Phonology: how the sounds of Akkadian are expressed in cuneiform script

Writing systems are almost never unambiguous representations of a language. Think of the many different ways to pronouce the sequence "ough" in English, for instance, or the fact that "there" and "their" sound the same. And our alphabet hardly begins to capture regional variation. Whether we say "bath" with a long or a short a we all spell it the same way. The cuneiform script and Akkadian language are similarly mismatched - but unlike English, which is spoken around the world, Akkadian is a dead language which we can only really know in its written (mostly cuneiform) form.

The sounds of Akkadian

Although there have been no native speakers of Akkadian for perhaps two thousand years or more, we can still have some idea of how it must have sounded. It helps, first of all, that Akkadian belongs to the same, Semitic, language family as Hebrew, Arabic and other languages of the Middle East that are still widely spoken today. It also helps that cuneiform was sometimes also used to write other languages, such as Ugaritic, which also used alphabetic scripts, or Old Persian which is the direct ancestor of modern Farsi. Nevertheless, we can only guess at rhythm, pitch, and intonation, and we have little idea about the exact sounds a speaker of any Akkadian dialect would have made. How differently did a village farmer speak from an urban priest? We will never know.

Vowels, diphtongs and glides

But all is not completely lost. We can distinguish four basic vowels - a, e, i, u - and some scholars think that there may have been an o sound too. There are also two semi-vowels (also called glides) that cannot form syllables on their own. We write them alphabetically as w and y.

Vowels may be short (as in "bat", "bed", "big", "but") or long (as in southern British English "bath" "bathe", "beet", "boot"). In cuneiform, syllables are often split in the middle the vowel. A writing ṭu ur 𒂅 𒌨 does not necessarily signal a long vowel, just that the scribe chose to write the syllable with two signs rather than one (ṭur 𒄙). If a scribe really means to write a long vowel, then he may write an extra vowel sign in the middle of the syllable: tu u ur 𒌅 𒌋 𒌨 ("turn around!").

Sometimes an extra vowel is written, however, to show the reader how to read the preceding sign. This is particularly important when distinguishing between i and e, as cuneiform often uses the same CV, VC, and CVC signs for these vowels. You can see this in the reference list of open signs. For instance lētum "cheek" and lītum "power" could both be written li/le tum 𒇷 𒌈, so in order to avoid confusion, the scribes would write le e tum 𒇷 𒂊 𒌈 and li i tum 𒇷 𒄿 𒌈 respectively.

Two different vowels together - a diphtong - often contract to a long vowel, so that only the last of the two is written. For instance, itÅ«ar "he turns round" (written i tu a ar 𒄿 𒌅 𒀀 𒅈, where here the extra vowel sign shows that the vowel change is deliberate), can contract to itâr (written i ta ar 𒄿 𒋫 𒅈 or just i tar 𒄿 𒋻).

When a vowel and glide are next to each other, they often elide too, so that a vowel plus y is written as a long vowel. For instance dayyānum "judge" is often written da a a nu um 𒁕 𒀀 𒀀 𒉡 𒌝. But a vowel plus w always becomes a long "u", so "she gave birth", which ought to be iwlid, is always written ūlid (u lid 𒌋 𒀖 or u li id 𒌋 𒇷 𒀉).

Consonants and glottal stop

There are 18 identifiable Akkadian consonants. Written in modern alphabetical order, the first 17 are: b, d, g, h (sometimes written ḫ), k, l, m, n, p, q, r, s, á¹£, š, t, á¹­, z. While most of these are familiar, we should note that h (or ḫ) is hard, as in "loch", and q is really an emphatic k, made with the back of the tongue touching the back of the roof of the mouth. Similarly, á¹£ (as in "its") and á¹­ (as in the disapproving "tut-tut") are emphatic counterparts to s and t, while š is pronounced "sh", as in "shoe". Finally, there is a weak consonant, called aleph and written ʾ, which is pronounced like the glottal stop in southern British English (e.g., the silent t in "what?").

In order to understand how cuneiform represents them, it is useful to group these consonants into families:

These groupings matter, because cuneiform script does not always distinguish between the consonants in each group, especially at the end of a syllable. You can see this, for instance, in the reference list of open signs. It is important to be aware, for example, that the signs 𒁍 𒊌 𒈝 could be read bu uq lum for buqlum "malt" or pu ug lum for puglum "radish". Both are correct readings; it is only possible to tell from the context which word is actually meant.

When a double consonant is written, it represents a double consonant in the language, and cannot mean a single consonant. So ṭup pa am 𒁾 𒉺 must always mean ṭuppam "tablet" and not ṭūpam (which in this case is not a word in Akkadian anyway). Conversely, a double consonant may not necessarily be written double. So ṭu pa am 𒂅 𒉺 𒄠 is just as correct as ṭup pa am as a writing for ṭuppam "tablet". In this case, because there is no Akkadian word ṭūpam, the scribe does not have to worry about ambiguity.

The glottal stop ʾ may stay hard or may elide with a neighbouring vowel, depending on the context. Hard alephs are always written with 𒀪 or 𒄴 (also used for vowel plus h), whatever the vowel(s) surrounding them. Soft alephs do not show up in cuneiform, or may become a long vowel. For instance, daʾmum "dark" may be written da aʾ mu um 𒁕 𒄴 𒈬 𒌝 or da a mu um 𒁕 𒀀 𒈬 𒌝. Both are correct.

For more detail, go to the Sound rules page.

Change over time

Both cuneiform script and the sounds of the Akkadian language changed over the millennia. They also differed from place to place, most obviously between Assyrian in the north, and Babylonian in the south. We further divide them both into Old (early second millennium) Middle (late second millennium) and Neo (early first millennium) varieties. For Babylonian we also recognise Late dialect and script (late first millennium) and a Standard version, used for literary texts and other scholarly writings. There were also several peripheral varieties of Akkadian, written to the west, south, and east of Mesopotamia.

It is relatively straightforward to identify different dialects of Akkadian, distinguished not only by their phonology but also by their grammar and vocabulary, and the repertoire of cuneiform signs that was favoured. For instance, towards the mid-second millennium BC the "m" sound at the end of nouns disappeared and words beginning with "w" lost their initial consonant. The word wardum "slave" for instance, became ardu, as can be seen by changing syllabic writings (from wa ar du um 𒉿 𒅈 𒁺 𒌝 to ar du 𒅈 𒁺). These are just two features of many that mark out Old Babylonian Akkadian from later dialects such as Standard Babylonian. But just as it is not necessary to master the entirety of cuneiform all at once, dialectal variation is not a matter that beginners of Akkadian have to worry too much about.

Content last modified on 05 Dec 2016.

Eleanor Robson

Eleanor Robson, 'Phonology: how the sounds of Akkadian are expressed in cuneiform script', Knowledge and Power, Higher Education Academy, 2016 []

Back to top ^^
© Higher Education Academy, 2007-11. Since 2015, SAAo is based at the Ludwig-Maximilians-Universität München, Historisches Seminar (LMU Munich, History Department) - Alexander von Humboldt Chair for Ancient History of the Near and Middle East. Content released under a CC BY-SA 3.0 [] license, 2007-20.
Oracc uses cookies only to collect Google Analytics data. Read more here []; see the stats here []; opt out here.