Implicational hierarchy¶

A linguistic ranking in which the presence of a higher property predicts the presence of lower prerequisites in a defined comparison population.

Version: v1 · 2026-10-25 · History
Domain-specific #: 14401
Origin domain: Linguistics
Subdomains: Linguistic Typology, Syntax, Morphology → Linguistics
Aliases: Linguistic Implicational Hierarchy

Core Idea¶

A linguistic implicational hierarchy ranks coded properties in a defined comparison population. The proposed rule says that if a language or paradigm has a higher-ranked property, it also has the lower properties beneath it. If the properties are ordered A < B < C < D, the combinations admitted by an exact chain are the lower prefixes: none, A, A+B, A+B+C, and A+B+C+D. A case with C but without B would be a gap against that proposal. The ranking has empirical content only after one specifies what counts as the same case, how each property is coded, and whether the proposal is absolute or a tendency.^[1]^[2]

This relation is used in unlike parts of linguistic typology. A primary relative-clause strategy may directly relativize only subjects, or extend down an ordered list of noun-phrase functions.^[1] Grammatical-number paradigms may distinguish singular and plural, add a dual, or further add a trial.^[2] These are not the same linguistic objects, but both can be assessed by a directed prerequisite ordering. The number proposal has known qualifications; drawing a logically exact matrix does not turn every observed typological pattern into an exceptionless law.^[2]

Structural Signature¶

Sig role-phrases: comparison population — ranked properties — fixed feature coding — directional prerequisite rule — gap assessment. The rule's strength, absolute or statistical, must also be stated.

Comparison population. Cases must be comparable at a declared grain: a language and its primary relative-clause strategy, or a specified number paradigm. A language's several relativization strategies cannot be mixed into one fictitious row.^[1] Larike subject prefixes cannot be silently combined with its free pronouns.^[2]
Ranked properties. Properties are ordered by the claimed prerequisite relation. In the accessibility case they are NP functions such as subject and direct object; in the number case they are available number distinctions.^[1]^[2] The arrow points from the more restrictive property to what it presupposes, not from historical earlier to later.
Feature coding. An explicit test says when a property is present. Keenan and Comrie distinguish direct object relativization from passive promotion followed by subject relativization.^[1] Malouf and colleagues print separately labeled subject-prefix cells for singular, dual, trial and plural.^[2]
Directional implication. Higher presence predicts lower presence. This is what excludes arbitrary feature combinations. The relation is asymmetric: availability of a plural does not, by itself, predict a dual.^[2]
Gap assessment. An observed higher property lacking a lower one tests the proposed hierarchy. For an absolute proposal one verified gap is a counterexample; for a robust tendency, exceptions require an explicit rate and scope rather than being silently erased.^[1]^[2]

What It Is Not¶

A mere ordered list of features is not an implicational hierarchy. The order must constrain which combinations should occur. Nor is a collection of frequent co-occurrences enough until direction and coding are declared: plural is more common than dual, but the claimed relation is specifically dual implies plural, not just a frequency ranking.^[2]

The same diagram does not establish a causal developmental path. Malouf and colleagues investigate one learning-based explanation of the number hierarchy; the hierarchy is the empirical generalization to be explained, not that learning model itself.^[2] A formal chain outside linguistics may instantiate a more general Prime-level ordering or implication relation. This named entry concerns empirical linguistic properties and the practice of checking language inventories.

An individual language may use more than one relative-clause strategy.^[1] Keenan and Comrie's subject-anchored lower-prefix constraint concerns a primary strategy; their secondary strategies can occupy a different segment of the Accessibility Hierarchy.^[1] Passive promotion can make an object-relative meaning expressible without direct object relativization by the primary strategy.^[1] Treating all these routes as one language-wide binary capability would change the case being tested.

Scope of Application¶

The hierarchy applies to cross-language syntactic comparisons when the ranked items are grammatical functions and the observation unit is a specified relative-clause strategy. Malagasy's primary strategy illustrates a subject-only prefix; Keenan and Comrie also present strategies reaching lower NP functions.^[1] The same structural test applies to morphology when the ranked items are available singular, plural, dual and trial distinctions within a comparable paradigm. Malouf and colleagues discuss English, Upper Sorbian and Larike as contextual number systems drawn from different grammatical domains; they are not three rows of one fixed-paradigm comparison.^[2] Their Larike human second-person subject-prefix row supplies the four-way worked paradigm.^[2] A cross-language inventory test must declare a common comparison category and coding rule before treating such systems as comparable rows.

Other proposed phonological or lexical hierarchies need their own feature definitions and evidence. The frozen source asserts an exceptionless nasal-phoneme chain, but Maddieson's original inventory survey reports a strong, nonabsolute association and discusses classification-sensitive Irish nasals.^[3] The basic color-term proposal is a nearby branching progression: a four-term inventory may have green or yellow, so that full case is not one total linear lower-prefix sequence.^[4] Those examples mark where this article's simple chain stops; neither is smuggled in as a third confirmed instance.

Clarity¶

The hierarchy makes three distinctions visible. First, an arrow is a conditional claim, not a statement that adjacent properties are the same thing. Second, an allowed inventory is not a guarantee that a language of that type exists; the matrix says which combinations the proposal permits. Third, a counterexample must be judged against the same coding rule and population that produced the ranking. A gap manufactured by switching from primary to secondary relativization, or from subject prefixes to free pronouns, tests a different assertion.^[1]^[2]

The notation A < B also risks hiding direction. Here B requires A; it does not mean A causes B or that every A-case proceeds to B. A subject-only relative-clause strategy is therefore a complete permitted case, not an incomplete language.

Manages Complexity¶

With four binary features there are sixteen possible on/off vectors. A strict four-step prerequisite chain admits only five lower prefixes. One can therefore summarize a potentially large table of language-by-feature observations with an order, a coding rule, and a list of gaps. This compression is useful because it exposes the exact disputed cells: whether an apparent object relative is direct or promoted, or whether a trial label in a reported paradigm denotes a true trial rather than a paucal.^[1]^[2]

Compression has a cost. A single yes/no row can conceal optional forms, overlapping strategies, heterogeneous grammatical domains and typological exceptions. The article should return to the underlying paradigm or construction whenever one of those distinctions affects an implication.^[1]^[2]

Abstract Reasoning¶

Given a newly documented paradigm, first fix the observation unit and code each proposed property. Then read the vector against the declared order. If the vector is a lower prefix, it is compatible with the hierarchy; that alone does not prove the hierarchy over the population. If a higher property lacks a lower prerequisite, inspect the source and coding. A real gap either falsifies an absolute version or becomes an exception that a tendency must count and explain.^[1]^[2]

In the number example, a documented true trial without dual would challenge the trial⇒dual step.^[2] A paucal by itself would not, because it is a different number category.^[5] In the syntactic example, an apparent object-relative clause must be checked for passive promotion before it is entered as direct object relativization by a primary strategy.^[1]

Knowledge Transfer¶

The literal procedure transfers between linguistic subfields: declare features, fix the unit of comparison, rank by proposed prerequisites, and inspect observed gaps. Syntax and morphology fill those roles with different constructions and data. The portable logic of a chain appears elsewhere, but using the linguistic term there would be an analogy unless the question still concerns language properties and typological evidence. Transfer also requires resetting the empirical strength; a well-supported implication in one subfield does not lend its evidential status to another.^[1]^[2]

Examples¶

Canonical: primary relative-clause strategy¶

Keenan and Comrie's Malagasy examples contrast a grammatical subject relative with a starred attempt at direct object relativization.^[1] A passive construction can promote the object to subject and then form a subject relative; that does not extend direct object relativization by the same primary strategy.^[1] Thus subject-only is an allowed prefix for this strategy.

Mapped back: The population is relative-clause strategies compared across languages; the ranked properties are direct relativizability of NP functions; the coding rule separates direct subject/object relatives from promotion; the directional rule predicts that a primary strategy reaching direct object also reaches subject; the starred Malagasy object construction is a boundary observation, not a gap against its subject-only prefix.^[1]

Applied: Larike grammatical-number agreement¶

Malouf, Ackerman and Seyfarth present a Larike human subject-agreement-prefix table with separate singular, dual, trial and plural columns, including a second-person row.^[2] They discuss English and Upper Sorbian as contextual number systems in different grammatical domains.^[2] The specified Larike row fills the permitted four-feature inventory.^[2] The authors say the broader number hierarchy is robust but not quite exceptionless.^[2]

Mapped back: The worked unit is Larike's human second-person subject-prefix paradigm; the properties are availability of singular, plural, dual and trial in that row; the proposed implication is trial⇒dual⇒plural; the gap test for a broader cross-language claim asks whether a separately documented, comparably coded true trial occurs without dual.^[2] It does not imply a morpheme-count inequality or mandatory use of the most specific form in each utterance.

Structural Tensions¶

T1: Exact chain vs empirical exceptions. An exact lower-prefix matrix is easy to test and explain, but the number hierarchy is described as not quite exceptionless. Weakening every violation away makes the claim empty; insisting that a tendency has no exceptions falsifies the source's account. Diagnostic: Is this article making an absolute universal or a measured tendency, and where are its verified gaps recorded?^[2]

T2: Simple feature rows vs faithful coding. A small yes/no matrix makes a large survey tractable, yet secondary relative-clause strategies and passive promotion change what “object relativization” means. Coding them together creates false support or false violations. Diagnostic: Can each positive cell be reproduced from one declared strategy and construction?^[1]

T3: General rule vs particular hierarchy. The reusable implication test helps compare syntax and morphology, while a named Accessibility Hierarchy or Number Hierarchy supplies its own ranked items and evidential limits. Collapsing them loses the particulars; treating them as unrelated loses the shared reasoning form. Diagnostic: Are the roles mapped without borrowing one field's feature definitions or evidence for the other?^[1]^[2]

T4: Domain autonomy vs Prime reduction. The exact prefix rule resembles general hierarchy and order, but this entry also requires linguistic feature coding, cross-language comparison and grammatical counterexamples. Reducing it to a generic chain loses the subject/strategy and number/paradigm questions; claiming a wholly separate logic would duplicate the more general order relation. Diagnostic: Does a proposed typed parent genuinely cover all instances and necessary roles, or merely share the word “hierarchy”?

Structural–Framed Character¶

The entry is structural within linguistics: the relation is not an evaluation that one language is better than another. Its ranking is a testable claim about feature availability, and the possible-prefix matrix follows from that claim. Yet the observation unit, feature boundaries and significance of an exception depend on linguistic analysis and human descriptive practice. A researcher must decide whether a relative clause is direct or promoted and whether a reported number form is a true trial; those decisions are not supplied by the abstract matrix.^[1]^[2]

In this article, the term is used for linguistic typology. Its vocabulary can travel as formal implication; the named linguistic entry is recognized only when empirical language properties populate the roles. Importing the same diagram into an unrelated field carries the Prime-level skeleton, not evidence that the linguistic hierarchy itself governs that field. Its character: a structural domain abstraction whose mathematical form is portable, while its identity and truth conditions remain framed by linguistic comparison.

Structural Core vs. Domain Accent¶

The skeletal relation is an ordered set of directional prerequisites and a test for higher-without-lower gaps. That skeleton is intelligible without language data and overlaps live Prime concepts such as Hierarchy and Order. The domain accent is not decorative terminology: linguistic categories must be coded, strategies and paradigms held fixed, and cross-language evidence assessed. A proposed hierarchy may be a robust typological tendency rather than a mathematical law.^[1]^[2]

The named entry therefore does not clear a new Prime bar by pointing to syntax and morphology alone; those are unlike settings within linguistics. Removing the linguistic feature and evidence frame leaves an asymmetric hierarchy of prerequisite levels. A distinct Prime would require nonlinguistic cases and a duplicate test beyond that existing core. The curated typed DAG edge is decompose to Hierarchy; independent review and an isolated official graph build confirmed the placement.

This entry is a decomposition of Hierarchy.

The live Linguistic Universals Prime concerns broad cross-language claims and their explanation; a preliminary implicational hierarchy need not yet meet its full evidence and explanatory signature. The curated strict decompose edge to Hierarchy retains ranked levels and directed cross-level constraint when linguistic coding is removed. Order and Network are inherited through that Prime's existing edges rather than repeated here. An isolated official graph build confirmed the placement without a cycle. The entry does not instantiate Markedness merely because higher-ranked distinctions are rarer: markedness has its own default and contrast claims, and frequency alone does not prove them.

Relationships to Other Abstractions¶

Current abstraction Implicational hierarchy Domain-specific

Parents (1) — outgoing typed relations

Implicational hierarchy is a decomposition of Hierarchy Prime

A linguistic implicational hierarchy retains ranked levels and asymmetric cross-level prerequisites when its linguistic frame is removed.

Hierarchy paths (4) — routes to 4 parentless roots

Implicational hierarchy → is a decomposition of Hierarchy → is a kind of Network

Show alternative paths (3)

Neighborhood in Abstraction Space¶

Implicational hierarchy sits in a sparse region of the domain-specific corpus (65^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (2769 abstractions)

Nearest neighbors

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

Not to Be Confused With¶

Noun Phrase Accessibility Hierarchy is one concrete syntactic ranking.^[1] Greenberg's Number Hierarchy is a morphological ranking.^[2] Neither is the general linguistic relation being defined here. Basic color-term progression contains a green/yellow branch and cannot be substituted for one total chain.^[4] Nasal-inventory implication needs source-specific feature classes and exception accounting; the frozen article's categorical matrix is source-held.^[3] Causal acquisition models may explain a hierarchy but are not the ranked observation itself.^[2]

References¶

[1] Edward L. Keenan and Bernard Comrie (1977). “Noun Phrase Accessibility and Universal Grammar,” Linguistic Inquiry 8(1), 63–99, especially §§1.2.2 and 1.3.1, pp. 67–70, examples (10a–e). https://www.jstor.org/stable/4177973 . Publisher identity checked; original-paper text inspected through a nonpublisher OCR reproduction at https://www.scribd.com/document/346593223/Noun-Phrase-Accessibility-and-Universal-Grammar . Earlier authors' LSA conference abstract (1972), printed pp. 190–191, indexed at https://www.lsadc.org/files/1972_searchable.pdf; whole PDF too large to open in review tool. 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

[2] Robert Malouf, Farrell Ackerman and Scott Seyfarth (2015). “Explaining the Number Hierarchy,” printed p. 1500 and footnote 1. https://scottseyfarth.com/docs/MaloufAckermanSeyfarth.pdf . First-party research paper inspected in full; Larike human subject-prefix table is reported from Laidig (1993) and Siewierska (2013), not original Larike fieldwork by these authors. 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 ↩28 ↩29

[3] Ian Maddieson (1984). Patterns of Sounds, printed p. 65 (Irish primary-nasal classification) and p. 69 (197/200 palatal/velar-nasal implication). https://kcr.digitalwords.net/media/txt/maddieson.i%3A1984%3Asounds.pdf . Original book full text inspected; nasal branch used here only as a boundary, not a categorical positive. registry ↩a ↩b

[4] Paul Kay, “Universality of Color Categorization,” author-hosted paper, PDF pp. 2–5. https://linguistics.berkeley.edu/kay/univ.col.cats.complt.pdf . Firsthand account of the Berlin–Kay proposal and subsequent revisions; four-term green/yellow alternatives. registry ↩a ↩b

[5] Surrey Morphology Group, “Number,” section on paucal systems. https://www.smg.surrey.ac.uk/features/morphosyntactic/number/ . Institutional typological reference consulted for the paucal-versus-trial distinction. registry ↩