Nominal Morphology[]
Sumerian nouns are composed of a simple (one or two syllable) root and a number of prefixes and suffixes. These markers can convey number, case, and can be used to produce semantic varients (i.e. “king” “kingship”). Unlike IE or Semitic languages, the case markers are placed at the end of a nominal phrase. That is, an entire phrase is treated like a single noun unit.
For instance, if the “son of the king” was the agent who built a temple,
| dumu lugal.ak.e | [(dumu lugal.ak).]e |
“dumu” is the son who built the temple, “lugal” is the king, “-ak” is the genitive marker, and “-e” marks the agent. The entire nominal phrase is dumu lugal.ak, meaning “son of the king.” The agent marker “-e” comes at the end of this phrase. That is, the entire phrase is the agent, not just the unmodified son.
In addition, genitives and other cases can be “stacked” at the end of phrases:
sa’a dumu lugal.ak.ak
“sa’a” means cat. The first “-ak” is applied to the “dumu,” making “cat of the son,” and the last one applies to the “lugal,” making “cat of the son of the king.”
Thus, in long noun phrases, all nouns except the one at the end are unmarked, with the case markings stacking at the end of the phrase.
Noun Chain[]
The sum of the nominal roots and the grammatical affixes is the noun chain. It can be summarized in the following schematic:
| 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Derivational Morphemes | N(1) | N(2) | Gen. | Pro. | Pl. | Case |
Most of these components may be freely withheld or removed, depending on the situation, but the components must appear in this order.
Derivational Morphemes[]
There are only two of these affixes. The first is “niĝ-.” It creates nouns out of verbs:
| ba “to bestow” | niĝ.ba “gift” |
However, the morpheme’s operation is not so simple. First, “niĝ-” seems, in some cases, to induce no change of meaning in the word. Second, the usage can be lexicalized in some words (it no longer operates according to the usual rules and takes on a meaning of its own). Third, it can operate in adjectives as well; it probably changes the adjectives to nouns, but it might sometimes produce a superlative instead.
The second morpheme is “nam-.” It produces abstracts out of nouns:
| lugal “king” | nam.lugal “kingship” |
It can also operate in compound nouns with an unknown usage.
Some Sumerologists argue that these two affixes are not true derivational morphemes and that “niĝ-” simply means “thing” and “nam-” means “to be.” Additionally, they argue that the affix “nu-,” which marks professions, is the only true derivational morpheme. There is some evidence, however, that “nu-” is a dialectical variant of “lu,” which means man, thus disqualifying it as well.
Nominal Roots[]
N(1) and N(2) are the two slots for the nominal roots. The second slot, only in use when there are two nominals in the nominal phrase, is used in the possessive constructions or as an adjective:
| lugal | gal | “great king” | lugal | kalam. | ak | “king of the land” |
| N(1) | N(2) | N(1) | N(2) | Genitive |
Genitive[]
As shown in the second case above, the placement of N(2) is convenient, because it is directly followed by the genitive marking slot. The genitive is not considered a “case” in the sense of the other Sumerian cases, because it does not occupy the slot and it is adnominal, in that it modifies nouns within the noun phrase and does not have a larger role in the sentence. There is another type of genitive construction known as the anticipatory genitive. It only occurs in poetry or other literature (i.e. not inscriptions). The possessor is put at the front of the phrase and the possessum (thing possessed) is suffixed with the third-person singular possessive pronoun:
| lugal kalam.ak | OR | kalam.ak lugal.bi | “king of the land” |
| Normal Genitive | Anticipatory Genitive |
It is impossible to find an adequate translation for the second construction in English, so the normal genitive is assumed.
Pronouns[]
The possessive person pronouns are added to this slot:
| Person | Singular | Plural |
|---|---|---|
| 1st | ĝu10 | me |
| 2nd | zu | zu-(e)-ne-(ne) |
| 3rd an. | ani | an-ne-ne |
| 3rd inan. | bi | bi-(e-ne) |
For instance:
| nin.ani “his lady” | lugal.me “our king” |
Plural Markers[]
The plural markers are in slot six. The default number of any noun is singular. To make it plural, the most common form, only for animate nouns, is “-ene.” There is no plural for inanimate nouns, so they can be singular or plural, depending on the context. Some nouns are singular in form (have no plural marker) but use plural verbal constructions and have a collective meaning (e.g. eren “troops”). There are several other ways of expressive the plural. When a plural noun is modified by an adjective, that adjective is doubled: dumu.gal “great son” dumu.gal.gal “great sons” Sometimes “-ene” is added in addition to the doubled adjectives: dumu.gal.gal.ene In addition, nouns without adjectives can become plurals by doubling. An “entirety” or “totality” or meaning is assumed. Just as in the adjectives, “-ene” may or may not be present: lugal.lugal “all kings” lugal.lugal.ene The difference in meanings is not known.
There are two more plural markers (note that they act as adjectives and not as affixes). “hi-a” means “various” or “of mixed sorts.” Thus, “udu hi-a” means “various small cattle.” “meš” is a form of the word “to be” and has evolved as a “makeshift” plural for the inanimate. It is used primarily in Akkadian texts as a plural marker for sumerograms. Another makeshift plural used in Akkadian is “bi-ene,” a mix of the 3rd person inanimate possessive pronoun and the plural marker. It simply expressed the plural in inanimate nouns.
Case Markers[]
The case marker is the last of the suffixes in the noun chain. There are approximately nine cases in Sumerian, two of which are involved in subject-verb-object relations (the direct cases), six of which are used to produce adverbal phrases (the adverbial cases), and one of which relates nouns (the adnominal cases, there are two if the genitive is counted):
| Case | Suffix |
|---|---|
| Ergative | -e |
| Absolutive | -Ø |
| Dative | -ra |
| Comitative | -da |
| Ablative/Instructive | -ta |
| Allative | -(e)še |
| Locative 1 | -a |
| Locative 2 | -e |
| Equative | -gin |
The ergative marks the agent (subject) of transitive verbs. The absolute marks the patient (object) of transitive verbs and the subjects of intransitive verbs:
| lugal.e iri.Ø mu.n.hul.Ø | “the king destroyed the city.” |
| king.ERG city.ABS destroyed | |
| lugal.Ø i.gin.Ø | “the king went” |
| king.ABS went |
The dative primarily marks the beneficiary of an action (dumu.ani.ra “for his son”), but it can also express the locative with animates (diĝir.ra “upon the god”).
The comitative expresses accompaniment (en.da “with the lord”).
The ablative/instrumental expresses either motion away from (e-gal.ta “away from the temple”) or the means by which an action is achieved (lu.ene.ta “by means of the men”).
The allative expresses motion towards (iri.(e)še “towards the city).
The locative 1 expresses location in or on (e-gal.a “in the temple”) while the locative 2 (aka the “locative-terminative”) expresses proximity (e-gal.e “next to the temple).
The allative and the locatives can also mark the object or goal of a cognitive verb.
There are also a number of lexicalized uses of the adverbial cases. Most animate nouns cannot take the special case endings by themselves. They end up using “makeshift prepositions” and the specific case in combination. These prepositions were once only body parts, but they evolved into markers of position in time and space:
| bar “exterior/because of” | da “side/next to” |
| igi “eye/before” | eĝer “back/behind” |
| murub “waist/in the midst” | šag “heart/inside” |
| ugu “forehead/before” | zag “side/outside of” |
Some words have unknown correlations between the original meaning and the preposition:
| en-na “[unknown origin]/until” | ki “earth/in, from” |
| mu “name/for” | nam- “[abstract prefix]/for the sake of” |
The general structure of a phrase involving these makeshift prepositions is PREP.NOUN.(GEN).CASE
For example:
| bar e.bi.ak.a | because-of ditch.[pronoun].GEN.LOC. “because of that ditch” |
That is, there is a preposition, followed by the animate noun, an optional “bridging” genitive, and the directional case. The directional case may be any of the ablative, allative, or the two locatives.
Gender[]
Sumerian has two genders: the animate and the inanimate. These terms are really just approximations, as the correlation between grammatical and natural genders are imperfect in Sumerian. All divinities and humans are animate while everything else is inanimate. Gender is not marked on the nouns, but it appears in pronouns and verbal referencing.
Pronouns[]
Sumerian marks the main arguments of the verb using verbal affixes; therefore, pronouns are generally scare, although they are used to emphasize a participant or indicate a topic shift.
Personal Pronouns[]
The personal pronouns have only one direct case; it acts as the subject of transitive and intransitive verbs. This is clearly a nominative case (as opposed to the more common ergative); however, no accusative form exists for the person pronouns, so the system is incomplete. The dative, comitative, terminative, and equative cases are also attested: Singular Plural 1 2 3 1 2 3 Nominative ĝa2-e za-e (a/e)-ne me-(en)-de3-(en) me-en-ze2-en e-ne-ne Dative ĝa2-a-ra/ar za-a-ra/ar e-ne-ra e-ne-ne-ra Comitative (a/e)-da za-(a/e)-da e-ne-da e-ne-ne-da Terminative a-(a/e)-še3 za-(a/e)-še3 e-ne-še3 e-ne-ne-še3 Equative ĝa2-(a/e)-gin7 za-(a/e)-gin7 e-ne-gin7 e-ne-ne-gin7
Possessive Pronouns[]
These have been discussed in the verbal chain section.
Reflexive Pronouns[]
The paradigm of reflexive pronouns is very incomplete. Only the absolute forms are known. The root is ni2, to which the possessive pronouns are added. Positional case endings may be added as well: Singular Plural 1st ni2-ĝu10 2nd ni2-zu 3rd an. ni2-(te-a-ni) ni2-te-a-ne-ne
in. ni2-bi ni2-ba/bi-a
Interrogative Pronouns[]
The interrogative pronouns have a couple of strange reversals from the other pronouns. First, the animate is indicated with a ‘b’ (usually the sign of the inanimate) and the inanimate with an ‘n’ (the marker of the animate). In addition, this pronoun class has an ergative-absolute rather than a nominative-accusative split. The interrogative pronouns can occur with certain suffixes, just as personal pronouns and the enclitic copula: Ergative a-ba-(a) “who?” Absolute a-ba “who?” a-na “what?”
Relative Pronouns[]
The closest thing to a relative pronoun are the particles “lu” (“man”) and “niĝ” (“thing”). They are both usually used in the same context as a relative pronoun: lugal lu e du-a “ the king who built the temple”
Adjectives[]
It is likely that all adjectives in Sumerian are, in fact, reduced verbs. There is only a small class of “true” adjectives with no known corresponding verbs. Most adjectives are simple verbal roots with a “nominalizer” suffix “-a.” Some adjectives have a null ending. The distribution of these two is not understood. As mentioned before with the derivational affixes, adjectives can be modified by the prefixes “niĝ” and “nam.” However, the meanings are not completely understood.
Adverbs[]
These are formed from both verbal and nominal roots with the suffix “-bi.” Sometimes this follows the nominalizing suffix “-a”: gal.bi “greatly” dug.bi “tenderly” ul.a.bi “rapidly” There is also the suffix “-(e)še,” homonymous with the allative case ending, creates adverbs having to do with manner: u.de.eš “as the day” gal.eše “grandly”
Modal and Temporal Adverbs[]
i3-gi4-in-zu moreover, what’s more i3-ge-en truly, in fact a-na-aš-am3 / a2-še3 how is it (that) a-da-lam (but) now i3-ne-eš3 now
Interrogative Adverbs[]
These are composed of the stem “me(n)” and a directional case suffix: me.a “at where?” me.še “to where?” me-na-am3 “when?”
Verbs[]
The verbal system is the most controversial aspect of the reconstructed Sumerian grammar. The Sumerian verb is composed of a root and a chain of various affixes. The Sumerian verb has the peculiar quality of cross-referencing all of the adverbial components of the sentence, as well as the agent and patient. ...