David Damrosch’s The Buried Book is a lively account of the early archaeological triumphs and shenanigans in and about Iraq & the British Museum, particularly in regards the rediscovery of the “Great Epic of Gilgamesh.”
In a section talking about the decoding of the cuneiform in which the poem was written Damrosch says, A “major complication in the process … was that cuneiform had originally been developed in southern Mesopotamia by people who spoke Sumerian, an ancient language completely unrelated to any other known language. The script had then been taken over by speakers of Akkadian, which became the most commonly written language for much of Mesopotamian history. Yet the Akkadian scribes continued to learn Sumerian as they mastered the script, and they often employed Sumerian loan words amid their Akkadian texts. It is as though, in reading an English text we would often have to pause and determine whether pain meant ‘suffering,’ as in English, or ‘bread,’ as in French.
“Conversely, a sign might have the same meaning in Akkadian as in Sumerian but a completely different sound: when used to mean ‘sky,’ the star symbol is pronounced an in Sumerian, but shamu in Akkadian. Names in particular could be tricky, for Assyrian names often included Sumerian elements, along with Akkadian symbols. This would lead George Smith [a self-taught linguist responsible for the first translation of Gilgamesh], for example, to misread the name Gilgamesh as ‘Izdubar’; he didn’t realize that what looked like two Akkadian characters, iz and du, were actually Sumerian signs pronounced ‘giz-ga’ or ‘gil-ga.’ He then guessed incorrectly on the final syllable, which was Akkadian as he assumed, but which can be pronounced either ‘bar’ or ‘mesh.’ … The reading of ‘Gilgamesh’ was finally established twenty-five years later by Smith’s friend and successor Theophilus G. Pinches, in an article triumphantly entitled ‘EXIT GISTUBAR!’”