successful result, not a failure to mint a kind.
The practical criterion is not abstract minimalism. A candidate earns admission only when users gain a claim, comparison, repair, stop condition, reliance condition, or boundary they cannot recover by composition without material loss. That keeps parsimony tied to FPF's work-facing purpose rather than to a taste for small vocabularies.
SoTA-Echoing
Current ontology-engineering practice favors modularity, reuse, explicit competency questions, and controlled admission of new terms over unchecked class growth. A.11 adapts that practice to FPF: the admission question is not merely "can a class be defined?" but whether the candidate changes admissible claims, boundaries, or work-facing use inside the FPF pattern system.
Constructional-ontology and BORO-like source lines add a second discipline: identity, construction, dependency, and part-whole distinctions must be recovered before a convenient term becomes a kind. FPF keeps that source discipline without importing a classical top-level taxonomy as-is; U-kinds remain tied to accepted ontics, slot discipline, and action-facing pattern use.
Relations
- Builds on:
E.24.UK,A.6.P,A.6.RCD,A.8,C.3,F.8,F.18, and direct subject patterns. - Coordinates with:
E.24.CDfor candidate detection andE.24.PUBwhen a publication form or structural name created the admission claim. - Does not replace: universal-core testing in
A.8, typed claim quantification inC.3, or naming discipline in Part F.
A.11:End
Last Updated: 2026-09-10 — upstream FPF commit a87d0ef4 (github.com/ailev/FPF)