Partitive Construction¶
A nominal construction that selects a member or portion in relation to an expressed set or entity whole through language-specific linkage.
Core Idea¶
A partitive construction is a nominal expression that selects a member or portion from an expressed set or entity treated as its whole. It combines a part-selecting expression, a whole-denoting nominal and a language-specific linking form. English one of these cats selects from a set; half of a cookie selects a portion of an entity. Latin pars militum (“part of the soldiers”) expresses the relation with a genitive whole instead of English of.[1][2]
This is a construction and interpretation, not one universal case ending or a rule that the whole must always carry a definite article. De Hoop's original analysis distinguishes set and entity partitives in English and explicitly accepts half of a cookie. A Finnish object can bear partitive case for negation or an unfinished situation without expressing this nominal whole–part construction.[1][3]
Structural Signature¶
- Selected part expression. A numeral, fractional expression or part-denoting noun identifies the member, subset or portion. Without a selected part, the phrase only names the whole.[1][2]
- Expressed whole nominal. A noun phrase identifies the set or entity from which the part is understood. In half of a cookie, a cookie is the entity whole even though it is indefinite; in pars militum, militum names the soldiers as a group.[1][2]
- Linking realization. A grammatical form joins selector and whole: English of or Latin genitive in these cases. The exponent varies by language and construction; the part–whole relation need not have one universal morpheme.[1][2]
- Part–whole interpretation. The selected referent is interpreted as a member or portion of the expressed whole. This semantic relation distinguishes the construction from a word merely bearing a similarly named case ending.[1][2][3]
De Hoop's English analysis further distinguishes a selector that takes a set argument from one that takes an entity argument. That distinction helps diagnose her data, but it is not imposed here as a universal semantic taxonomy of every language's partitives.[1]
What It Is Not¶
A partitive construction is not the same as partitive case. Official Finnish usage guidance gives partitive-marked objects in negative clauses and unfinished or open situations. Such an object may occur without an overt whole nominal and selected part. Its morphology alone does not establish the construction defined here.[3]
It is not restricted to a definite embedded nominal. Earlier descriptions of English of-partitives emphasized definiteness, but de Hoop treats half of a cookie as well formed and explains the entity/set distinction. Nor is every Latin genitive partitive: Allen and Greenough state a partitive-genitive rule for part words and also note ex/de alternatives in some partitive expressions. The form has to be read with its actual whole–part interpretation.[1][2]
The closest near miss is a Finnish partitive-case object licensed by negation alone: the grammatical label “partitive” is present, but no nominal part is selected from an overt whole. A bare amount with no expressed whole likewise does not supply this particular nominal construction.[3]
Scope of Application¶
In English nominal syntax and semantics, one of these cats illustrates selection of a member from an expressed set. Half of a cookie illustrates an entity portion, including an indefinite whole. De Hoop analyzes when the selector expects a set or an entity and shows why a blanket definiteness rule misses some cases. These are English constructional facts in her account, not a cross-language law.[1]
In Latin, Allen and Greenough §346 identifies part words followed by a genitive of the whole, including pars militum. Its morphology differs from English of, while the selected-part and expressed-whole roles remain present. The same grammar also permits ex/de in specified alternatives, so the genitive is not claimed as the only possible Latin exponent.[2]
Clarity¶
Identify the part expression, the whole nominal, the linker and the intended whole–part reading. Then ask whether the part selects individuals from a set or a portion of one entity. In English, one of these cats and half of a cookie are both partitives, yet the first requires a set argument for one, while the second permits an entity such as an indefinite cookie.[1]
Do not infer the construction from morphology alone. A Finnish partitive object under negation is a case choice in a clause, not automatically an overt nominal part–whole phrase. Conversely, Latin pars militum expresses the whole through genitive marking rather than a free English-style preposition.[3][2]
Manages Complexity¶
The four-role test compresses diverse surface forms into a common constructional question without erasing language-specific grammar. It separates (a) the part selected, (b) the set or entity whole, © the linker licensed by the language, and (d) the resulting interpretation. The test avoids treating every occurrence of of, genitive or partitive morphology as equivalent.[1][2][3]
Within English, the set/entity distinction narrows a misleading single rule. De Hoop's well-formed half of a cookie shows why a surface definiteness filter can reject a valid entity partitive. The analysis helps explain data only under its language- and determiner-specific hypotheses.[1]
Abstract Reasoning¶
Given a candidate nominal, locate the proposed part selector and overt whole. Determine whether the whole supplies a set of members or an entity with portions, then check that the selector can take that kind of argument. Verify the language's linking form and the intended part–whole interpretation. If one role is missing, do not classify the phrase merely because it contains an of-like word or partitive-labeled case.[1][2]
For English half of a cookie, half selects a portion of the entity denoted by a cookie; its indefiniteness is no objection to de Hoop's entity-partitive reading. For Latin pars militum, §346's part noun and genitive group supply a different morphological realization. A Finnish negative-clause object calls instead for a case-licensing explanation unless an independent whole–part nominal is present.[1][2][3]
Knowledge Transfer¶
The role map transfers between the English prepositional and Latin genitive cases: both express a selected part relative to a named whole. The exact preposition, case form and semantic selection conditions do not transfer automatically. A source for one language cannot be used to decree the other's morphology or grammaticality.[1][2]
Other uses of “partitive,” especially Finnish object case, require a fresh constructional test. The wider part–whole relation is conceptually portable, but this named nominal construction remains a linguistic form-and-interpretation pattern. No independent nonlinguistic instance of the named construction is asserted.[3]
Examples¶
English entity partitive: half of a cookie¶
In de Hoop's attested example, half selects a portion, a cookie supplies the whole entity, and of links them. The indefinite article does not invalidate this reading. It is an entity partitive in her English analysis, not an example of selecting one member from a plural set.[1]
Mapped back: selected part expression → half; expressed whole nominal → a cookie; linking realization → of; part–whole interpretation → half of the cookie entity.
Latin partitive genitive: pars militum¶
Allen and Greenough §346 gives pars militum, “part of the soldiers.” Pars denotes the part and genitive plural militum denotes the group to which that part belongs. The Latin genitive realizes the link; the expression does not need English of.[2]
Mapped back: selected part expression → pars; expressed whole nominal → militum; linking realization → Latin genitive of the whole; part–whole interpretation → a part drawn from the expressed soldiers group.
Structural Tensions¶
No intrinsic two-objective trade-off defines this construction. Set and entity partitives are semantic variants in de Hoop's English account, not goals that compete within one expression. Linguists may compare rival constraints on their distribution, but that theoretical dispute is not itself a structural tension required for nominal partitive membership.[1]
Structural–Framed Character¶
The entry is structural within a language-specific frame. Evaluative weight: the part–whole reading is an interpretive/grammatical fact, not a quality score. Human-practice dependence: speakers' variety-specific forms and interpretations matter; an analyst does not confer grammaticality by choosing a label. Institutional origin: grammar books describe usage and research papers propose analyses, but neither institution creates all partitive constructions. Vocabulary travel: “partitive” names a case, a genitive use and nominal constructions, so the word must be typed. Import versus recognition: recognize a part selector, expressed whole, linker and reading in the language at issue before borrowing a rule from another. Its character: a recurrent linguistic construction with variable exponents and a domain-bound whole–part interpretation; the English/Latin comparison does not establish a substrate-neutral Prime.[1][2][3]
Structural Core vs. Domain Accent¶
The core is a selected nominal part related to an overt set or entity whole through grammatical linkage. English of and Latin genitive are language accents; the part/whole roles and licensed combination are indispensable. De Hoop's set/entity analysis sharpens the English cases but is not part of a universal morphology claimed for every language.[1][2]
The live Grammar entry is an independent prerequisite: remove a natural-language system of forms and combinations and the nominal construction is undefined, while Grammar has many patterns without partitives. Its inherited Prime System ancestor supplies a portable organized-pattern idea for Grammar, not a taxonomic genus that turns this one phrase into a system. A more general part–whole Prime is a separate future question; this entry has only linguistic positive cases and does not clear that bar. Grammatical Case cannot subsume English of, and Mereology studies formal part–whole relations rather than this linguistic form.
Instantiates / Related Primes¶
This entry presupposes Grammar.
The reviewed graph records strict composition/presupposes to Grammar. The selector, whole and linking form require a variety-specific grammar; Grammar can exist without this partitive pattern and is not a parent genus of a single nominal phrase. There is no subsumption edge asserted here.[1][2]
Grammatical Case is relevant to Latin and Finnish morphology but does not cover the English prepositional instance. Grammatical Relation concerns syntactic functions such as subject or object, not this phrase identity. Grammatical Category is a feature-value dimension. Mereology concerns formal part–whole theory. Prime Relation is a broad comparison about relata, not the nearest object-level genus of a nominal construction.
Relationships to Other Abstractions¶
Current abstraction Partitive Construction Domain-specific
Parents (1) — more general patterns this builds on
-
Partitive Construction presupposes Grammar Domain-specific
A partitive nominal construction presupposes a language's forms and licensed combinations for expressing its part–whole relation.Without a natural-language grammar, the part selector, expressed whole and English prepositional or Latin genitive linkage have no licensed linguistic form. Grammar exists without partitives and is a whole variety-specific system, not a taxonomic genus of one construction. The edge is a strict prerequisite rather than subsumption.
Hierarchy path (1) — routes to 1 parentless root
- Partitive Construction → Grammar → System → Composition → Gestalt Principles → Holism
Neighborhood in Abstraction Space¶
Partitive Construction sits in a sparse region of the domain-specific corpus (100th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Grammatical Case & Reference (10 abstractions)
Nearest neighbors
- Adjunct (Grammar) — 0.76
- Absolute Construction — 0.75
- Fallacy of Division — 0.73
- Direct Product — 0.73
- Fallacy of Composition — 0.73
Computed from structural-signature embeddings · 2026-10-08
Not to Be Confused With¶
Finnish partitive object case: may be licensed by negation or open aspect without an overt selected part and whole. A universal definiteness constraint: rejected by de Hoop's half of a cookie. Every Latin genitive: only a genitive of a whole after an appropriate part expression bears the cited reading. A bare measure or amount: no expressed whole–part relation has been established. Mereological part–whole relation: a semantic component, not the language-specific nominal construction itself.[1][2][3]
References¶
[1] Helen de Hoop, “A Semantic Reanalysis of the Partitive Constraint”, Lingua 103 (1997): 151–174, especially examples (1), (10)–(13) and the set/entity discussion at printed pp. 160–161. Original research article; its English semantic analysis is not generalized here as a universal rule. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s ↩t ↩u
[2] J. H. Allen and J. B. Greenough, “Allen and Greenough’s New Latin Grammar for Schools and Colleges” (1903), §346 “Partitive Genitive,” opening rule, §346a.1 pars militum and §346c alternatives. Online transcription of the original grammar. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q
[3] Institute for the Languages of Finland, “Objekti hän luki kirjaa, kirjan milloin partitiivi, milloin ei?”, original page title “Objekti: hän luki kirjaa, kirjan (milloin partitiivi, milloin ei?)”; “Ohje lyhyesti,” negation and unfinished/open-situation sections. Official usage guidance. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j