When written examples of an ancient dead language are discovered, how do the translators assign a sound to each symbol?
Let’s say a large stone with strange symbols is discovered. From the acompanying picture below the text, they can determine that the message is along the lines of “Queen Lissa led her army to a great victory.” Translators discern that any of the symbols that are underlined are a name. The symbols *&@@! are translated as “Lissa.” But how do they know that the asterik was pronounced as an “L” sound? Are they really speaking the language as it would have been spoken by the writer?
They can only guess, but it is an educated guess based upon the languages that are spoken in the region today and their various etymologies. Since languages are all related to each other by varying degrees, it is possible to get somewhat of a rough idea, but not specifics.
For instance I’ve heard speculation that people in Victorian England sounded closer to modern Americans than to what British sound like today. Whatever they really sounded like, we can assume they would have sounded close enough to one or the other pronunciation that we’d have some chance of understanding them.
In general, consonant sounds are a bit more stable than vowels, so the “I” and the “A” in Lissa might not be remotely correct for pronunciation today, but there’s a good chance the “L” and “S” are similar. Vowels drift alot from region to region and time to time, so there’s no guarantee that what is consistantly an E today will be an E on an ancient tablet.
Fortuitous discoveries like the Rosetta Stone have been crucial in giving meaning to heiroglyphs. This provides an excellent base for the process of unscrambling the sound that Thaumaturge describes.
1.) Comparison: For a long time, Egyptian hieroglyphics were indecipherable. Enter the Rosetta Stone, which was inscribed with the same message in Greek and two Egyptian scripts, Demotic and Hieroglyphic. The meaning became immediately evident from the Greek, and the Demotic (a script which had already been deciphered) gave the clues to the pronunciation. There was still some guesswork involved (hieroglyphs, like modern-day Arabic, Hebrew, and other scripts, was written without vowels), but it’s a start. Cite
2.) Historical evidence: Take the example of Linear B, a script found on inscriptions in Greece and the Mediterranean islands. For a long time it remained undeciphered, mainly because nobody quite knew which language it represented. Linguists tried the “brute force” approach, trying to match the symbols to known words in ancient languages such as Hittite, Sumerian, and Basque. Eventually somebody (a certain Michael Ventris) tried matching it with Ancient Greek, and wonder of wonders, things started to make sense. It turned out that Linear B had a partial descendant on the form of the Cypriot script, and was used to write an early dialect of Greek. Cite, Cite.
Exact pronunciation is tough to nail down anyway. As **Thaumaturge ** said, letter-to-sound correspondence is far from fixed, especially with vowels. The best clues to how a word was sounded at the time come from contemporary writings (Noah Webster, for example, gives a lot of context about pronunciation during his time). Poetry is also important, since rhymed words give good clues about pronunciation.
With ancient Egyptian (I’m sticking to that because it’s the only dead language I’ve studied), a lot of it is guesswork. The first problem, of course, is that none of the vowels were even written down. So transliterating the consonants is obviously much easier. With the consonants, I think it’s been a combination of what they eventually evolved into in Coptic (which survives as a liturgical language), and what they were transliterated as in other languages like Greek and Latin. We’re not really sure how all of the consonants were pronounced, but we’re sure about most of them and have decent guesses for the rest. As for the vowels, since they weren’t written, we can only find out what they were in specific instances, usually through transliterations made by other ancients in their own languages.
All of this, however, is somewhat separate from the question of how we actually pronounce ancient Egyptian. When the language is transliterated into Roman letters for scholarly study of literature, only the consonants are written. When we pronounce the words, we generally turn the consonants transliterated j and w into “ee” and “oo” respectively, or “y” and “w.” We pronounce the ones transliterated as A and a (usually transliterated as a thing that looks like a 3 and an ayin, but written thusly in ASCII) as “ah.” In between other consonants, we usually insert the letter “e” when speaking.
Of course, when names of kings and such are written for the general public, or for a greater academic audience beyond that which knows Egyptian, they are written differently. So some of the letters scholars of the language write with diacritical marks are written in a fashion more standard to English (“dj” instead of underlined d or ASCII D, “sh” instead of s with hachek or ASCII S, etc.). To further complicate things, many names of gods are written according to the way they were transliterated by the Greeks and Romans: DHwty or Djehuty becomes Thoth, Ast becomes Isis, wsjr becomes Osiris. Many cities are known by their Greek, Roman, Arabic, or modern names. So in summary, how we pronounce Egyptian in English isn’t totally arbitrary, but it’s a pretty complicated story. But we don’t fret too much over it, since there aren’t any ancient Egyptians around to correct us, and the language was spoken for several thousand years, so their pronunciation probably changed quite a bit anyway.
Sometimes certain facts about human pronunciation quirks aid in the decipherment of scripts as well as the other way around. Take the aforementioned breaking of Linear B by Ventris for example. (Unfortunately, a very poor quality site is given in Daver914’s first link.)
Ventris noted that on several tablets that certain symbols were crossed off and replaced by another. Certain combinations of such “fixes” occured more often than others. This gave a clue that those symbols represented similar sounds. This was a key basis for the construction of a table of sounds that lead to breaking the script.
It also helped a lot that the underlying language was Archaic Greek and that Ancient Greek was well documented. So a span of “merely” 500 years had to be bridged to get a general picture of pronunciation.
BTW, the first word Ventris “decoded” was basically “tri-po-de”, next to a pictogram of a three legged pot. I.e., a tripod.
I highly recommend reading a book on the decipherment of Linear B. An amazing tale of a language detective at work. If Ventris had only not died so young…
Note that the Egyptian scripts were also decoded using the knowledge of Coptic, the language of the Coptic Church in Egypt which appears to be a surviving remnant of the original Egyptian language. So that also gives a big leg up in pronunciation.
Etruscan, OTOH, may have no surviving relatives, so the few fragments of that language may never be deciphered, let alone a pronunciation given.
In the case of Mayan, some people in Central America still speak a language that is related to the Mayan spoken language. I can’t accuratly discribe how they were able to relate the sounds to the various parts of the glyphs, but it took several years, and the language is now pretty much totally understood.
This book is an excellent look at the subject and it discusses most of the languages mentioned so far in this thread. The author mentions that there is more than one way to decypher an ancient language.