Parthood

Type: Operator / Relation

Working Definition: Parthood is the context- and time-indexed relation under which a thing x is part of a whole w under a disclosed unity basis. Proper Parthood is the same relation restricted to cases where x and w are distinct. A unity basis fixes what makes w one whole within a bounded modeling domain.

Function in the Grammar: Parthood states how a whole is structurally organized without collapsing that question into Containment, Membership, location, Representation, material constitution, or institutional classification. PART_OF is the generic, reflexive relation. PROPER_PART_OF excludes the case where x equals w. DIRECT_PART_OF marks a part at the modeling granularity immediately beneath w; a claim at that granularity does not by itself establish parthood at a coarser one. COMPOSES states when a set of parts forms the whole under the unity basis's coverage criterion. Every part of the whole must overlap some member of the set. Listing parts that are each individually part of the whole is not sufficient on its own. In an OWL implementation, binary relations such as properPartOf, locatedIn, memberOf, and represents can remain ordinary object properties. The six-argument PART_OF assertion itself generally requires reification, for example a ParthoodAssertion node carrying mode, unity basis, context, and modeling domain. Closed-world enforcement of required fields generally needs a constraint layer such as SHACL rather than OWL alone.

Formal Grammar Representation

<part-of-statement> ::= PART_OF(
  part: <part-thing>,
  whole: <whole-thing>,
  mode: <part-mode-set>,
  unity_basis: <unity-basis>,
  context: <context-id>,
  modeling_domain: <modeling-domain>
)

<part-thing> ::= physical-object | organism-component | document-segment
              | software-component | quantity-of-stuff
              | other identifiable individual admissible as a part

<whole-thing> ::= physical-object | organism | document
               | software-system | other identifiable individual
                 admissible as a whole

<part-mode-set> ::= "{" <part-mode> ("," <part-mode>)* "}"
<part-mode> ::= improper-part | component | portion | segment | ingredient
             | structural-part | functional-part | legal-instrument-part

Mode-typing constraint:
  improper-part IN M(x, w)  <=>  x = w
  -- reserved for the case where a whole is counted as a part of itself;
     it does not co-occur with a substantive mode, and no substantive
     mode applies when x = w

<unity-basis> ::= structural-principle | functional-principle
               | legal-instrument-principle | ConstitutiveRule
               | other disclosed principle fixing what makes
                 the whole one whole

<modeling-domain> ::= bounded regime fixing which unity bases and
                     part-thing / whole-thing types apply
                     (named modeling_domain, not domain, to avoid
                     collision with rdfs:domain in an OWL rendering)

PART_OF(x, w, M, u, c, d)

Derived relations:

PROPER_PART_OF(x, w, M, u, c, d) :=
  PART_OF(x, w, M, u, c, d) AND x != w

DIRECT_PART_OF(x, w, M, u, c, d) :=
  PROPER_PART_OF(x, w, M, u, c, d) AND
  NOT EXISTS y: PROPER_PART_OF(x, y, _, u, c, d)
              AND PROPER_PART_OF(y, w, _, u, c, d)

COMPOSES(parts: P, whole: w, unity_basis: u, context: c, modeling_domain: d) :=
  for every xi in P: PART_OF(xi, w, Mi, u, c, d)
  AND for every y such that PART_OF(y, w, _, u, c, d):
        OVERLAPS(y, P, u, c, d)

OVERLAPS(y, P, u, c, d) :=
  exists xi in P, exists z:
    PART_OF(z, y, _, u, c, d) AND PART_OF(z, xi, _, u, c, d)

Semantic Constraints

A PART_OF statement is syntactically well-formed when all six arguments are identifiable, the mode set is non-empty, and a unity basis is named. Whether the relation actually holds is a separate question the syntax cannot settle.

Every relation defined above shares one context argument c, rather than a distinct context per pair. Time is fixed as a function of that shared context alone:

t = RELEVANT_TIME(c)

PART_OF(x, w, M, u, c, d) holds iff:
  UNITY_BASIS_APPLIES(u, w, c, d, t)
  SATISFIES_PART_CRITERIA(x, w, M, u, c, d, t)

DIRECT_PART_OF, COMPOSES, and OVERLAPS each pass the same c into every nested PART_OF or PROPER_PART_OF call. Every claim inside those definitions therefore evaluates at the one snapshot t = RELEVANT_TIME(c). No separate time-equality premise is needed inside them.

An undisclosed unity basis leaves a claim incomplete rather than false. The claim cannot be evaluated until a basis is supplied. A disclosed basis that does not govern w, or whose criteria x fails to satisfy, leaves the claim well-formed but not holding. This is the same distinction Count-As draws between syntactic well-formedness and Validity.

SATISFIES_PART_CRITERIA is supplied by the unity basis for every substantive mode. This grammar does not decide that content, in the same way Count-As's SATISFIES(x, Conditions(g)) delegates its content to the governing basis g. The one exception is the improper-part mode. Its criterion is fixed by this grammar as x = x, always satisfied, rather than left to u.

Reflexivity, antisymmetry, and transitivity hold for PART_OF at the single shared time t = RELEVANT_TIME(c):

t = RELEVANT_TIME(c)

PART_OF(x, x, {improper-part}, u, c, d) iff
  UNITY_BASIS_APPLIES(u, x, c, d, t)                            [reflexivity]

PART_OF(x, w, M1, u, c, d) AND PART_OF(w, x, M2, u, c, d)
  AND same (u, c, d)
  =>  x = w                                                     [antisymmetry]

PART_OF(x, y, M1, u, c, d) AND PART_OF(y, w, M2, u, c, d)
  AND same (u, c, d)
  =>  EXISTS M3: PART_OF(x, w, M3, u, c, d)                     [transitivity]

Reflexivity requires only UNITY_BASIS_APPLIES, not a separate SATISFIES_PART_CRITERIA check. The improper-part criterion is x = x. It is trivially satisfied whenever the basis applies at all, so no additional condition can fail.

Antisymmetry's premises need not share a mode. Suppose both PART_OF(x, w, {structural-part}, u, c, d) and PART_OF(w, x, {functional-part}, u, c, d) are asserted under the same basis, context, domain, and time. Neither premise uses {improper-part}, so the mode-typing constraint already commits both premises to x != w. Antisymmetry then derives x = w from the same two premises. The premise set is jointly inconsistent, not sequentially retyped. No assignment can satisfy x != w and x = w at once. A modeler who reaches this pair of claims must reject at least one of them, rather than accept both and relabel the mode afterward.

Transitivity guarantees some mode set M3 for the coarser claim. It does not guarantee M3 = M1 or M3 = M2. The mode applicable at the coarser granularity may need its own check against SATISFIES_PART_CRITERIA.

This transitivity is restricted to a single, fixed unity basis, context, and modeling domain, all evaluated at the one shared time t = RELEVANT_TIME(c). It follows the classical mereological treatment Simons surveys and is not a departure from that tradition. Transitivity is not guaranteed across a change of unity basis or modeling domain. A DIRECT_PART_OF claim at one modeling granularity does not by itself establish PART_OF at a coarser one. SATISFIES_PART_CRITERIA must be checked again at the coarser level.

Formal Pattern

Direct-part pattern:
  t = RELEVANT_TIME(c)
  x != w
  UNITY_BASIS_APPLIES(u, w, c, d, t)
  SATISFIES_PART_CRITERIA(x, w, M, u, c, d, t)
  NOT EXISTS y: PROPER_PART_OF(x, y, _, u, c, d)
              AND PROPER_PART_OF(y, w, _, u, c, d)

  THEREFORE PROPER_PART_OF(x, w, M, u, c, d)
  THEREFORE DIRECT_PART_OF(x, w, M, u, c, d)
  THEREFORE PART_OF(x, w, M, u, c, d)

Fixed-basis chaining pattern:
  t = RELEVANT_TIME(c)
  PART_OF(x, y, M1, u, c, d)
  PART_OF(y, w, M2, u, c, d)
  same (u, c, d)
  -- t is shared automatically: both premises carry the same c,
     and t = RELEVANT_TIME(c)

  THEREFORE EXISTS M3: PART_OF(x, w, M3, u, c, d)

Composition pattern:
  for each xi in P: PART_OF(xi, w, Mi, u, c, d)
  for every y: PART_OF(y, w, _, u, c, d) => OVERLAPS(y, P, u, c, d)

  THEREFORE COMPOSES(P, w, u, c, d)

PART_OF is the primitive relation in this grammar. A claim may establish it directly by satisfying UNITY_BASIS_APPLIES and SATISFIES_PART_CRITERIA under any mode; the chaining pattern above is one sufficient route, not the only one. PROPER_PART_OF, DIRECT_PART_OF, and COMPOSES are fully defined by the := clauses in Formal Grammar Representation. The Direct-part and Composition patterns above restate those definitions in derivable form; they do not add alternative, weaker sufficient conditions beside the definitions.

Core Relations

Typical Questions

• What exact thing is claimed to be part of what whole?
• What mode set is claimed — component, portion, segment, ingredient, structural-part, functional-part, legal-instrument-part, or several of these together?
• What unity basis is disclosed, and does it actually govern this whole in this context and modeling domain?
• Are the part's criteria under that basis satisfied, or only asserted?
• Is the claim about generic Parthood, Proper Parthood, or a Direct Part at a stated granularity?
• Is a candidate composition claim supported by a coverage check, or does it only list some parts without showing every part of the whole overlaps the set?
• Is the candidate part instead merely contained in, located on, attached to, or a member of the candidate whole?
• Does a claimed chain of PART_OF relations hold the unity basis, context, and modeling domain fixed across every link?
• If the whole's part-structure changes, is a persistence claim about the same whole being made — and if so, does it belong to Identity or Continuity rather than to Parthood?

Examples

Distinctions

Generic Parthood ≠ Proper Parthood ≠ Direct Parthood. PART_OF is reflexive: a whole is a part of itself. PROPER_PART_OF excludes that case. DIRECT_PART_OF further restricts to the immediately subordinate granularity, with no intermediate part between x and w. A statement using the bare relation name without specifying which of these three is meant is underspecified.

Parthood ≠ Class subsumption. A particular bedroom instance may be part of a particular residence. Bedroom is not thereby a subclass of Residence. "Primary," used of one bedroom among several, is a role a part occupies in a given case, not a member of the part-mode set and not a class distinction.

Parthood ≠ Containment. A chair can be inside a room without being part of it; moving the chair changes nothing about the room's structure. A room may instead be a functional part of a residence: removing it changes the residence's structure, not merely its contents.

Parthood ≠ Membership, as a modeling commitment. A trustee may be a member of a board without being a component of it the way a wheel is a component of a bicycle. Whether some membership relations are themselves mereological is contested in the literature this entry draws on. This grammar stipulates, rather than settles, that Membership does not entail Parthood unless a stated unity basis says otherwise.

Parthood ≠ Location or attachment. A detached garage may be located on a parcel without being part of the residence built on that parcel. A utility connection may serve a building without being part of it. LOCATED_ON and SERVES do not by themselves satisfy SATISFIES_PART_CRITERIA for any unity basis.

Parthood ≠ Representation. An address designates a location; it does not stand in a part-whole relation to anything. A listing represents an asset or a bundle of rights; the listing is not part of the property it represents.

Parthood ≠ Material constitution, as a modeling commitment. Clay may constitute a statue without being a detachable component of it. This grammar treats material constitution and Parthood as disjoint relations. That disjointness is a modeling choice; the mereological status of constitution remains contested in the broader literature, and this entry does not resolve that debate.

Parthood ≠ Count-As, though the two may compose. PART_OF(Garage_1, Building_1, {structural-part}, ...) describes organization. COUNTS_AS(RaisedHand_1, Vote, ...) creates an institutional classification under a governing basis. The two are not the same relation. One may supply a condition the other needs. An incorporation-by-reference clause may need to count as making an exhibit part of an agreement. Only then can a legal-instrument-part claim hold.

Parthood ≠ Constitution. Constitution in this encyclopedia names an institutional governing instrument, not material constitution. The word "constitutes" names three distinct relations: material constitution, institutional constitution (Count-As and Constitutive Rule), and Parthood. Naming all three here is what keeps this entry from reintroducing the ambiguity it exists to resolve.

Common Failure Modes

Minimum Viable Test Case

w  = Building_2
x1 = Room_7
M1 = {functional-part}
u  = ResidentialLayoutCode_2024
c  = Context(Listing_9_Inspection)
d  = RealEstate

t = RELEVANT_TIME(c)

u states: a functional-part room is an enclosed, permanently roofed
space bounded by the building's exterior or load-bearing walls,
accessible from the building's interior circulation.

UNITY_BASIS_APPLIES(u, w, c, d, t)
SATISFIES_PART_CRITERIA(x1, w, M1, u, c, d, t)
  -- Room_7 is enclosed, permanently roofed, bounded by Building_2's
     exterior walls, and accessible from Building_2's interior
     circulation. All four stated conditions hold.

PART_OF(x1, w, M1, u, c, d)
DIRECT_PART_OF(x1, w, M1, u, c, d)
  -- no intermediate part sits between Room_7 and Building_2

Result:
  Room_7 is a direct part of Building_2 under the residential-layout
  unity basis. This does not establish that Room_7 sits inside
  Parcel_9 independently of Building_2, that the listing for
  Building_2 includes Room_7 as a represented feature, or that
  Building_2 remains the same whole if Room_7 is later removed.

Composition check:
  P = {Room_7, Room_8, Hallway_3}

  COMPOSES(P, w, u, c, d) holds only if every part of Building_2
  under u overlaps some member of P. If Building_2's foundation
  slab is a part of Building_2 under u and overlaps no member of P,
  COMPOSES(P, w, u, c, d) fails until the omission is corrected.

Failure case, disclosed basis, unsatisfied criterion:
  x2 = Garage_5
  LOCATED_ON(Garage_5, Parcel_9)

  PART_OF(x2, w, {structural-part}, u, c, d)

  UNITY_BASIS_APPLIES(u, w, c, d, t) holds; u governs Building_2 here.
  SATISFIES_PART_CRITERIA(x2, w, {structural-part}, u, c, d, t) fails;
  Garage_5 is detached and is not bounded by Building_2's exterior
  walls, so u's stated criterion is not met.

  The claim is syntactically well-formed and its basis is disclosed
  and applicable. It does not hold. The reason is an unsatisfied
  criterion, not an undisclosed basis.

Incompleteness case, undisclosed basis:
  PART_OF(Garage_5, Building_2, {structural-part}, unstated, c, d)

  This statement is incomplete rather than false. It cannot be
  evaluated until a unity basis is supplied.

Intellectual Lineage

Parthood formalizes the mereological relation studied by Stanisław Leśniewski (Podstawy ogólnej teoryi mnogości, 1916) and systematized in axiomatic form by Henry S. Leonard and Nelson Goodman, "The Calculus of Individuals and Its Uses" (1940). Peter Simons, Parts: A Study in Ontology (1987), surveys this tradition and its later developments, including time-relative parthood and restricted, non-universally-transitive part relations. Those developments belong to the mereological literature itself; this entry does not introduce them. Classical treatments typically hold generic parthood reflexive, antisymmetric, and transitive, with proper parthood as the distinct irreflexive restriction, and this entry follows that structure rather than departing from it. COMPOSES follows Leonard and Goodman's fusion-style composition, requiring that every part of the whole overlap some member of the composing set, rather than merely listing parts. What this entry adapts from Count-As, elsewhere in this encyclopedia, is not context- and time-indexed parthood, which the mereological tradition already supports, but the six-argument grammar form — thing, mode, unity basis, context, modeling domain — used to express it within this encyclopedia's shared operator vocabulary.

Cross-References

Required

• Context: supplies RELEVANT_TIME and the situational frame every PART_OF and COMPOSES claim is evaluated against
• Constitutive Rule: may supply or support a unity basis when the basis is itself rule-governed, as in a legal-instrument-part mode
• Count-As: the concept Parthood is most often mistaken for; also the relation PART_OF may compose with directly, as in the incorporated-exhibit example
• Constitution: named to resolve the three-way lexical collision around "constitutes" — material, institutional, and mereological

Consequential

• Identity: marks the persistence boundary Parthood does not settle; this encyclopedia's current entry does not yet adjudicate bare part-replacement persistence, a coverage gap this entry states rather than fills
• Continuity: marks the same boundary for institutional traceability across time
• Transformation: may result when a change in part-structure composes with a change in the whole's condition or standing

Evidential

• Validity: evaluates whether a disclosed unity basis actually governs and its criteria actually hold, the same well-formed/holds distinction Semantic Constraints draws for PART_OF
• Evidence: may support a contested SATISFIES_PART_CRITERIA claim, for example a survey, blueprint, or inspection record showing whether x meets w's stated criteria