make it difficult to distinguish a sequence of independent words from an instance of
compounding; in practice, much reliance is placed on instances borrowed into
Akkadian as single words. Such head-initial compounds include ones in which the
modifier is a noun: e 2 – masˇ“(animal) pen” (“house” + “goat”); a verbal noun: ki-tusˇ
“dwelling” (“place” + “living”); a verbal adjective: di-til-la“verdict” (“decision” +
“finished”); or an adjective: lugal“king” (“person” + “great”). This last example is
written as a ligature in the reverse sequence (gal+ lu 2 ), lending some support to the
suggestion that the Sumerian noun phrase was once more head-final. Another, no
longer productive type of compound provides further support for this possibility, being
itself head-final, as in an-edin“high plains” (“heaven” + “plains”).
As these compounds are typically written logographically, it is difficult to specify any
phonological changes that accompanied their formation. However, the limited
evidence (such as e 2 –masˇ also being written amasˇand phonographic writings of ki-tusˇ
as ku-tu-usˇ) suggests that vowel assimilation as a result of compounding was much
more extensive than our transliterations indicate.
THE NOUN PHRASE
Verbs aside, the basic building blocks of a Sumerian clause are noun phrases, consisting
of as little as a noun (or pronoun), serving as the head of the phrase, and a case marker,
indicating the end of the phrase. These two elements can, however, be separated by
sequences consisting of the noun’s modifiers, a determiner and a plural marker. To put
this more bluntly: the Sumerian sequence corresponding to of my great kingswould be
king great my PLURAL of.
Before looking in more detail at these phrases, the earlier discussion of bound
morphemes needs to be revisited, distinguishing between affixes that are restricted in
terms of their host and clitics that are less so. Briefly, possessive ’sin English can be
described as a genitive case marker, genitive being the grammatical case primarily
concerned with possession, which can also be indicated by the preposition of, English
prepositions often corresponding to case marking in other languages. A typical genitive
affix occurs bound to the noun that is the possessor (as in Latin regis“of the king” from
rex“king”). In contrast, ’sis a clitic that occurs bound to whatever is the last word in
the noun phrase describing the possessor: the king’s brother, the king of England’s brother,
the king who was fat’s brother, and so on.
A Sumerian noun phrase can end with as many as three different clitics: a
determiner and plural marker as well as a case marker. Not all the determiners are,
however, bound. In fact this description can be applied to only some of the possessives,
although the convention in transliteration is to apply it to all of them, as well as to the
demonstrative determiners. The evidence that at least some of the possessives, those
which are vowel-initial in their fullest forms (such as /ani/ “her/his”), were phono-
logically dependent comes from writings that resume a preceding consonant (such as
shag 4 – ga-ni“her/his heart”) and reductions after a vowel (such as dumu-ni“her/his
child”). The same analysis can be applied to the plural marker, the next possible
constituent in the noun phrase, whose fullest form is /ene/ and which is restricted to
phrases that have a human noun as their head.
Two broad uses of case marking can be identified: to specify the syntactic relation-
ship between a noun phrase and a verb, and to specify syntactic relationships within a
–– Graham Cunningham ––