Duolingo’s Yiddish Course.

Jordan Kutzik writes for the Forward about Duolingo’s new Yiddish course:

Although the course is significantly shorter than many of the site’s 39 other languages, Duolingo Yiddish is still massive. Altogether, it encompasses 70 sections called “skills,” with each skill featuring five levels. The 350 levels have three to six lessons each. With every lesson requiring at least five to seven minutes, the roughly 1,300 lessons will take a minimum of 250 hours for the average student to complete. […]

The course gives a thorough overview of Yiddish grammar. Taught through a series of exercises built like a video game to incentivize memorization, Duolingo Yiddish begins with standard greetings, home and food vocabulary and regular day-to-day topics from telling time to describing family members, shopping trips and vacations. Specifically Jewish vocabulary is introduced fairly late, with the first such lesson, on Shabbos, appearing about halfway through. […]

Duolingo is known for prompting students to translate funny and even bizarre sentences, and its Yiddish edition doesn’t disappoint. The sentence “di yidn zenen mid” (the Jews are tired) is destined to become a meme on Twitter and “mayn vayb iz keynmol nisht tsufridn” (my wife is never pleased) sounds like the opening of a classic albeit decidedly dated Borscht Belt routine. “Ver voynt in an ananas untern yam?” will get a laugh from many younger millennials who grew up watching “SpongeBob Square Pants.” It translates to the first line of that show’s theme song: “Who lives in a pineapple under the sea?” […]

Duolingo Yiddish’s “incubator stage” was launched some five years ago, but the course was effectively redesigned from scratch when a new group of volunteers took over the project in 2019. Before work could begin in earnest, the crew, made up of young native Yiddish-speakers who grew up in Hasidic and Yiddishist homes, had to decide what form of the language to teach. Yiddish courses and textbooks typically use a standardized dialect often called “YIVO Yiddish,” which is unfamiliar to the vast majority of contemporary Yiddish speakers, who hail from Hasidic communities. Israel and Isac Polasak, the project’s most active volunteers, grew up in a Satmar Hasidic community and wanted the course to focus on the dialect spoken by Brooklyn Hasidim. They soon hit a snag, however.

“I realized that we couldn’t make a course with Hasidic Yiddish because it wasn’t standardized,” Isac Polasak said. “It wasn’t a matter of better or worse. We just needed a standard to work with.”

The twins put out a call for volunteers who knew formal Yiddish grammar to join the project. Meena Viswanath, 32, a civil engineer and scion of a prominent Yiddishist family, saw their Facebook post and soon became the project’s grammar maven, editing the lessons and double checking spelling. […]

While all of the participants agreed to base the course on standard literary Yiddish, they reached an impasse when deciding on how the language should be pronounced. The Polasaks and other Hasidic volunteers wanted to go with a pronunciation that matched the Yiddish they learned at home while Viswanath and others wanted to teach the so-called “YIVO pronunciation” traditionally used in academic institutions because it more closely corresponds to standard Yiddish spelling. Unable to come up with solution, the group decided to punt, releasing a survey in November 2019, which garnered 6,000 responses, with half of participants voting for the Hasidic pronunciation.

“Ultimately it’s a good compromise,” Viswanath told me in Yiddish. “Most Yiddish-speaking communities use this pronunciation and speakers of Hasidic Yiddish did most of the work on the course. There were also fewer resources for students to learn it.”

Indeed, Duolingo Yiddish is the first major course that reflects how the majority of Yiddish speakers pronounce the language today. Its transliteration scheme bears this out. The Yiddish word for “and,” און, which in the northern dialects and standard Yiddish is pronounced as “un,” is transliterated in the Duolingo course as “in.” The greeting שלום־עליכם is similarly transliterated as “shulem alaykhem” with the “ay” pronounced like the word “eye” as opposed to the “sholem aleykhem” found in textbooks.

For those wishing to dig deeper into contemporary Brooklyn Yiddish, a special skill set featuring 35 lessons on Hasidic Yiddish teaches grammatical forms not covered in most courses (for instance the southern Yiddish forms of “you,” “you all,” and “yours” ets and enk) and vocabulary you won’t find in most Yiddish dictionaries. Among them are borrowings from English like “mufn” (to move, i.e. change residences) and “vakn” (to walk), terms from the old country not familiar to most non-Hasidic Yiddish speakers like shtayt (slow), tshionish (skinny) and bundatsh (French toast, a borrowing from Hungarian) and more contemporary slang like “Hak a lebn” (literally: chop a life, i.e., “have a blast”). And for those thanking a storekeeper in Williamsburg, the Hasidic way of saying “thank you” is not “a dank” but “sh’koyekh,” a shortening of the Hebrew term yasher koyekh (literally: “may your strength be firm.”)

I love ets and enk. We discussed the prospective course, and Meena Viswanath, last year, and it sounds like it turned out very well. And for those interested in investigating Hasidic Yiddish further, here is a set of videos, courtesy of rozele.


  1. Is Yiddish „enk“ related to the Bavarian „enk“ (dialect for „euch“)?

  2. That’s my guess, but surely someone will actually know.

  3. David L. Gold says

    @Vanya and LH

    Yes, the source is presumably Austro-Bavarian.

    The forms are

    nominative עץ (ets)
    oblique (dative and accusative) ענק (enk)
    possessive adjective ענקער (enker)
    verb ending -עטס (ts-)

    They are used when addressing at least two persons each of whom is addressed with the familiar pronoun, דו (du).

    At least since the nineteenth century they have been used in a minority of topolects. Earlier texts provide evidence for wider use.

  4. David Marjanović says

    nominative עץ (ets)

    That’s interesting. The Bavarian dialects that have [ẽŋk] for the oblique have [es] for the nominative, without [t]. Throughout Bavarian, the clitic form is [s]; however, the verb ending is [t͡s].

  5. Christopher Culver says

    If a course focuses on the literary standard, the student can at least enjoy all the great Yiddish literature and film. But are the American Hasidic communities actually open to talking to outsiders (especially non-Jews), so that if a course teaches modern Hasidic Yiddish the student can actually use it in practice? My only contact with the ultra-Orthodox has been hitchhiking in Israel. The Haredim are great supporters of hitchhiking and they would stop for me, but once I was in the car they weren’t at all open to conversation, even when the journey lasted a long while. I could only assume it was because they wanted to limit their exposure to things from outside their own community.

  6. Presumably the students can use it with each other, and I doubt Hasidic Yiddish is so distinct they can’t use it with anyone but Haredim. At any rate, they polled prospective users and the vote was for that pronunciation, so there you go.

  7. @ David M.

    One possibility is that Yidish ets was originally *es, which became ets /ec/) under the influence of the verb ending.

    S.v. enker the Grimms note this passage: “herr Hans, wo habt esz enker schäfle? wo sind enker arme leut?” ( Frey garteng. 39) and add “so auch in Östreich, Kärnten, Tirol.”

    Is sz merely the earlier form of the scharfes S ? Or could sz have represented *[c] in older German spelling?

    I notice, for the first time, Östreich. The Yidish for ‘Austria’ is bisyllabic everywhere: עסטרײַך (estraykh), which I had till now throught was a Yidish innovation.

  8. 1-Christopher Culver’s story above reminds me of one of my first summer jobs, as a lifeguard at an indoor pool in Montreal. Among our clientele were children from two Jewish day camps, segregated by gender, along with the adults supervising them: the boys were from ultra-orthodox families, and not only did most speak Yiddish among themselves, there were a number of Yiddish monoglots among them, with one of the adult supervisors translating what I and the other lifeguards had to say.

    One fine day, while the children were leaving and heading for the locker room, one of the Yiddish monoglot children, in the presence of a bilingual adult supervisor, came to me and had a question. Now, I have never studied Yiddish, but I already had some German and knew of some ways in which Yiddish and German differed, and thus I understood the child’s question perfectly, and actually managed to give an answer in something the child understood perfectly. While the child was VERY pleasantly surprised, and otherwise satisfied with my answer, the adult supervisor was likewise surprised. But NOT in a pleasant way. His behavior and his colleagues’ dramatically changed thereafter: while before this incident it had been common for the supervisors to stay away from the pool and to keep an eye over things from a distance, after this incident one of the supervisors was ALWAYS within earshot of me whenever the children were. And I was the only lifeguard given this special treatment. Plainly, they were afraid I might say something, directly to the children (audible gasp), in my very limited (Schoolboy German-influenced L2) Yiddish.

    Meaning that Christopher Culver’s suspicions, above, ring very true for me.

    2-On the Yiddish pronoun /ets/: David Gold’s suspicion about its origin rings very true to me: phonologically similar pronouns and person-marking morphology are very liable to influence one another.

  9. Andrew Dunbar says

    > Its transliteration scheme bears this out. The Yiddish word for “and,” און, which in the northern dialects and standard Yiddish is pronounced as “un,” is transliterated in the Duolingo course as “in.” The greeting שלום־עליכם is similarly transliterated as “shulem alaykhem” with the “ay” pronounced like the word “eye” as opposed to the “sholem aleykhem” found in textbooks.

    This had me thinking they must’ve made the course in transliteration rather than in Hebrew script, which surprised me since I’ve been doing several DuoLingo courses and they all use the native script.

    So after I finished my current English-for-Thai-speakers lesson, I did the Yiddish placement test and there is no transliteration to be seen. Which makes me glad. But also makes me puzzled about the above quote.

    Oh and my placement test got me to level four in the first “skill”, on the alphabet. Thanks to my tourist German and interest in all alphabets and writing systems.

    Now if only Duolingo would make a Thai for English speakers course!

    (And Albanian, Armenian, Basque, Burmese, Cantonese, Estonian, Georgian, Icelandic, Khmer, Latvian, Lithuanian, Maltese, Mongolian, Tamil, and Tibetan of course)

  10. the pronunciation skirmishes over the Duolingo course were bitter, and may mark the emergence of a new camp in the long-running yiddish language war: non-cradle-tongue speakers & learners who aren’t part of the hasidic world but prioritize speaking a yiddish that’s closer to the largest active* dialects, and generally reject the utility of the YIVO-yiddish “standardization” project (at this point a nominal dakh-shprakh but in practice a distinct dialect used by almost no speakers of other dialects).

    i’m more or less part of that camp, but was one of the few advocates of the 3rd pronunciation option: southeastern yiddish, which (in its voliner/volhynian version, the standard theatrical dialect) has a history as a trans-regional pronounciation in common use (unlike either hasidic** or YIVO yiddish) and is the heritage dialect of most descendents of yiddish speakers.

    from what i’ve seen so far (i started the course at the beginning to be able to look at the whole thing but am only at the start), pronunciation is the course’s only deviation from YIVO-yiddish. that leads to some truly ridiculous moments, like seeing װוּ, a spelling only found in the YIVO sphere (everyone else has continued to use װאו, avoiding the need for a diacritic by using a silent “mekhitse alef”), pronounced as /vi/, the (historical & current) majority pronunciation that YIVO rejected for the northeastern dialect’s variant /vu/. similarly, it gives “yidish” as ייִדיש, which again has a YIVO shibboleth diacritic (ידיש, אידיש, and even יידיש being the alternatives – to some extent but not entirely tracking a a pronunciation without the initial glide). i’ll report in as i go if interesting things turn up…

    i’m glad at least there’s a supplement with actual satmar yiddish! and not just because of ets & enk, which i do like. if my memory serves, they’re only one of quite a few sets of pronouns used in various dialects – all of course excluded from the charmed circle of YIVO-yiddish. i think tracking them all down would be a great excursion in this epic dataset… but it won’t be me any time soon, alas.

    * and historically largest, for that matter. YIVO-yiddish is oriented towards northeastern yiddish (mostly because YIVO was in vilna, and bought into the idea that litvaks were more “cultured” than other yiddish speakers), which was spoken by around 1/3 of pre-attempted-genocide yiddish speakers.

    ** for Duolingo-drama purposes, “hasidic yiddish” is pretty specifically the dialect spoken by the satmar hasidim**, which is unterlander yiddish, rather than what’s spoken by other courts (i think mainly polish yiddishes). unterlander yiddish (from what’s now eastern slovakia, northwestern romania, and far southwestern ukraine***) is its own idiosyncratic animal – as i understand it, it’s closest to polish yiddish (and so in the mid-eastern a/k/a central dialect cluster) pretty distinct.

    *** currently the largest hasidic court (counting its two branches together) – but that only puts it at ~20% of the hasidic world, though a larger percentage of new york’s hasidim.

    **** in contrast to oyberlander yiddish, from what’s now western slovakia, parts of northern hungary, and a slice of far eastern austria. it’s an east-west division not an altitude one; no real relation to the goyish geographic notions of “lower hungary” and “upper hungary”.

  11. @rozele,
    in its voliner/volhynian version, the standard theatrical dialect — how did that happen? How universal was that? In Israel, the last great names of Yiddish theater were the duo of Dzigan and Schumacher, both from Łódź; did they speak Voliner Yiddish on stage?

  12. I was looking for Dzigan skits, and found this one (Yiddish with Hebrew subtitles). Dzigan was one of these comedians who could be hilarious with a single eyebrow (the XXL nose helps too.) Anyway, the other actors are Gideon Singer, who was from Brno, and Carol Feldman, from Tulcea (in SE Romania). Are they all speaking some standard dialect, or are they each using their native variety?

  13. David L. Gold says

    @ Andrew Dunbar. “This had me thinking they must’ve made the course in transliteration rather than in Hebrew script, which surprised me since I’ve been doing several DuoLingo courses and they all use the native script.”

    You are on target. Had they begun the project, as I believe they should have, with Yidish in Yidish letters rather than in romanization, they could have used the Standardized Yidish Spelling of 1937, which is tied to no single pronunciation and is therefore suitable as the basis for teaching any pronunciation, just as you can teach people to read and write English or French or Spanish, etc. and chose whichever pronunciation you want them to acquire without making any changes in the spelling that you teach them.

    The text I am now writing, for example, can be pronounced in any of a number of ways, so that if, for instance, you are teaching a British pronunciation, you would tell your students to pronounce the word been, used above, /bi:n/, and if you are teaching an American pronunciation, you would tell them to pronounce it /bIn/.

    So too in Yidish. The most frequent word for ‘synagog’, for example, is spelled with three letters (שול) — shin vov lamed — each of which represents one of the three phonemes of the word, yet the spelling does not tell you how each of the phonemes should be realized, which is left up to the teacher.

    @ Christopher Culver and Etienne. You too are on target. The ultra-Orthodox do not want more contact with outsiders; they want as little as possible because they fear that outside influences will lead to defections (search for the string “leaving ultra-Orthodox families youtube”).

  14. Is sz merely the earlier form of the scharfes S ? Or could sz have represented *[c] in older German spelling?
    The first assumption is correct. It’s just how the Grimms write “ß”, and “sz” is still one of the conventional ways to write “ß” when the letter isn’t available.

  15. David Marjanović says

    S.v. enker the Grimms note this passage: “herr Hans, wo habt esz enker schäfle? wo sind enker arme leut?” ( Frey garteng. 39) and add “so auch in Östreich, Kärnten, Tirol.”

    Is sz merely the earlier form of the scharfes S ?

    Yes, and it’s used here to differentiate the word from es “it” by marking the vowel as “long”, which is the only way to indicate [e] as opposed to the [ɛ] of es. I’ve actually seen it spelled elsewhere.

    (In most Bavarian dialects, vowel length is wholly predictable, and in the remainder nearly so; conversely, Standard German and various dialects reshuffled the vowel qualities to agree with the quantities after the latter were reshuffled around the end of the Middle Ages, but the qualities were never reshuffled in Bavarian.)

    Or could sz have represented *[c] in older German spelling?


    I notice, for the first time, Östreich.

    That’s how it’s pronounced in (much of?) Germany; never in Austria, though.

    “sz” is still one of the conventional ways to write “ß” when the letter isn’t available

    It seems to have died out in the last 30 years, though, and doesn’t seem to be mentioned anywhere in the spelling reform of 1998/2005.

  16. PlasticPaddy says

    Is ß still called “ess-zet” then or is there an alternative name? Oh, i see Hans has “scharfes ess”.

  17. David Marjanović says

    “Scharfes S” is the name I grew up with.

  18. I’ve encountered “scharfes S”, but I personally use “es-zet” and so do most people I know.
    It seems to have died out in the last 30 years
    Possible. It’s not frequently used, and I am a dinosaur anyway 🙂

  19. oyberlander yiddish, from what’s now western slovakia, parts of northern hungary, and a slice of far eastern austria

    In other words, the dialect spoken by the former Jewish community centered around Pressburg/Bratislava? There isn’t that much Austria left east of Vienna, “a slice of far eastern Austria” would just be a few small towns in Burgenland.

  20. I’ve encountered “scharfes S”, but I personally use “es-zet” and so do most people I know.

    The way I learned it more than 50 years ago, the letter ß is always called [es’tsɛt]; scharfes s, on the other hand, is a term of popular phonetics for voiceless [s], whether it is spelled ß, ss or s. And if you had to use a typewriter lacking ß, you used ss instead, never sz.

    In Germany this is always trisyllabic, usually with [ə] in the second syllable, although there is also a variant with [ɛ] (which sounds pedantic and unnatural to my ears).

  21. David Marjanović says

    The way I learned it more than 50 years ago, the letter ß is always called [es’tsɛt];

    (You mean [ɛs’tsɛt].)

    scharfes s, on the other hand, is a term of popular phonetics for voiceless [s], whether it is spelled ß, ss or s.

    That doesn’t work in Austria. 🙂 Over here, [z] apparently never developed*, and the length contrast for consonants never disappeared**, so that allows scharfes S for ß, Doppel-S for ss and rundes/normales S for s (“round” in contrast to the obsolete ſ, langes S). Thanks to a round of word-final shortening of long consonants, ß doesn’t line up with /sː/ even in Austrian Standard German:


    [s ~ z̥] (free variation, always voiceless)

    niesen (admittedly that’s irregular)

    * It’s a completely exotic sound that I had to learn to articulate for French and English.
    ** Except in Carinthia, where the whole sound system got reinterpreted in Slovene terms. However, while voiced [b d g] developed, [z] did not; instead, s, ss, ß have simply merged.

    In Germany this is always trisyllabic

    I’ve never heard a German say it that way. I wonder what the regional variation in Germany is, then.

    there is also a variant with [ɛ] (which sounds pedantic and unnatural to my ears)

    It is! For phonotactic reasons, in front of /r/, you’d expect [ɐ], which indeed is the middle vowel in Austria.

  22. scharfes s, on the other hand, is a term of popular phonetics for voiceless [s], whether it is spelled ß, ss or s.
    Yes, that’s actually what I remember from school (I wasn’t sure anymore, too long ago).
    And if you had to use a typewriter lacking ß, you used ss instead, never sz.
    Where I have seen SZ is in book titles and headings that are all caps. I have e.g. a German Shakespeare edition from about the 1950s where “Measure for Measure” is MASZ FÜR MASZ.
    Trisyllabic with schwa in the second syllable for me as well. While I have encountered the disyllabic variant in writing (something I associate with older books or poetry), I don’t remember ever consciously having heard it. But swallowed schwas are not unusual.

  23. David L. Gold says

    I see several differences between Yidish usage and Austro-Bavarian usage. To take this sample sentence from the Grimms’ dictionary:

    “herr Hans, wo habt esz enker schäfle? wo sind enker arme leut?”

    1. Habt shows that the ending is -t. In Yidish it is -ts (pronounced [c]). Hence, ‘where do you have…?” is vu hots ets…? OR avu hots ets…

    2. herr Hans tells us that esz can refer to one person. Yidish ets is a plural familiar form that refers to two or more people, each of which is individually addressed by du.

    3. The Yidish equivalent of enker arme leut is enkere oreme-layt. Thus enker is a singular form only (vi heyst enker kind? ‘what’s your child’s name?’ vi heyst enker tate?’ what’s your father’s name?’ vi heyst enker mame? ‘what’s your mother’s name?”).

    If the noun is plural, the form must be enkere: enkere kinder, enkere tates, enkere mames.

    And the imperative form in Yidish ends in -ts, as in, kinder, kumts esn!

  24. David L. Gold says

    The German bisyllabic form occurs at least in nineteenth-century publications, for example,

    Julius Franz Borgias Schneller: Geschichte von Oestreich und Steiermark (Dresden, 1828)

    János Nepomuk und Jozsef Mailáth: Geschichte des östreichischen Kaiserstaates (Hamburg, 1850)

    also as a family name: Oestrich ~ Östreich

    as in Christian Östreich (1867 Neu Isenburg – 1951 Starkenburg).

    This article is available online:

    Karl Roth: Oesterreich oder Oestreich. In: Kleine Beiträge zur deutschen Sprache, Geschichts- und Ortsforschung. 1, 1850, S. 179–182.

  25. PlasticPaddy says

    Herr Hans is a squire or person of quality, so is addressed using a plural form by lower ranking persons. Although Germany never had a successful revolution, this usage is no longer current. Some people with aristocratic names use the titles, say Graf down to Freiherr/frau/in and if you don’t know them, you could address them using Herr/Frau/Fräulein + the title but that would also be the case with Doktor and maybe some other professional titles, including but not limited to Wachtmeister. It is usually best to say Jawohl, Herr Leutnant (heel clicking is optional), otherwise your interlocutor might become angry and break his sword over your head.

  26. PlasticPaddy says

    Regarding how one should address Karl-Theodor Maria Nikolaus Johann Jacob Philipp Franz Joseph Sylvester Buhl-Freiherr von und zu Guttenberg, one should be aware that both “Freiherr zu Guttenberg” and “Herr Doktor (zu) Guttenberg” could be problematic, and one should probably try “Karl-Theodor”, even if one has not yet been properly introduced.

  27. @David L. Gold

    Habt shows that the ending is -t. In Yidish it is -ts

    The Austro-Bavarian ending is also -ts, at least in the dialects I am familiar with. Perhaps Grimm’s transcription wasn’t that accurate? David M. certainly knows more.

  28. Andrej Bjelaković says

    While we’re on Bavaria, David, why is Lothar Matthäus pronounced /maˈtɛːʊs/ and not something like /ˈmathɔʏ̯s/?
    Or, if you want, why is maˈtɛːʊs spelt like that?

  29. David Marjanović says

    Why talk to Dr. Googleberg at all? I don’t think he has anything interesting to say…

    “herr Hans, wo habt esz enker schäfle? wo sind enker arme leut?”

    1. Habt shows that the ending is -t.

    I strongly suspect the whole sentence, except the words of interest, has been rendered in Standard German or nearly so, perhaps to avoid distracting from the topic, or simply to avoid the very tricky issue of how to represent a Bavarian sound system with Standard German spelling conventions.

    Strikingly, -le isn’t Austro-Bavarian at all. It’s Swabian.

    Similarly strikingly, the representation of Leute can hardly be Austro-Bavarian either. The loss of the final -e is indeed universal there, but it’s much more widespread; eu, however, has merged into ei, except in eastern Austria where it has somehow escaped the general unrounding and become [ɶ].

    Likewise, sind isn’t Austro-Bavarian; it’s been wholly replaced by the last surviving “present subjunctive” form seien, pronounced [san ~ sãn].

    In the indicative and the imperative, the 2pl verb ending is [ts] in all forms of Austro-Bavarian I’ve ever encountered. I find the Alemannic [d̥] very noticeable.

    herr Hans tells us that esz can refer to one person.

    That was two hundred years ago, when the 2nd person plural (elsewhere uppercase Ihr) was the most polite form of address. Nowadays it’s flat-out impossible – and so is combining Herr/Frau with a first name.

    Thus enker is a singular form only

    Indeed its use for the plural strikes me as rather odd. Perhaps Grimm was citing from memory and couldn’t quite remember which -e to drop?

    And the imperative form in Yidish ends in -ts, as in, kinder, kumts esn!

    That’s universal in Austro-Bavarian, and a striking omission from the quote; it’s a strong reason to think it’s been adapted to illustrate only the topic of the dictionary entry.

    Anyway, here’s how I’d pronounce the quote in my (non-rhotic) dialect, the total lack of idiomaticity aside:


    Germany never had a successful revolution

    1918 kinda was, though. It was a bit low-key, but less so than Austria’s.

    maybe some other professional titles



    I think that’s obsolete. (But Inspektor, with /ʃ/, might not be in Austria.)

  30. David Marjanović says


    Greek/Latin eu gets identified with German eu, but Greek aio/Latin aeu is not.

    Bavarian has nothing to do with it, this is all spelling-based. So is, for most people who do that, the use of [ɛː] which not many people have in their sound inventory; I pronounce him with [eː] unhesitatingly.

  31. PlasticPaddy says

    [ɛː] is interesting. Is it clear that, before the late 1800s, any modern (i.e. not MHG) dialect (apart from the Standard one) expressed the Bären/Beeren distinction by applying [ɛː] to the former word? I suppose there must have been one (otherwise it would have had only a historical justification to be in the standard).

  32. David Marjanović says

    Along the Rhine, people do have a wide-open [æː] natively. The first time I heard someone from Frankfurt speak (I overheard a casual conversation in Standard German), this threw me off so far that it took me a second or two to remember that this was German and I was supposed to be able to understand it! Then I understood it again…

    Bären never had that, though. I think I’ve even seen a prescriptivist say pronouncing it with [ɛː] is hypercorrect. Its spelling, if it’s historical and not purely a graphical device to distinguish it from Beeren, must date from the probably short period after its historical /ɛ/ had already been lengthened to [ɛː] in its open syllable (and in the monosyllabic word Bär*), but not yet lifted to /eː/.

    * The lengthening of open syllables came from the north and hasn’t reached Switzerland. The lengthening of monosyllabic words that don’t end in too many consonants** came from Switzerland and hasn’t reached most of Dutch or the currently-or-formerly Low-German-speaking areas.

    ** Maybe there’s a requirement for every content word to contain at least three moras, whether distributed over two syllables or just one… I’d need to postulate that the vowels of stressed open monosyllables are allophonically overlong, but I can actually live with that, because it would fit the general dependence of vowel length on stress.

  33. Matthäus
    Fans of Borussia Mönchengladbach, like me, pronounce that name Judas anyway 🙂
    To the point, German orthography normally doesn’t distinguish between cases where a vowel digraph designates a diphthong or two separate vowels. The latter is the case mostly in some loan words, e.g. Mosaik “mosaic, tesselation” has “a?i”, not a diphthong, and across morpheme boundaries, e.g. in beurkunden “to document, certify” with “@?u”. You simply need to know that beinhalten is “to contain” (be-inhalten) and not “to hold a leg” (*bein-halten).

  34. David Marjanović says

    Oh yes. The Dutch have the courage to use ë and ï; farther east, that somehow never caught on (outside of two or three surnames like Groër /ˈgroːɐ/). And so, you need to know that Batterien and Belgien don’t rhyme, and that Kopien now almost exclusively rhymes with the former (-/ˈiːn/) but used to rhyme with the latter (-/ɪɛn/ south, -/jɛn/ north).

  35. Kopien now almost exclusively rhymes with the former (-/ˈiːn/) but used to rhyme with the latter (-/ɪɛn/ south, -/jɛn/ north).

    I’m officially giving up on German.

  36. Kopien now almost exclusively rhymes with the former (-/ˈiːn/) but used to rhyme with the latter (-/ɪɛn/ south, -/jɛn/ north).
    Interesting. I have [-‘i:@n] in both Batterien and Kopien, and that’s also the pronunciation Wiktionary notes. I have heard the former pronounced with [-i:n], but never consciously the latter. Duden and Wiktionary note your Southern pronunciation as Austrian, so that lines up. So for me, the difference to Belgien is in the stress (on the stem) and in the length (short “i” in Belgien).

  37. David Marjanović says

    Interesting. The presence of a separate schwa evidently has repercussions throughout the phonotactics…

    I’m officially giving up on German.

    Think of it this way: the model has shifted from a Latinate copia to French copie, helped by the increased frequency of the verb kopieren.

    (…Now that photocopies are much rarer than 20 years ago, that might change again, who knows.)

  38. Think of it this way

    No, I get how it happened, but I used to feel at least I could pronounce German when I read it.

  39. The presence of a separate schwa evidently has repercussions throughout the phonotactics…
    Just to be clear, the schwa is only there in the plural, it’s part of the plural suffix, not of the stem.

  40. David Marjanović says

    That’s it, though: you interpret the words as Batteri(e)-en, Kopi(e)-en with an underlying vowel in the plural suffix. Up south, we interpret them as Batterie-n, Kopie-n with no vowel in the ending, because -en endings are generally interpreted as representing (mostly syllabic) /n/ unaccompanied by a vowel.

  41. Okay, then we’re on the same page. I think it’s also relevant here that in many dialects in the Northern part of Germany (also in parts of Dutch, including colloquial Standard Dutch), the historical suffix -en has become /@/, i.e. dropping the /n/, not the vowel.

  42. David Marjanović says

    Ah yes. Also in all of Alemannic in the widest sense (I think).

  43. Kriĉjo, the course was primarily designed by American Hassidim, so there’s your answer to that one

  44. @Y:

    i don’t actually know the history of teatr-yidish – i’ll ask shane baker if he knows the next chance i get (he speaks it as his main dialect because of learning from the last pre-war generation of actors; they let him teach theater electives at YIVO but give him side-eye about it). i suspect it has to do with some combination of dialect prestige and founder effect (early modern yiddish theater having been particularly strong in bukovina and galitsye), but also some notion of wide intelligibility across the “polish/central” and “ukrainian/southeastern” dialects (i.e. the majority of yiddish speakers, with the stereotype of litvaks as snooty and dry making them even more marginal than their smaller numbers).

    my understanding is that voliner was the basic “neutral” theater dialect, but that a lot of different dialects turned up as part of characterization (theater english is similar, though with more range of the “neutral” base allowed these days). that’s definitely what i hear in yiddish film dramas. in the dzigan clips i’m hearing some differences, but all southern (polish/ukrainian) yiddish. i hear dzigan himself as more polish, but i’d expect comic actors would have more leeway to stick closer to their home dialects. and i don’t know the tells for lodzher yiddish well enough to pick it out for sure.


    i don’t know the dialect maps well enough to be sure, but yes: bratislava would be the biggest place in the oyberlander zone (though i’m not sure where budapest falls beyond the “assimilated to magyar” layer), and the part that’s now in austria is a very thin strip that wouldn’t ever have been much, populationwise.

    @David L. Gold:

    every standardized yiddish spelling (in every alphabet) claims to be transdialectal. none of them are, except for salomo birnboym/birnbaum’s lovely and maddening transliteration scheme* – and it only manages it by being a fully philological project that transcribes based on protovowels, which makes it more or less unusable unless you’re a historical linguist.

    the problem isn’t /u/ vs /i/ for ו or /aj/ vs /a:/ vs /ej/ for ײַ, or even that ױ is /ej/ in the northeast when it’s /oj/ in the south but /oj/ in the northeast when it’s /ou/ in the south – it’s that זון is /zʊn/ for both “son” and “sun” in the northeast, but /zin/ for “son” and /zɪn/ for “sun” in the south (to give one classic example). only birnboym, with his “zjjn” / “zjn” spellings gives any trace of that distinction (among others) – everybody else just shuffles which parts of which dialects’ pronunciations they’re gonna claim don’t matter.

    * birnboym’s “yiddish: a survey and a grammar” came back into print a few years ago, and is tendentious and informative and fascinating and (by me, anyway) fun. it’s got everything, including a 40-page section of 76 short texts covering 800 years and as many dialects as you want to argue they represent, given in his transliteration system.

  45. I enrolled in the Duolingo Yiddish course inspired by this thread, thank you. When I joined, the evening of the post, there were about 12K participants and now there are over 50K.

    For me it helps with my Duolingo Hebrew course because the Yiddish course pronounces much more of the text. Apparently there is not good enough automatic text to speech software for Hebrew and perhaps for Yiddish as well. Even though the words are quite different, because the Yiddish is written using a Hebrew alphabet, hearing it pronounced seems to help me convert text to sound. I know the alphabets and sounds aren’t exactly the same in the two languages but it’s better for me than trying to guess the Hebrew sounds without the hints.

    IANALinguist, apologies for any mangling of the jargon.

  46. Yiddish orthography is straightforwardly phonetic, so automatic text to speech should be as easy as any language. Pointed Hebrew is likewise phonetic without exceptions. However unpointed Hebrew, used in most contemporary texts, is impossible to do right without understanding the context, which no software can do at present.

  47. Stu Clayton says

    @rozele: but give him side-eye about it

    This is a neat expression I’d never read or heard. It reminds me of scheel ansehen.

  48. January First-of-May says

    Pointed Hebrew is likewise phonetic without exceptions.

    IIRC the so-called “kamats katan” (i.e. kamats representing /o/) is unpredictable or nearly so (I’m not sure if there are any minimal pairs), and the distribution of pronounced/unpronounced shva is predictable but very complicated. In some fonts there are also some esoteric edge cases involving the interaction of holam and shin/sin dots.

    (Wikipedia also mentions that vav-dagesh is indistinguishable from shuruk, though, again, I’m not sure if there are any minimal pairs.)

  49. Yiddish orthography is straightforwardly phonetic

    except where it isn’t* – which, for YIVO orthography used as directed, is pretty much everywhere outside the strictest YIVO classroom. i already mentioned the contrastive /i/ /ɪ/ distinction (both notated as “ו”), which operates everywhere except the northeastern dialects; another example is “אָ” notating both /u/ and /ɔ/ (and possibly /o/ as well) in satmar/duolinguo yiddish and elsewhere.

    and that’s not even getting into the actually existing use of the YIVO system, or the other, much more widely used orthographies. to point out just one messy area: א is officially silent in most orthographies; in YIVO’s, it only appears word-initially before non-alef vowels; in other orthographies, א is also used to clarify strings of repeated letters (distinguishing “װאו” from “ואװ”, for example, where YIVO has introduced וּ to mark the vowel); but in practice, both א & אַ are used for /a/*** by almost everybody.

    i’m sure a text to speech program could be built to deal with yiddish as well as one could be built to handle english, but i’m not convinced the tasks would be all that different.

    * i’m ignoring, outside these footnotes, the fact that in the currently most-used orthographies (and also in YIVO’s) almost all words of hebrew/aramaic origin are spelled according to their etymology**, while their suffixes and prefixes are spelled phonetically. except of course when they aren’t (plurals in particular, plus the odd lexified phrase).

    ** which mostly sorta predicts their pronunciation – but only if you know how they’d be pronounced in hebrew or aramaic (no, we [mostly] don’t use vowel pointing in yiddish, why do you ask?), and the patterns of sound change that lead to their yiddish cognates.*****

    *** in older texts, this may be a result of movable-type typographers saving their limited stock of אַ for situations where they needed to make it clear that the letter was not אָ, which was also often given as א where no ambiguity would arise. but nobody ever had that problem in handwriting or linotype, and we don’t have it on computers****, and the practice also appears in all three of those modes of writing.

    **** though we do have that problem on phones that think the jewish alphabet is the “hebrew” alphabet, and that “hebrew” means israeli (the language, not the citizenship), and so don’t give us anything but the unpointed glyphs israelis (the citizenship) favor.

    ***** this is why the only good 20thC yiddish spelling reform was the soviet one (the first round, before they went after the word-final letterforms).

  50. David Eddyshaw says

    IIRC the so-called “kamats katan” (i.e. kamats representing /o/) is unpredictable or nearly so

    Not really, as far as the pronunciation reflected in the Masoretic text is concerned, once you take stress into account (and stress is, of course, always marked.)
    Kamatz only represents /ɔ/ in closed unstressed syllables, where /a:/ doesn’t occur*. The corresponding phoneme became /o/ (represented by kholam) in stressed syllables, at some stage between the Hexapla period and the invention of the Tiberian pointing.

    There are a few edge cases with shwa, although again they can almost all be disambiguated by looking at stress; the matter is more complicated than with kamatz, though, because the actual distinction between shwa-as-actual-vowel and shwa-as-marker of vowellessness disappeared quite early in the Tiberian tradition (and left the bgadkpat phenomenon as contrastive, which it wasn’t up until that point.)

    *Here I’m covertly assuming contrastive vowel length, which admittedly is not marked in the Tiberian system; however, as I’ve remarked eslewhere, there seem to be excellent reasons for supposing that it existed in the TIberian scholars’ actual pronunciation of Hebrew, because otherwise the rules for stress sandhi are difficult to understand in any natural way.

  51. David Eddyshaw says

    There is an interaction between shwa vocalisation and the realisation of kamatz in unstressed syllables because of the question of whether syllables are closed or not, I must admit.

    Thus /qa:tə’la:/ “she killed” could be mistaken for /qɔt’la:/ be reading the shwa as mute (though in verbs with bgadkpat consonants you can reconstruct where “vocalic” shwa must have been.)

    However, the Tiberians must have made a distinction in the actual vowels themselves at the time that they wrote the cantillation marks, because the stress-dissimilated form before a word with initial stress in close combination is /’qa:təla:/, whereas a form of the shape /qɔt’la:/ does not become /’qɔtla:/, but instead loses stress altogether and is linked to the following word with maqqeph [this, despite the fact that /’qatla:/ actually is possible for an underlying stress pattern, incidentally.]

  52. (Not exactly Yiddish. This is as good a place as any:) I just found an 1837 book, a bilingual Hebrew/German (“ashkenazit”) edition of part of the Old Testament—with the German written in Hebrew letters. It’s not Yiddish: פאָן von instead of פון fun, איין ein indefinite article, plus other orthographic peculiarities, like גאטט Gott and זיא sie. I’d never heard of such a thing.

  53. Tiberian

    Every single time I read it as Tibetan.

    I am starting to wonder if there indeed was a Tibetan Hebrew somewhere sometime.

  54. John Cowan says

    Probably not. Tibetan Irish, however, is definitely a thing.

  55. @Y:
    ooh! that does look interesting!
    is it a word-by-word gloss in german in jewish characters? or a full translation? (i don’t have the hebrew to tell…) and the commentaries look like they aren’t given in the vernacular…

    either way, i wonder if it’s a late edition of a sefer composed in the period when german-in-jewish was relatively common?

  56. David Marjanović says

    The few lines I’ve deciphered made me wonder if the text was just mechanically transcribed from a contemporary rendering of Luther’s translation, but Google says it’s close but not identical. Still, it’s not at all vernacular, with dative -e (deinem Volke) preserved.

    Further peculiarities: lots of word-final z, perhaps to get around the fact that s and m are really hard to tell apart in this font; ü rendered as i, which is not surprising, because Karlsruhe is in a large unrounding area; ä, however, rendered as א with three dots underneath, which is really amazing; every -en spelled out; spurious contemporary th rendered as such; double consonant letters sometimes transcribed double, sometimes single; א used as a general vowel-length marker when there’s no silent h to transcribe.

    In short, whoever wrote the German Bible text definitely had the Latin-letter orthography in mind, but may not have been precisely transcribing an existing edition. I haven’t tried to decipher anything in the commentaries.

  57. I think it’s just a plain translation, e.g. Joshua 2:23 (apologies for transcription errors, I don’t know German): “Die beiden Männer gingen sodann vom Gebirge herab, setzten über den Jarden, kamen zu Jehoschua Sohn Nun, und erzählten ihm alles was ihnen begegnet war.” Is that passable/correct/idiomatic German?

    Proper names are written in Rashi script. <ä> is transcribed as אֶ, which is not used in Yiddish. /ts/ when written <z> is transcribed צ, but when written <tz> is transcribed טצ. Final s is transcribed as ז, i.e. /z/; is that a Yiddishism?

  58. David Marjanović says

    Oh, yes, it’s definitely a translation, not a gloss or anything.


  59. Proper names are written as in Hebrew, not a transcription of the German equivalent.

  60. i don’t know enough about westerly yiddishes and german-in-jewish conventions to know about the ז-. with easterly ones, in relation to english at least, it’s more the other way around, soundwise and in spelling (i was just reading in bessie tomashevski’s memoir about “מיסעס פּאלעק’ס גרענד האָטעל” [“mrs. pollock’s grand hotel”] in baltimore).

  61. In older Israeli slang, a derogatory term for Ashkenazi Jews was /vuzvuz/, imitating Yiddish speakers who didn’t know Hebrew. That’s where I got the idea that Yiddish, even the predominant eastern dialects, voiced the final s.

    This translation is apparently a continuation of a project started by Moses Mendelssohn, who published a German translation of the Pentateuch in Hebrew letters.

    This book was published in Karlsruhe. Mendelssohn was born in Dessau and lived in Berlin.

  62. David Marjanović says

    The German Wikipedia article is explicit:

    Seit 1771 von einer Krankheit geschwächt, versuchte Mendelssohn, sich bei der Übersetzung der biblischen Psalmen (erschienen 1783 und korrigiert 1788) zu erholen, und begann mit den Vorarbeiten zu seiner deutschen Übersetzung des Pentateuch. In hebräischen Buchstaben neben dem Urtext abgedruckt und ausführlich auf Hebräisch kommentiert, sollte sie Juden die Bibel und gleichzeitig die deutsche Sprache näher bringen; sie erschien von 1780 bis 1783.

    “Weakened by an illness since 1771, Mendelssohn tried to recover by translating the biblical psalms (published in 1783 and corrected in 1788), and began with preparatory work for his German translation of the Pentateuch. Printed in Hebrew letters next to the original text and commented in detail in Hebrew, it was meant to familiarize Jews with the Bible and at the same time with the [Standard] German language; it appeared from 1780 to 1783.”

    That sounds exactly like what we’re looking at. The English article commenting on the elegance of the German style also fits.

  63. My question is, who continued his work and translated the rest of the OT.

Speak Your Mind