Noun Phrases

20.1 The Structure of the Noun Phrase

The Alashian verb phrase has the following overall structure:

[demonstrative] [quantifier] [adjective(s)] [noun] [genitive phrase] [adjective(s)]

Proto-Semitic appears to have been fairly strongly head-initial, as are most of Alashian's closer relatives. However, modern Alashian has a decidedly mixed structure, with demonstratives and quantifiers preceding the noun and genitive phrases following it; adjectives may come either before or after the noun, depending on its state.

20.2 State

20.2.1 Construct State

The construct state marks the head of a genitive construction. Construct forms historically appear to be variants of non-construct forms that appeared due to the different stress pattern of genitive constructions versus the noun in isolation elsewhere; in genitive constructions the stress pattern of the head and modifiers tend to become more closely connected, which over time can result in divergent phonetic development.

In modern Alashian, the construct serves two primary purposes: to form 'compounds' (i.e., tightly-bound endocentric or exocentric noun-noun compounds with distinct lexical meaning) and to form possessor-possessee constructions with inanimate heads.

Construct 'compounds' are similar to noun-noun compounds in Indo-European languages. The head appears in the construct state, followed by the modifier in either the absolute or determinate states. These compounds are pluralized by pluralizing the head (e.g., yūbil hān “bus” (lit. “carrier of people”) → yūbilē hān “buses” (lit. “carriers of people”)), although in casual usage it is not uncommon for some more frequent compounds to be treated as though they were a single word (e.g., yūbilannien “buses”, as though the singular were yūbilān).

Constructs are also the standard means of expressing looser genitive relationships whenever the possessor is non-human. Constructs with an animate possessor can be seen in archaic language (particularly in the Bible), but such phrases that are still in common usage tend to be viewed more as compounds nowadays than as possessor-possessee constructions.

Multiple constructs may be chained together to create complex genitive phrases, in which cases all nouns other than the last one appear in the construct state.

If the head is a compound noun phrase (conjoined with υε- ve- “and”), only the last noun will appear in the construct state; all others will be in the absolute state, even if the construct as a whole is definite.

Note that if the construct noun ends in a short vowel and the following word begins with either a vowel or the definite prefix ha-, the short vowel will drop: τζέδρε čedre “center” → τζέδρε μεδινώ čedre medinā “center of a city” → τζέδρ αμμεδινώ čedr (h)ammedinā “center of the city”.

20.2.2 Determinate State

The determinate state marks a noun as being definite. It can appear in any syntactic role other than the head of a genitive construction.

When a noun in the determinate state follows a noun in the construct state, the entire construct is made definite.

The determinate prefix ha- becomes n- when the nominal stem begins with /ʔ/ or /h/. However, this n- becomes 'un- when following a noun in the construct state that ends in a consonant or a phrasal/nominal preposition ending in a consonant, or 'an- when following a primitive preposition ending in a consonant.

After the clitic prepositions λι- li-, βι- bi-, and χι- xi-, the prefix ha- becomes -a- and n- becomes -an-.

All proper nouns are considered to be in the determinate state, even if it is not explicitly marked with the determinate prefix ha-/n-. Personal names are always marked, hence why nearly all Alashian names appear to begin with (h)a-. If both a first name and last name are given, the determinate prefix is only added to the first word.

Country names, although always determinate, are usually not marked in standard Alashian, although in colloquial/dialectal usage they often are. However, if the country name follows a noun in the construct state, it will always appear explicitly marked.

Generic nouns, and especially generic abstract nouns, usually appear in the determinate state, whereas in English they will be indefinite.

20.2.3 Partitive State

The partitive state is unique to Alashian, not found in the other Semitic languages. It ultimately derives from the Old Alashian preposition min “from” (Proto-Semitic *minay-), which cliticized to the following word ¹ and eventually was reanalyzed as a prefix rather than a preposition. It serves a number of different functions.

By itself, the partitive state indicates a partial or indefinite quantity, usually translated as “some”. This meaning is most often seen on direct objects or in existential expressions; subjects in the bare partitive state are quite rare, instead employing a construct with riḏmuos “a number of”.

However, a partitive state noun in a possessive construction may appear in any position. The construction Y-part of-X is translated as “some of X's Y”.

The partitive is also consistently used after a number of contruct quantifiers: κάλ kal “all [of]”, κάν kan “how many?”, φάλγε falge “half [of]”, ήμα 'ēma “any [of]”. If the quantifier is a concrete noun (e.g., “a cup of tea”), the partitive is optional. The partitive is not used with σιλυλλή silullē “many”, however.

In what is clearly a frozen usage from when the partitive prefix still meant “from”, the standard of comparison in a comparative construction (i.e., the 'Y' in 'X is bigger than Y') is expressed using the bare partitive state.

If the standard of comparison (or some other partitive construction) requires a nominal construct, the form is more interesting. This 'partitive construct' behaves like the 'compound' forms described above in section 20.2.1, so that the entire construct phrase behaves like a single noun in the partitive state. Notice below, for instance, how the construct ληλούτ αζζάμαν lēlūt hazzaman “summer nights” takes on the partitive form μιλληλούτ αζζαμανιήν millēlūt hazzamanien, as though the partitive prefix mi- and plural suffix -ien were being added to the 'stem' lēlūt hazzaman.

20.2.4 Absolute State

The absolute state is used for all nouns that are not construct, determinate, or partitive. They are inherently indefinite.

When a noun in the absolute state follows a noun in the construct state, the entire contruct is indefinite.

The absolute state is the citation form of nouns. This is the form nouns will appear in when completely devoid of syntactic context, such as in dictionaries or on signs.

20.3 Number

Alashian has two productive numbers: singular and plural. However, this simple model is complicated by existence of four different types of plurals: external, internal, dual, and gentilic-plural.

The singular is used for lone nouns as opposed to multiple nouns. It is also used for mass/uncountable nouns (known as “singularia tantum”, or 'singular only') and many abstract nouns. Such singular nouns take singular agreement on verbs and adjectives.

In addition, the singular is used with nouns modified by a numeral more than ten. Here, however, adjective and verb agreement is plural, keeping in line with the phrase's semantics.

The external plural, consisting of a suffix such as -ien or -uoš in the absolute state, is the most common means of forming plural forms. Nouns with external plurals always take plural adjective and verb agreement.

A handful of nouns, known as “pluralia tantum” or 'plural only', have an external plural but no singular, even though they refer to a singular object. These nevertheless always take plural agreement.

When accompanied by a numeral between two and ten, the external plural will be used.

The dual, marked by -ī in the absolute state, is a special type of external plural used with nouns that commonly occur in pairs. The dual was once a distinct number in Alashian, separate from the singular and plural, but in modern Alashian it is simply a relic ending used to mark the plural on certain nouns. Just like other external plurals, nouns with dual endings take plural agreement on both adjectives and verbs.

Some nouns have both a dual form and an external plural, but always with distinct semantics, with the dual form typically preserving the original meaning and the plural having acquired a metaphorical one. For instance, the noun ιάδ yad “hand” has the dual form ιαδεί yadī “hands” and the plural ιαδιήν yadien “shares, lots”.

The internal plural (also called a broken plural, discontiguous plural, or collective plural) is morphologically singular but semantically (and sometimes synactically) plural. Internal plurals have no explicit plural markings. If an internal plural refers to a human, it takes plural adjective and verb agreement, just like other plural forms.

If the internal plural refers to anything other than a human, attributive adjective and verbal agreement is feminine singular, while predicate adjective agreement is plural and matches the gender of the noun in the singular. Pronouns referring to such a noun may be either feminine singular or plural, with a general preference for plural forms.

Internal plurals are not used alongside numerals. Even if the usual plural form of a noun is internal, it will always appear as an external plural when modified by a numeral between two and ten.

The gentilic plural is a special suffix -ēn that forms the names of many different nations and peoples. Such nouns always take feminine plural agreement in both adjectives and verbs.

20.4 Case

20.4.1 Case in Nouns

Although Proto-Semitic appears to have had a well-developed case system with a three-way nominative-accusative-genitive contrast, by the time of the earliest extant Alashian texts this system had nearly completely been lost, other than some debatable instances in texts that were already deliberately archaizing. Modern Alashian now makes no morphological case distinctions in its nominal system, although a few syntactic traces of this older system remain.

Like a number of the Canaanite languages, Alashian has a nota accusativi, a specialized preposition serving as a marker of the accusative case. It only appears when the direct object is human; otherwise the direct object is unmarked. The nota accusativi is τα ta before a consonant and τ' t before a vowel or the definite prefix ha- (the combination of which, t ha-, is pronounced /tʰa/).

Proto-Semitic used the bare accusative as an adverbial marker of time and destination. In modern Alashian, this survives as the use of bare, prepositionless noun phrases that serve as duratives. If definite, they may also mark when an action occurred or will occur.

The accusative of destination does not have an unambiguous descendent in modern Alashian ² . However, a number of 'transitive' verbs of motion can been seen in older texts, where an original directional adverb was reinterpreted as an actual direct object. Such usage is no longer encountered in the modern language. The verb γ̄άλαλ ǧalal “enter” seen in the citation below requires the directional preposition ιλ 'il “towards” in the modern language.

The epenthetic *a- and *u- that can appear in between the two nouns of a construct or a preposition and its object are the only morphological remnants of Semitic case marking (the accusative and nominative cases respectively). The tightly-bound nature of these two environments encouraged the preservation of the intervening case marker while they were entirely lost in all other positions.

20.4.2 Case in Pronouns

The Alashian pronominal system has a more robust case system, with direct reflexes of the Proto-Semitic nominative and possibly accusative ³ pronouns and an innovated genitive pronoun series based on the preposition λι- li- “of” (originally, “to”).

The nominative case is used for the subject of a sentence as well as for resumptive pronouns.

The accusative case is used for the direct object of a sentence, as well as after prepositions. When an accusative personal pronoun follows a preposition, it carries an emphatic sense (in contrast to the more typical declined prepositions); there is no such emphatic meaning for other pronouns.

Genitive pronouns mark possession, and can be used in either attributive or predicate position. They always follow the noun they modify. In predicate position, it is more common to hear the 'full' declined forms of li- rather than the short forms.

Disjunctive personal pronouns (used in situations devoid of syntactic context) consist of the full accusative forms in the first and second persons and the full nominative forms in the third person. This includes dislocated/emphatic pronouns, clefts, and verbless elliptical constructions.

These disjunctive pronouns are also the basis of the 'pseudo-partitive' used when the standard of comparison in a comparative construction is a pronoun or after quantifiers requiring partitives. This merely consists of the mi- prefix added directly to the pronoun form with appropriate gemination at the boundary. Non-personal pronouns use the accusative as a base.

20.5 Adjectives

20.5.1 Ordering

Typically, an attributive adjective in the absolute state will appear before the noun it modifies, while an attributive adjective in any other state (determinate, partitive, construct) comes afterwards.

However, any adjective modifying a nominal construct (whether the whole construct, the head, or the modifier) must come after the entire construct. It can thus sometimes be ambiguous what exactly the adjective is modifying if other agreement clues (e.g., gender) are not able to help.

Adjectives that are modifying any sort of pronoun always come afterwards, even if the pronoun has an indefinite meaning.

Adjectives that are attached to their head with the intervention of some modifier, such as the comparative κιυ kyu “more”, always follow the noun. In addition, such adjectives always appear in the absolute state, regardless of the state of the head noun.

20.5.2 Construct Adjectives

Alashian allows attributive adjectives to appear in the construct state and take their own genitive modifier, in effect allowing a nominal construct to function as an adjective. These are often equivalent to English compound adjectives such as “dark-haired”, “two-winged”, “good-hearted”, “low-fat”, “oil-rich”, etc., but also to some phrasal adjectives such as “full of smoke”.

Construct adjectives agree their their head in gender, but appear in the construct state rather than whatever state their head is in. The noun that is in turn modifying the adjective will appear afterwards, typically in the absolute state.

In colloquial usage, it is not uncommon to hear construct adjectives replaced by a normal adjective plus a prepositional phrase beginning with βι- bi-.

20.5.3 Secondary Predicate Adjectives

Secondary predicate adjectives are adjectives that describe the state of the another noun while the action of the primary predicate is taking place, as in “He came home tired”. In Alashian secondary predicate adjectives behave in much the same way as typical predicate adjectives: they show gender and number agreement, but are always in the absolute state. The only difference is that secondary predicates must always be preceded by the preposition βι- bi-.

20.5.4 Adjectives versus Stative Verbs

There often exists some semantic overlap between basic Alashian adjectives and their cognate stative verbs (e.g., the adjective τζείλ čīl “cold” versus the verb τζείλ čīl “be cold”, or ραττούβ rəthūb “wet” versus ράτταβ rəthab “be wet”). The choice of forms generally comes down to the animacy of the noun: animate nouns prefer stative verbs (or stative participles attributively), while inanimate nouns prefer adjectives. Naturally, this distinction only applies for concepts that have both an adjective and a stative verb.

However, secondary predicates never employ stative verbs when there is an adjectival equivalent, as could be seen in example 111, where the adjective κάβιδ kabid “tired” was used rather than the verb κάβαδ kabad “be tired”.

20.5.5 Numerals

Numerals constitute a special subclass of adjectives with their own unique behavior. In the modern language ⁴ , they do not agree in gender or number with the noun they modify. In fact, they generally do not modify nouns at all; they form adjectival constructs, with the quantified noun forming the genitive component of the construct. That is, a numeric construction like θάλυττετ εισιήν ṯaluttet 'īsien “three people” is grammatically identical to an adjectival construct like τήβ λιήβ tēb lieb “good-hearted”, except that the numeric construct is generally headless, while the adjectival construct is more often attributive (though the headless form τηβιήν αλλιήβ tēbien hallieb “the good in heart, the good-hearted [people]” does exist as well, but is far less common).

As constructs, numeric expressions can be made definite by switching the modifying noun to the definite state.

Only the numerals '1' and '2' are somewhat exceptional. '2' is also a construct, but unlike other numerals, it shows gender agreement. '1', however, is a true attributive adjective, and so shows full gender, number, and state agreement, and also follows normal adjective ordering rules.

Numerals can also serve a pronominal function, as in “the two of us”, “the three of you”, “the four of them”, etc. These sorts of quantified pronouns can be formed in several different ways.

For low numbers (primarily 2-6, although it is grammatical for numbers up to 10), possessive suffixes may be added directly to the construct form of the numeral. These forms take verbal agreement matching the possessive suffix, so that θινείν/θαττείν ṯinīn/ṯəthīn “the two of us (m/f)” will take first person plural agreement, for example.

Another means of forming quantified pronouns which can be used with any numeral is to take the numeral in the determinate state (i.e., the 'counting' form with the definite prefix added) and follow it by a genitive pronoun. This sort of construction will always take third person agreement.

20.6 Pronouns and Proforms

20.6.1 Personal Pronouns

The full-form nominative and accusative pronouns are generally only seen in three circumstances: disjunction (discussed above in section 20.4.2), marking the arguments of the copula в̄εί vī, and marking emphasis.

If the subject or complement of в̄εί vī “be” is a pronoun, it will always appear in its full form, never as a clitic.

If the subject of a verb is emphasized, the full-form nominative pronoun may appear in addition to the clitic.

If the direct object of a verb is emphasized, the full-form accusative pronoun may appear instead of the accusative clitic. This may cause the nominative clitic to reappear in its place, however; compare the following two sentences, one with a non-emphatic object and one with an emphatic object.

Full-form pronouns are also always used when an adverb is modifying the pronoun. This can be considered a type of emphasis.

20.6.2 Reflexive Pronouns

For the majority of native Semitic roots, reflexive verbs are formed simply by conjugating a root according the nitkatab paradigm. Nitkatab is inherently reflexive, and so requires no external marking of a direct object.

However, newer verbs following the concatenating European-type paradigm do not have reflexive forms, nor do quadriconsonantal roots of any origin. These form reflexives periphrastically, by using the reflexive pronoun as the direct object. The reflexive pronoun behaves syntactically as a noun, so it almost always follows the verb and a nominative clitic usually accompanies the verb.

However, the reflexive pronoun may be used with nitkatab verbs for emphasis. When this is done, the pronoun must be preceded by the nota accusativi τα ta, even though it never appears with the reflexive pronoun elsewhere; ta appears to be serving a transitivizing role here, since nitkatab verbs are normally unable to take direct objects.

20.6.3 Possessive Pronouns and Suffixes

The main purpose of the genitive pronouns is naturally to indicate possession. However, they also serve a number of dative functions, hence why they were first introduced as the genitive/dative series. In particular, the genitive/dative pronouns are used to mark the indirect objects of verbs of mental state such as “seem”, “look [like]”, “sound [like]”, “interest”, and so on.

Curiously, with these same verbs, if the indirect object is a noun rather than a person pronoun, the preposition χι- xi- is always used, never λι- li- as the pronoun would suggest. For instance, modifying example #135 above, the sentence “Mathematics doesn't interest Assufkyā at all” would be rendered Αμμαθιματικώ λω ι'εντερεσεί Χασσυφκιώ βεήμα μυώδ Hammaṯimatikā lā y'enteresī Xassufkyā be'ēma muod.

The possessive suffixes are used to indicate the possessor of a handful of common nouns, particularly kinship terms. However, they are incompatible with nouns that have any other modifier such as a numeral or adjective; in such an environment, genitive pronouns must be used instead.

20.6.4 Demonstrative Pronouns and Adjectives

Demonstrative pronouns are also fairly noun-like syntactically, in that they cannot ever take the place of clitic pronouns as, for instance, personal pronouns can. They always take third person agreement. The masculine singular is the default form used with an indeterminate referant.

Demonstrative adjectives, however, have unique behaviors. The noun they modify cannot be in the absolute state; by default they must be in the determinate state. They always precede a noun in the determinate or construct states, unlike typical adjectives, but follow nouns in the partitive state.

20.6.5 Resumptive Pronouns

Alashian makes frequent use of two types of resumptive pronouns, copular and relative.

In copular sentences containing the verb “be”, resumptive pronouns are often used whenever the subject noun phrase is considered 'heavy'. More specifically, they may be used when the subject NP:

• Includes a relative clause.
• Is modified by any postposed adjectives (but not if the adjective(s) are exclusively preposed).
• Is compound (i.e., includes the conjunction υε- ve- “and”).
• Is a determinate construct (but not if the construct is indeterminate, unless adjectives are postposed).

The resumptive pronoun simply consists of a third person personal pronoun or demonstrative pronoun matching the subject's gender and number. Personal pronouns are used if the subject is animate, demonstrative pronouns if inanimate.

Resumptive pronouns are also always found in embedded clauses whenever the embedded clause is the object of a preposition in the matrix clause. The Alashian general purpose subordinator δε de cannot be preceded by a preposition, so the preposition appears in the embedded clause along with a resumptive personal pronoun or prepositional inflection. An expression such as “the room that I am in” is thus rendered “the room that I am in it”.

20.6.6 Pseudo-Pronouns

The use of possessive suffixes with quantifiers or other non-nominal words results in what are termed 'pseudo-pronouns', nominal constructions that have pronominal meaning. One type, numeric pseudo-pronouns, was previously discussed in section 20.5.5. This same formation can be used with other quantifiers, such as κάλ kal “all [of]”, σιλυλλή silullē “many [of]”, and κάν kan “how many [of]”. The syntax is the same as with numerals: forms with possessive suffixes show verbal agreement matching the possessive, while forms with genitive pronouns always show third person agreement.

In formal language, pseudo-pronominal forms of είθ 'īṯ “there is” and λήθ lēṯ “there is not” are used as general affirmatives and negatives: ειθεί/είθ λιή 'īṯī/'īṯ lie “Yes, I am/did/will”, λήθικ/λήθ λάκ lēṯik/lēṯ lak “No, you aren't/didn't/won't”, etc. These forms are rarely used in colloquial speech other than in one frozen construction consisting of these + a present participle, emphasizing that an action is taking place right at the present moment or that is going to take place momentarily.

20.7 Prepositions

In many respects, Alashian prepositions are construct-like, but not necessarily actual constructs. In Old Alashian, there were two classes of prepositions: the clitics—βι- bi-, λι- li-, χι- xi- as in the modern language, plus κα- ka- “like, as”—and the constructs. Over time, however, and aided by the development of ADJ-N word order, a third class emerged—the modern primitive prepositions—which have more in common with the clitics than with the constructs.

Construct prepositions, in both Old and Modern Alashian, form a true nominal construct with their direct object, and this entails all of the usual syntactic properties of constructs; for instance, head-first N-ADJ order is mandatory, since Alashian does not permit adjectives to intervene in the middle of a construct. In contrast, the clitic prepositions and the modern primitive prepositions allow for an unrestricted noun phrase as their object; that is, all of the same rules that govern independent noun phrases, such as the subject or object of a verb, also apply to the object of one of these prepositions. Thus both head-first and head-final word orders are allowed, and even adverbs can intervene between the preposition and object.

The clitic prepositions behave just as primitive prepositions syntactically, the only difference being that they cannot occur independently. They are always prefixed to the first word of the noun phrase constituting their object, whether it be a noun, adjective, or adverb.

Prepositions do not allow personal pronouns as their objects without some modification. As previously discussed, clitic and primitive prepositions may decline (a remnant of an older clitic pronoun that fused to the preposition), while the construct prepositions incorporate themselves into possessive formations, either with genitive pronouns or with possessive suffixes. For emphasis, clitic prepositions may take their special emphatic/extended forms and primitive prepositions may be followed by disjunctive pronouns; construct pronouns have no form marked as emphatic, although possessive suffixes are markedly non-emphatic.