Complement (linguistics)

In linguistics, a complement is a syntactic constituent—typically a word, phrase, or clause—that completes the meaning of a predicate by providing essential information necessary for the grammatical and semantic coherence of a sentence, often functioning as an argument of a verb, adjective, noun, preposition, or postposition.^[1][2] Complements differ from adjuncts in their obligatoriness, as they are licensed by the head element and cannot be omitted without rendering the construction incomplete or infelicitous.^[3] They play a central role in clause structure across languages, enabling the expression of complex relationships such as predication, identification, or description.^[4]

Types of Complements

Complements are categorized based on their syntactic position and function relative to the head:

• Subject complements: These follow linking verbs (e.g., be, seem) and describe or identify the subject, such as noun phrases or adjective phrases (e.g., "She is a teacher").^[2]
• Object complements: These provide additional information about the direct or indirect object of a transitive verb, often adjectives or nouns (e.g., "They elected her president").^[2]
• Complement clauses: Embedded clauses that serve as arguments of matrix predicates, functioning semantically as propositions, events, or facts (e.g., "I know that she left"); these exhibit cross-linguistic variation in structure, including tense-aspect-mood marking and subject dependence.^[4][2]
• Adjectival, nominal, and prepositional complements: These complete adjectives (e.g., "aware of the issue"), nouns (e.g., "the destruction of the city"), or prepositions, which typically require them for full valency.^[1]

In frameworks like Systemic Functional Grammar, complements are analyzed as part of the clause's residue, contributing to experiential and interpersonal meanings while interacting with mood and theme structures.^[3] Cross-linguistically, complementation strategies vary, influenced by the semantics of the matrix predicate (e.g., factive vs. non-factive verbs) and hierarchical scales of integration, from fully clausal forms to reduced nominalizations.^[4] This diversity underscores complements' importance in typology, syntax, and semantic compositionality.

Definition and Overview

Core Definition

In linguistics, a complement is a syntactic element—such as a word, phrase, or clause—that supplies essential information to complete the meaning of a head, typically a verb, noun, or adjective, thereby distinguishing it from optional modifiers that merely add extra detail.^[5][6] This role ensures that the predicate or expression achieves semantic and grammatical completeness, as complements are often required by the head's subcategorization frame, which specifies the types of arguments it selects.^[7] The term originates from the Latin complementum, meaning "that which fills up or completes," and entered English grammar usage in the 19th century to describe these completing structures.^[8][9] Key characteristics of complements include their necessity for grammaticality or full semantic interpretation; for instance, certain verbs demand a complement to form a valid sentence, rendering the construction ungrammatical without it.^[10] Subcategorization frames, as formalized in generative syntax, determine these requirements by encoding the head's selectional properties for specific complement types, such as nominal phrases or clauses. A classic example is the obligatorily transitive verb "resemble," which requires a direct object complement like "her sister" to complete its meaning ("She resembles her sister"), whereas the incomplete form "*She resembles" violates grammaticality.^[11] Similarly, in nominal constructions, complements complete the head noun, as in "the destruction of the city," where "of the city" is indispensable for specifying what was destroyed.^[12] Predicative complements represent a subtype that follows linking verbs to describe the subject, further illustrating how complements integrate essential attributes into the clause structure.^[13]

Distinction from Modifiers

In linguistics, a key criterion distinguishing complements from modifiers (often termed adjuncts) is obligatoriness: complements are required to complete the valency or subcategorization frame of their head, such that their omission results in ungrammaticality or semantic incompleteness, whereas adjuncts provide optional additional information and can be deleted without affecting core grammaticality.^[14] For instance, in the sentence "She placed the book on the table," "the book" functions as a complement (obligatory direct object) and "on the table" as an adjunct (optional locative modifier), as "*She placed on the table" is ill-formed while "She placed the book" remains grammatical.^[15] This distinction underscores that complements saturate the argument structure of the head, often fulfilling essential semantic roles such as patient or theme, while adjuncts merely modify without such necessity.^[16] Several syntactic diagnostic tests further delineate complements from adjuncts. The do-so substitution test replaces the verb phrase with "do so," allowing adjuncts to remain but stranding complements, which leads to ungrammaticality: for example, "She ran quickly, and he did so slowly" succeeds with "quickly" as an adjunct manner modifier, but "*She preferred tea to coffee, and he did so to juice" fails with "to coffee" as a complement.^[6] Similarly, the coordination test permits conjoining like categories but rejects mixing complements and adjuncts: "She placed the book on the table and in the drawer" coordinates two adjuncts successfully, yet "*She preferred tea and quickly" ill-formedly mixes a complement ("tea") with an adjunct ("quickly").^[15] Subcategorization mismatches also render structures ungrammatical when complements are absent or inappropriate, as in "*She arrived the airport" violating the verb's frame, unlike the optional adjunct in "She arrived early."^[16] Within theoretical frameworks like X-bar theory, complements and adjuncts differ structurally: complements attach as sisters to the lexical head within the bar-level (X'), directly completing its projection, whereas adjuncts adjoin to the intermediate bar-level (X') as optional modifiers, allowing recursion and iterability. This configuration explains why complements resist reordering or iteration—e.g., "*state of Kansas of Maine" is impossible—while adjuncts permit stacking, as in "a state near Kansas by the ocean."^[11] Such attachment rules, formalized in X-bar schema, ensure complements are integral to the phrase's core structure, contrasting with the peripheral role of adjuncts.

Types of Complements

Predicative Complements

Predicative complements, often referred to as predicatives, are syntactic elements that connect to the subject or object of a clause via a copular or linking verb, such as be, seem, or become, to express attribution (describing a property) or identification (specifying an identity or role).^[17] These complements complete the predication by providing essential information that cannot be omitted without rendering the clause incomplete or semantically odd.^[17] For instance, in identificational constructions like She is a teacher, the noun phrase a teacher identifies the subject's role, whereas in attributive cases like She is happy, the adjective happy attributes a state to the subject.^[18] Syntactically, predicative complements typically occupy a position immediately following the copular verb, resulting in a subject-copula-predicative complement (S-Copula-PredC) order in basic clauses.^[17] They can manifest as noun phrases (e.g., He became president), adjective phrases (e.g., The food tastes delicious), or even clausal structures (e.g., The problem seems to be solved).^[18] This positioning distinguishes them from other complements, as they are licensed specifically by copular verbs and integrate into the verb phrase to form a unified predicate.^[17] In theoretical terms, predicative complements contribute to the overall predicate in frameworks like predicate logic, where they encode properties, relations, or identities that apply to the subject or object, often treated as part of a higher-order predicate structure.^[19] Unlike appositives, which provide non-essential renaming or elaboration without requiring a copula (e.g., John, my brother, arrived), predicative complements demand the linking verb for grammaticality and semantic coherence.^[17] This obligatory copula presence underscores their role in core predication rather than mere supplementation.^[17] Predicative complements encompass subject complements as a key subset, where the focus is attribution to the clause's subject via the copula.^[18]

Subject and Object Complements

Subject complements following raising and certain linking verbs such as appear and seem (raising verbs) or become (a change-of-state linking verb) attribute properties or states to the subject. For instance, in the sentence "She appears tired," the adjective "tired" serves as the subject complement, describing a perceived state of the subject "she." These structures are distinct from predicative complements in basic copular contexts by relying on verbs that subcategorize for a small clause complement, where the subject originates within the embedded clause and raises to the matrix subject position for case assignment.^[20] In generative syntax, subject complements form part of a tenseless small clause [NP XP], lacking inflectional elements like tense or agreement, as proposed in early analyses of predication. A key structural property is the raising test, which demonstrates the embedded origin of the subject; for example, the ungrammaticality of "*Tired appears she" contrasts with the well-formed "She appears to be tired," highlighting the small clause's role before raising. This analysis ensures uniform treatment of subject-predicate relations across verbal complements.^[20] Object complements, by contrast, follow direct objects and ascribe attributes, roles, or identities to them, completing the verb's meaning in constructions subcategorized by verbs like call, make, consider, and elect. In "They elected her president," the noun "president" functions as the object complement, specifying the role assigned to the direct object "her." These complements typically realize as adjectival or nominal predicates within a small clause structure [NP XP], such as [her president], embedded under the matrix verb.^[20] Structural tests for object complements include pseudopassivization, where the direct object raises to subject position, as in "She was elected president," preserving the small clause interpretation and confirming the complement's predicative role. Another example is "We consider the plan feasible," where "feasible" predicates a property of "the plan" in the small clause [the plan feasible]. Verbs permitting object complements are limited to those that semantically license resultative or ascriptive predication, distinguishing them from simple transitive verbs.^[21] The formalization of subject and object complements as small clauses emerged in 20th-century generative grammar, with initial predication-based approaches in Williams (1975) treating them as co-indexed NP-XP relations without full clausal status, later refined by Stowell (1981) into binary-branching constituents to account for constituency and case requirements. This development resolved earlier debates on whether such complements were reduced relatives or flat VPs, establishing small clauses as maximal projections of non-verbal predicates.^[21]

Syntactic Functions

As Verb Arguments

In syntax, complements function as obligatory arguments that saturate the valency of a verb, ensuring the predicate's argument structure is fully realized for grammaticality.^[22] Transitive verbs, for instance, require a direct object complement to complete their valency, as in She ate the apple, where "the apple" serves as the patient argument; without it, the sentence remains incomplete (She ate).^[23] Ditransitive verbs extend this requirement to two complements, typically a direct object and an indirect object, as seen in He gave her the book, where "her" (recipient) and "the book" (theme) both obligatorily fill valency slots.^[24] Theta theory further elucidates the role of complements by assigning them specific semantic roles, or theta roles, such as agent, patient, theme, or goal, which verbs lexically specify to link syntax and semantics.^[25] In the sentence John gave Mary a gift, "John" bears the agent theta role as the external argument, while the complements "Mary" (goal) and "a gift" (theme) receive internal theta roles, satisfying the verb's thematic requirements.^[26] These roles ensure that complements are not merely syntactic fillers but bearers of essential semantic content, with violations leading to theta criterion breaches in well-formed sentences.^[25] Subcategorization frames complements as lexically determined by the verb, specifying the category and number of required arguments to avoid ungrammaticality.^[22] For example, the verb put subcategorizes for a noun phrase (NP) direct object and a prepositional phrase (PP) indicating location, yielding She put the book on the table; omitting the PP results in deviance (She put the book).^[27] This lexical specification distinguishes obligatory complements from optional adjuncts, which can be added without affecting core valency, such as the adverbial "quickly" in She put the book on the table quickly.^[28] Within Government and Binding (GB) theory, the argument structure of verbs positions complements as projections from the verb phrase (V'), where they occupy sister positions to the verb to fulfill subcategorization and theta assignment.^[25] This projection principle mandates that lexical properties, including complement requirements, be preserved across syntactic levels, as in the D-structure representation of The cat chased the mouse, where "the mouse" projects as the object complement from V'.^[27] In ergative-absolutive languages, such structures may align differently, with transitive subjects absorbing ergative case while objects remain absolutive, though the core verb-complement relation persists.^[24]

In Nominal and Adjectival Constructions

In nominal constructions, complements serve to complete the semantic and syntactic requirements of the head noun, often appearing as prepositional phrases or clauses that are lexically specified by the noun. For instance, deverbal nouns like "arrival" typically require an "of-genitive" complement to specify the event participant, as in "the arrival of the guests," where the prepositional phrase fulfills the noun's valency similar to its verbal counterpart but adapted to nominal syntax.^[29] Clausal complements also occur with nouns denoting cognitive states or propositions, such as "the fear that he would fail," in which the that-clause provides essential propositional content licensed by the head noun.^[29] These complements are obligatory for nouns with high semantic weight, distinguishing them from optional post-nominal modifiers like relative clauses.^[29] Adjectival complements, by contrast, follow the adjective to satisfy its subcategorization frame, commonly taking the form of prepositional phrases or clausal structures that elaborate on the adjective's meaning. Predicates of awareness or evaluation, such as "aware," license prepositional complements like "aware of the risk," where the PP specifies the theme or source essential to the adjective's interpretation.^[30] Clausal complements are prevalent with epistemic or deontic adjectives, including finite that-clauses (e.g., "sure that it works") for propositions with tense and modality, or non-finite to-infinitives (e.g., "eager to succeed") that often involve subject control or raising for co-referentiality between the matrix subject and the clause.^[30] Semantic categories influence complement choice: epistemic adjectives favor that-clauses (67% in corpus data), while dynamic ones prefer to-clauses (21%).^[30] Within the determiner phrase (DP), nominal complements integrate structurally below specifiers and determiners, merging directly with the head noun to form the core projection before adjuncts or further modification. For example, in complex noun phrases like "the destruction [of the city] [by invaders]," the PP complements attach to the noun "destruction," ensuring the DP's internal hierarchy positions them closer to the head than peripheral elements.^[31] This attachment pattern maintains the noun's argument structure without disrupting the phrase's overall determinative function. Adjectival complements similarly embed within adjective phrases (APs), positioning after the head adjective in post-predicative or extraposed constructions, such as "it is certain [that the plan will succeed]," where the clause integrates as a dependent to complete the AP.^[30] In dependency grammar, complements of nouns and adjectives are analyzed as required dependents that link the head to its arguments, contrasting with optional modifiers and determiners, which function as non-essential satellites. For nouns, complements like prepositional phrases depend directly on the head (e.g., "of" in "student of physics" depends on "student"), fulfilling grammatical valence without altering the head's category.^[32] Adjectives exhibit parallel dependency relations, with complements such as adverbs or nouns (e.g., "epoch" depending on "last" in "last epoch") treated as subordinates that specify the adjective's scope.^[32] Determiners, often viewed as specifiers in phrase-structure approaches, are reinterpreted here as complements when obligatory, emphasizing dependency over hierarchical bars.^[32] This framework highlights complements' role in linear precedence and semantic integration, analogous to verbal arguments but tailored to non-verbal heads.^[32]

Broader and Comparative Uses

Extended Interpretations

In semantics, the term "complement" extends beyond strict syntactic arguments to encompass elements that serve as truth-value bearers, particularly in the case of sentential complements. These are clausal structures embedded under predicates of propositional attitudes, such as "believe [that P]," where the complement clause denotes a proposition—a primary bearer of truth and falsity—that the higher verb evaluates or relates to. This interpretation aligns with structured propositions, where the semantic value of the complement is a complex object composed of the sentence's meaning and its possible-worlds truth conditions, distinguishing it from non-propositional complements like nominal objects.^[33] In pragmatics, complements function as discourse completers by integrating into the broader conversational context, filling gaps in shared knowledge or resolving presuppositions raised by the matrix clause. For instance, a complement clause may provide evidential support or background elaboration that advances the discourse coherence, as seen in constructions where the embedded content pragmatically strengthens the overall utterance meaning through inference rather than pure semantics.^[34] This role highlights how complements contribute to the dynamic updating of common ground in dialogue, complementing the propositional content with interactional relevance.^[35] The concept of complements has undergone historical expansion within Chomskyan generative grammar, originating in the 1960s with foundational work on phrase structure and subcategorization but evolving significantly in the 1980s to incorporate functional projections. Early frameworks treated complements as obligatory sisters to heads in X-bar structures, but later developments introduced Complementizer Phrases (CPs) and Inflectional Phrases (IPs), where complements occupy specifier or complement positions within these extended projections—such as wh-complements in CP for interrogative embeddings. This broadening allowed for a more nuanced account of clause embedding and movement, tracing back to Chomsky's initial proposals on syntactic transformations.^[36] Theoretical extensions in frameworks like Head-driven Phrase Structure Grammar (HPSG) further refine complements as non-head daughters that saturate the valence list of the lexical head, specifying the syntactic and semantic requirements for combination. In HPSG, a head's valence feature encodes the list of complements needed to form a saturated phrase, with daughters unifying to discharge these obligations, as in verb phrases where object NPs fulfill the subcategorization frame.^[37] However, this expansive usage has drawn critiques for over-broadening the category, potentially conflating complements with specifiers; for example, in antisymmetric approaches, traditional specifiers are reanalyzed as iterated complements, challenging the binary head-complement relation in X-bar theory and complicating empirical tests for argumenthood.^[38] A representative example of such embedding is the nominal complement in "the fact [that it rained]," where the sentential clause acts as a complement to the noun "fact," conveying propositional content within a larger syntactic structure. This construction underscores the semantic depth of complements in licensing factive presuppositions, an aspect often underexplored in introductory syntactic accounts that prioritize surface forms over embedded truth-conditional contributions.^[39]

Cross-Linguistic Variations

Cross-linguistic variations in complements reveal significant typological differences in how languages structure and realize verbal arguments. In head-final languages such as Japanese, complements precede the verb, reflecting a consistent OV order where heads uniformly follow their dependents, unlike the head-initial VO pattern common in languages like English.^[40] This reversal arises from syntactic movements that derive the surface order, ensuring complements appear before the verb in phrases like "Dana-ga hon-o kat-ta" (Dana-NOM book-ACC buy-PAST).^[40] Similarly, serial verb languages like Akan employ serial verb constructions (SVCs) where multiple verbs chain together as a single predicate, sharing complements such as objects without coordination markers or embedding.^[41] For instance, in Akan, a construction like "She took the fish sold (it)" shares the object "the fish" across verbs, treating it as a unified complement in a monoclausal structure with single tense and negation scope.^[41] Case marking further diversifies complement realization, particularly in ergative languages where alignments differ from nominative-accusative systems. In Basque, an ergative-absolutive language, transitive objects receive absolutive case (marked by zero morphology), while subjects take ergative case, as in "Medikua-k pirata-Ø beldurtzen du" (doctor-ERG pirate-ABS frighten PRES.3SG), positioning the absolutive complement as the theme without accusative marking.^[42] This structural case assignment by little v for absolutive complements contrasts with inherent case views, evident in non-finite contexts where ergative drops.^[42] In Romance languages like Spanish, complements often manifest as clitics, such as the dative "le" functioning as an indirect object pronoun attached to the verb, as in "Le di el libro" (to-him/her gave-1SG the book), which cannot be coordinated or modified and precedes tensed verbs in proclisis.^[43] Dialectal variations, like leísmo, extend "le" to animate direct objects, doubling full NPs marked by the differential object marker "a".^[43] Non-configurational languages introduce additional variability through flexible word order, complicating complement identification. In Warlpiri, an Australian language, free constituent order relies on case marking rather than hierarchical position to distinguish subjects and objects, allowing complements to appear in any linear position while maintaining argument roles via ergative-absolutive morphology.^[44] This non-configurationality contrasts with fixed-order languages, as complements are identified through adpositional or case cues rather than syntactic configuration. Complementizers also vary; in German, the complementizer "dass" (cognate with English "that") is obligatory for finite embedded clauses, unlike optional zero complementizers in English, enforcing verb-final order in subordinates without a null option.^[45] Despite these variations, universal tendencies persist in requiring valency satisfaction, where predicates demand specific complements to fulfill argument structures across languages. All languages exhibit robust transitive-intransitive distinctions, with verbs following semantic hierarchies (e.g., effective action verbs more likely transitive than relation verbs) to predict complement patterns.^[46] However, complement types diverge in polysynthetic languages, where arguments like objects are often incorporated as affixes within the verb, obviating separate nominal complements and aligning with the polysynthesis parameter that limits free noun use in argument positions.^[47] This incorporation satisfies valency morphologically, differing from analytic languages that rely on independent phrases.^[47]