The sounds of long-dead languages
Published: 14 October 2017 (GMT+10)
According to the evolutionary view of language, the processes involved are supposed to be naturalistic and undesigned.
However, the Bible tells us in Genesis that God created Adam with language. Thereafter, Adam’s descendants defied God’s command to spread out and fill the earth following Noah’s Flood. On the plain of Shinar (Sumeria/Babylonia), they began to build a city, called Babel, and a tower, the top of which would be ‘unto heaven’. At that time, perhaps only 100 or so years after the Flood, they were all still part of the same society, speaking the same language. As a judgment upon their rebellion, God confused their language, so that they would not be able to understand one another, and would scatter over the whole earth. This creation of different languages was thus a sudden, miraculous event.
One of our readers, H.H., wrote that he was puzzled about something in the article Languages of the post-Diluvian World which discusses the spoken and written word in a post-Flood world.
How can we know the sounds of a long-dead language such as ancient Sumerian? The article goes into the variations in phonetic sounds which may be noted in written artefacts but where does the actual sound come from?
The article’s author, Dr Murray Adamthwaite, responds:
This is a frequently asked question, and the answer is fairly straightforward: assigning phonetic values to signs just is the task of decipherment. If signs cannot be given a sound value they have not in that case been deciphered.
There are at least two aspects to the problem:
Enough texts need to be available so that one can do a sign count. If that count comes to around 30 or so it can be concluded that the script is alphabetic. If 80 or more then it is syllabic (a consonant plus a vowel or vice versa, e.g. Linear B—mostly consonant plus vowel in that case). Cuneiform runs into several hundred signs (although only about 120 are commonly used), so it is both syllabic and logographic (where a sign denotes a whole word). Egyptian hieroglyphs also have several hundred signs, but there the issue is that signs can be single consonant, bi-consonantal, or even tri-consonantal; and then a whole range of determinatives, signs which indicate the category to which a word belongs.
One of the problems encountered with cuneiform was that of polyphony (where one sign can have a number of different sound values), and homophony (where the same sound value can be represented by several different signs). When the decipherment was first announced in the early 1850s there was widespread skepticism, and the Royal Asiatic Society in 1857 conducted a test, whereby a newly-acquired prism of Tiglath-Pileser I (now in the British Museum) was given to four different scholars acting independently. Their resultant decipherments and translations were sufficiently alike that the society could authoritatively announce that the script had indeed been deciphered, and the underlying language was therefore open for analysis.
How was it done?
In the case of both hieroglyphs and cuneiform the decipherers began by focusing on proper names which had come down to us via Greek sources and the Bible (these names in the Hebrew Bible, e.g. Sennacherib, turned out to be much more accurate renderings than in Greek sources). Also, of immense value were bilingual and trilingual texts, where at least one was known, or could be worked out. In the case of hieroglyphs, scholars had the Rosetta Stone (found during Napoleon’s military incursion into Egypt), a decree of Ptolemy V in Greek, Demotic, and hieroglyphs. Because scholars knew Greek well, from that they worked on the hieroglyphs. In the case of cuneiform they had the inscription of Darius the Great on a rock face in the Zagros Mountains at Behistun. The Old Persian part was worked out using the Avesta, the sacred text of the Zoroastrians, and from thence to the Akkadian/Babylonian section. It sounds easy, but along the way there were many, many headaches.
Since the work of the early decipherers there have, of course, been many refinements. Take Sumerian, for example: homophony is a major problem with its script, and at one time it was proposed to explain the phenomenon by the theory of tonality, whereby one sign with, e.g., the value ‘sag’ has a different tone than another sign with ostensibly the same value (analogous to Mandarin). But that theory has now been abandoned. However, another theory is that Sumerian had a range of nasalised sounds (such as you get in French), which seems to be enjoying a currency today. Hence the work goes on.
I repeat what I said at the outset: deciphering scripts just is the task of assigning sound values to various signs, and although the bulk of the work was accomplished in the 19th century, the work of refinement still continues.
Interesting article. It must be so difficult to figure out ancient languages.
It made me wonder how easy or hard it would be, if thousands of years hence, people found English texts (and English were no longer spoken), for the decipherers to figure it out (and assuming perhaps that there was only one other written language extant from this period, for them to compare with).
It seems to me that English must be really difficult for foreigners to learn, let alone for future people to decipher. Look at all the words which sound the same but are spelt differently, or spelt the same but are sounded differently ... let alone all the irregular verbs ... and so on and so forth. Oh, plus shortening words to 'don't' or shouldn't etc.
Thanks for this. You cover the painstaking process of deciphering these documents and use the term "sound value", presumably as a lead towards meaning. However, there is also the question of what it might have actually sounded like. How can we be sure what the Koine sounded like, or even the Hebrew of the writers' day? I realise that this may not impact meaning but it could. Thanks.
In response to Mr A’s question the following distinction needs to be made: consonants are fairly constant, albeit within the range of a certain phoneme, say /p/, there can be certain phonetic variations, e.g. ‘p’ as in ‘pot’ or ‘p’ as in ‘top’. Phonetics deals with the sounds actually heard; phonemes are the sounds thought to be heard—and represented by the same written symbol. Then some consonants can be nasalised, e.g. as with ‘ng’ in the French drink cognac.
The other part of the equation concerns vowels. They change like the proverbial weather between both dialects and languages. What we call accents have much to do with the pronunciation of vowel sounds. In regard to ancient languages the consonants can be ascertained with a fair degree of certainty, but the precise pronunciation of vowels is a little less certain, which is another way of saying that the precise accent remains obscure. However, because Akkadian is written with a syllabic script the vowel sounds are known, even if the precise accent is less well known.
Another problem involves the inadequacy of the script. The cuneiform script was taken over from the Sumerians by the Akkadians, and as usual, it proved inadequate for writing what is a Semitic language. So for example the vowel ‘o’ is not represented separately in the cuneiform script, but I believe must have existed, as virtually all other Semitic languages have an ‘o’ sound. Quite possibly the ‘o’ sound was subsumed under the ‘u’ sound, such that ‘lu’ covered also for ‘lo’ (although I must stress that this is my own view).
I could say much more, but I hope this helps.
Murray R. Adamthwaite