The Saganaki of Madness.

Nick Nicholas has posted about Karamanlidika orthography, an absolutely fascinating account of ways in which Turkish words were written in Greek script with greater or lesser degrees of phonetic accuracy. He explains why that was difficult and gives striking examples of how hard the writing system of Greek can be anyway (Greek-speakers pronounce Δάντης ‘Dante’ as /ðandis/ or /ðadis/, and λούμπεν ‘Lumpenproletariat’ as /luben/; compare the uncertainty of Russians about when to pronounce е as ё /yo/). I might not have posted it here, because I just posted yesterday about the return of Ἡλληνιστεύκοντος, and I really shouldn’t turn LH into an affiliate of that fine site. But when I read this passage, about writing [ʃ] with a sigma surmounted by three dots in a triangular pattern (clearly modeled on the Arabic system), I couldn’t resist sharing it:

I’ve asked Peter [Mackridge], and he’s sent me a sample from the satirical comedy Το σαγανάκι της τρέλας. (Not “The frying pan of madness”, let alone “The saganaki of madness”: contemporary Greek σαγανάκι “small frying pan”, and any dish prepared in a small frying pan, like fried cheese, is a diminutive of σαγάνι < Turkish sahan “copper dish”. The Turkish word here is the unrelated sağanak: “The storm of madness”.) The comedy is attributed to Rigas Feraios, and was published in Lia Brad Chisacof. 2001. Ρήγας. Ανέκδοτα κείμενα, Athens. the text is published alongside the manuscript, and he has sent me two instances of the novel diacritic in question […]

Talk about your linguistic coincidences! As a lover of the cheesy wonder that is saganaki, I will never think of that play as anything but The Saganaki of Madness. And have I mentioned how happy I am that Ἡλληνιστεύκοντος is back? I look forward to many more tidbits about Greek linguistic history, and I urge you to add it to your RSS feed or bookmark it or whatever people do these days.


  1. Copying Arabic diacritics into another script also occurs in the Arte para ligeramente saber la lengua arauiga by Pedro de Alcalá (1505). He represents Arabic tha with Latin c surmounted by the three dots of the Arabic letter, dhal with d with a single dot above. (Kha is represented by h with the entire Arabic letter above).
    (page 39 on)

  2. I do enjoy seeing on Greek-language world maps such place names as Αντίς Αμπέμπα.

    There’s a restaurant named Kaspi (כַּסְפִּי) in Tel Aviv. The name on the awning is written in Hebrew, but also as ḴAṢ̣P̣I — adorning the Latin script with the Hebrew niqqud, for giggles.

    A Greek-styled restaurant near Jaffa is named Parakalo (i.e. “You’re welcome”). The name on the building appears, perversely, as PΛRΛΚΛLΘ.

  3. January First-of-May says

    adorning the Latin script with the Hebrew niqqud, for giggles

    A large list of such examples (with annotations in Hebrew) is linked in the comments of this Language Log thread.

  4. Thanks! The final link to these examples is here.

    It’s making my eyes bleed, and not in a pleasant way.

  5. Hebrew Comic Sans will do that.

  6. J.W. Brewer says

    “The Saganaki of Madness” is such a great title that it may need the characters and plotline it deserves, which the 18th century text may not quite provide. Perhaps it starts with a few Miskatonic University students deciding on a whim to go to Arkham’s one Greek restaurant for what is intended to be a light-hearted evening out, but then …

  7. January First-of-May says

    To add to J.W. Brewer’s idea, I’m completely unfamiliar with saganaki the dish (though it sounds delicious), but the word itself feels interestingly exotic – far too exotic to be simply Greek or even Turkish.

    I want to say that my first thought for what it could be is Sumerian (like Annunaki), but it’s probably actually Japanese (like Nagasaki).

  8. J.W. Brewer says

    Here’s an interesting bit of (purported) history being recreated for educational purposes decades after the (purported) fact, viz. the quest for the perfect linguistic utterance for the server of saganaki to exclaim while lighting it on fire, with numerous alternative candidates being tried and found wanting. By the time I lived in Chicago a bit over two decades after these purported events, a bunch of restaurants in the same neighborhood all offered the same saganaki ritual (with similar flames and exclamation), and I do not know if others dispute this particular establishment’s claim to be the fons et origo of the practice.

  9. @J.W. Brewer: When he was a kid, my cousin used to ask the waiters at the restaurants in Greek town where the “Opa!” came from. I don’t think any of them ever knew. He was local to Chicago, so he must have eaten there many times without ever learning the answer.

  10. > how hard the writing system of Greek can be anyway (Greek-speakers pronounce Δάντης ‘Dante’ as /ðandis/ or /ðadis/, and λούμπεν ‘Lumpenproletariat’ as /luben/; compare the uncertainty of Russians about when to pronounce е as ё /yo/

    Aren’t these different phenomena? Greek has /n/ (ν), /t/ (τ) and /d/ (ντ) and they’re one-to-one with the orthography, it’s just that /nt/ and /nd/ don’t exist phonotactically. This part of the orthography isn’t ambiguous like the Russian example. (I don’t know about word/morpheme boundaries, though.)

    Of course, ambiguity exists elsewhere in modern Greek orthography, like η, ι, υ, ει, οι, or υι for /i/.

  11. Well, sort of. You can treat the contrast as a separate phoneme /d/ or as a phoneme cluster /nt/, and it can be varyingly realized as [nt], [nd], [ⁿd], [n͡d], or [d]. The same is true of the labial and velar analogues.

  12. Do any speakers have a /nt/-/d/ distinction? I don’t know the details about the е/ё thing, but I assume they’re phonemically completely distinct, but there’s some confusion because people tend to leave the dots out.

    FWIW, GT pronounces even σύνταξη (syntax) with what sounds like something close to [d] to me (possibly prenasalized), although I’m not sure if that’s considered multi-morphemic, synchronously speaking.

  13. January First-of-May says

    Russian е/ё confusion is, IIRC, due to a sound change from around the 16th century that didn’t really affect the spelling, and that one 18th century lady proposing a letter that reflected the still-transparent relation (the relation is somewhat less transparent now, because, with one exception, ѣ never participated in it, which is of course obscured in post-1918 spelling, but it’s still pretty clear that there is a relation).

    It has entirely nothing to do with what you’re describing for Greek; if anything, I’d say that the closest Russian equivalent is probably the reduction of unstressed vowels (though of course it isn’t reflected in the orthography), or maybe the old-fashioned treatment of initial /h/ as /g/ in loanwords (which meant that the names Hilbert and Gilbert both ended up as Гильберт).

  14. January First-of-May says

    because, with one exception, ѣ never participated in it

    …which actually makes “does the root have ё in the relevant environment” a pretty good test for the yat’ – though of course there are exceptions in both directions, and loads of untestable, and/or very hard to test, cases (семь, as in “7”, is an especially hard one – there are forms with ё, but they’re extremely obscure).

  15. “Opa!” is the first thing the Russians would have reached for and maybe Polish too?

  16. because, with one exception, ѣ never participated in it

    Not sure what you mean by “never participated in it,” but there are certainly more than one word in which ѣ is pronounced /yo/: звѣзды and гнѣзда come immediately to mind.

  17. Aren’t these different phenomena?

    Yes, of course; I didn’t say they were the same, I just said they could be compared as instances where native speakers can’t properly read their own writing systems.

  18. In 2011 I took a trip to Greece with four old friends. Our excellent driver/guide pronounced ντ as [d] and μπ as [b]. I had to get used to Olibia, Adirio, Sidagma Square, etc., and to translate for other members of our group. The modern Greek pronunciation of Thucydides θu ki ‘ði ðis was particularly baffling though I suppose somewhat closer than modern English to ancient Attic.

    On Aegina, when we were on our own, someone recommended a restaurant called “Baba”. When we got there, I said “Here it is, but it’s closed.” My companions objected: “That doesn’t say ‘Baba'”. The sign read “ΜΠΑΜΠΑ”, and I again had to explain.

  19. January First-of-May says

    Not sure what you mean by “never participated in it,” but there are certainly more than one word in which ѣ is pronounced /yo/: звѣзды and гнѣзда come immediately to mind.

    It was the former that I meant by the “one exception”. I’m not sure where I would have heard that it was the only one, admittedly.

    In any case, it was sufficiently regular that it’s still fairly clear that, back when the relevant sound change happened (…not sure where I got the 16th century date from either; Wikipedia says it was probably earlier than that), the sound represented by ѣ was still pronounced differently enough from the one represented by е for the former to be unaffected by the change.

  20. > native speakers can’t properly read their own writing systems

    I don’t think this describes the Greeks. Greeks turn “lumpen” into “luben” because of their phonology, not because of their writing system. Their writing system, quite reasonably, just doesn’t make distinctions that their phonology doesn’t make (in this particular case).

  21. David Marjanović says

    “Opa!” is the first thing the Russians would have reached for and maybe Polish too?

    It’s in FYLOSC too.

    Slap a [h] on it, and suddenly it looks far less exotic.

  22. it’s probably actually Japanese (like Nagasaki).

    Actually, there is such a word in older Japanese:

    ‹さがなき(性なき)›子供たちは、・・[性=悪い性根] (modern) やんちゃな
    saganaki (modern “yanchana”) kodomotachi = mischievous/naughty children

  23. Dainichi asked, a month ago:

    > Do any speakers have a /nt/-/d/ distinction?

    No, and that’s the point of what I was writing, I guess: speakers older than 50 do [nd], speakers younger than 50 do [d] (there is also a dialectal isogloss in play), and noone ever says [nt]…

    … except that is, speakers highly fluent in Italian or Turkish in the past, who did differentiate /nt/ and /d/ and /nd/ in their Greek (or at least in their Italian and Turkish loanwords). The Turkish speakers used diacritics to make those distinctions. The previous Italian speakers used Italian orthography. But Modern Greek speakers simply don’t.

    … I said noone ever says [nt]: there is an exception, and it’s not a surprising one: the Greek spoken in Southern Italy.

  24. January First-of-May says

    and that one 18th century lady proposing a letter that reflected the still-transparent relation

    Apparently later research suggests that the lady in question (Dashkova) probably didn’t actually propose the letter attributed to her, and the actual inventor might well be Karamzin, as had been suspected for a long time before someone turned up the Dashkova anecdote and then (later) someone else turned up some (just barely) pre-Karamzin-quote instances that apparently eventually turned out to be misdated by a few years.

    Over most of the 18th century, the actual letter used for that sound (in the rare instances that a distinct letter was used at all) was a completely different one (technically a digraph, namely їô – exact spellings varied but this was a common version, and in particular the one used by Karamzin to explain his new letter), conceptualizing it as /jo/ rather than as an allophone of /e/.
    Maybe we’d have been better off in terms of consistency if їô (or some simpler variant) persisted as the letter of choice. No way to confuse that with anything else, especially as ô wasn’t used anywhere in Russian otherwise.

    (Apparently the aforementioned Dashkova made a big proposal for it in 1783, which failed because the guy in charge of approving it was against any new letters that aren’t in the Bible. But she also proposed the letter h for an allophone of /g/ that had basically merged back (and/or maybe never split in the first place) in the dialect that became the OTL modern standard, and I’m not sure I want to imagine where we’d be if that caught on…
    The proposal had at some point been reinterpreted as suggesting Karamzin’s letter, but apparently it in fact did no such thing.)


  1. […] Hat looks at how some Turkish-speaking Christians transcribed the Turkish language in the Greek […]

Speak Your Mind