Head-Marked Dependency Construction¶

A grammatical construction whose overt relational marker belongs syntactically to the head side of a dependency.

Version: v1 · 2026-10-25 · History
Domain-specific #: 14397
Domain group: Humanities
Origin domain: Linguistics & Semiotics
Subdomains: Morphosyntax, Linguistic Typology → Linguistics & Semiotics

Core Idea¶

A head-marked dependency construction places overt morphosyntactic marking of a grammatical relation on its syntactic head side.^[1]^[2] In a clause, the verb is the head relative to its arguments under the analysis used by Nichols and Bickel; in a possessive noun phrase, the possessed noun is the head relative to the possessor.^[1]^[2] The test concerns one construction and relation at a time. A language with one head-marked construction need not qualify as consistently head-marking as a whole language.^[3]

The marker's syntactic role matters more than its visible position. It may be an affix or a separate grammatical word.^[1]^[2] Tiwi has a separate uninflected pronoun marking possession that stays with the possessed noun under word-order inversion, so WALS treats it as head marking.^[2] Conversely, WALS treats a case-inflected Bororo dependent as dependent marking even when it cliticizes toward the head.^[2] Neither adjacency nor phonological attachment alone settles locus.

Structural Signature¶

Sig role-phrases: identified grammatical dependency → syntactic head and dependent → overt relational formative → head-side syntactic locus → construction-level classification.

Dependency and parse. Identify the relation within a particular phrase and which member is head. Without that analysis, “head side” has no testable referent. WALS explicitly identifies the verb as clause head and the possessed noun as possessive-phrase head for its surveys.^[1]^[2]
Overt relational marker. The formative must reflect the relation between the members. An affix, clitic, or qualifying separate word can do so, but unmarked order alone is not overt head marking. WALS excludes word order from its clause-locus classification.^[1]
Head-side syntactic locus. Determine whether the marker belongs to the head-side expression. A possessed-noun agreement form can qualify; a genitive on the possessor is dependent marking. A case-inflected dependent that moves next to the head does not change syntactic side merely through proximity.^[2]
Local scope. The result classifies the tested construction. Whole-language type requires a further selection and combination of relations, and language-specific alternative constructions can differ.^[2]^[3]

What It Is Not¶

Agreement is one frequent realization, not the complete identity. WALS gives Fijian possessive marking on the possessed noun that does not agree in the possessor's person or number.^[2] Double marking has relational marking on both head and dependent; Southern Sierra Miwok possession is the source's example.^[2] Dependent marking locates the relevant marking on the non-head; Chechen genitive on a possessor illustrates it.^[2]

The article does not say that any word next to a head is a head marker. WALS gives Bororo as a headward-migrated but syntactically dependent case form, and Chamorro as a floating, phrase-position marker.^[2] The same caution applies to its clause examples of floating markers.^[1]

Scope of Application¶

The entry applies when the analyst can identify a head-dependent relation and an overt marker of that relation in a bounded construction. It covers a verb with argument-relation marking in a clause and a possessed noun with possession marking in a noun phrase.^[1]^[2] It also covers a separate-word marker assigned syntactically to the head, as in Tiwi, rather than requiring every exponent to be inflection on the lexical head.^[2]

WALS's whole-language typology samples the default or least restricted possessive construction and direct or primary object marking in a transitive clause, then combines those two cardinal results. “Consistently head-marking” in that survey means both are head-marked.^[3] That survey rule must not be read back into every local case or treated as a universal classification of every construction in a language.

Clarity¶

For a proposed instance, ask four questions in order: What relation is marked? Which member is the syntactic head under the declared analysis? Where does the overt relational formative belong syntactically? Does the proposed classification apply only to this construction? The last question prevents an agreement suffix on one verb from being turned into an unsupported whole-language label.

A useful near miss is Bororo possession in WALS: its case-inflected pronominal dependent can cliticize to the head, but the source classifies it as headward-migrated dependent marking.^[2] Tiwi contrasts with it because its separate uninflected possessive formative immediately precedes the possessed noun in both orders of possessor and head, evidence that the formative belongs to the head side.^[2]

Manages Complexity¶

Grammatical relations can be expressed through agreement, case, a possessive affix, a separate formative, or no overt marker. Classifying by syntactic locus keeps these realization types comparable without pretending they are identical. WALS's clause and possessive surveys use the same head/dependent/both/neither distinctions while defining the relevant relation and sampling conventions for each construction.^[1]^[2]

The local test also prevents mixed patterns across a language from being hidden. A language may mark one construction on its head and another on its dependent. WALS therefore builds its whole-language type only after testing its two chosen cardinal relations.^[3]

Abstract Reasoning¶

The reusable inference is about locus, not a specific affix: once the relation and syntactic roles are established, classify the overt formative by which side it belongs to. Changing from a clause verb to a possessed noun changes the carrier but retains the test. Changing only surface order need not change the result, as the Tiwi inverted pair shows.^[2]

This is a defeasible analysis: a better account of the formative's syntactic function may change the classification. The Bororo example shows why a phonological-host shortcut can fail. The proposed locus must be supported by grammatical evidence rather than inferred from a diagram or adjacency alone.^[2]

Knowledge Transfer¶

The clause and possessive cases transfer the same question across unlike constructions: is an overt marker of the dependency assigned to its head? Tzutujil expresses subject/object relations on a verb; Acoma expresses possession on the possessed noun.^[1]^[2] The roles remain comparable even though the relations, languages, heads and marker forms differ.

Transfer to a whole-language claim is conditional. WALS's language-level measure is a further combination of two specific default relations; it is not the result of renaming this local pattern.^[3] A generic dependence relation outside linguistics also lacks the morphosyntactic marker and syntactic head test that make this entry literal.

Examples¶

Canonical: Tzutujil transitive clause¶

Nichols and Bickel show a Tzutujil clause in which the verb agrees with subject and object while the displayed words for boys and rock have no overt markers of those clause functions.^[1] They accordingly locate the marking of these argument relations on the verb, the clause head.^[1] Mapped back: transitive clause → verb head and noun-argument dependents → overt subject/object agreement on the verb → head-side locus for the displayed relations. The example does not by itself prove the marking of every Tzutujil construction.

Applied: Acoma possessive phrase¶

WALS gives an Acoma phrase glossed as a brother's house.^[2] The possessed noun has a person/number form agreeing with the possessor noun; the outer possession relation is marked on the possessed head.^[2] The possessor noun's own internal person morphology does not make it a genitive marker of the outer dependency.^[2] Mapped back: possessive phrase → possessed-noun head and possessor dependent → overt relational agreement on the possessed noun → head-side locus. This example changes the syntactic construction and marker carrier from the Tzutujil clause.

Scope check: Tiwi separate-word possessive marker¶

WALS places a Tiwi uninflected possessive pronoun with the possessed noun in two opposite orders of possessor and head.^[2] That syntactic association supports head marking even though the marker is a separate word.^[2] Mapped back: possessive dependency → possessed head → separate relational formative staying with that head → head-side locus. Tiwi checks the identity's realization boundary; it is not a third distinct construction family.

Structural Tensions¶

The source material supplies classification boundaries rather than an inherent two-sided optimization tension. A clause and a possessive phrase can be compared through a common locus rule while keeping their different syntactic relations and sampling choices explicit. Broadening the rule to every physical head-adjacent marker would misclassify Bororo; narrowing it to inflection would miss Tiwi.^[2] The disciplined middle is a syntactic-locus test with stated source evidence.

WALS's whole-language aggregation gains a surveyable comparison by selecting two cardinal relations, but a local construction analysis retains details that a single language label cannot report.^[3] These are different levels of description, not evidence that one local test can substitute for the other.

Structural–Framed Character¶

Vocabulary travels within linguistics: the head/dependent distinction and marker-locus question work in clauses and possessive phrases, including bound and separate formatives. Evaluative weight is low: head marking describes where a grammatical relation is marked; it does not rank it above dependent, double or zero marking. Institutional origin is limited: WALS authors choose the two sampled relations and default-construction convention for a comparative map, but their survey does not create the Tiwi or Acoma syntactic pattern. Human practice is constitutive of the carrier: speakers use grammatical constructions and analysts justify a head/dependent parse from language evidence, so the test cannot be detached from linguistic practice. Import versus recognition: applying the same syntactic-locus test to another attested clause or phrase recognizes a literal instance; calling a sensor on the “head” of a machine head marking would import a metaphor without a grammatical formative. Its character: a reusable linguistic locus pattern whose literal instances require construction-specific grammatical evidence and whose whole-language extension has a separate sampling rule.

Structural Core vs. Domain Accent¶

The portable skeleton is a marker assigned to one side of a directed relation. The linguistic differentia is essential here: syntactic head versus dependent, an overt morphosyntactic formative, and a particular grammatical construction. Removing these would turn the entry into generic relation or dependency language and erase the Tiwi/Bororo distinction. The Tzutujil/Acoma pair establishes transfer within linguistics, not a substrate-neutral Prime. A broader side-located relational signal might be a future Prime only if unlike nonlinguistic carriers support the same role and inference test. No such cross-domain evidence is in this packet, and no current Prime is assigned as its owner.

The live Agreement (linguistics) entry is a neighbor because agreement realizes some head-marked cases, but Fijian nonagreeing marking and Tiwi's separate formative defeat an exact-ownership claim.^[2] The live Double-Marking Language entry names a both-loci and language-level pattern, not this local head-side classification. The reviewed typed parent is Grammatical Relation by composition/presupposes: the marker-location test needs an analyzed syntactic relation, while grammatical relations can exist with dependent, double or zero marking. Graph validation and this post-edit text remain pending.

This entry presupposes Grammatical Relation.

The linguistic head/dependent analysis is related to the live Dependency Grammar and Grammatical Relation entries. Their roles help frame the test, but neither alone asserts the location of overt relational marking. Prime Dependency is a broad directed-reliance concept, and Prime Markedness concerns default versus marked contrasts; this article does not claim that every grammatical head-marked instance satisfies either Prime's full signature. Grammatical Relation is a prerequisite, because head-side location is undefined without a relation between head and dependent; this is not a claim that the local construction is a subtype of all grammatical relations.

Relationships to Other Abstractions¶

Current abstraction Head-Marked Dependency Construction Domain-specific

Parents (1) — outgoing typed relations

Head-Marked Dependency Construction presupposes Grammatical Relation Domain-specific

Locating a grammatical relation's overt marker on its syntactic head presupposes an analyzed head-dependent relation.

Condition / exception Requires an identified grammatical head-dependent relation and overt relational formative; surface adjacency or phonological host alone cannot establish head locus.

Hierarchy path (1) — routes to 1 parentless root

Head-Marked Dependency Construction → presupposes Grammatical Relation → presupposes Relation

Neighborhood in Abstraction Space¶

Head-Marked Dependency Construction sits in a moderately populated region (55^th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Formal Grammar & Syntactic Structure (18 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25

Not to Be Confused With¶

Head-marking language. The frozen title is a language-level typology. This article describes a local construction; the original member still has whole-language, English-analysis, geography and history branches open.
Agreement. Agreement can realize head marking, but nonagreeing head marking exists.
Dependent or double marking. A marker on the dependent alone, or on both members, has a different locus profile.
Headward migration. A dependent's case marker remains dependent marking merely after cliticizing toward a head.
Floating marking. A formative positioned by phrase edge or another word does not automatically have head locus.

References¶

[1] Johanna Nichols and Balthasar Bickel, “Locus of Marking in the Clause,” The World Atlas of Language Structures Online, §§1 and 1.1–1.5, especially Tzutujil example (1), separate-word formatives in §1.4 and floating marking in §1.5. https://wals.info/chapter/23 registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k

[2] Johanna Nichols and Balthasar Bickel, “Locus of Marking in Possessive Noun Phrases,” The World Atlas of Language Structures Online, §§1.1–1.5, especially Acoma example (1), Fijian (2), Chechen (4), Southern Sierra Miwok (5), Tiwi (8), Chamorro (10) and Bororo (12). https://wals.info/chapter/24 registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s ↩t ↩u ↩v ↩w ↩x ↩y ↩z ↩27

[3] Johanna Nichols and Balthasar Bickel, “Locus of Marking: Whole-language Typology,” The World Atlas of Language Structures Online, §1 definition and Tzutujil example (1). https://wals.info/chapter/25 registry ↩a ↩b ↩c ↩d ↩e ↩f