The complexities of the Sumerian logosyllabary can be summed up as follows:
1 The signs could either be logograms or syllabograms (or both).
2 Because of the logographic origins of the writing system, words that were pronounced the same were often written with different signs. In other words, there were many homonyms. Thus, for example, the spoken word gu could mean “shout,” “thread,” or 12 other things, each of which was written with its own logogram. The English use of two, too, and to is a pale reflection of the same idea.
3 On the other hand, many signs were polyvalent. In other words, the same sign could stand for two or more morphemes that meant different things and were pronounced in different ways. For example, the sign known to Sumerologists as apin could be read either apin (“plow”) or engar (“farmer”). But the different morphemes were generally related in meaning, and would have shared a pictogram in proto-cuneiform. The correct word would be deduced from context, as in English, where context can usually tell us whether house should be pronounced [hauz] and mean “to provide shelter for,” or be pronounced [haus] and mean “a place where people live.”
4 To reduce the chance of confusion brought on by the polyvalence of the signs, the Sumerians used determinatives and phonetic complements. Determinatives provided clues as to what kind of meaning a word would have, while phonetic complements spelled out part of a word syllabically so as to indicate which reading of a polyvalent logogram was being used.
The system may appear baroque to the modern mind, and especially to those used to an alphabet. Attaining literacy in cuneiform required hard work, as attested by the long word lists that student scribes studied and copied. Throughout its history Sumerian cuneiform retained aspects of having been a first invention. There was no earlier writing system to learn from, no pre-existing concept of literacy to compare with, no idea that a writing system could be more or less efficient or learnable. Yet it more than filled the need for which it had been invented. When Sumerian administrators first put reed to tablet it is doubtful that anyone dreamed that their invention would be used to write history, literature, and science texts. It took some 500 years, yet eventually that is precisely what happened.
Sumer occupied only the southernmost part of Mesopotamia, a land of growing city-states in the first half of the third millennium. The language of the Sumerians was unrelated to any other language that we know of. It was a language isolate, like Basque, with no known relatives. East of Sumer was the land of Elam, where the Elamite language was also a language isolate. North of Sumer, in what is now central Iraq, was the land of Akkad. The Akkadians spoke a Semitic language related to Hebrew, Aramaic, and Arabic, but entirely unrelated to Sumerian. The linguistic diversity of the region may be reflected in the biblical story of the tower of Babel, which describes the building of a ziggurat (temple) at Babylon, near the Sumer–Akkad border.
Despite differences of language and culture, the elites of Akkad adopted many Sumerian customs, including the building of ziggurats. Even before Sargon I (2334–2279 BC) conquered Sumer and instituted Akkadian as the language of administration, Akkadian scribes had worked out ways to adapt Sumerian cuneiform to their own language. Quite simply, they took the meanings of some of the logograms and supplemented them with the sounds of the syllabograms. Thus the sign for “reed,” Sumerian gi, was now pronounced qanuum, or, when it meant “to render,” taarum. But it could also still stand for the syllable [gi] in a word spelled out syllabically! The original unifying pronunciation that made sense of the use of this one sign for three purposes was lost. In other words, while the three uses of the sign had all been pronounced [gi] in Sumerian, Akkadian word lists tell us they were pronounced [qanu:m], [ta:rum], and [gi] in Akkadian.
This apparently nonsensical arrangement would have made sense to Akkadian scribes, however, since they would have been bilingual in Akkadian and Sumerian. In order to learn to read and write, the earliest Akkadian scribes had had to learn Sumerian. From then on it became a tradition: an educated person learned Sumerian. This pattern has been repeated many times throughout history. A language that is written becomes dominant over other languages in its region, inspiring a great deal of second-language study by speakers of other languages.
An important contribution of the Akkadians to the cuneiform writing system was the expansion of the syllabary, with the result that Akkadian, unlike Sumerian, was written mostly in syllabograms with the occasional logogram thrown in, rather than vice versa. This expansion and adaptation brought about the full maturation of the script.
The Sumerians had tended to be imprecise about the match between syllabograms and spoken syllables. Like English, Sumerian allowed closed syllables – syllables formed of a consonant plus a vowel plus another consonant, schematically notated CVC. But the number of possible CVC syllables is quite large, and it was impractical to have a separate symbol for each one. The Sumerians tended to write CVC sequences as simply CV, leaving it up to the reader to recognize from context what consonant had been left out. This may be one reason for the apparently large number of homonyms.
Syllabaries are naturally more apt for some languages than for others, as the number of possible syllables varies from language to language. Most syllabaries in the history of writing have consisted of only – or almost only – CV syllabograms, with a few additional V syllabograms for vowel-initial words. Languages whose spoken syllables are also entirely or almost entirely of CV form are well served by a syllabary, while those with more complex syllables are not (such as English, whose longest single syllables are strengths, phonologically CCCVCCC, and twelfths, CCVCCCC). If written syllabically, these languages must resort to representing either more vowels or fewer consonants than the spoken word actually contains.
Akkadian possessed closed syllables with a vengeance. Furthermore, as a Semitic language, it entrusted the core meaning of its words to its consonants. In Semitic languages most word cores consist of three consonants, with the addition of vowels and/or the doubling of consonants providing inflectional information. So, for example, you conjugate a verb by changing the vowels. This is like English sang being the past tense of sing, but much more systematic. Thus the sequence of three consonants p-r-s was the core of the verb meaning “to render a decision,” but the present tense form of the verb was iparras (doubling of the second consonant being part of the inflection) and the past tense was iprus. The infinitive was paraasu. A spelling system that did not allow Akkadian scribes to indicate all the consonants in a word like iprus – with its VC.CVC structure – would have forced them to leave out the essential meaning of the word. So they began to use some syllabic signs for VC sequences. They could then break up closed CVC syllables into two signs, one CV and the other VC. Thus iprus could be spelled ip-ru-us. There was some ambiguity in this, as the sequence ru-us could be used to spell a syllable pronounced [ru:s] (with a long vowel) or [rus] (with a short vowel). The trouble of inventing and remembering a symbol for every CVC syllable was averted, however.
The Akkadians also systematized the appearance of the signs. In Akkadian cuneiform the individual wedge-shaped lines that made up the signs all had their heads at the left (for horizontal or diagonal signs) or at the top (for vertical signs). The effect can be seen in the contrast between the second-column and third-column versions of the signs given in figure 2.2. The Akkadian versions (on the right) draw from a smaller and more organized repertoire of shapes.
After the fall of Sargon’s dynasty around 2100 BC, Sumerian once more became the administrative language of the unified land called “Sumer and Akkad.” Later dialects of Akkadian known as Babylonian and Assyrian eventually replaced Sumerian, but Sumerian continued to be studied by scribes and scholars, and a considerable portion of the extant literature in Sumerian was collected by the late Assyrian king Assurbanipal (668–627 BC) and housed in his library at Nineveh. Akkadian also went on to produce a large corpus of legal, scientific, and literary texts, such as the law code of Hammurabi, obs
ervations of the planet Venus (plate 1), and the Epic of Gilgamesh (originally a tale about a Sumerian hero, but put into its final form by Akkadians). In fact, due to the durability of clay tablets – which were preserved rather than destroyed whenever a city was put to the torch – the period of Akkadian cuneiform has left us more primary documents than any other time up to the invention of the printing press.
From the seventh century onwards cuneiform was rivaled by the Aramaic alphabet, and with the conquest of Mesopotamia by Alexander the Great (336–323 BC) it lost its official support. The last known cuneiform text is an astronomical tablet dating to AD 75.
In its heyday, however, cuneiform covered a wide area. The invention of writing constituted a revolution in bureaucratic technique, and like any other successful revolution, its effects could not be contained. Soon after the development of proto-cuneiform, in the Uruk III period at the end of the fourth millennium BC, early writing inscribed on clay tablets also appeared in the land of Elam, east of Sumer, now part of Iran. It is not entirely clear how much the appearance of writing in Elam owes to the development of writing in Sumer. Most likely, the idea of using writing for accounting purposes had reached the people of Elam, but they may or may not have seen proto-cuneiform before they set out to develop their own script. The script, known as proto-Elamite, is thus far undeciphered. Like proto-cuneiform, it was used for concise administrative accounts and not for lengthy prose. Unlike proto-cuneiform, it did not develop further, and so we don’t have more elaborated versions of the script to use for decipherment, nor do we know what language the proto-Elamite scribes spoke.
Instead of continuing to adapt and develop their own script, the people of Elam later borrowed the cuneiform script of their Mesopotamian neighbors and used it for over 2,000 years. At first they wrote in Akkadian, then adapted the script to their own language, which we can now identify as Elamite, a language with no known relatives. Unlike the Akkadians, the Elamites had no particular loyalty to the ancient Sumerian way of doing things. They considered logograms unnecessary complications: why learn a separate sign for a word if you can just spell it out with a syllabary? They ruthlessly cut down the number of signs, arriving at about 130, compared to the roughly 600 signs required to write Akkadian. Most of the signs were syllabic – either CV (or just V), VC, or sometimes CVC. As in Akkadian, most CVC syllables did not have their own sign and were written CV-VC.
Interestingly, the number of logograms gradually increased over time, a trend that is noticeable in other phonologically written languages too. At first the most important consideration is to develop a system that can be easily learned, and so additional signs that can only be used for a single morpheme are an unnecessary complication. After a while, though, literacy is taken for granted, and it becomes more important to write efficiently. A single sign for a common word can be written more quickly than the individual syllables or phonemes (consider how annoying it would be to write out “dollar” every time, rather than “$”), and so the Elamites came to use about 20 common logograms.
Cuneiform also spread north. In an area covering northern Syria, northern Iraq, and part of Turkey, the Hurrian civilization flourished for about a thousand years, starting toward the end of the third millennium BC. Like the Elamites, the Hurrians borrowed syllabic cuneiform from the Akkadians, cutting the number of signs to achieve an efficient syllabary. A related language, Urartian, readapted the script, preserving more of the logograms, in the period 830 to 650 BC.
Cuneiform was also adapted to the Hittite language of Asia Minor in the middle of the second millennium BC. Hittite was an IndoEuropean language, a member of the same linguistic family as English. Like the Urartians, the Hittites chose to borrow a good number of the ancient Sumerian logograms along with the Akkadian syllabary.
Thus the original Sumero-Akkadian cuneiform was used for or adapted to languages of five different families: Sumerian and Elamite (both isolates), Akkadian (Semitic), Hurrian and Urartian (related to each other but not to other languages), and Hittite (Indo-European).
The idea of cuneiform – a script consisting of impressed wedge-shaped signs – spread yet further. Old Persian, the Indo-European language that came to share and then dominate the Elamite language area, is the language of a number of cuneiform-type royal inscriptions from the Achaemenid dynasty, which ruled Persia from 559 to 331 BC and included the famous kings Cyrus, Darius, and Xerxes. Luckily for later decipherers of cuneiform, Darius had a fondness for commemorative inscriptions; but the Old Persian script was not otherwise much used. It was midway between an alphabet and a simple syllabary. Some signs were simple consonants (C), some simple vowels (V), and some a syllabic CV combination, for a total of 36 phonologically based characters, plus a word divider, 5 logograms, and numerals. The wedge-shaped lines that make up the signs were selected from those of Akkadian cuneiform, but the actual signs were completely different.
There is one final use to which the cuneiform style of writing was put. The last version of cuneiform to become known to the modern world was discovered in 1929 at what is now Ras Shamra on the coast of Syria. The site was once the Canaanite city of Ugarit, which flourished between the fifteenth and twelfth centuries BC.
The Ugaritic script consisted of only 27 to 30 signs, depending on the context in which it was used. A script with so few signs is almost bound to be an alphabet – with each sign representing a single consonant or vowel – rather than a syllabary. And indeed it was an alphabet. Within a year of its discovery it had been deciphered; not only was it an alphabet, but it turned out to be a transliteration into cuneiform-type characters of the ancient Semitic alphabet which spawned the Phoenician, Hebrew, and Arabic scripts, and which ultimately lies behind our own alphabet as well. The letters were shaped entirely differently from those of the other Semitic alphabets, but the names of the letters and the alphabetical order were the same. Ironically, the oldest evidence for the ancient alphabetical order that we still follow when we recite “A, B, C” is from a Ugaritic cuneiform tablet of about 1400 BC, and not from letters related in shape to ours or to any of the other Semitic-derived alphabets.
And so the method of writing impressed cuneiform signs on clay tablets was used for logograms, for syllabaries, and even for an alphabet. This remarkably adaptable and long-lived tradition finally died in the early years of our own era, and the very fact that such writing had once existed was forgotten.
Meanwhile the Mesopotamian tablets lay buried in the shifting sediments of time. Whole cities were abandoned and forgotten, along with the public inscriptions their kings and conquerors had erected in them. Not a word of any form of cuneiform could still be read, and when in the seventeenth century AD reports reached the West of a wedge-shaped form of writing on the ancient stones of the Middle East, some scholars refused to believe that it could actually be a kind of writing. Today we can read the inscriptions of Darius the Great, the astronomy of the Babylonians, the laws of the Akkadian kings, and the myths of the Sumerians. How did this happen?
The ego of the Persian king Darius (521–486 BC) played a not insignificant role in the decipherment of cuneiform. Darius had his exploits and decrees inscribed in various public places in and around his new capital city, Persepolis. His son Xerxes (486–465 BC) followed his example. To ensure that no one missed the point, the inscriptions were written out in three languages: Old Persian, Elamite, and Babylonian Akkadian.
But even the glory of kings passes. In the seventeenth century AD no one could read a word of Darius’ and Xerxes’ boasts. No one knew who had written them, what three languages they were in, or even that they were in three different languages.
The first accurate copies of Persepolis inscriptions reached the hands of Western scholars between 1772 and 1778, published by Carsten Niebuhr, a Dane who had spent time in Persepolis meticulously copying the cuneiform inscriptions. He was the first to realize that there were three different wedge-like scripts represented. He was also able to confirm that the writing ra
n from left to right, and he assembled a list of the individual characters in the Old Persian script. But there was still no way to know what the inscriptions actually said. How could anyone go about deciphering a text in an unknown language with no readable translation?
The first part of the answer took the form of an inspired guess: if you don’t know what the text says, what do you think it should say? Taken too far this method can be disastrous, and many false decipherments of forgotten scripts have been based on such an error (a sixteenth-century Dutch doctor, for example, tried hard to prove that the language recorded by ancient Egyptian hieroglyphs was in fact Dutch). But it is common knowledge that monumental public inscriptions were generally made at the behest of kings and made frequent mention of those kings in their texts. It is also true that the names of kings are sometimes preserved in the records of other peoples. Thus we owe to the Greeks the memory of the names of the Achaemenid Persian kings as well as many of the Egyptian pharaohs. Both the decipherment of cuneiform and that of Egyptian hieroglyphs began with the names of these monarchs.
The German school teacher Georg Grotefend (1775–1853) was the one to make the inspired guess. On the basis of later Persian inscriptions written in the recently deciphered Sassanian script, he guessed that there would be a repeating formula of the style “so-and-so, the king, king of kings, son of so-and-so, the king, king of kings.” And sure enough, the texts did contain repeated phrases (hypothesized to be “the king, king of kings”) interspersed with sections of text short enough to be names. If his hypothesis was correct, then one text mentioned “x, the king, king of kings, son of y, the king, king of kings,” while another text mentioned “y, the king, king of kings, son of z.” Clearly, then, z was not a king. So who were x, y, and z?
The Writing Revolution Page 4