Martin Worthington of SOAS has put online an archive of recordings of “modern Assyriologists reading ancient Babylonian and Assyrian poetry and literature aloud in the original language.” The About page explains:

It is intended to serve several purposes, some for Assyriologists, and some for the wider public. First, it aims to foster interest among students of Babylonia and Assyria in how these civilisations’ works of verbal art were read aloud in the past, and how they should be read aloud today.
Second, it provides a forum in which scholars who have theories about Babylonian and Assyrian pronunciation, metre, etc. can present a concrete example of how their theories sound in practice. […]
Third, as a record of the ways in which contemporary scholars read Babylonian and Assyrian, it will some day serve a historical function. Many great Assyriologists, including some who had influential theories of Babylonian metre and phonology, passed into history without leaving a single recording of how they read Babylonian and Assyrian. This archive will provide at least some record of how scholars read Babylonian and Assyrian in the twenty-first century.
Finally, but not least, the questions which students of ancient languages most frequently hear from laymen are: “How did they sound? And how do you know?”. This website is meant to serve as an introduction to these issues, providing the public with some idea of how modern Assyriologists think Babylonian and Assyrian were pronounced.

The recordings are here; start by clicking on the open-book symbol at the right of the recording you’re interested in, then open the “listen to the recording” link in a new tab or window, and you can hear the recording while following along in the transcription and translation. It’s a lot of fun, and the ones I’ve listened to have sounded quite convincing (in the sense that they sound like an actual language being spoken rather than the kind of stilted, exaggerated reading you often get from professors trying to read Ancient Greek or Old English aloud—of course, it is extremely unlikely that any of the readers actually sound like an ancient Babylonian). Via Languages Direct; they remark “you are struck, more than anything, by a feeling of familiarity through the similarity in sound of Babylonian to its modern offspring such as Arabic,” which doubtless has to do with the fact that the readers are familiar with Arabic but have never heard any speakers of Ancient Babylonian. Thanks for the link, Sashura!


  1. Just to point out that Arabic isn’t a “modern offspring” of Babylonian, but a sister language.

  2. Good point.

  3. Very cool! These are all Akkadian language recordings, there are no recordings of Sumerian, those would be extremely interesting as well.

  4. aqilluqqaaq says

    Babylonian and Assyrian could be sisters; or maybe Akkadian and Eblaite; Babylonian and Arabic are more like third cousins once removed.

  5. I remember a “multicultural exhibit” at a Universiy I worked at, where foreign students were asked to showcase “their culture”, and a group of Arab students had put him a poster saying (INTER ALIA) that Arabic was an ancient language deriving from Babylonian.
    Strange to see the same mistake here: is there some kind of folk myth (in the Arab world?) relating Arabic to Babylonian?

  6. aqilluqqaaq says

    I think the reasons include the fact that Arabic preserves some archaic (proto-Semitic) phonological and morphological features, including the three nominal case endings which it shares with Old Akkadian, though these are subsequently lost in Neo-Babylonian. Some of the archaic phonological archaic features of Arabic (the interdentals, among other things) are already lost or attenuated in Akkadian. The innovations are obvious though (e.g. indefinite nunnation, definite article ’l-).

  7. “Third cousin once removed”, yes, fair enough.
    Presumably this is the same false assumption that is accountable for many English speakers believing English is descended from Latin or Greek, i.e. that if a modern language and an ancient language are related, the former must be a daughter of the latter.

  8. “many English speakers believing English is descended from Latin or Greek”: where are they hiding? I’ve never met one.

  9. I’ve come across this idea pretty often. Maybe you move in better educated circles, or maybe (judging from your email address) the misconception is more common in the US than the UK.

  10. languagehat,
    Just wanted to say thanks for your response on a previous post, and in return I’ll offer up that in mid-December there’ll be a new anthology out edited by Otto Penzier,titled: The Greatest Russian Stories of Crime and Suspense.
    Not in Russian, although I know many of the people here could read it that way!

  11. marie-lucie says

    I once had a conversation in a Greek restaurant with a Greek waitress who assured me that Greek was descended from Latin, which she had studied in school.
    Many French people still think that French is descended from Gaulish (after all, we have known about nos ancêtres les Gaulois since the second grade).

  12. I have heard of people who think English is descended from Latin, but much more common is the belief that English is descended from German. That’s the same cousin-ancestor confusion that shows up in people claiming evolution is about humans descending from apes.
    But part of it could also be the similarity between “Germanic” and “German”, which causes people to think Germanic languages come from German. The “German”-“Germanic” confusion shouldn’t occur in non-English languages, which have very different words for the two.

  13. I once met a student at Cambridge University who told me that the English word “woman” was descended from a Greek word. I am sorry to say that (if memory serves after 30+ years) he was the son of a prominent classicist.

  14. Is there any reason not to call an apelike common ancestor of the humans and the living apes an ape?

  15. Is there any reason not to call an apelike common ancestor of the humans and the living apes an ape?
    That was commonly done already the late 19th century. There was only general awareness of the facts of biological inheritance (geese don’t give birth to horses), although there was much speculation about diachronic relatedness based on morphological comparisons. Nothing was known about the mechanisms of heredity such as we know them today, and viewpoints have become more statistical.
    Nowadays, I think that even to call your “common ancestor” a proto-ape or a proto-human would be too wooly a formulation – and it would sound teleological. It seems that clades are one of the more up-to-date ways of understanding such matters. David Marjanovic is the guy to answer your question professionally.

  16. David charges by the hour?

  17. This made me think of this clip ( where Arne Torp, of Oslo University, gives a very lively rendition of a Proto-Norse reconstruction of the Atlakvida. (There is no evidence that the Atlakvida was actually formulated in Proto-Norse [although the matters that it deals with date from that period, which is why this particular piece was chosen], but this is what it would have sounded like – phonologically – if it had. This kind of reconstruction immediately raises some tricky questions: for instance, the Old Norse passive -sk [modern Scandinavian -s(t)] probably didn’t exist in Proto-Norse, but it’s been reconstructed to sound as if it had, namely as *-sik.)

  18. Overacting.

  19. He does get carried away toward the end, but it’s a lot of fun, and less than two minutes long. Here‘s the direct link.

  20. In the US there is a long tradition in middle school of teaching kids to recognize Latin and Greek roots – mostly so we can do well in spelling bees and standardized tests. But I can’t remember a teacher ever explaining how those roots got into English in the first place. That pedagogical approach probably plays a role in encouraging people to suppose English must be descended somehow from Greek and Latin.

  21. I believe I was taught that German and English were descended from Greek, in contrast to the Romance languages, which were descended from Latin. Even at the time I knew better. I was also taught by two separate teachers that Slovak was Slavic, whereas Czech was Germanic. Again, I knew better and did not hesitate to advertise that fact. Which did not win me points with my fellow students or my teachers.
    Here’s a very silly recording of the Iliad in ancient Greek.

  22. Here’s a very silly recording of the Iliad in ancient Greek.
    Indeed it is, but in Stephen Daitz’s defense you could say that he is at least trying to cope with known facts of Ancient Greek prosody. Corresponding facts about Babylonian and Assyrian prosodies are, I suspect (though I don’t know) much skimpier. That might largely account for these very interesting recordings sounding “like an actual language being spoken.” That is, not too different from whatever languages the reciters habitually speak.

  23. Uh, if nobody minds my asking, where did English come from? And for that matter, why doesn’t Teh Wiki consider German to be a Romance language?

  24. English came from Old English came from Anglo-Saxon came from proto-Germanic came from proto-Indo-European, is how I learned it anyways. German is not a Romance Language because it is not descended from a dialect of Latin.

  25. there is an entry on Ishtar on the Worthington’s page. Can someone tell me how Ishtar came to be Aphrodite and lose the hunting/trapping connection on the way? Did the say AY-shtar, then Ai-star-tah, then s turned to ph? But what next?

  26. This Wiki page has a very informative chart outlining the historical development of Germanic languages:
    This one has a family tree of Romance languages:

  27. marie-lucie says

    Adding to what TMK said: English received a massive infusion of Romance vocabulary, first through the “borrowing” or adoption of Norman French words after the 1066 conquest and later through the adoption of learned words based on Latin. Examples of the first are words for meats (beef, mutton, veal) as well as a host of miscellaneous others (chief, journey, judge, laundry, prince(ss), etc), many of which replaced the Old English equivalents. Examples of the second are conduct, construct, composition, education, destiny, general, mystery, mysterious, predestination, temporary, vocabulary, etc, many of which had also been taken from Latin into French with minimal adaptation. These two sets of Romance additions or replacements are the reason why most of English vocabulary is so different from that of German nowadays, although some of the basic vocabulary is the same or only slightly different (eg hand, finger, foot, nose, begin, end, sing, swim, sleep, etc).

  28. marie-lucie says

    Sashura: Ishtar and Aphrodite are names of two goddesses in two different cultures. The names have nothing to do with each other, but their functions and attributes were very similar, so it was possible to say, for instance, “The Greeks worship Ishtar under the name of Aphrodite”. Similarly, the Greeks and the Romans were aware that Athena and Minerva were the same goddess, Phoebus and Apollo the same god, Zeus and Jupiter the same father of those deities, etc. This is one reason why many deities were known under several different names, and the names of equivalent deities were rarely related between one culture and another.

  29. Ah, thank you Modesto Kid and m-l. So it seems that English and Latin descended from a common ancestor, much like people and monkeys.Here is a “PIE chart” and here is a simpler one. Now if only I could get the Asian, Semitic, and African languages to fall into place so easily.

  30. Similarly, the Greeks and the Romans were aware that Athena and Minerva were the same goddess, Phoebus and Apollo the same god, Zeus and Jupiter the same father of those deities, etc.
    I would say rather “the Greeks and the Romans decided that Athena and Minerva were the same goddess, Phoebus and Apollo the same god, etc.” In other words, there was a deliberate syncretism of previously independent sets of divinities. I think you will find that the attributes of those “same” gods and goddesses sometimes differed quite substantially.

  31. marie-lucie says

    LH: Fair enough! “same” was an elastic term. The Romans “recognized” their own gods where they could, not only in Greece but for instance in Gaul. That practice enabled the conquered territories to keep their religious observances intact or merge them with the Roman ones, and avoided religious divisions through the empire. The problem with the Jews and early Christians is that they did not want to equate their single God with any of the Greco-Roman gods, and therefore their religion was incompatible with the official Roman one.

  32. John Emerson says

    As I understand the Greek and Roman gods themselves, as well as the non-Greek non-Roman gods (and as far as that goes, the Madonna) so that the identifications just amount to putting an umbrella over an even larger number of manifestations, rather than claiming a one-to-one identity.
    And yes, I’m fine, but I only have an hour a day on the internet, which is greatest time suck in the history of creation.

  33. marie-lucie says

    Nijma: Now if only I could get the Asian, Semitic, and African languages to fall into place so easily.
    Many linguists would like to get this result too! Language classification is not as simple as it might seem. The Indo-European languages have been throughly studied over the last two hundred years (building on much older traditions too) and their classification is uncontroversial at this point, but they are not the only ones spoken in Europe (and Asia!): Basque is a family all by itself, and there is also Sami (= Lapp), Finnish and Estonian which are related to some native languages of North Asia (Hungarian is also distantly related to those, but is a more recent addition to the European linguistic landscape; so is Turkish), as well as the Caucasian languages which straddle the Southern border between Europe and Asia. Semitic is a fairly compact family, related to other languages of North Africa and Southwest Asia. The languages limited to Asia or Africa include several large families in each continent, as well as some language isolates (unrelated to any others), and some of the groups are controversial.
    Among the native languages of the Americas, there are also a number of different groups, as well as considerable controversy about how many should be recognized (at least one family appears to be related to a North Asian one). I was about to forget Australia, which has one very large group and another smaller group which may consist of several unrelated languages. The Pacific Ocean with all its islands has one huge group (Austronesian) which ultimately came from Asia (Taiwan), and some other smaller groups, especially in Papua-New Guinea which has the largest density of languages in the world.

  34. –the problem with the Jews and early Christians is that they did not want to equate their single God with any of the Greco-Roman gods, and therefore their religion was incompatible with the official Roman one.
    If only Julian did not die from the arrow wound on the battlefield, European civilization might’ve ended up with the syncretic Greco-Roman-Judeo-Christian religious cult.

  35. So succinct, m-l! Thanks, I’ll file that.

  36. Not as old, but just as cool: the Kantilena (Wiki, manuscript) as it probably would have sounded to its contemporaries.

  37. David Fried says

    Just wanted to say “thanks” for an hour of nerd entertainment on a Saturday (!) night; first trying to spot the Semitic roots I know from Hebrew in Gilgamesh, and work out a little of the structure; then watching Fruitcake the Dane, or whatever he called himself, do Proto-Norse; and then Liberakos ho rhapsodos annihilate Ancient Greek. I’m fairly sure that in Ancient Greek prosody, long vowels were about twice as long as short, not eight times. . . In all fairness, Fruitcake and Liberakos are plainly two of the bravest human beings who ever lived. . .

  38. I’ve seen English described as being descended from a language invented by Norman squaddies for picking up Anglo-Saxon barmaids. In other words, a creole. To that extent, I suppose its Romance ancestry, even if somewhat bastardised, should be allowed. Or maybe not, I don’t see any Romance syntax or inflection in there, just a bunch of vocabulary.

  39. marie-lucie says

    chris y: Several languages have been described (not by linguists) as having been “invented” by this or that group. It is not as easy as it might seem to “invent” a new language, as opposed to agreeing on a limited indispensable vocabulary for communication on specific topics (this is how pidgins are formed, and they are often short-lived). Communication between “squaddies” and “barmaids” would hardly lead to the formation of a new language adopted by the whole population of a country.
    Syntax is not a very reliable indicator of relatedness, since it can change with close contact with another language, but inflection usually is. So are function words which casual “inventors” might not think about, such as articles, question words, or the verb “to be”.
    There is an extensive literature on creoles and their development, and the development of English doesn’t quite qualify.

  40. Trond Engen says

    DE: Arne Torp, of Oslo University, gives a very lively rendition of a Proto-Norse reconstruction of the Atlakvida
    Here‘s the same poem in Old Norse. It’s not hard to hear which language it was written in.
    AJP Snowcrown:
    I may add the not completely irrelevant fact that he’s the father of our national actress superstar Ane Dahl Torp. They did a funny little film together for the TV language program Typisk Norsk a few tears ago. That one seems to have dropped out of the net, but here‘s one she did. Not on Proto-Norse, though, but on, eh, strategies of avoidance.

  41. Interesting, Trond. Thanks for that. Too bad the other one’s disappeared.

  42. It also bears remembering that the Romans recognized that their gods were ‘the same’ as the Germanic gods—so Tacitus tells us that the Germans worship Mercury primarily (Odin being another name for the god of craft and language), and creating for us a funny equivalency between Wednesday and Mercredi.

  43. what, you mean Wednesday is from Odin?

  44. marie-lucie says

    Sashura, this origin is well-known. “Odin”, “Wotan”, and a few similar forms are just variants of the same name in different Germanic languages. Except for Monday, Saturday and Sunday, the other English day-names are based on those of the Germanic gods considered the same as the Roman ones. Perhaps in Germany the identification of Odin with Mercury was resisted, since Wednesday is “Mittwoch”, literally “Midweek”.

  45. Except for Monday
    What? Monday is named after Moon, Lundi is named after Luna.

  46. “Moon” is not a Germanic god.

  47. Oh I see. But it is the Germanic word for the celestial body associated with Luna, right?

  48. Yes.

  49. marie-lucie says

    “Luna” is just the personification of “luna”, the ordinary Latin word for ‘moon’.

  50. John Cowan says

    Perhaps in Germany the identification of Odin with Mercury was resisted, since Wednesday is “Mittwoch”, literally “Midweek”.

    There is an archaic and dialectal Wodenstag, just what you’d expect. Dienstag is half-borrowed from Low German or Dutch; the pure High German forms Ziestag, Ziestig are dialectal.

  51. David Marjanović says

    Perhaps in Germany the identification of Odin with Mercury was resisted

    No – instead, Christian missionaries got rid of them both.

    literally “Midweek”

    Except for the gender reassignment surgery that brought Woche f. into agreement with Tag m..


    Dienstag comes, says, Wikipedia, straightforwardly from Mars Thingsus, the Romanized avatar of OHG Ziu. In the diocese of Augsburg, the missionaries got rid of him too, and his day used to be Aftermontag.

    Still speaking of Tuesday, Ertag was formerly widespread in Austria and Bavaria; sidestepping the Romans, it continues the Greek Ares and is blamed on the Goths, who probably confused Ares and Arius. Likewise, Thursday was Pfinztag in at least some of the same places, from Greek pente; compare Pfingsten “Pentecost”.

  52. Aftermontag

    Very logical system:

    Monday, Aftermonday, Midweek, Aftermidweek, Friday, Afterfriday, Sunday.

  53. PlasticPaddy says

    Was the original system Moon Day, Mars Day, Mercury day, Jupiter Day, Venus day, Saturn Day, Sun Day? Or was this Latins and their influence on the rest?

  54. The original system was Sun Day, Moon Day, Nergal Day, Nabu Day, Marduk Day, Ishtar Day, Ninib Day.

  55. John Cowan says

    PlasticPaddy: Well, the Babylonian version was the original: days named after planets named after gods. Everyone else copied it, using their own gods. The order of the planets follows a pattern. If you sort the planets in order of decreasing distance from the Earth (Saturn, Jupiter, Mars, Sun, Venus, Mercury, Moon), then each successive day comes three planets after the preceding one, treating the list of planets as cyclic.


    In the Old English (and Hobbit) calendar, the first month of the year was Afteryule and the last month Foreyule (modernized); the middle two months were Forelithe and Afterlithe.
    The Hobbit week was calqued on an Elvish original, and the days were Sterday (from the stars), Sunday, Monday, Trewsday (from the Two Trees), Hevensday, Mersday (from the sea), Highday. The last was “the chief day of the week”, so although the names obviously parody ours, they are offset by two days.

  56. January First-of-May says

    Everyone else copied it, using their own gods.

    Except the Jews, who, presumably lacking sufficient gods, just named the non-Sabbath days in order.

    (Similar naming practices went on in other monotheistic societies later; e.g. Slavic days of the week are essentially named “Nowork, Afternowork, Second, Middle, Fourth, Fifth, Sabbath” – Russian having later replaced “Nowork” by “Resurrection” as the original came to mean “week”.)

  57. Estonian goes “Firstday”, Secondday”, “Thirdday”, “Fourthday”, “Friday”, “Washday”, “Holyday”, in what looks like mostly Slavic with two Germanic outliers (wonder how that came about exactly; the two Germanic ones are not exact cognates with Finnish so they could have come later from Low German).

    Amusingly, despite no actual cognates, most of the English names could be also folk-etymologized without too much violence into Oneday, Twoday, (Wednesday), Thirdday, Fiveday, Sabbathday, Se’enday. (We don’t speak about Fourday.)

  58. PlasticPaddy says

    Or if your clergy thinks fasting is good for you, you have “first fast”, “between fasts” and “fast”.as wed-fri.

  59. David Eddyshaw says

    Arabic has First, Second, Third, Fourth, Fifth, Meeting, Sabbath.

    Icelandic has Sunday, Monday, Third, Midweek, Fifth, Fasting, Washing (a pragmatic people.)

  60. @John Cowan: Tolkien made a little-noticed error in laying out the hobbit calendar. The One Ring is destroyed on March 25, which becomes New Year’s Day for the Fourth Age calendar. This was to match up with the Old English calendar in the Sixth Age. However, the English New Year was an approximation on the spring equinox, which does not fall in March (i.e. today) on the hobbit calendar, but rather on 1 April.

  61. John Cowan says

    The New Year was in fact March 25 Julian, which is why income-tax day in the UK is April 6 (it should be April 7, but the necessary adjustment was overlooked in 1900). It’s also notable that the Fellowship sets out on December 25 (which of course had no special significance then) and Frodo and Sam’s journey lasts exactly three months.

  62. Three months plus two days of Yule.

Speak Your Mind