Adjective is a lexical, part-of-speech category. The words belonging to this category modify nouns in a noun phrase structure, assigning an attribute to the referent or specifying its properties. Deverbal adjectives can take arguments of their own and assign case to them.

Adjectives typically have comparative and superlative forms, and can further be modified by adverbial expressions and degree words. They can also be predicates in nominal copular clauses. At the same time, adjectives typically do not form a homogenous class, what is more, in many languages adjectives cannot be distinguished either from nouns or from verbs based on morphosyntactic and distributional criteria. This means that they do not even form a distinct open class in these languages.

Most of the adjectives denote property concepts (e.g. dimension, physical properties, age, value, etc.), but in many languages adjectives (mostly derived forms) can be used as relational and locational modifiers. Intensional adjectives (e.g. "alleged") are a special subtype: they cannot be used as predicates.

If more attributive adjectives are combined in the same construction, the following relative word order can be observed which reflects the hierarchy of various semantic types: opinion/value >> size >> age >> shape >> color >> origin/nationality >> material >> purpose (for more, slightly different hierarchies , cf. Dékány 2021: 16-17)

The term adposition covers at least prepositions and postpositions, and thus refers to a word class which can be distinguished from nouns, verbs and adjectives, and which is sometimes considered a lexical part-of-speech category and sometimes a functional one. Prepositions are lexical elements positioned before their complements, while postpositions follow their complements. Semantically they are relational: the most frequent and numerous subtypes are spatial in meaning, either locative or directional (e.g. in, on, at, into, onto, between etc.); others can be temporal (e.g. after, since) or have more abstract meanings (e.g. for). The phrases headed by adpositions can be predicates (John is in the garden), arguments (John fell in(to) the lake) or free adverbials (John had a coffee in the kitchen) in the clause.

Adpositions are semantically and morphologically closely related to oblique case suffixes, adverbs, as well as verbal particles (or preverbs) in languages that have them, therefore these are sometimes all considered to belong to the same category.

Languages vary as to whether they allow adpositional phrases to modify nouns.

Adverb is the name of an open part-of-speech class. Adverbs are modifiers of verbs, adjectives, other adverbs, or even larger constituents, such as entire verb phrases or sentences. Sentence adverbs (e.g. unfortunately, probably) commonly express the speaker’s attitude toward the event of the proposition, while modifiers of verbs or verb phrases commonly express time, place, direction, manner, or other circumstances (e.g. yesterday, home, slowly). Modifiers of adjectives and adverbs usually express degree (e.g. very, extremely). Manner and degree adverbs may have a comparative or superlative form.

Despite the semantic diversity adverbials may present cross-linguistically, from a syntactic point of view they can be considered to belong to the same structural category as adpositional expressions and oblique case-marked nouns do.

Agreement is a systematic correspondence between the formal properties of two or more elements within a linigusitic domain. Agreement may involve a variety of grammatical categories, for instance number, person, gender, case, definiteness, etc. The domain of agreement may also vary, i.e. a verb (or rather a predicate) can agree with its subject, or a modifier can agree with the head it modifies. Noun phrase internal agreement is often referred to as concord in theoretical literature. The formal manifestation of agreement markers is usually affixation.

When the reference of an expression depends on an antecedent within the same sentence or in the textual context, we speak about anaphors, or anaphoric reference. For instance, the relation between a pronoun and its lexical antecedent is anaphoric, if they are coreferential in the text. The anaphoric use of a definite article marks the noun phrase as having a well-established reference in the discourse, which means that the referent was already mentioned previously.

Aspect is one of the categories of the TAM system, alongside tense and mood. An important distinction is drawn between grammatical aspect and lexical aspect (alternative terms for this contrast include: viewpoint vs situation and outer vs inner aspect). Lexical aspect, also known as Aktionsart, is an inherent property of a verb. It is customary to distinguish between different verb classes: states (own), activities (walk), accomplishments (paint a picture), achievements (buy). Lexical aspect is related to telicity. Grammatical aspect, on the other hand, describes “the internal temporal constituency of a situation” (Comrie 1976). Thus, grammatical aspect expresses temporal reference, similarly to tense. The most common distinction here is between perfective and imperfective aspect. Perfective aspect is used for events that are conceived as bounded, without reference to any flow of time during. Imperfective aspect covers situations conceived as existing continuously or repetitively as time flows. Further distinctions can also be made, e.g., differentiating continuous from progressive aspect, and habitual aspect. Cross-linguistically, the expression of aspectual distinctions is often interrelated with the marking of tense and mood.

The so called belong-construction is a predication about to whom a discourse-given entity belongs, e.g. The book is John’s/yours. This is to be distinguished from predicative possession, in which HAVE-possessive predication introduces a new discourse entity and establishes the relation of this entity to another one, which is notionally a possessor (e.g. John has a book.) However, the construction has often been characterized as a subtype of predicative possession (e.g. Stassen 2009: 28-30), although some typological works classify it completely separately, cf. the so-called genitive predicates, listed among minor clause types with nonverbal predicates in Dryer (2007a). You find the construction under the term Possessor predication in the Typological Database of the Ugric Languages.

The subject of a belong-construction is normally given, and as such definite and/or topical, and the predicate establishes a relation between this entity to another one, which is notionally a possessor. Cross-linguistically, the structure of belong-constructions seems to correspond to copular clauses, in which the predicate part either appears as a nominal predicate (English The book is mine), or as a locative one (French: Le livre est à moi.) in terms of syntax.

Case is a grammatical category of nominal expressions. Morphological case marking is used to encode the grammatical function of noun phrases in a sentence. When case is assigned by virtue of the structural position of an expression, this constituent is claimed to receive a structural case. When case is assigned by virtue of the semantic function the relevant expression fulfill, we can speak about inherent cases (e.g. constituents functioning as a "goal" are assigned dative case in certain languages). This basic distinction is also referred to as a distinction between grammatical and semantic or adverbial cases.

From a syntactic point of view, oblique case-marked nouns, adpositional expressions and adverbials can all be considered to belong to the same structural category

Uralic languages have a rich inventory of case suffixes and postpositions.

There are more than one point of view from which clause types can be discussed. According to the type of their predicates and their internal structure, verbal and nonverbal predicates can principally be distinguished. Based on the valency-properties of the verb, verbal predicates can further be classified: the most fundamental distinction that is usually made is between transitive and intransitive predicates. As more relevant for the present database, we can make a further distinction among nominal predicates, adjectival predicates, and locative predicates within nonverbal predication. A common feature of these patterns is that they typically make use of a copula verb. For this reason they are also called copular clauses. The copula itself is rather a functional word, but very often takes the form of verbs that more or less correspond to the English be. (See also Copula, and Copular clauses)

Co-compounding is a strategy of phrasal coordination, in which the members denote closely associated concepts and are also strongly bounded structurally, and prosodically. They consist of exactly two morphologically parallel juxtaposed nominals. The members may form an additive and exhaustive list (e.g. father-mother), a generalizing one (such as a pair of opposites: day-night), a collective one (the properties of the two words cover a set of referents with similar properties (e.g., milk butter for ‘dairy products’), or a synonymic one.

Compounds are constructions in which a noun directly modifies another noun. However, lexical compounds must be distinguished from syntactic compounds. Lexical compounds behaves more like a single word, often with an idiosyncratic meaning which is not necessarily predictable from the meaning of the components. Syntactic compounds are formed via a productive syntactic construction.

Conjunction words are functional categories that form one of the closed classes among the parts of speech, with a limited number of members. Coordinating conjunctions connect words, phrases, or clauses. Even though they typically appear between the elements they conjoin, they are structurally closer to one of the conjuncts, either to the one they precede, or to the one they follow. In the case of paired conjunction words, these functional elements either precede each of the conjuncts, or they follow each of the conjuncts.

Coordinating conjunction can also be expressed by the simple concatenation of the conjuncts, without a linking element, or by a comitative strategy involving some kind of adpositional expression.

Constituents of the same category can be coordinated by various strategies. Coordination can be phrasal and clausal, and can express semantically a conjunction (e.g. coffee and tea) or a disjunction (e.g. coffee or tea). It may involve a single coordinator (monosyndatic), or multiple coordinators (bisyndetic or polysyndatic). Co-compounds and comitative expressions can also be used to form coordinated structures.

The term comes from the Latin word copula ‘link’ and is generally used for elements that connect the subject and the predicate in a clause. The typical properties associated with copulas are that they are semantically ‘light’ (or they may even be completely devoid of lexical meaning), they are used as connecting elements in contexts where the (lexical) predicate is nonverbal, and in many well-known languages they carry verbal inflection; e.g. English be as a typical representative. Copulas have been often defined as ‘dummy’ elements that serve the sole purpose of carrying verbal morphology, however, there are languages in which the copula does not carry inflection and also languages in which a copula is present in places where it is not needed as the host of inflectional morphology.

Languages vary as to their employment of copulas in various contexts with nonverbal predicates. Many languages have at least some cases with nominal predicates where a connecting element is not present (see e.g. Hungarian, where a verbal connector is missing in 3rd person present tense indicative, but is needed otherwise), these clauses have sometimes been described as lacking a copula altogether and sometimes as clauses involving a zero copula (in contrast to full copulas).

A common definition, or rather description, of ‘copular sentences’ characterizes them as those sentences whose main verb is be (the copula) and its equivalents across languages. In copular sentences, the lexical predicate is not a verb, but a nominal, adjectival or adpositional expression, and if there is a linking element between the subject and the non-verbal predicate, it is generally semantically very ‘light’, i.e. a copula. Generally sentences without an overt linking element of the above type are also grouped with those that contain a (full) copula, as they often require a copula under some conditions.

Copular clauses are often categorized on the basis of the nature of the predicate, and with nominal predicates generally at least predicational (e.g. John is a doctor), specificational (e.g. John is my best friend), and equative (e.g. Cicero is Tully) clause types are distinguished with the possible addition of a fourth, identificational type (e.g That is John) of copular clause.

Definiteness means that the referent of the noun phrase is identifiable in the discourse, for both the speaker and the addressee. Definiteness is a universal semantic and pragmatic notion, but its grammatical realization is a language specific property. Depending on the actual logical-philosophical approach, definiteness is associated with concepts such as uniqueness, inclusiveness, and familiarity, but the basic pragmatic notion that seems to cover its crucial property is identifiability: the speaker signals that the hearer is able to assign a referent to a certain noun phrase, that is, the hearer can identify the referent of the definite noun phrase. A typical realization of this function in grammar is the use of a definite article, but there are many languages that lack articles and make use of other strategies to encode definiteness.

The article is a functional category forming a closed class of words in a given language. The definite article is a fully grammaticalized category encoding definiteness on a syntactic level, which means that it encodes the semantic-pragmatic notion of referential identifiability (see also Definiteness). The range of its usage, however, can vary from one language to another, that is to say, the grammatical encoding of definiteness by an article may segment the semantic field at different points in different languages. For instance, there are languages that only makes use of anaphoric articles and do not need extra marking of situational uniqueness.

Deixis concerns those areas of grammar that basically rely on the extralinguistic context of the utterances. This means that the interpretation of deictic expressions depend on the contexts in which they are used. Linguistic forms encode time, place and person deixis besides discourse deixis. Deictic expressions must always have a reference point, a so called deictic center, which is prototypically the speaker and/or the addressee, and the time and place in which the expressions are uttered.

Demonstratives encode directly accessible reference: the hearer is instructed to match the referent of the noun phrase with some object which is either identifiable/visible in the context, or which is known on the basis of previous discourse. These are also referred to as deictic and anaphoric function, respectively. Pragmatically speaking, the former can be described as exophoric use, while the latter as endophoric use (including the anaphoric and discourse deictic uses). A further function of demonstratives can be characterized as recognitional use. In this context, demonstratives modify discourse-new entities, whose reference must be familiar for the hearer through a special shared knowledge of the interlocutors. All these uses imply referentiality and demonstratives are necessarily definite.

The overall function of demonstratives is to coordinate the focus of attention between interlocutors. Depending on the referent's position with respect to the deictic center, languages typically distinguish between proximal and distal demonstratives. However, distance neutral forms are also attested, as well as three-term and four-term deictic systems. From a syntactic point of view, demonstratives can be used as pronouns as well as modifiers within a noun phrase. Apart from pronominal and adnominal demonstratives, many languages employ so called identificational demonstratives in copular clauses, which in many descriptions are not properly distinguished from pronominal demonstratives.

There is a further, special use of demonstratives, often called indefinite or specific indefinite demonstratives, which rather reflect the speaker's attitude to the denotatum, than a proper deixis or anaphoric meaning.

Determination takes place through determiners, i.e. functional elements that relate to referential properties of a referent.

In a broader definition, determiners include articles, demonstratives, as well as certain quantifiers. These modifiers identify the referent or determine its position within a set whose reference is well-established. (See also Definiteness)

Discourse particles form a closed class among the parts of speech. These elements contribute to the expressive content of an utterance, rather than to its propositional content. Their basic function is to help to organize a discourse by adding information about the epistemic states of the speaker or other discourse participants with respect to the content of an utterance. They provide clues about their knowledge and belief systems thus establishing a link between the propositional content and its context of use.

Ellipsis is the omission of clausal material that is necessary for the interpretation; the missing material is recovered from the antecedent or from the context. There are numerous types of ellipsis acknowledged in theoretical syntax, including verb phrase ellipsis, noun phrase ellipsis, sluicing, gapping, striping, etc. For example, in John saw Alice but Bill didn’t, the second conjunct is understood as ‘Bill didn’t [see Alice]’, with the ellipted verb phrase being understood based on the antecedent. Nominal ellipsis targets the noun head, potentially with some of its modifiers. Suffixes that would normally belong to the head noun, may remain overt and be phonologically supported by the last overt element in the noun phrase.

Theoretical analyses of ellipsis vary significantly depending partly upon whether a constituency-based or a dependency-based theory of syntactic structure is assumed. A further question, raised within the constituency-based approaches, is whether ellipsis is to be understood as a syntactically present, but phonologically unpronounced structure or as a silent pronominal form substituting the missing material. The latter analysis is related to the ones suggested for pro-drop and null arguments.

Evidentiality is a grammatical category that encodes the source of information on which the statement is based. Languages employ evidential systems of different sizes: differentiating between firsthand and non-firsthand information; further distinctions may also be made, e.g., reported, inferential, visual, sensory evidentials. Evidentiality may interact with aspect and tense as well as with mood and modality. It has been observed that evidentiality contrasts can be neutralized under negation and evidentials are only used in main clauses. Furthermore, it has been noted that evidentials easily spread in language contact.

Existential constructions introduce a new participant into the discourse, an indefinite theme argument, which is usually associated with a location. Existentials do not predicate about individuals, but have a presentational meaning (i.e. they are thetic, rather than categorical statements). The construction in most languages has a so called existential verb or ‘be’-like copula. In some languages, there is also an expletive proform in the construction, like in English There is a book on the table.

Furthermore, existential and locative clauses are often treated as derivationally related in terms of syntax. In this approach, the difference can be captured by the identifiability of the theme. Accordingly, existential constructions are only used when the theme is non-identifable. Note however that existential clauses often makes use of a specialized existential verb or copula and/or have a word order different from locative clauses. Existential clauses often have special negation strategies too.

Predicative possession often follows the pattern of existentials with a [+human] location. (See also Predicative possession)

Existential quantifiers are pronominal elements or modifiers that form expressions denoting at least a minimal, non-zero amount or number of the quantified expression, e.g. some water, a boy. (See also Quantification)

According to the traditional approach, finiteness is a property of the verb. Non-finite verbs differ from finite ones with respect to numerous morphosyntactic and syntactic properties, such as argument structure (including the morphosyntactic encoding of the arguments), expression of TAM and agreement markers, negation, valence and voice distinctions. A great deal of variation is observed across languages with respect to the (morpho)syntax of non-finite clauses; for this reason, finiteness has been treated as a scalar concept in the typological literature. In the generative literature, especially in the Government and Binding theory, finiteness has been considered to be crucial for the licensing of the subject’s nominative case; in more recent works within Minimalism, finiteness is viewed as a sentential property rather than a verbal one. (See also Nominalizations)

Focus is a basic notion of Information structure (see also Information structure). Focusing is an operation on a set of alternatives in the discourse. Focus of assertion, or identificational focus, specifies an alternative from the set as that one for which the background part of the utterance holds. This operation may also be quantifying over the set of alternatives, negating alternatives, or scaling alternatives on a scale of likelihood.

The focus typically conveys new information, but the correlation between focus-background on the one hand, and new vs. given information on the othar is not a rule. Fucus has such uses as correction and confirmation of information, in which cases focused constituents can easily be discourse given.

There are various ways of focus marking, such as pitch accents, word order, specific syntactic constructions like cleft sentences.

Indefinite noun phrases are typically used to introduce a new entity into the discourse domain. It might refer to any of the entities that satisfy the description of the noun phrase, the addressee does not have to identify it uniquely. Indefiniteness can be specific, or non-specific. Specific indefinites encode noun phrases whose reference cannot be directly identified, but there is an entity or a set with a well-established reference to which it is related or in which it is a member. Indefiniteness is often encoded by a specialized determiner, the indefinite article, but quantified noun phrases can also be indefinite.

Since subjects in existential clauses needs to be discourse-new, only indefinite nominals are grammatical in existentials.

The article is a functional category forming a closed class of words in a language. Indefinite articles encode indefiniteness, but there are languages in which this grammatical element is completely absent, or is only used to mark specific indefinites (see also Indefiniteness, Specificity). Noun phrases with an indefinite article can also have a generic reading (e.g. A lion is usually dangerous.).

In general, the indefinite article is either identical in form with the numeral 'one', or can be derived from this numeral historically. In many languages they can only be distinguished by different stress patterns (indefinite articles are usually unstressed), or by their different syntactic positions. Unlike definite articles, indefinite articles are typically restricted to the use with singular count nouns. Definite and indefinite articles do not form a natural pair and there are many languages that only make use of one of them.

Information structure can be defined as the grammatical reflection of information packaging, the pragmatic structuring of a proposition in a given discourse. This includes the grammatical distinctions made between known (discourse-old) and new (discourse-new) information in a certain proposition, where old information is topical and new information is focal in sentences. (See also Focus and Topic)

A group of lexemes describe entities with a prototypically unique referent, which can be identified quite independently of the discourse situation, based on the common knowledge of the interlocutors about the world (e.g. The sun has risen.). Inherently unique nouns often seem to behave as proper names in syntax.

Locative expressions refer to a place, thus, they first and foremost comprise a semantic category. Formally, they can be expressed by adpositional phrases, case-marked nominals, or locative adverbs. Those phrases that refer to a location are semantically relational in nature, as they express the spatial relation or configuration between an entity in the clause (the Figure) and their complement (the Ground). For example, in a sentence such as John is in the garden, the subject John is the Figure that is located with respect to the Ground the garden and the nature of the relation is that John is contained within the space of the garden.

Locative expressions often develop temporal and more abstract, not spatiotemporal meanings while maintaining their formal properties, therefore, an expression with the locative preposition in can also be used with a temporal or a non-spatiotemporal reference (e.g. The meeting is in an hour, John left in a hurry).

Locative expression are predicates in locative clauses, a subtype of nonverbal predicational patterns, in which a theme is linked to a location. Locative expressions also serve as adnominal modifiers; in this case they contribute to the referential properties of the noun phrase by locating the head noun in a spatio-temporal or cognitive dimension. (See also Modification and Nonverbal predication)

Modifiers within the noun phrase narrow the denotation, i.e. they specify the head noun descriptively, or restrict the domain of reference of the noun phrase. They may appear as adjectival modifiers, nominal expressions, adpositional phrases or adverbs, as well as clausal constituents, such as participial expressions or finite relative clauses (the latter are also characterized as complex modifiers due to their internal structural complexity). Modifiers can be both arguments and adjuncts depending on the nature of their relation to the head they belong to. Arguments are lexically selected dependents of argument-taking nouns and adjectives. Nevertheless, prototypical modifiers are adjunct-like attributive adjectives.

Modifiers can be restrictive and non-restrictive, admitting that they are often not easy to distinguish from each other. Restrictive modifiers restrict the denotation of the noun, thus contributing to the identification of the referent. Non-restrictive modifiers provide additional information about something with an already well-established referent. Consequently, uniquely referring noun phrases, such as e. g. proper name can only be modified by non-restrictive modifiers.

In the database three types of adnominal modification are distinguished: qualitative, quantitative and locative, following partly the taxonomy of Rijkhoff (2001). Qualitative adnominal modifiers specify the properties of the referent, quantitative adnominal modifiers contribute information about its quantity or its relation to a set of referents, typically expressed by specialized lexemes, such as e.g. numerals (see also Quantification), while locative adnominal modifiers contribute to the identification of the referent by locating the head noun in a spatio-temporal or cognitive dimension. This kind of modification is typically expressed by demonstratives, possessor phrases, locative adpositional phrases, or restrictive relative clauses.

As for the linear ordering of the constituents, modifiers either precede or follow the head noun, and this parameter is a language-specific one for each modifier-type. Word order variation, however, can also be observed within a single language. There can be productive morphological devices or dedicated syntactic strategies in a language to convert various expressions into structurally convenient modifiers.

Mood is a grammatical category, expressed on verbs, that is used to encode the speaker’s attitude towards the assertion. This means that mood is broadly related to modality. Some examples of moods attested across languages are conditional, (ir)realis, indicative, interrogative, imperative, subjunctive, and optative. The grammatical category of mood is interrelated with the other members of the TAM system, tense and aspect.

When a clause has a predicate that is instantiated by a nominal expression, we are dealing with a nominal predicate. Nominal predicates present a subtype of nonverbal predicational patterns and at the same time, in many languages, appear in copular clauses. In copular clauses the presence of a functional word, a copula is required beside the predicate in order to form a grammatical sentence with the subject. Nominal predicates, however, are often special in some languages in that they may not require the presence of such a copula, but can appear juxtaposed to a lexical subject without any (overt) connecting element.

Nominal and adjectival predicates are often similar in their distributional properties and with respect to the use of the copula, while in some languages they exhibit partially distinct morphology and syntax. If the nominal expression is non-referential, it rather assigns some properties to the subject or classify it to a certain set. In this case it is closer to adjectival predicates. When the nominal expression is referential and/or definite, we can speak about an equative, or equational clause (e.g. Peter is the head of the department.)

In the broad sense, the term nominalization is used to cover cases in which a word that does not belong to the part-of-speech of nouns is used as a noun (or as the head of a noun phrase). In the stricter sense, it applies to morphological processes where the input is an adjective or a verb and the output is a noun. These are deadjectival and deverbal nominalizations, respectively: generosity is an example of the former, while growth is an example of the latter. Within the class of deverbal nominalizations, it is customary to distinguish between event nominalizations (growth, running) and argument-denoting nominalizations (driver, opener). In the framework of generative syntax, nominalizations have been analyzed as so-called ‘mixed extended projections’: syntactic structures consisting of functional projections with different categorial specifications, e.g., in the case of deverbal nominalizations, a nominal layer tops off a verb phrase. (see also Finiteness)

Nonverbal predication covers cases in which the predicate in a subject-predicate configuration is not a verbal element but a nominal, adjectival, or a locative expression, the latter appearing in the form of an adpositional or adverbial expression. Beyond these three basic types, i.e. nominal, adjectival, and locative predicates, Dryer (2007a) also lists some minor clause types with nonverbal predicates, among them the so-called genitive predicates (e.g. This money is mine.) which corresponds to the belong-construction in the present database (see also Belong-construction).

According to Payne's alternative typology (1997), nonverbal clauses can be divided into the following types: equation (e.g. She is my mother.), proper inclusion (e.g. Kurumaku is a hunter.), attribution (e.g. She is intelligent.), locative predication (e.g. The cat is on the mat.), existential predication, (e.g. There are wild cats in Africa.), and possessive predication (e.g. Tom has a car.).

Noun is a part-of-speech category. The words belonging to this lexical class typically denote entities, i.e. they are names of persons, things, places, etc. From a morphosyntactic point of view, they can be characterized by their nominal inflection paradigm, such as number and case affixes, as well as personal agreement markers.

Nouns are heads of noun phrases that may also include modifiers, such as adjectives, and can be determined by various types of determiners, such as articles. If they function as a predicate, they are normally accompanied by a so-called copula (see also Copular clauses, Nominal predicate).

Nouns can be classified from various semantic points of view: whether they are abstract or concrete nouns, count or mass nouns, animate or inanimate, inherently unique, inherently relational, etc. These inherent properties may affect their morphosyntactic and distributional properties when they appear in a structure. Grammatical gender is also a typical inherent lexical property of nouns that assign them to (often arbitrary) formal classes. Languages usually have a distinct set of so-called proper names whose denotata are fixed. They can be distinguished from common nouns on formal, distributional criteria as well.

Nouns can assign case (e.g. genitive case) to a possessor expression; furthermore, deverbal nouns can take arguments of their own and assign case to them.

Noun phrases are headed by a noun, and noun phrases usually serve as arguments in a clause (e.g. subjects, objects). Number and case are the most important nominal categories that are usually expressed by morphological exponents.

Number is a typical nominal category that usually surfaces in a morphological marking. Number marking, however, can be semantically restricted or structurally conditioned. For instance, certain lexical classes (e.g. collective or mass nouns) might resist plurality, while others only appear in plural form. The presence of quantifiers may also prevent the head noun from being morphologically marked for number.

Besides plural, many languages have a distinct form to mark duality. Number marking may have allomorphs to encode the so called associative plural (where a certain referent and its associates, e.g. family members are meant), furthermore, the marker can be different if the noun is possessed or is case-marked.

Adnominal modifiers and determiners may agree with the head noun in number and gender. Noun phrase-internal agreement phenomena are referred to as concord. Number and gender marking can also appear on non-agreeing modifiers within a noun phrase, but in this case the head noun is assumed to be ellipted.

Number is a basic grammatical category of nominal expressions. Nouns, or rather noun phrases, can be marked as singular or plural. Many languages also encode dual number.

Marking of associative/familiar plural may show morphological and distributional differences from ordinary plurals, and can often appear on nouns with specific lexico-semantic restrictions (e.g. the noun must refer to a human entity).

Numerals (just as quantifiers in general) contribute quantity information to the noun phrase. Cardinal numerals indicate how many referents the noun phrase denotes, while ordinal numerals identify the referent according to its order in a reference set. Whereas ordinal numerals are usually derived from cardinals, their syntactic properties often differ.

In partitive constructions the head noun's reference is interpreted as being the member or a part of a set of individuals that are already present in the universe of discourse. These constructions are usually specific indefinite noun phrases (see also Specificity).

Partitive constructions can also be conceived as a kind of quantification, although they are typically realized either as a possessive strategy (e.g. three of the dogs), or as an adpositional strategy (e.g. three from the dogs). Depending on the contexts, other quantified expressions can also have a partitive reading (see also Quantification).

Person is a nominal grammatical category which expresses the relationship of the participants of the speech act to the discourse. The speaker is marked by the first person, the addressee by the second person, while third person marks anyone/anything else being talked about that is neither the speaker nor the addressee. The most common grammatical distinction encoded together with person is number (see also Number).

Possession is a relation that can be expressed both on phrase level (see also Possessive construction) and on clause level (see also Predicative possession and Belong-construction). Although the morphological marking of the possessor and the possessee might be identical in these constructions, they usually show crucial differences in other aspects, such as agreement, word order and the referential status of the participant noun phrases. The most relevant semantic properties of possession that might influence its appearance or cause variation between alternative structures are (in)alienability and a potentially distinct encoding of permanent vs. temporary possessive relation.

Possessive construction or adnominal possession is a construction in which the head noun is modified by a noun phrase whose denotatum is a possessor. Possessive construction is typically not restricted to express possessive relation in the sense of ownership, but encode relations such as kinship (e.g. Masha’s sisters), body parts (e.g. Masha’s eyes), part-whole relations (e.g. the bicycle’s wheels), physical or mental state (e.g. Masha’s fever), attribute (e.g. Masha’s patience), and various abstract relations (e.g. Masha’s birthday). Some further special functions can be observed in the so called identifying possessives (e.g. the city of Rome), and in non-referential possessive constructions (e.g. woman's hat).

There is a considerable variation cross-linguistically in marking possessive relationship: some languages only mark the possessed noun, some the possessor, others mark both members or neither of them. These strategies are called head-marking, dependent-marking, double marking, and simple juxtaposition, respectively. When there is a connecting element which does not belong to either of the members, but only marks the relation itself is traditionally called an ezafe-construction. While head-marking can be limited to encode that the head noun is possessed, in many languages person/number agreement also appears on the head noun with the possessor. Pronominal possessors may have distinct forms, these may also agree with the possessed noun in gender and number, and in some languages it is sufficient to mark pronominal possessors merely by an agreement suffix on the head noun. Multiple possessive strategies may co-exist in one and the same languages. Their distribution is often conditioned by a semantic contrast which can be characterized as alienable vs. inalienable possession. Inalienable relationship most commonly involves kinship and part-whole relations.

Possessive constructions not necessarily, but prototypically are definite, since the referent of the possessed noun is existentially presupposed and identified via its relation to the referent of the possessor.

Predicative possession (or HAVE-possessive predication) introduces a new discourse entity and establishes the relation of this entity to another one, which is notionally a possessor (e.g. Peter has a dog.). Predicative possession is closely related to existential and locative clause structure, as it often follows the pattern of existentials with a [+human] location. At the same time, possessors might bear a special case (such as e.g. dative) instead of a locative case, and the neutral word order can also differ from those found in existential and locative clauses in the same language.

Note that some languages predicative possession is expressed by a transitive verb (like English have), or make use of both strategies, the existential one and the one with a transitive verb.

Pronouns form one of the closed classes among the parts of speech, having a more or less fixed inventory and a limited number of members. At the same time, the category of pronoun is used as a cover term for a series of functional words that may have quite different grammatical properties and interpretation. Pronouns are typically single-word expressions, but behave as phrases from a structural point of view. At the same time, their distribution can be nominal, adjectival as well as adverbial.

A fundamental distinction can be drawn between referential pronouns and quantificational pronouns. Referential pronouns include personal pronouns (e.g. I, me), demonstratives (e.g. this, that), possessive (e.g. mine), reflexive (e.g. myself) and reciprocal pronouns (e.g. each other). Deictic pronouns can also have temporal and locative function (e.g. now, then, here, there). The other major group of pronouns consists of quantificational expressions, also called indefinite pronouns (e.g. someone, everything), interrogative and relative pronouns (e.g. who, which).

Referential pronouns, such as personal pronouns or demonstratives pick their referents via deixis, or else, are used anaphorically, referring to entities already mentioned in the discourse. These are used when the speaker assumes that the addressee is able to identify the intended referent without the help of a noun phrase with more descriptive content. The basic semantic content of personal pronouns is limited to number, (gender) and person, while demonstratives may express relative distance beyond indexing. Reflexive and reciprocal pronouns (such as e.g. himself, or each other) are logically and syntactically bound to a local antecedent in the clause structure.

Quantificational pronouns needs a domain of quantification in which they can be interpreted, e.g. universal pronouns, such as everybody, or everything, refer to all members of a group of referents that are relevant in the discourse context.

Pronouns normally occur by themselves without any modifiers. Modification of pronouns are severly restricted, and it is even more holds for referential personal pronouns, than for quantificational expressions. At the same time, most pronouns can be modified by relative clauses and to a limited extent by other (mainly locative) adnominal modifiers as well.

This phenomenon covers cases where certain types of pronouns may be omitted under discourse-pragmatic conditions (the pronoun can be inferred from the context). The term comes from the generative syntactic tradition (see Chomsky 1981; see also Perlmutter 1971, Rizzi 1982). This is a formal approach to the long-standing observation that some languages, e.g, Italian, allow for the possibility of a silent (i.e., unpronounced), referential, definite subject of finite clauses. Such languages are referred to as ‘pro-drop languages’ or, more recently, ‘null subject languages’ (see Roberts 2019: Ch.3), unlike English, for example, which does not allow such subjects and is thus classified as a ‘non-null subject language’. This difference across languages is referred to as ‘null-subject parameter’. Recent literature has further refined the formal typology of pro-drop: a distinction is made between ‘consistent null subject languages’, ‘partial null subject languages’ and ‘radical null subject languages’ (see Roberts 2019: Ch. 3; see also Biberauer et al. 2010, Barbosa 2019). The phenomenon has received considerable attention in the typological literature, too (see Dryer 2013). More broadly, pro-drop is related to null arguments (including null objects), ellipsis and anaphora.

Proper names are nouns with special lexical properties: they are rigid designators in the sense of Kripke (1972). Unlike common nouns, proper names are inherently referential, they refer directly to single individuals, not in virtue of an associated descriptive content. What belongs to the class of proper names in a given language is not obvious, there being no independent lexico-semantic definition for proper names. Proper names, as already definite and unique, normally can only take non-restrictive modifiers (see also Modification). When a restrictive modifier is used with proper names, one of the following interpretations emerge: there are more possible referents with the same name (e.g. Pliny the Younger), or one intends to refer a characteristic subsection/subset of the property set denoted by the otherwise unique referent, e.g. the young Herodotus.

Quantifiers (including numerals) are expressions whose meanings involve the notion of quantity they contribute quantity information to the noun phrase. (Clause-level quantification is not considered here.) In a phrase such as three monks, the cardinal number is the quantifier, and the monks is the quantified expression. Major types of quantifiers are numeral expressions, degree quantifiers (mid-range quantifiers such as many, several, etc.), universal quantifiers (e.g. each, every), existential quantifiers (e.g. something), free choice items (e.g. any), negative quantifiers (e.g. nothing), interrogative quantifiers (e.g. which?). Existential quantifiers denote at least a minimal, non-zero amount or number of the quantified expression (e.g. some water, a boy), while universal quantifiers refer to an exhaustive amount or number of the quantified expression (e.g. all the water, every boy). In between existential and universal quantification lie a variety of mid-range quantifiers: numeral expressions refer to natural numbers, while degree quantifiers indicate that the quantity under discussion is below or above the standard (e.g. many/much, a few, little), or else, it meets the standard (i.e. enough, sufficient). Degree quantifiers usually can be modified by various intensifiers, or further degree modifiers, and also have comparative and superlative forms.

The majority of these elements have both pronominal and adnominal, i. e. modifier use. The latter use is also characterized as determiner quantification. The morphosyntax of quantification is not unitary and no coherent grammatical category of quantifiers can be defined cross-linguistically, it is rather their semantics that suggest that quantifiers constitute a special class of modifiers.

Partitive constructions can also be classified under the notion of quantification, although they are typically realized either as a possessive strategy (e.g. three of the dogs), or as an adpositional strategy (e.g. three from the dogs).

Reference is basically a pragmatic phenomenon: speakers uses linguistic expressions to identify entities in the world. Nouns have denotations, meanings that are inherent in the lexical entries. When nouns are used in discourse, however, they might refer to entities that are identifiable either in the situation context, or in the linguistic context (as already present or just introduced into the discourse). Noun phrases can be referential, when they refer to a referent or a set of referents in the world.

There a variety of linguistic devices to encode referentiality in grammar, the most typical are determiners. (See also Definiteness, Determination.)

Relative clauses are subordinate clauses that modify a noun phrase. The pivot, which is often a noun phrase, is a constituent semantically shared by the matrix clause and the relative clause. Although the relative clause is connected to surrounding material by the pivot, the semantic and syntactic role of the pivot in the relative clause is in principle independent of its role in the matrix clause. For example, in The mouse that I caught _ yesterday was hungry, the mouse is the subject of the matrix clause but it corresponds to the direct object of the relative clause. It has been observed that languages differ with respect to what grammatical roles can be relativized (see the so-called Accessibility Hierarchy, Keenan & Comrie 1977; Comrie & Keenan 1979). Relative clauses can be classified by a number of criteria: (i) the kind of modification (restrictive, non-restrictive/appositive, degree/amount), (ii) the hierarchical status of RC (embedded within a DP, correlative), (iii) the presence/absence of head (headed, free/headless), (iv) the hierarchical position of head (externally or internally headed), (iv) finiteness (finite, participial), etc.

Specificity is a semantic-pragmatic notion. Indefinite noun phrases can be specific or non-specific, depending on interpretation: whether they denote a particular entity, or speak of an arbitrary member of the class described by the noun phrase (see also Indefiniteness). Noun phrases with a numeral, a degree-quantifier or existential expressions are ambiguous between indefinite specific and indefinite non-specific reading, while universal quantifiers are normally specific. More types of specificity are distinguished in semantics. To pick up two such types, in the case of partitive specificity the referent of the indefinite expression is part of a set already introduced into the discourse (e.g. My son's friends came over last night. A boy broke the window.), while in the case epistemic specificity at least the speaker has some knowledge about the referent of the indefinite (e.g. A boy broke the window. I know him, he is Billy Smith.).

Importantly, there are syntactic environments that only allow non-specific arguments. Predicates that assert the being, or coming into being/ appearance of an entity, cannot have a specific subject with an existential presupposition. (The phenomenon is called (in)definiteness effect in syntactic theory.) At the same time, other predicates (typically those expressing a state) require their argument to be definite, or at least specific. Nominal and locative copular clauses usually belong to this kind of predicates. Note that non-specific argument cannot be topicalized.

Under certain circumstances definite noun phrases can also have a non-specific, non-referential reading (Lyons 1999: 165-178).

Tense is understood as a grammatical category that deictically refers to the time of the event or state denoted by the verb in relation to some other temporal reference point. Tense is typically marked on the verb by an affix (‘simple tense’), although it can be expressed with a syntactic construction; the latter is referred to as ‘periphrastic tense’, cf. English goes vs will go (alternative terms: synthetic vs analytic tense). An important distinction is the one between absolute time reference (the event is anchored with respect to the speech time) and relative time reference (the event is anchored with respect to a point established in the discourse; see also Aspect). The former is exemplified by the present, past and future tenses, whereas the latter covers verb forms or periphrastic constructions like the pluperfect (“past-in-the-past”) or the “future-in-the-past”. The expression of tense is closely related to the expression of aspect and mood; thus, they are treated as a tense–aspect–mood (TAM) system.

Topic is a basic notion of Information structure (see also Information structure) The topic constituent is the entity (or set of entities) that the speaker identifies as the one about which the comment is given. This means that the topic is what the sentence is about. Predication (or comment) is the part that says something about the topic. Categorial sentences have a topic-comment/predication structure, while thetic sentences are presentational statements without a topic.

Topics are usually identified by intonational patterns, word order, or even by special markers in some languages. Topics prefer to appear at the periphery of the sentence, and they prefer the left periphery over the right one.