Cuneiform script and Sumerian writing

When people living in southern Mesopotamia (now southern Iraq) toward the end of the 4th millennium BC created documents by inscribing cuneiform characters on clay tablets, they took the first step in the history of writing. The language represented by those characters is thought to be Sumerian, on the basis of writings from a slightly later era. The excavation and deciphering of Sumerian writing began with European and American researchers in the latter half of the 19th century, and a Sumerian grammar was published already in the 1870s. Every time I look back at the history of Sumerian research, I marvel at the speed of progress in the research.

Sumerian is a language isolate not belonging to any linguistic family. Attempts have been made to relate it to Turkish, Hungarian, and even Japanese, but each attempt failed to find a scientific foundation and ended in fruitless labor. In terms of typology, Sumerian is an agglutinative language in which prefixes and suffixes are attached to verb and noun stems (Japanese is also such a language), and is an ergative-absolutive language in which the subject of transitive verbs is marked by the ergative case, and the subject of intransitive verbs and object of transitive verbs are marked by the absolutive case. Subject-object-verb is the general order for writing sentences.

Documents written in Sumerian are quite diverse. There are inscriptions recording the achievements of kings; administrative and economic documents dealing with income and expenditure accounts of the royal household; documents involving the buying and selling of dwellings, farm land, and slaves; documents recording court rulings; lists of vocabulary on themes such as trees, animals, fish. and hides; epics singing the praises of heroes such as Gilgamesh; and myths related to Enlil, the chief god of the Sumerian pantheon, Inanna, the goddess of love and war, and Enki, the god of wisdom. These writings were made over a long period from the latter half of the 3rd millennium BC to the first half of the 2nd millennium BC, so when using them to reconstruct Sumerian grammar, it is usual to divide the time span into periods: 2500-2300, 2100-2000, and 2000-1700 BC.

Sumerian orthography

Sumerian is written in cuneiform characters. They are usually logograms (standing for words), but can also function as phonograms (standing for sounds). For example,, when used as a logogram can mean water (pronounced "a" as in father), but when the same character is used as the phonetic symbol for "a" (discarding the meaning and using only the sound), it is a postposition (similar to the Japanese particles ni and de), or a verb prefix. Try imagining kanji used not only as logograms, but also as phonograms like Japanese Man'yōgana characters. Also, Sumerian writing doesn't use punctuation marks and there are no spaces between words. Because of this type of writing system, it is no surprise that scholars come up with different interpretations of the same sentence.

Furthermore, in Sumerian writing, all the elements required to make a grammatically correct sentence weren't always necessarily written down. When the expected prefixes or suffixes are not present, for example, we cannot tell whether they were left out because they were not required, or whether they were omitted in writing even though they were grammatically necessary. The question could be easily solved if only we could ask a native speaker, but for Sumerian, a language which ceased to be a vernacular at around the beginning of the 2nd millennium BC, this is impossible. There is no one to provide information and no continuity of knowledge. Often linguists can gain considerable help in interpretation by comparing languages from the same linguistic family, but for Sumerian, a language isolate, that path is also closed. A well-known Assyriologist once said, "There are as many Sumerian grammars as Sumerian grammarians," and the situation has not changed much in the half century since that quote.

Writings of budding scribes

From the start of research on the Sumerian language one encounters many difficult problems that seem impossible to overcome, and there have been many times when I have taken a pessimistic view of the possibility of reconstructing Sumerian grammar. And now there is one more difficulty we must keep in mind. Many epics and myths that make up the valuable group of texts used in the reconstruction of Sumerian grammar were written on clay tablets in the first half of the 2nd millennium BC, at a time when Sumerian could no longer be heard in the streets of Mesopotamian towns (the third period of writing mentioned above). Moreover, these tablets were not copied from the original texts in order to be faithfully passed on to future generations. Instead, they were written by children whose mother tongues were languages other than Sumerian (Semitic languages such as Akkadian), and who were practicing to become scribes. They wrote down sentences that were dictated to them, or sentences that were memorized beforehand. We know that Sumerian at that time was a language acquired through study, because we have half-cursing conversations between children such as "Your tongue isn't suited to Sumerian" and "I can hold a conversation in Sumerian," or "How will a scribe who knows no Sumerian be able to translate correctly?"

Sumerian compound verbs

I have been trying to understand how Sumerian sentences are constructed (syntax), often comparing Sumerian to various other languages. The topics of my research include Sumerian compound verbs, argument structure of verbs such as "have" and "get" and "stand" (intransitive) and "stand" (transitive), focus-marking constructions, and how sentences are combined (are they in a subordinate relationship or juxstaposed?). Here we will take a look at compound verbs, which make up a large majority of all Sumerian verbs. A Sumerian compound verb is a noun and verb pair forming a new verb, such as "head" + "give" = "attack" and "tree" + "touch" = "offer (an offering)." There are an especially large number of compound verbs like the first example, comprised of nouns representing body parts, such as "eye," "ear," "neck," "head," "hand," "foot," "heart," and so on, and verbs (the meaning of the verb itself is not always necessarily clear). In Sumerian, there is no simple verb representing basic visual activities such as "look" or "watch." In both cases the noun "eye" is employed with some verb to express the required meaning.

Example:

"eye" + "open" = "look"

In a similar way, there are no single-word verbs to represent mental activities like "think" and "forget". Instead we find the noun "ear," considered by the Sumerians as the seat of intelligence, and some kind of verb.

Example:

"ear" + verb (meaning unknown) = "forget"

When "hand" or "foot" is actually involved in the action, that noun becomes the nominal element of the compound verb.

Example:

"hand" + "approach" = "receive"
"foot" + "follow" = "step on"

Compound verbs expressing feeling involve the noun "heart."

Here is a question for you as you start to see a pattern in Sumerian compound verbs.
Question: What is the meaning of "heart" + "grab"?
Answer: "be angry"

Comparisons with noun incorporation

Compound verbs aren't completely unique to the Sumerian language. The linguistic phenomenon of forming new verbs by combining nouns and verbs is generally known as noun incorporation, and we can see this in more than a few languages throughout the world. For example, Takelma (now extinct), a former native American Indian language spoken in California, combined the verb meaning "find" with body part nouns to form new verbs as can be seen in, "nose" + "find" = "smell," "ear" + "find" = "listen," and "breast" + "find" = "think." In Walmatjari, an Australian Aboriginal language, we find the example of "ear" + "stand" = "listen." Also, the Uto-Aztecan language, Ute, combines the noun for "hand" with various verbs to form the verbs for "squeeze," "knead," and "slap."

Here, the question arises as to whether noun incorporation and Sumerian compound verb formation are the same phenomenon. Semantically, both are similar in the fact that they both combine noun and verb components to form a verb with a new meaning, but I believe they differ in terms of syntax. In the noun incorporations given above, the nouns become part of the newly formed verb and lose their individual salience both semantically and syntactically. As a result they have no syntactic status of their own. Therefore, in the Takelma example of "ear" + "find" = "listen," the component "ear" cannot be modified by an adjective such as "good." In comparison, in Sumerian compound verbs, the nouns representing body parts, in spite of losing their semantic salience, are marked as absolutive, meaning that they are the direct object of the verb. That shows that the noun components retain their own syntactic status. The nouns in Sumerian compound verbs can take an adjective, and at times their semantic salience is restored. For instance, the "eye" of the compound verb "eye" + "open" = "look" can take adjectives such as "clear" or "kind." Accordingly, it can be said that the difference between noun incorporations and Sumerian compound verbs lies in whether the noun components retain their unique syntactic status or not.

Future research

Over the past few years I, along with researchers from the University of Pennsylvania, have been working on a project to construct a textual corpus by adding morphological and syntactical analyses to digitalized Sumerian texts. While Sumerian interpretation and computer programming problems have meant that we have had difficulties moving forward, we hope to complete the first stage of the corpus in the near future. By bringing that tool to bear on this ancient language with its many difficulties, I hope to conduct more detailed research and to find more detailed evidence as I strive for more rational interpretations of Sumerian grammar. Reconstructing grammar is admittedly a low-profile job. However, reminding myself that is the core of Sumerian research, I spend every day buried in a mountain of cuneiform documents.

Fumi Karahashi
Associate Professor of Ancient Oriental Studies, Faculty of Letters, Chuo University

Born in Fukushima Prefecture. Graduated from the College of Comparative Culture, Tsukuba University. Obtained her Ph.D. with Sumerian Compound Verbs with Body-Part Terms (University of Chicago). Worked as a lecturer teaching Akkadian and Sumerian at University of Barcelona, University of Michigan, University of Chicago and University of Pennsylvania, before coming to the Faculty of Letters at Chuo University in 2007 to teach ancient history. Major research focus is Sumerian grammar. Is interested in Mesopotamian and Greek literature and also collects textual and iconographic materials related to East and West cultural exchange.