A verb is a part of speech that typically expresses an action, occurrence, or state of being. It is often identified by its ability to convey tense—such as past, present, or future—through inflectional changes or auxiliary elements in many languages.^[1] For example, in languages with rich morphology, verbs like those meaning "walk" may form past tense as "walked," while others change more irregularly, such as "go" to "went." Verbs encompass not only physical actions but also mental processes (e.g., "think"), perceptual states (e.g., "seem"), and existence (e.g., "be").^[2] The English term "verb" derives from the late 14th-century "verbe," borrowed from Old French "verbe" and ultimately from Latin verbum ("word"), rooted in the Proto-Indo-European wer- ("to speak"), highlighting its historical link to core expressive elements of language.^[3] In syntax across languages, verbs typically form the core of the predicate, often requiring agreement with subjects in features like person and number (e.g., singular "runs" vs. plural "run"), and they may combine with auxiliaries to express complex tenses, aspects, or moods. This role is crucial for indicating temporality, viewpoint, and other grammatical relations in communication.^[1] In linguistic typology, verbs are generally an open word class, permitting the creation of new forms to reflect evolving usage, such as neologisms derived from other parts of speech.

Overview

Definition and characteristics

A verb is a fundamental part of speech in linguistics that denotes events, encompassing actions, processes, states, or changes of state.^[4] This category contrasts with nouns, which primarily refer to entities, by focusing on dynamic or static situations involving participants. Semantically, verbs encode temporal relations and event structure, often specifying whether an event is ongoing, completed, or habitual through aspectual distinctions.^[5] Syntactically, verbs function as the head of the verb phrase (VP) and typically occupy the core position in the predicate of a clause, linking the subject to additional arguments or complements.^[5] They exhibit valency, determining the number and type of obligatory or optional arguments required, such as agents, patients, or themes, which shapes sentence construction—for instance, transitive verbs like "hit" require an object, while intransitive verbs like "sleep" do not.^[6] Verbs also govern agreement with subjects in person, number, and sometimes gender, as seen in languages like Spanish where the verb form changes based on the subject's features (e.g., hablo for "I speak" vs. hablan for "they speak").^[7] Morphologically, verbs are distinguished by rich inflectional paradigms in many languages, marking categories like tense (past, present, future), mood (indicative, subjunctive), voice (active, passive), and aspect (perfective, imperfective).^[8] For example, English verbs conjugate for tense and person in the present (e.g., walks vs. walk), while more synthetic languages like Swahili inflect for multiple categories simultaneously. This inflectional capacity sets verbs apart from non-inflecting categories like adverbs. Semantically and syntactically, verbs often allow for derivation into other forms, such as nominalizations (e.g., "run" to "running"), reflecting their central role in event conceptualization.^[9] Cross-linguistically, the verb category is considered virtually universal, though its realization varies; some languages lack distinct verb forms or rely on auxiliary systems, yet verbs consistently anchor temporal and aspectual information in clauses.^[10] In theoretical frameworks like generative grammar, verbs project hierarchical structures (e.g., VP within IP or TP) that encode these properties, underscoring their role in sentence interpretation.^[7]

Role in sentences

Verbs function as the head of the verb phrase (VP), which forms the predicate of a clause and expresses the primary action, state, event, or relation in a sentence. In syntactic structure, the verb determines the overall organization of the clause by projecting its arguments—such as subjects and objects—and specifying their thematic roles, thereby licensing the necessary complements to form a complete proposition. For instance, in the sentence "The cat chased the mouse," the verb "chased" heads the VP, requires a subject (agent) and an object (patient), and conveys the event's directionality and completion.^[11] As the syntactic head, verbs impose selectional restrictions on their complements, dictating the category and semantic properties of elements they combine with; for example, transitive verbs like "devour" select noun phrases as direct objects, while intransitive verbs like "sleep" do not permit them. This head-driven projection principle ensures that the verb's subcategorization frame shapes the clause's hierarchical structure, as outlined in phrase structure grammars where sentences (S) consist of a noun phrase (NP) specifier and a VP complement. Verbs also bear inflectional morphology that marks tense, aspect, mood, and voice, anchoring the sentence temporally and modally—e.g., "walks" indicates present habitual action in third-person singular contexts.^[12][13] In sentence production and comprehension, verbs play a pivotal role by integrating lexical meaning with syntactic relations, facilitating the mapping from conceptual structure to linear word order. Psycholinguistic studies demonstrate that verbs are often retrieved early in sentence formulation, guiding the selection of arguments and influencing processing efficiency; for example, verbs with high argument complexity, like ditransitives ("give"), demand more planning resources than simple intransitives. This centrality underscores verbs' function in establishing coherence across clauses, including coordination and subordination, where they maintain aspectual and temporal continuity.^[14][15]

Functional Classification

Lexical verbs

Lexical verbs, also referred to as main verbs or full verbs, constitute the core class of verbs that carry substantial semantic content, denoting actions, states, processes, or events in a sentence.^[16] Unlike functional categories such as auxiliaries, lexical verbs express rich, complex meanings that contribute directly to the propositional content of utterances, such as "eat" for consumption or "run" for physical motion.^[17] They form an open lexical class, allowing for the continual addition of new members through borrowing, derivation, or coinage, which contrasts with the finite, closed inventory of auxiliary and modal verbs.^[18] Syntactically, lexical verbs exhibit robust inflectional properties, including marking for tense (e.g., present "plays" vs. past "played"), aspect (e.g., progressive "is playing"), mood, person, and number, particularly in languages like English where third-person singular present forms add an -s suffix.^[16] These verbs typically head verb phrases and license arguments such as subjects and objects, enabling them to function as the predicate nucleus in clauses, as in "She devours the book rapidly."^[18] Semantically, they profile dynamic relationships that unfold sequentially over time, distinguishing them from more static categories like nouns or adjectives.^[18] In functional classification, lexical verbs differ from auxiliaries in their inability to stand alone as the sole verbal element in a tensed clause; auxiliaries, such as "be" or "have," provide grammatical support for tense, aspect, or voice but lack independent lexical meaning and must co-occur with a lexical verb.^[17] For instance, in "She has eaten," "has" is auxiliary while "eaten" is the lexical verb bearing the core semantics of consumption.^[17] This distinction underscores lexical verbs' role as the semantic backbone of sentences, with auxiliaries serving a subordinate, structural function.^[19] Cross-linguistically, lexical verbs may incorporate classifiers or deictic elements in sign languages like American Sign Language, enhancing their expressive capacity for events involving spatial or referential details.^[16]

Auxiliary and modal verbs

Auxiliary verbs, also known as helping verbs, assist the main lexical verb in a clause by contributing grammatical information related to tense, aspect, voice, or mood, without carrying primary semantic content themselves.^[20] In English, the core auxiliary verbs are be, have, and do, each serving distinct functions: be forms progressive aspects (e.g., "She is running") and passive constructions (e.g., "The book was read"), have marks perfect aspects (e.g., "They have finished"), and do supports negation, questions, and emphasis in simple tenses (e.g., "Do you understand?").^[21] These verbs differ from lexical verbs in that they cannot stand alone as predicates and lack independent lexical meaning, instead embedding within the verb phrase to modify the main verb's grammatical properties. Modal verbs form a specialized subclass of auxiliaries that encode modality, expressing notions such as possibility, necessity, permission, ability, or obligation, thereby situating the proposition relative to the speaker's judgment.^[22] In English, prototypical modals include can, could, may, might, must, shall, should, will, and would, which precede the base form of the main verb without inflection for tense, person, or number (e.g., "She can swim" for ability; "You must leave" for obligation).^[23] Unlike primary auxiliaries like be and have, modals do not participate in progressive or perfect constructions and exhibit defective paradigms, lacking non-finite forms such as infinitives or participles.^[24] Linguistically, auxiliaries and modals occupy a unique syntactic position, often analyzed as functional heads in the verb phrase that project less structure than lexical verbs, enabling phenomena like subject-auxiliary inversion in questions (e.g., "Is she running?") and do-support in non-affirmative contexts.^[25] Cross-linguistically, while no universal formal definition exists, auxiliaries are commonly identified by their inability to bear independent arguments and their role in periphrastic constructions, distinguishing them from serial or light verbs in languages like those with complex verb serialization.^[26] In theoretical frameworks, such as those treating auxiliaries as a unified category, be, have, do, and modals share properties like tense marking (except modals) and negation sensitivity, supporting their classification as a cohesive system rather than mere affixes or separate particles.^[27]

Valency and Argument Structure

Valency

In linguistics, valency refers to the inherent capacity of a verb to govern a specific number and type of syntactic arguments, known as actants, which are obligatory elements required to complete the verb's semantic structure. This concept, borrowed from chemistry where it describes an atom's bonding potential, positions the verb as the central organizing element of a clause, akin to a nucleus that binds its satellites. Lucien Tesnière introduced the term in his seminal 1959 work Éléments de syntaxe structurale, defining valency as the number of actants a verb can attach, emphasizing that these actants—including the subject—are semantically anchored in the verb's meaning and essential for expressing a complete predication.^[28][29] Tesnière distinguished actants from circonstants, the latter being optional adjuncts that provide circumstantial details (such as time, place, or manner) without contributing to the verb's core valency; for instance, in "She eats an apple in the kitchen," "an apple" is an actant while "in the kitchen" is a circonstant. Verbs are classified by their valency based on the number of actants: avalent verbs require none (e.g., "It rains," where the subject is non-referential); monovalent verbs take one (e.g., "She sleeps," with only a subject actant); bivalent verbs require two (e.g., "He eats rice," subject and direct object); and trivalent verbs need three (e.g., "She gives him a gift," subject, indirect object, and direct object). Tetravalent verbs, involving four actants, are rare and typically occur in specific constructions. This classification highlights valency as a lexical property, though it can appear to vary through syntactic operations like passivization, which reassigns actant roles without altering the verb's underlying requirements.^[30][31][32]

Types by number of arguments

Verbs are classified by the number of arguments they require, a property central to their valency or subcategorization frame in syntactic theory. This classification typically includes intransitive verbs, which take one core argument (the subject); transitive verbs, which take two (subject and direct object); and ditransitive verbs, which take three (subject, direct object, and indirect object). These categories reflect the semantic and syntactic roles the verb assigns to its participants, influencing clause structure across languages. Intransitive verbs license only a single argument, usually realized as the subject, and do not permit a direct object. For example, in English, the verb sleep in "The child sleeps" requires no object, as adding one like "*sleeps the bed" results in ungrammaticality. This type is common for verbs denoting states or one-participant events, such as arrive or die, and is prevalent in many languages, where intransitive constructions often form the basis of basic clauses. Cross-linguistically, intransitives may split into unaccusative (e.g., fall, where the subject is a theme) and unergative (e.g., run, where the subject is an agent) subtypes based on semantic role distinctions. Transitive verbs, by contrast, obligatorily take two arguments: a subject (typically an agent) and a direct object (often a patient or theme). An English example is break in "The boy broke the window," where omitting the object yields an incomplete sense in many contexts, though some transitives allow object drop under specific pragmatic conditions. This class encompasses verbs of caused change or transfer, like hit or give (in its two-argument use), and transitivity is a key parameter in typological studies, with languages varying in whether objects are marked (e.g., via case) or unmarked. High transitivity correlates with events involving volitional agents affecting patients, as analyzed in discourse-functional frameworks. Ditransitive verbs require three arguments: a subject, a direct object, and an indirect object, often encoding transfer events with a goal or recipient. In English, give exemplifies this in "She gave him a book," where both him (indirect object, recipient) and a book (direct object, theme) are necessary; rephrasing as "She gave a book to him" uses a prepositional phrase for the indirect role, but the core three-argument structure remains. Other examples include send or tell. Ditransitives are less common than monotransitives but crucial for understanding double-object constructions, which in some languages (e.g., via dative alternation) allow flexible ordering of objects based on discourse prominence. Semantically, they often involve possession transfer, with the indirect object bearing a beneficiary or goal role. Some verbs exhibit variable valency, functioning as ambitransitive or labile, alternating between intransitive and transitive uses without morphological change—for instance, English open in "The door opens" (intransitive) versus "She opens the door" (transitive). This flexibility highlights how argument structure can be influenced by lexical semantics and constructional patterns, though strict ditransitives rarely reduce to fewer arguments without altering meaning. Such classifications inform computational linguistics and language acquisition models, where verb argument patterns predict syntactic behavior.

Copular and impersonal verbs

Copular verbs, also referred to as linking or equative verbs, function primarily to connect a subject to a non-verbal predicate, such as an adjective, noun phrase, or prepositional phrase, thereby expressing relations of identity, attribution, or location without denoting an action or event.^[33] In linguistic terms, they form copular clauses where the semantic content resides in the predicate complement rather than the verb itself, which often contributes minimal or no independent meaning beyond linkage.^[34] For instance, in English sentences like "She is happy" or "That is a book," the verb "is" serves as the copula, equating the subject to an adjectival or nominal complement.^[17] Beyond the prototypical copula "be," English employs a range of verbs that can act as copulas in specific contexts, including change-of-state verbs like "become" (e.g., "The weather became stormy") and perception-based verbs such as "seem," "appear," "feel," "look," "sound," and "taste" (e.g., "The soup tastes delicious").^[17] These verbs exhibit variable valency but typically require a subject and a predicative complement, distinguishing them from full lexical verbs that assign thematic roles like agent or patient.^[35] Semantically, copular constructions often encode existential, locative, or predicative meanings, with cross-linguistic typologies showing polysemy between copular and locative/existential functions in many languages, including English.^[36] Impersonal verbs, by contrast, are characterized by their lack of a specific referential subject, often employing a dummy or non-referential element like "it" in English to fulfill syntactic requirements, resulting in constructions with zero or minimal arguments (avalent or monovalent valency). They typically express general states, atmospheric conditions, or mental impressions without assigning a thematic role to a personal subject, as seen in examples such as "It rains," "It snows," or "It seems likely."^[37] Scholarly analyses identify two main categories: weather verbs (e.g., rain, thunder) that denote natural phenomena impersonally, and verbs of perception or cognition (e.g., seem, happen, occur) that introduce clausal complements without a nominative argument.^[37] In modern English, true impersonal verbs are limited compared to historical stages like Old English, where a broader class—including verbs of desire like "long" or "hunger"—appeared in impersonal dative-subject constructions (e.g., "Me hungers," meaning "I am hungry").^[38] These evolved toward personal constructions by the Middle English period, with impersonal uses persisting mainly in fixed expressions.^[39] Overlap exists between copular and impersonal functions, as certain copulas like "seem" can adopt impersonal syntax (e.g., "It seems to rain"), highlighting their shared low-argument structure and role in non-agentive predication.^[40]

Valency marking and alternations

Valency marking encompasses the grammatical mechanisms languages employ to signal the number, type, and roles of arguments a verb requires or permits. In morphological terms, this often involves case affixes on noun phrases to distinguish core arguments like subjects and objects, as seen in languages such as Finnish, where transitive verbs mark direct objects with accusative or genitive case.^[41] Verb agreement systems further encode valency through polypersonal affixes that cross-reference multiple arguments, as in Algonquian languages where verbs inflect to indicate subject and object features.^[42] Syntactically, valency is marked via obligatory complements or adjuncts, with prepositional phrases serving this role in Romance languages like French, where verbs such as penser à (to think about) require a specific preposition to realize their two-valent structure. These marking strategies are central to valency theory, pioneered by Lucien Tesnière in his 1959 work Éléments de syntaxe structurale, which analogized the verb's capacity to govern actants (arguments) to chemical valency, emphasizing the verb as the structural nucleus that obligatorily "connects" a fixed set of dependents.^[43] In cross-linguistic perspective, morphological marking predominates in agglutinative languages, where affixes directly alter or indicate valency; for example, in Turkish, the verb root gel- (come) takes a causative suffix -tir/-dir to form getir (bring), increasing valency from one to two arguments via explicit affixation.^[44] Conversely, analytic languages like English rely more on word order and auxiliary verbs for marking, though residual case (e.g., genitive) and agreement (subject-verb number) still play roles.^[28] Valency alternations refer to systematic variations in a verb's argument realization, allowing the same lexical item or related forms to exhibit different valencies while preserving core semantics. These alternations can be non-morphological, as in English conative alternation, where hunt the deer (transitive) alternates with hunt for the deer (intransitive with preposition), reflecting optional object incorporation.^[45] Beth Levin's 1993 classification of English verbs into over 70 alternation classes, based on shared syntactic behaviors tied to semantic roles, highlights patterns like the locative alternation: load the truck with hay (apply-location) versus load hay onto the truck (location-apply), where argument types shift between theme and goal.^[45] Morphological alternations involve affixation to explicitly change valency, common in languages with productive derivations. In Bantu languages like Swahili, the applicative suffix -il-, as in pika (cook) becoming pikila (cook for), increases valency by adding a beneficiary argument.^[46] Decreasing alternations, such as passivization, reduce overt arguments; for instance, in German, active Der Koch bereitet das Essen (the chef prepares the food) passivizes to Das Essen wird bereitet (the food is prepared), demoting the agent via morphological marking on the auxiliary. Causative-inchoative alternations, where an intransitive verb gains a causer argument, often feature dedicated morphology in Indo-European languages, as in French ouvrir (open, intransitive) versus causative faire ouvrir (make open).^[47] These patterns underscore how alternations link syntax to semantics, with Levin's framework demonstrating that verb classes predict participation in specific alternations based on event structure.^[45]

Agreement

Subject-verb agreement

Subject-verb agreement is a core syntactic phenomenon in which the finite verb in a clause morphologically matches its subject in grammatical features such as person, number, gender, and sometimes case, thereby establishing a formal link between the subject and predicate. This process is central to sentence structure across many languages, facilitating clarity in grammatical relations.^[48] The primary features of subject-verb agreement include number (singular or plural), person (first, second, or third), and gender (masculine, feminine, or neuter), though not all languages mark every feature. In languages with rich verbal inflection, agreement ensures that the verb's form reflects the subject's properties, reducing ambiguity in argument structure. For instance, mismatches can lead to ungrammaticality, as seen in error production studies where speakers erroneously pluralize verbs after singular collective nouns like "the team."^[49][50] In English, subject-verb agreement is relatively limited, primarily affecting present-tense verbs through the third-person singular -s suffix, while other persons and tenses show minimal distinction except for the copula be. Examples include "The cat runs" (singular subject, singular verb) versus "The cats run" (plural subject, plural verb); agreement here targets number, with person less overtly marked. Theoretical analyses challenge the traditional assumption that all tensed English verbs agree in both person and number with their subjects, arguing instead that agreement is optional or absent in many cases, such as with modals (she can, not she cans) or infinitives. Mismatches often arise from notional (semantic) plurality rather than strict grammatical number, as in "A number of students are" versus "The number of students is," highlighting a tension between form and meaning.^[51][52][53] Cross-linguistically, agreement patterns vary significantly by language family and typology. In Romance languages like Spanish, verbs inflect for both person and number across tenses, as in "Yo hablo" (I speak, first-person singular) versus "Hablamos" (we speak, first-person plural). Arabic, a Semitic language, incorporates gender agreement alongside person and number, particularly in the past tense; for example, "Kataba" (he wrote, masculine) contrasts with "Katabat" (she wrote, feminine) for third-person subjects. English, by comparison, exhibits shallower agreement morphology, lacking consistent gender or full person marking, which aligns with its analytic tendencies. These differences influence processing and acquisition; for instance, learners of English as a second language from agreement-rich L1s like Spanish may overapply features, leading to errors in number matching.^[54][55] Theoretical debates in linguistics often frame subject-verb agreement as a syntactic operation within generative grammar, where features on the subject "probe" and agree with the verb via feature checking. However, psycholinguistic research reveals that agreement computation can be attracted by intervening nouns, causing errors like "The key to the cabinets are" instead of is, suggesting interactive rather than strictly hierarchical processing. In pro-drop languages like Spanish or Arabic, agreement is crucial for subject recovery when the subject is omitted, underscoring its role in discourse cohesion. Overall, while universal in many languages, the scope and triggers of agreement reflect parametric variation in universal grammar.^[56][57]

Object and other agreement phenomena

Object agreement, also known as object-verb agreement, is a grammatical phenomenon in which a verb morphologically inflects to match the phi-features (such as person, number, gender, or class) of its direct or indirect object, either in addition to or instead of agreeing with the subject. This type of agreement is rarer cross-linguistically than subject-verb agreement and tends to occur in languages with rich verbal morphology, such as those in the Bantu, Uralic, or Austronesian families.^[58] Object agreement often correlates with the syntactic position or semantic properties of the object, such as its definiteness, animacy, or topicality, and may be restricted to specific clause types like transitives without overt subjects.^[59] Cross-linguistic studies reveal that object agreement is typically conditioned by hierarchical constraints within the agreement system. For instance, Edith Moravcsik's seminal analysis of 45 languages found that verbs agree with objects only if they also agree with subjects, and object agreement is more likely when the object is pronominal or definite, suggesting a universal implicational scale where subject agreement implies object agreement but not vice versa.^[60] In Bantu languages like Chichewa, the verb exhibits subject prefixes and object suffixes simultaneously, as in "chi-li-teng-a" (it [class 7] is selling it [class 5]), where "chi-" marks the class 7 subject, "li-" the class 5 object, and "teng-a" is the verb root "sell".^[61] Similarly, in Khanty (a Uralic language), object agreement on the verb occurs when the object is topical and definite, often overriding subject agreement in certain contexts.^[62] These patterns indicate that object agreement serves to index arguments for discourse prominence or to resolve ambiguity in pro-drop systems.^[63] Theoretical accounts of object agreement often invoke minimalist frameworks, where phi-features on the verb probe for matching features on the object via Agree relations, potentially competing with subject agreement. In some analyses, object agreement is treated as a post-syntactic operation, applying after spell-out to ensure morphological realization without affecting core syntax, as argued in cases where agreement fails to influence word order or case assignment.^[64] For example, in Hindi-Urdu, a mixed agreement system allows object-verb agreement in ergative constructions when the object is specific, leading to attraction effects in processing where interveners influence agreement computation.^[65] Beyond direct object agreement, verbs may exhibit agreement with indirect objects, obliques, or even adjuncts in specific languages. In Palauan (Austronesian), verbs show phi-agreement with direct objects via suffixes, even in the presence of subjects, highlighting valency-based triggers.^[66] In Romance languages, a related phenomenon is past participle agreement, where the participle agrees in gender and number with a preceding direct object clitic, as in Italian "Le ho lette" ('I have read them' [fem. pl.]), conditioned by the object's preverbal position and absence of the auxiliary's own agreement.^[67] In American Sign Language, agreement verbs spatially index both subject and object locations, but object agreement is restricted to animate or definite referents, demonstrating modality-specific constraints on phi-feature valuation.^[68] These "other" agreement phenomena underscore the interplay between syntax, morphology, and semantics in verbal inflection.

Tense, Aspect, and Modality

Tense

In linguistics, tense is a grammatical category that expresses the location of an event or state in time relative to a reference point, typically the moment of utterance or another temporal anchor. It primarily functions through morphological markers on verbs, such as inflections or auxiliaries, to indicate temporal relations like past, present, or future. Unlike aspect, which concerns the internal structure of events, tense focuses on their external positioning along a timeline.^[69][70] Tense systems are distinguished as absolute or relative. Absolute tense anchors the event time directly to the speech time (deictic center), as in English where the present tense denotes simultaneity with utterance ("I eat now") and the past tense denotes anteriority ("I ate yesterday").^[69] Relative tense, by contrast, locates the event relative to another event or reference point, not necessarily the utterance time; for example, in English narrative sequences, "She entered the room and sat down" uses past tense relative to the prior event.^[69] Some languages, like certain Bantu languages, rely predominantly on relative tense systems, where forms indicate "before" or "after" a matrix event without absolute coding.^[69] Cross-linguistically, tense marking varies widely. According to the World Atlas of Language Structures (sample of 222 languages), approximately 42% distinguish past from non-past tenses with no remoteness distinctions, often treating present and future as a single category, as in Japanese or Arabic.^[71] About 60% mark past tense in some form (including remoteness distinctions), while future tense marking is present in about 50% of languages, frequently expressed via modal verbs or auxiliaries rather than dedicated inflection, reflecting its prospective rather than retrospective nature; the remaining languages rely on context or adverbs.^[71][72] Languages with no overt tense marking, such as Mandarin Chinese or Yucatec Maya, use adverbs or context instead.^[69] Tense is typically realized morphologically, but analytic constructions are common in isolating languages. For instance, in Indo-European languages like German, tense affixes alter the verb stem (e.g., gehen "go" vs. ging "went"), while in creoles or pidgins, free auxiliaries may serve this role. Semantic theories, such as those in Reichenbach's framework, model tense as a relation between event time (E), reference time (R), and speech time (S), where past tense orders E before R (aligned with S), and future after.^[69] This model has influenced analyses of tense-aspect interactions across language families.^[70]

Aspect

Grammatical aspect is a verbal category that encodes the internal temporal structure or constituency of a situation, focusing on how the event unfolds or is distributed over time, in contrast to tense, which locates the situation relative to the moment of speaking. This distinction emphasizes that aspect deals with situation-internal time rather than external deictic time. For instance, aspect may indicate whether an action is viewed as complete, ongoing, repeated, or having repercussions in the present.^[70][73] The foundational distinction in aspectual systems is between perfective and imperfective aspects. Perfective aspect portrays a situation as a bounded whole, often implying completion or wholeness without detailing its internal phases, as in the English sentence She wrote the letter, where the action is presented holistically. Imperfective aspect, conversely, highlights the internal composition of the situation, such as its duration, repetition, or progression, as in She was writing the letter, drawing attention to the ongoing process. This binary opposition is central to many languages, particularly Slavic ones, where verbs often come in perfective-imperfective pairs, such as Russian čitat' (imperfective, 'to read') and pročitat' (perfective, 'to read [completely]').^[73] Comrie (1976) outlines a cross-linguistically motivated hierarchy of aspectual categories, starting with the broadest contrast: perfective > imperfective, under which subtypes like progressive (emphasizing temporariness and ongoing activity, e.g., English be + -ing forms) and habitual (indicating repeated or characteristic actions, e.g., habitual in West African languages like Igbo using suffixes) are subsumed. The progressive subtype is prominent in English and other Germanic languages, marking actions in progress at a reference point, while habitual aspects appear in languages like Navajo, where prefixes denote customary behaviors. This hierarchy reflects implicational universals, where languages marking narrower categories like progressive typically also mark the broader imperfective.^[73] A distinct category is the perfect aspect, which relates a past situation to a later reference point, often conveying resultative, experiential, or persistent relevance, as in English I have visited Paris (experiential perfect) or The door has been opened (resultative). Unlike perfective, the perfect does not view the situation as bounded but as having current effects; it is grammaticalized in Indo-European languages via auxiliaries like have but absent or differently encoded elsewhere, such as in Mandarin Chinese through particles like le. Iterative or frequentative aspects, marking repeated occurrences within a situation, occur in languages like Finnish (e.g., suffixes like -ele- in lukea 'to read' yielding 'to read repeatedly'). Cross-linguistically, aspect is realized morphologically (affixes in Russian), periphrastically (auxiliaries in English), or lexically (verb pairs in Chinese), with variation in how languages prioritize aspect over tense—aspect-dominant in Slavic and Creole languages, tense-dominant in English and Romance.^[73][70]

Mood and modality

In linguistics, mood refers to a grammatical category of the verb that encodes the speaker's attitude toward the propositional content, such as its factuality, desirability, or necessity, often through inflectional morphology.^[74] This category is distinct from but overlaps with modality, which encompasses the broader semantic domain of possibility, obligation, permission, and evidentiality, expressed not only via verbal moods but also through modal auxiliaries, adverbs, or lexical items.^[75] While mood is typically grammaticalized as a finite verbal inflection, modality can be realized periphrastically or lexically, allowing languages to convey nuanced speaker perspectives without always altering the verb stem.^[76] Common moods across languages include the indicative, which marks factual or realis assertions; the subjunctive, signaling hypothetical, counterfactual, or non-factual scenarios; the imperative, used for commands or requests; and less frequent ones like the optative for wishes or the jussive for exhortations.^[74] For instance, in Spanish, the subjunctive mood appears in subordinate clauses expressing doubt or emotion, as in Ojalá que llueva ("I hope it rains"), where the verb llueva contrasts with the indicative llueve for a factual statement.^[77] Cross-linguistically, mood systems vary in complexity: Indo-European languages like Greek distinguish four moods (indicative, subjunctive, optative, imperative), while Salishan languages such as St'át'imcets (Lillooet) employ a subjunctive mood in contexts like questions, conditionals, and future tenses to restrict conversational backgrounds to non-factual propositions.^[76] In contrast, isolating languages like Mandarin often lack dedicated mood inflections, relying instead on particles or context for similar effects.^[75] Modality in verbs frequently involves modal auxiliaries that select a main verb complement, expressing epistemic (judgment of truth), deontic (obligation or permission), or dynamic (ability) senses.^[78] In English, core modals like must convey necessity (She must leave), may possibility (She may leave), and can ability (She can leave), with these verbs lacking non-finite forms and showing subject-auxiliary inversion in questions.^[79] Semi-modals such as have to or be going to extend this system, blending modal and aspectual meanings, as in She has to leave for external obligation.^[79] Beyond auxiliaries, some languages grammaticalize modality directly in verbal morphology; for example, in Turkish, evidential moods distinguish confirmed (-miş) from directly witnessed (-di) events, integrating epistemic modality into the tense-mood-aspect system.^[74] Theoretical frameworks, such as those in Palmer's typology, classify modal systems into propositional (epistemic) and event (root) modalities, highlighting how moods like the irrealis in many Austronesian languages bundle non-factual meanings across these domains.^[75] The interplay between mood and modality underscores their role in clause typing and illocutionary force, where moods often align with sentence types: indicative with declaratives, imperative with commands, and subjunctive with non-assertives.^[80] In functional linguistics, as explored by Bybee, the grammaticalization of mood from lexical sources (e.g., verbs of volition becoming optatives) follows paths of semantic bleaching, where high-frequency modals fuse with main verbs to form inflections.^[81] Cross-linguistic studies reveal that languages with rich mood paradigms, such as those in the Bantu family, encode up to a dozen moods via suffixes, contrasting with analytic strategies in creoles where modality relies on invariant modals.^[74] This variation reflects universal cognitive categories filtered through typological constraints, ensuring verbs convey not just temporal location but also the speaker's epistemic stance.^[82]

Voice

Active and passive voice

In grammar, voice is a morphological or syntactic category of the verb that encodes the relationship between the verb and the participants in the event it describes, particularly the role of the subject with respect to the action or state. The active voice represents the unmarked or basic form of this category, in which the subject of the clause functions as the agent or actor performing the action expressed by the verb.^[83] In transitive clauses, this typically results in a structure where the agent-subject precedes the verb, followed by the patient or theme as the direct object.^[83] For example, in English, the sentence "The author wrote the book" exemplifies active voice, with "the author" as the agent-subject initiating the action on the patient-object "the book."^[83] Active voice is the default construction in most languages for expressing straightforward agentivity, aligning semantic roles like agent and patient with canonical syntactic positions to facilitate clear event depiction.^[83] The passive voice, by contrast, is a derived construction that shifts the focus by making the patient or recipient of the action the subject of the clause, while the original agent is either relegated to an oblique prepositional phrase (often introduced by "by" in English) or omitted entirely.^[84] This reconfiguration demotes the agent from its privileged subject position, allowing the clause to highlight the undergoer of the event.^[84] In English, passives are morphologically realized through a form of the auxiliary verb "be" combined with the past participle of the main verb, as in "The book was written by the author," where "the book" (the patient) becomes the subject.^[84] When the agent is suppressed, as in "The book was written," the construction emphasizes the result or the affected entity, which is useful when the agent's identity is irrelevant, unknown, or to be de-emphasized.^[84] Cross-linguistically, passive voice exhibits considerable variation in form and application, though it is a widespread phenomenon in the world's languages.^[85] According to Keenan and Dryer (2007), a prototypical passive construction satisfies two core properties: (1) promotion of the transitive direct object (or an equivalent) to subject position, with associated syntactic privileges like case marking and agreement; and (2) optional or obligatory demotion of the original subject (agent), often to a non-subject role or suppression.^[85] For instance, in Latin, the active "Puella puerum videt" ("The girl sees the boy") becomes passive "Puer a puella videtur" ("The boy is seen by the girl"), promoting "puer" (boy) to subject while marking the agent with the ablative preposition "a."^[85] Not all languages possess a dedicated passive; those that do may limit it to transitive verbs, exclude agent phrases, or extend it to intransitives (e.g., deriving impersonal passives like German "Es wurde getanzt," "There was danced").^[85] Keenan and Dryer further observe that passives often correlate with verb classes, being more common with highly transitive, affected-patient verbs, and serve to adjust clause structure for discourse purposes.^[85] The primary function of the passive voice lies in its ability to manipulate information structure and topicality, reversing the default mapping of thematic roles to grammatical functions seen in active voice.^[86] In active voice, the agent typically maps to the subject (a high-prominence position), and the patient to the object; the passive inverts this, promoting the patient to subject to make it the topic of the clause when discourse context requires focus on the undergoer rather than the doer.^[86] This is particularly evident in scientific or formal registers, where passives foreground processes or results (e.g., "The experiment was conducted" instead of naming the researcher).^[84] Additionally, passives can express affectedness or adversity in some languages, such as Japanese benefactive passives like "Tomodachi ni purezento o agemaremashita" ("I was given a present by a friend," implying benefit to the speaker as patient-subject).^[85] Overall, while active voice prioritizes agency and directness, passive voice enhances flexibility in encoding participant salience and event perspective across languages.^[85]

Other voices (antipassive, middle)

The antipassive voice is a valency-reducing construction primarily attested in ergative-absolutive languages, in which the patient (P) argument of a transitive verb is demoted to an oblique role (often marked with a dative or locative case) or suppressed entirely, while the agent (A) is promoted to the single core argument, typically in the absolutive case.^[87] This results in an intransitive clause structure, with antipassive morphology often affixed to the verb, such as a dedicated suffix or a zero-marked form in some languages.^[88] Antipassives serve functions like focusing on the agent, indefinite or non-specific patients, or aspectual modifications, and they contrast with passives by preserving agent prominence rather than patient promotion.^[89] A classic example occurs in Chukchi (Chukotko-Kamchatkan family), where the transitive clause təkən-ga məlgən-ən-Ø-ə ("the man-ERG walrus-ABS kill-PRES-3SG.A/3SG.P"; the man is killing the walrus) alternates with the antipassive tək-ø-ə-n məlgən-ə ("man-INTR-kill-PRES-3SG walrus-DAT"; the man is walrus-hunting), demoting the patient to oblique status.^[87] Similar patterns appear in Inuktitut, where antipassive markers like -si- allow the agent to take absolutive case while the patient receives an oblique ending.^[88] The middle voice, by contrast, is a cross-linguistically variable category that typically encodes events with reduced transitivity, where the subject participates in multiple roles—often as both agent and patient or beneficiary—without requiring a distinct object.^[90] It encompasses a semantic continuum of "middle situations" involving low elaboration of participants, such as spontaneous events, bodily actions, or actions benefiting the subject, and is morphologically distinct from active and passive voices in languages that mark it. Seminal typological analysis identifies core middle functions including reflexive (action on oneself), reciprocal (mutual action), anticausative (inchoative change without external cause), and permissive (allowing an event), often unified by degrees of agentivity and participant individuation.^[90] In Ancient Greek, the middle voice is exemplified by forms like luomai ("I loose [for myself]" or "I ransom myself"), contrasting with the active luō ("I loose [something else]"), where the subject acts in its own interest or upon itself.^[91] Modern examples include Dutch zich wassen ("to wash oneself," reflexive middle) or English "The book sells well" (facilitative middle, implying the subject enables the event without full agency). Middle markers frequently overlap with reflexive pronouns or passive forms, leading to polyfunctionality in languages like Spanish or Japanese.^[90] While antipassives and middles both decrease verbal valency, they differ in alignment: antipassives align with ergative systems by adjusting patient accessibility, whereas middles often bridge active and passive semantics in nominative-accusative languages, sometimes evolving from reflexive constructions.^[88] Typological surveys indicate antipassives in about 20-30% of sampled languages, concentrated in families like Austronesian and Pama-Nyungan, while middles appear more broadly but with greater formal variation.

Non-finite Forms

Infinitives

An infinitive is a non-finite verb form that lacks tense, mood, and agreement with a subject, serving as the base or unmarked form of the verb and exhibiting hybrid properties between verbs and other lexical categories.^[92] Unlike finite verbs, which inflect for person, number, and tense to function as the main predicate of a clause, infinitives cannot head independent clauses and instead embed within larger syntactic structures.^[93] This non-finite status allows infinitives to retain verbal qualities, such as taking objects or adverbs, while adopting nominal or adjectival roles.^[94] Morphologically, infinitives vary across languages but are typically uninflected or minimally marked compared to finite forms. In English, the standard infinitive appears with the infinitive marker to (to-infinitive), as in to run, or without it in bare form after modal auxiliaries like can or perception verbs, as in She can run or I saw her run.^[94] This distinction reflects historical developments where the to-infinitive evolved from a prepositional phrase involving the dative preposition to, reanalyzed as a verbal complementizer by Middle English.^[95] In contrast, some languages feature inflected infinitives that agree in person or tense with a controller, such as Portuguese falares ('to speak' for second person plural) in embedded contexts.^[96] Syntactically, infinitives fulfill diverse roles, often as complements to verbs, adjectives, or nouns, or as subjects and adverbials. In English, they commonly act as direct objects of verbs like want or decide (I want to leave), subjects (To leave now would be unwise), or purpose modifiers (She studied to pass the exam).^[94] Bare infinitives pair with control or raising verbs, implying subject sharing (John tried to leave), while to-infinitives introduce non-obligatory control or ECM (exceptional case marking) constructions (I expect him to arrive).^[97] These functions highlight the infinitive's capacity to embed tense and aspect interpretively from the matrix clause, rather than expressing independent temporality.^[97] Cross-linguistically, the presence and form of infinitives exhibit significant variation, with not all languages possessing a dedicated infinitive category. English and most Germanic languages rely on a single infinitive marker, but Romance languages like Spanish and Italian display personal infinitives in specific contexts, such as with prepositions or in subordinate clauses (para que vengas evolving from inflected forms).^[96] Albanian lacks a true infinitive, substituting subjunctive clauses (të punoj 'to work') or adverbial participles (për të punuar) for similar functions, reflecting a typological preference for finite embedding.^[93] In agglutinative languages like Turkish, infinitives (-mek) nominalize verbs for complementation (gitmek istiyorum 'I want to go'), blending verbal and nominal morphology.^[98] Such diversity underscores that infinitives often arise from nominalized verb forms in language evolution, with their syntactic integration constrained by the host language's clause structure typology.^[99]

Participles and gerunds

Participles are non-finite verb forms that combine properties of verbs and adjectives, allowing them to modify nouns or participate in complex verbal constructions across languages.^[100] They typically derive from finite verbs through morphological marking, such as affixes indicating tense, aspect, or voice, and function adjectivally (e.g., "boiled egg" in English, where "boiled" describes the noun) or in periphrastic tenses (e.g., "has eaten" using the past participle "eaten").^[100] Typologically, participles vary widely: in Indo-European languages like Latin, they include present and perfect forms used in relative clauses or absolutes, while in Turkic languages, they often serve as finite-like predicates in subordinate clauses without nominalizing suffixes. This versatility positions participles as a cross-linguistically common category for expressing ongoing or completed actions relative to other events, as seen in their role in resultative constructions in languages like Evenki, where they denote states resulting from prior actions. Gerunds, by contrast, are non-finite forms that nominalize verbs, functioning primarily as nouns while retaining some verbal traits like taking objects or adverbs.^[101] In English, gerunds end in -ing and denote actions as entities (e.g., "Swimming is fun," where "swimming" acts as the subject noun), distinguishing them from present participles, which modify nouns or form progressive tenses (e.g., "the swimming child").^[102] Cross-linguistically, gerund-like forms appear as verbal nouns in Semitic languages, such as Arabic maṣdars that express abstract events (e.g., "reading" as kitābah), or in Romance languages where they overlap with infinitives in nominal roles but differ in allowing genitive possessors. Unlike participles, which emphasize adjectival modification, gerunds prioritize event nominalization, though boundaries blur in languages like Spanish, where the -ndo form serves both participial and gerundial purposes in adverbial phrases.^[101] The distinction between participles and gerunds hinges on their syntactic roles: participles integrate into adjectival or verbal periphrases, while gerunds embed verbal actions within nominal syntax, often heading noun phrases with determiners or modifiers. In typological surveys, this separation is not universal; for instance, some Uralic languages use converb-like forms that combine participial adjectival use with gerundial nominalization, challenging strict categorization. Both forms contribute to non-finite verb systems by enabling compact expression of subordination and modification, as evidenced in their frequent co-occurrence in analytic languages to build tenses or relative clauses without full finite verbs.^[101]