The grammar of the Hittite language has a highly conservative verbal system and rich nominal declension. The language is attested in cuneiform, and is the earliest attested Indo-European language.
Hittite distinguishes between two genders, common (animate) and neuter (inanimate). The distinction between genders is fairly rudimentary since it is made only in the nominative and accusative case, and the same noun is sometimes attested in both genders. It is still debated whether or not this reflects a prehistoric merger of inherited PIE masculine and feminine into a single common/animate gender or an archaic system in which there was already a common gender. Nouns referred to living beings (humans, animals and Gods) are usually found in the common/animate gender, but some inanimate objects actually have the common/animate gender. For instance, nouns in a-stem and t-stem are common/animate and, given how productive was the formation of words in the a-stem and t-stem, many words in Hittite indicating inanimate objects are actually in the common/animate gender in the nominative and accusative. The other gender, the neuter/inanimate, is referred to objects, including parts of the body, and abstract concepts or collective nouns, e.g. "family, assembly, troops, humanity". Some common examples of neuter declension are the u-stem nouns and the nouns formed by the suffixes -ÃÂtar, -eà ¡à ¡ar and the suffix for collective nouns -a(i)-. Words derived by common/animate gender roots through neuter suffixes are neuter.
The only reference to a female gender, which however does not erase the two-gender system "common-neuter gender", is the infix -(à ¡)à ¡ara-, used to indicate female gender for humans and deities.
The nominal system consists of the following 9 cases: nominative, vocative, accusative, genitive, dative-locative, ablative, ergative, allative, and instrumental.
Hittite declension system also distinguishes between two numbers (singular and plural) and shows indirect traces of a dual number; due to syncretism, the ending of ablative and instrumental in the plural coincide.
Hittite language is based on split ergativity: when a common/animate noun is the subject of a transitive verb, e.g., "The child eats the apple", the subject is marked by the nominative case. By contrast, when a neuter noun is the subject of a transitive verb, e.g., "The spear kills the soldier", the subject of the sentence is marked by the ergative case; hence, only neuter nouns show the ergative case in their declension, which means that common/animate nouns show 8 cases, while neuter nouns show 9 cases. Inflected adjectives always have the ergative case in their declension, but this case is used only when an adjective is referred to a neuter noun in the ergative case, i.e., followed by a transitive verb; consequently, adjectival declension shows 9 cases. Personal pronouns as the subject are always in the nominative case; the subject of an intransitive verb always take the nominative ending as well.
The basic scheme of suffixation is given in the table below, which is valid for almost all nouns and adjectives. The sample word shown is antuá¸«à ¡aà ¡, "man", a-stem noun (common/animate gender, thus a name without the ergative case). The letter "à ¡" is always pronounced as /s/, while "z" is always pronounced /ts/ and derives from an ancient */ti/ or */tj/. For instance, in mi-verb declension, the 3rd person singular ending in the present tense -zi according to Hoffner and Melchert comes from an earlier *-ti.
Given that the nominal root ends with the thematic vowel -a and some suffixes starts in a vowel, the final -a in the root is elided (e.g., attaà ¡, "father" > dat-loc. atti, not *attai). Only in some rare instances that represent an innovation, a dat-loc. in -ai is found; some of these instances are foreign words and foreign names.
Neuter nouns in the accusative singular take -n only if the thematic vowel is -a-, e.g., yukan (plough). All the other neuter nouns take -a (sometimes indicated as the "zero-ending" -â ). As for the neuter accusative plural, names belonging to the common/animate gender take -uà ¡, while names belonging to the neuter gender take -a. Only names in i-stem take the dative-locative in -ya. Vocative and allative case have no plural counterpart neither in Old Hittite (OH), nor in Middle Hittite (ME) and New Hittite (NE). As already stated, due to syncretism, the ending of ablative and instrumental in the plural coincide. The instrumental case has two possible endings (-it, -et) and, according to Hoffner (2008), -it is the oldest ending; it developed from an earlier ending -t through anaptyxis/ephentesis in environment of a preceding dental consonant. Some tracks of this presumed ending can be found in irregular instrumental endings, in names such as genu (knee) > genut, à ¡ÃÂkuwa (eye) > à ¡ÃÂkuwat.
In MH, two more plural suffixes were created for the nominative and accusative (nom. -eà ¡, -uà ¡, -aà ¡; acc. -uà ¡, -eà ¡, -aà ¡). Then, the allative and instrumental cases both merged with dative-locative -az(a). The use of old allative -a and old instrumental -it in NH are archaisms. In NH, all the three plural suffixes for nouns of common gender in the two strong cases (nominative and accusative) collapsed into -uà ¡, with only some exceptions. In OH, the original suffix of the genitive plural is -an; then, a new suffix -aà ¡ was coined in Late OH and displaced -an in NH.
Adjectives in a-stem share the same endings of noun declension; their gender depends from the gender of the noun they refer to (common/animate or neuter).
In the following examples, all forms not directly attested are put between brackets.
The adjective chosen is araḫzenaà ¡, araḫzenan ("external"), inflected for both genders.
Hittite is a head-final language, with subject-object-verb word order.
Hittite syntax shows one noteworthy feature that is typical of Anatolian languages: commonly, the beginning of a sentence or clause is composed of either a sentence-connecting particle or otherwise a fronted or topicalized form, and a "chain" of fixed-order clitics is then appended.
Yes-no questions were marked using prosodic features such as rising intonation. In writing, they were partially left unmarked: Scribes in Assyria and Babylonia who wrote Akkadian in cuneiform script (and later Hittites as well) sometimes indicated the interrogative intonation by a plene spelling of the vowel in the final syllable of the central word in the interrogative clause. Nevertheless, the use of this device in Hittite was rare and probably not codified.
When compared with other early-attested Indo-European languages, such as Ancient Greek and Sanskrit, the verb system in Hittite is morphologically relatively uncomplicated. There are two general verbal classes according to which verbs are inflected, the mi-conjugation and the ḫi-conjugation. The names are drawn by the ending of the first person singular in the present tense. Rose (2006) lists 132 ḫi-verbs and interprets the ḫi-mi oppositions as vestiges of a system of grammatical voice ("centripetal voice" vs. "centrifugal voice").
There are two voices (active and medio-passive), two moods (indicative and imperative), two aspects (perfective and imperfective), and two tenses (present for the present and future time and preterite for the past time); the difference between the present and the future can be disambiguated through an analytical construction. Additionally, the verbal system displays two infinitive forms, one verbal substantive, a supine and a participle.
Modality (e.g. "could, would, should, must") in conveyed in Hittite through modal particles and adverbs instead of modal verbs. Both verb class have some verbs that contain ablauting vowels according to fixed patterns.
The basic conjugational endings are as follows:
The ending -meni in the present tense is confined to nu-causative verbs; the ending -weni is identical in the present and imperative tense, thus the context helps disambiguating between the two tenses. In the preterite, -un is used with vocalic stem, i.e., a verbal root that ends in a vowel, which is then deleted. In the preterite, the ending -à ¡ta for 2nd person singular is used in a specific group of verbs, while the ending in -t is used in i-stem verbs; the ending -ta in the 3rd person singular is used in the i-stem verbs too.
In the imperative, the ending in -t is used in nu-causative verbs.
The infinitive -wanzi comes from a pre-Hittite locative in *-wenti or *-wonti showing vowel mutations and *-ti > -zi mutation; the supine -wan comes from a pre-Hittite locative without the final part of the ending, *-wen or *-won.
The Set I endings are default; the Set II endings are taken primarily by monosyllabic ablauting mi-verbs. Within Set I verbs, the Ib endings are taken by stems ending in -u.
A simple example of conjugation in the present tense is ḫarzi ('to have, to hold'); the verb belongs to the mi-conjugation verb class and is non-ablauting:
Hittite radical verbs, that being verbs formed via the affixation of the inflectional endings directly to the root, are sometimes marked by a mobile stress pattern wherein stress is applied to the root in the singular and to the suffix in the plural. For instance, according to the linguist Anthony Yates, the Hittite -verb was perhaps realized as in the third-person singular non-past active, whereas its third-person plural form was potentially realized as . Likewise, the -conjugation verb was perhaps realized as in the third-person singular, and the third-person plural form may have been realized as . However, other radical verbs display a type of fixed accent, such as , which was perhaps realized as in the third-person singular, whereas the third-person plural form was perhaps realized as .
According to Yates, synchronically, within Hittite, certain roots were inherently stressed whereas others lacked stress by default. When naturally unaccented roots such as were affixed with the likewise unstressed singular inflectional endings, the accent fell upon the leftmost syllable (i.e. the root). However, when these same unaccented roots were affixed with the naturally stressed plural forms, the accent was retained in the inflectional endings. In the case of roots with fixed accents, such as , due to a phonotactic prohibition against words with multiple stress placements, their interaction with the fixed-accent plural endings required the deletion of one stress marker. Ultimately, according to Yates, in such a scenario, the accent was preserved on the leftmost syllable, thus explaining the fixed-accent pattern for certain radical verbs. When an unstressed root was modified with an affixed marked by a fixed accent, such as , the stress always remained on the suffix. The interaction between the underlying accent in the suffix and plural endings was resolved via the leftward shift of the stress towards the affix. If these same affixes were applied to inherently stressed roots, the accent always shifted leftwards. Thus, the root , when combined with the affix and the ending , produced the form ().
The ablauting Hittite -verbs were inherited from the Proto-Indo-European root present and aorist, both of which display ablaut in the root. This original PIE ablaut paradigm yielded several different patterns of vowel gradation in Hittite:
There are examples of various radical -verbs that lack synchronic ablaut, though they may historically derive from originally ablauting paradigms. In certain cases, one ablaut grade was generalized to the entire term paradigm. For instance, Kloekhorst argues that the non-ablauting verbs ("to wipe") and ("to wipe") once formed a singular ablauting pair, itself reflective of Proto-Indo-European . Other non-ablauting Hittite verbs probably emerged due to the action of sound laws rendering both the strong and weak stem of the Proto-Indo-European term identical. For example, both stems of the originally ablauting aorist would yield the Hittite form , which is reflected in the non-ablauting verb ("to set fire to"). Besides radical verbs, there are also non-ablauting terms formed via certain affixes, such as fientive suffix , or the denominative suffix .
The negation adverb is natta ("not"); nÃÂwi translates "not yet", while lÃÂ translates "don't...!" in orders and prohibitions ("imperatival negative"); lÃÂ can be used with the imperative or, in NH, with the present with an imperatival negative meaning. Another use of lÃÂ is the "categorical negative", an emphatic negation in an obvious context which can be translated as "certainly not", e.g., "A blind man certainly doesn't see, a deaf man certainly doesn't hear, a lame man certainly doesn't run".
Negative adverbs are usually put right before the verb, in pre-verbal position; in rhetorical questions and emphatic questions, natta put at the beginning of a sentence, before the subject (if expressed).
The verb is conjugated in the present tense and belong to the mi-conjugation verbal class. In the first person plural, no *asweni or *esweni form is attested yet. The verb "to be" can be omitted in the present tense in sentences containing a predicate nominative (e.g., "I am a man"), thus creating a nominal sentence. Consequently, adjectives referred to the proedicate nominative take the nominative case as well.
Personal pronouns are inflected according to their case. They have an enclitic version as well, which can be used as a direct or indirect object (acc-dat). The third person has only the enclitical version and distinguishes between common/animate gender and neuter gender. The following table contains the nominative cases of all pronouns and the enclitic form (acc-dat).
Possessive pronouns have an earlier enclitic version and a later full/analytical version placed before the noun. Possessive pronouns are inflected according to cases and take the gender of the noun they refer to. The following table shows the nominative form:
Demonstrative pronouns are put before the noun; they are inflected according to the case and the gender of the noun they refer to. Hittite has a three-way system to indicate position: near to the speaker, near to the listener and far from both ("here-there-yonder").
The pronouns "this, that" in the nominative singular are kÃÂà ¡ and apÃÂà ¡; their plural is kà(later kà «à ¡) and apà(later apà «à ¡). Their neuter counterparts are kë and apÃÂt, plural kàand apÃÂ. Adverbs "here, there" (kÃÂ, apiya) are derived from demonstrative pronouns as well as kinun and apiya ("now, then/in the past").
Both ordinal and cardinal numbers in Hittite were often written with ciphers instead of syllables, which makes both the reconstruction of their pronunciation and their translation in context difficult. Hence, most number are indicated by the Arabic cipher and their ending, e.g. "one" in the nominative common/animate gender is known as "1-aà ¡". Number "one" was reconstructed by Goedegebuure (2006) as *à ¡ia-. Numbers from one to four are declined in Hittite.