Sfyria.

We’ve done whistled languages before here — Yupik in 2005, La Gomera most recently in 2011 — but it’s been a while, so herewith please find sfyria, according to the breathless BBC Travel report by Eliot Stein “one of the rarest and most endangered languages in the world” (what would reporters do without superlatives?):

Hidden deep in the south-east corner of the Greek island of Evia, above a twisting maze of ravines that tumbles toward the Aegean Sea, the tiny village of Antia clings to the slopes of Mount Ochi. There are no hotels or restaurants within 40km, and the hamlet is so remote that it doesn’t exist on Google Maps.

But as you travel here along a dizzying road from Karystos, through a mythical landscape of megalithic ‘dragon house’ stone tombs and giant Cyclopic boulders, you’ll hear an ancient siren song reverberating against the mountain walls. That’s because for thousands of years, the inhabitants of Antia have used a remarkable whistled language that resembles the sounds of birds to communicate across the distant valleys.

If you’re wondering where they got the “thousands of years,” voilà:

No-one can recall exactly how or when the villagers here began using sfyria – which comes from the Greek word sfyrizo, meaning ‘whistle’ – to communicate. Some residents speculate that it came from Persian soldiers who sought refuge in the mountains some 2,500 years ago. Others claim the language developed during Byzantine times as a secret way to warn against danger from rival villages and invading pirates. There’s even a belief that in ancient Athens, they’d post whistlers from Antia on the mountaintops as sentries so they could signal an imminent attack on the empire.

Remarkably, sfyria was only discovered by the outside world in 1969, when an aeroplane crashed in the mountains behind Antia. As the search crew went out to look for the missing pilot, they heard shepherds volleying a series of trilled scales back and forth across the canyons and became enchanted by their cryptic code.

The BBC would presumably not take “some residents speculate” as a reliable source for anything important, but hey, it’s just language. Carping aside, this sort of thing is always fun, and I thank Trevor for passing it along!

Comments

  1. From the Beeb article: “sfyria is effectively a whistled version of spoken Greek, in which letters and syllables correspond to distinct tones and frequencies.”

    From a comment in your previous post about La Gomera, quoting the UNESCO Intangible Cultural Heritage lists (2009): “The whistled language replaces each vowel or consonant with a whistling sound: two distinct whistles replace the five Spanish vowels, and there are four whistles for consonants. The whistles can be distinguished according to pitch and whether they are interrupted or continuous. With practice, whistlers can convey any message.”

    In other words, both of these languages are just substitution codes. They are not distinct languages but just ways of transmitting the underlying languages, Greek in the case of sfyria, Spanish in the case of La Gomera. They only differ from Braille, shorthand, or Morse code in that they are whistled rather than punched, written or telegraphed.

  2. dainichi says:
  3. Athel Cornish-Bowden says:

    By chance I saw your post a day after watching an interview with a young girl in La Gomera who had just had a whistled conversation with her friend several hundred meters away. I found it hard to believe that the very limited range of sounds distinguishable to my ears could express Spanish, but then I have similar problems with sign languages, so I can believe that it really does. Apparently whistling is now a required subject in La Gomera schools, and we saw a class in which it was being taught. You can see La Gomera from Tenerife, and I’ve long thought I’d like to go there, but although I’ve been to Tenerife several times (in the line of duty) I’ve not managed it. It seems to be a smaller version of what Tenerife would be like without the overdevelopment. Canary Spanish is much closer to Latin American — in pronunciation, vocabulary and grammar (no “vosotros”) — than anything you’d hear in Madrid, and I find it much easier to understand.

  4. Athel Cornish-Bowden says:

    I don’t know how the pronunciation of Greek in Evvia differs from the standard language, but maybe the BBC reporter heard it said as “Andias”, searched Google maps for “Andias, Greece” and got “We could not find Andias, Greece”.

    I’m also surprised that the final ς is dropped in the report. In all the Greek I’ve ever heard ς at the end of a word is always pronounced very strongly. I have a colleague in Chile whose name begins with Mp and ends with s: people there treat the Mp as P and ignore the s, but that’s Chile.

  5. Whistled version of English.

    (sorry for the lowbrow intervention — the story reminded me of this old childhood favorite)

  6. Really? So what am I looking at here?

    Hah! Oh, BBC, have you no shame?

  7. Athel Cornish-Bowden says:

    In the page at http://languagehat.com/silbo-gomero/ you said You can now (Aug. 2007) see a video describing the island of Gomera entirely in Silbo (!) at this page. Thanks, Jeremy!

    Unfortunately the link has gone the way of all flesh, and now (Aug. 2017) goes to a page offering to sell the domain.

  8. I think that’s the way of the new flesh, actually.

  9. IIRC, Aristophanes in the Lysistrata mentions Karystos as a place where the people are as rough and up-country as you can get.

  10. Web Archive (WayBack Machine) link to replace that dead Aug. 2007 link mentioned above:
    https://web.archive.org/web/20080509100845/http://www.kirchersociety.org/blog/2007/08/07/the-whistling-language-of-gomero-island/

  11. Thanks, Raven!

  12. dainichi says:

    > I’m also surprised that the final ς is dropped in the report

    I am in absolutely no way an expert on Greek, but if Αντιάς is the nominative of a masculine singular noun declined as described here

    https://en.wikipedia.org/wiki/Modern_Greek_grammar#Masculine_nouns

    , all other cases have Αντιά, so could that be the reason? The same article says that prepositions usually take the accusative, and names of towns are often used with prepositions, “in Antias”, “to Antias” etc. The reporter might just have heard Αντιά more than Αντιάς and assumed the former was the name. Not defending him in any way, just speculating.

  13. Yes, I think that’s probably the explanation. I’m not so much bothered by the reporter’s failure to realize that as by the fact that nobody involved with the publication of the piece knew it. I don’t expect reporters to know things, just to report what people say.

  14. It’s impossible to report what people say if you don’t know what they are talking about. Even trained stenographers (which reporters are not) make all kinds of mistakes with technical matters, and the names of small foreign villages definitely count as technical. I learned this when I gave what I thought was a fairly lucid answer to a question asked by the reporter for an in-house rag, and asked to see the copy: it was complete word salad. I rewrote it into what I had said, and all was well. But most interviewees don’t have that sort of good fortune.

  15. Trond Engen says:

    One thing is not being familiar enough with the language to spot a case form, another not being familiar enough with the geography to find the place on the map without the search function. But note that it’s BBC Travel, not e.g. BBC Culture or BBC Earth. Put together these suggest to me that the piece is not first hand reporting but rather a fictional travel based on some other travel guide or marketing material. Or maybe on participation in a commercial guided tour.

  16. Ooh, you’re even more cynical than I am — good call!

  17. if Αντιάς is the nominative of a masculine singular noun

    It’s feminine:
    nom., acc. Αντιά
    gen. Αντιάς

    Το ΒΒC έκανε αφιέρωμα στη «σφυριά», τη σφυριχτή γλώσσα της Αντιάς, που πλέον αργοσβήνει.
    […]
    Σύμφωνα με την Δήμητρα Χένγκεν, γλωσσολόγο που ταξίδεψε με τον ρεπόρτερ του BBC στην Αντιά,

    http://www.iefimerida.gr/news/354071/sfyria-bbc-ekane-afieroma-sti-sfyrihti-glossa-tis-antias-stin-eyvoia-poy-argosvinei

    And if σφυριά is transliterated as sfyria, why is βυσσανάδα vissanada rather than vyssanada?

  18. Athel Cornish-Bowden says:

    As υ is just one many ways of representing the same vowel sound, I would expect sfiria and vissanada (or maybe vissanadha). sfyria probably derives from classical notions of how Greek should be transcribed (but of course υ and ι are believed to have had different sounds in classical Greek)

  19. Trond Engen says:

    Ooh, you’re even more cynical than I am

    Probably too cynical, so I should apologize. Reading the piece again, it seems clear that he went there and spoke to people, and also that his photos are his own. What I do think is that he’s more dependent on having it served to him than his written account conveys. But that may well be a choice made for the sake of brevity. It’s not himself he’s portraiting.

  20. Trond Engen says:

    juha: It’s feminine:
    nom., acc. Αντιά
    gen. Αντιάς

    Why is it written Αντιάς on the map?

  21. Trond Engen says:

    It just struck me that the southern tip of Evboia is marked consistently as Arvanite on the linguistic maps I’ve seen. And it can’t be a scattering of Arvanite islands in an ocean of Greek. According to Encyclopedia of the World’s Minorities, there were 87 Arvanite villages in the Evboia (presumably as of 2005 or so). Whether or not Antia(s) is one of them, this could be important for the understanding of Sfyria.

  22. I think it’s more likely that those 85 villages are the historical Arvanite-speaking ones in the region (out of about 500 in Greece as a whole, says WP). Arvanitika is going under fast, and probably no one has surveyed them to see if anyone still speaks it in those villages.

  23. Trond Engen says:

    Says the Encyclopedia:

    The Arvanites are currently located in more than 500 villages in various provinces of Greece. There are about 87 Arvanites villages in the departments of Euboea, 84 in Attica, 70 in Corinth, 60 in Boeotia, 35 in Argolis, 20 in Messenia, 20 in Achaea, and seven in Phtiotis. Other settlements exist in the islands of Andros, Hydra, Poros, Spetsai, Ankistri, and Salamis, with a few more in other regions of Greece.

    Adding the numbers don’t get you above 400 across Greece, but 87 in Evboia still stands. The actual situation in those villages is obviously quite another matter. But I didn’t mean to suggest anything about the current arvanicity of Antia or any other village, just that the Arvanite settlement in southern Evboia could be relevant for the development of Sfyria. The fact that it’s a whistled form of spoken Greek suggests that Antia is traditionally Greek-speaking. Or if not, why Greek in the whistled language? Or is the Sfyria system language-independent and can be used to code vocal phonemes into whistling in any language?

  24. @Athel: By chance I saw your post a day after watching an interview with a young girl in La Gomera who had just had a whistled conversation with her friend several hundred meters away. I found it hard to believe that the very limited range of sounds distinguishable to my ears could express Spanish, but then I have similar problems with sign languages, so I can believe that it really does.

    50 years ago in Austin I was fairly competent in everyday ASL (American sign language). At one point I counted 300 signs that I knew. I don’t know what exactly you mean by “express” Spanish, but I know for a fact that ASL does not “express” English, nor is it a “substitution code … [sfyria and La Gomera] only differ from Braille, shorthand, or Morse code in that they are whistled rather than punched, written or telegraphed”, as Martin put it above.

    A sentence in ASL (as well as in the signing used in Germany) is a kind of suggestive skeleton that needs to be fleshed out in order to become spoken English. For example there are no “definite articles”in practical use, the meaning of these being expressed in different ways according to context (provided the meaning depends on definite articles at all, which it usually doesn’t once you get used to not having them, cf. Russian). There is little explicit subordination, ASL is even more paratactic than English.

    Possibly this is part of the “problems with sign languages” that you say you have ?

    I would expect sfyria and Gomera to be similar to ASL in these respects. But I know nothing about them.

  25. No, the whistles really are codes for spoken languages, with some redundancy of wording added in order to compensate for lost phonological distinctions. There are systems that use signs similarly, like Signed Exact English, which maps every English word onto a sign (an ASL sign where available, an otherwise unknown sign otherwise). When you see someone speaking English and signing at the same time, it’s usually SEE or something close to it rather than ASL.

    The drum talk of West Africa is an even more radical deformation of the corresponding spoken languages: it throws away everything but lexical tone and prosody. As a result, ordinary words have to be replaced by conventional phrases, and people who speak the underlying language need training to understand drum talk.

  26. Signed Exact English, which maps every English word onto a sign

    John, is it really the case that SEE maps each English word to one sign ? So that there would be one SEE sign to which “go” maps, and a different sign to which “went” maps ? I would have expected that “went” maps to a pair of SEE signs, just as the ASL for “went” is “yesterday”+”go”.

  27. Athel Cornish-Bowden says:

    I don’t know what exactly you mean by “express” Spanish, but I know for a fact that ASL does not “express” English, nor is it a “substitution code …

    I meant what the first reply above said:

    “The whistled language replaces each vowel or consonant with a whistling sound: two distinct whistles replace the five Spanish vowels, and there are four whistles for consonants. The whistles can be distinguished according to pitch and whether they are interrupted or continuous. With practice, whistlers can convey any message.”

    i.e. that Silbo is an encoding, not a language.

    On the other hand I’m well aware that sign languages are languages (and that, for example, American sign language is quite different from British sign language but similar to French sign language), not encodings, and if I suggested otherwise was unintentional. My difficulty with them is that they seem to represent a great deal of meaning with what seems to a non-signer to be a very limited range of gestures.

  28. Having first learned a language, English, in which case only barely still exists, I have often wondered about what it means to talk about the name of a person or place in a language in which proper nouns are declined. Does it depend on the language? Is the nominative automatically the name? What about languages (like French), where the “default” case is accusative? (I know the answer for French, but what about in general?)

  29. they seem to represent a great deal of meaning with what seems to a non-signer to be a very limited range of gestures.

    The signs “themselves” – which I take it you mean by “gestures” – are only part of the communication. The accompanying grimaces, emphatic body movements, slightly varied repetition, dramatic pauses etc help to “represent a great deal of meaning”. Just as they do in spoken English.

  30. What about languages (like French), where the “default” case is accusative?

    I’m not sure what you mean by that. Nouns of course have no cases at all in modern French. Are you referring to the fact that the modern French form is usually, not always, derived from the Old French oblique case? (The first exceptions that come to mind are Georges and Jacques.)

  31. marie-lucie says:

    RC, do you mean something like English Cicero vs French Cicéron? The French version of the names are based on what happened with other words of the same structure, where the Latin case endings (not just the accusative) were lost but the stem (occurring in most of the word forms) was preserved, as in Ciceronis, Ciceronem, Cicerone, preserving the stem Ciceron- which remained as the basic French form. This is not just for names but for ordinary words, as in mansio ‘estate’, mansionis, mansionem etc, Fr maison ‘house’, and with other stem-final consonants, e.g. flos ‘flower’ but floris, florem etc, French fleur.

    English imported lots of Frenchified Latin words, like education, but for some reason more recent borrowings go back to the Latin nominative, as in ratio and a few others.

    As for Georges and Jacques vs. Cicéron, they are common first names, as well as known from religious history and legend, while far fewer people were familiar with Latin political celebrities. The final s is supposed to be from the Latin masculine singular nominative, which may have been used as vocative also as the differences between old cases faded away. (Sorry, I did not check in a proper work on the history of French).

  32. David Marjanović says:

    Is the nominative automatically the name?

    Yes (in nominative–accusative languages), because that’s the form used in the simplest sentences like “X is Y”. (In ergative–absolutive languages, I’m sure the absolutive is automatically the name; I don’t see a lot of Basque placenames ending in -k.)

    What about languages (like French), where the “default” case is accusative?

    I’m not sure what you mean. Are you perhaps referring to the fact that moi, toi etc. are descended from oblique forms (in the way m-l explained)?

  33. Trond Engen says:

    Yes (in nominative–accusative languages), because that’s the form used in the simplest sentences like “X is Y”

    Is that the case (sorry) in languages with a vocative? Or at least an unmarked vocative?

  34. SEE maps each English word to one sign

    I oversimplified slightly: SEE actually maps English morphemes to signs (thus ‘boys’ is signed BOY + -S and ‘examination’ is signed EXAM + INE + TION) and it is possible to drop medial morphemes if no additional ambiguity is created (thus EXAM + TION is sufficient). I don’t know if -S plural and -S 3rd person singular are signed the same; I do know there are separate signs for AM, IS, ARE, WAS, WERE, which are all zero in ASL.

  35. In the languages with a vocative that I know, the nominative is still used in sentences of the type “x is y”. The vocative is used to address people, as in “Bill, be quiet”. As the nominative is used to quote names, one can say that the nominative form is the name.
    @m-l: a further example where the current French form if the noun goes back to the nominative is fils ; here the nominative “s” has survived even in pronunciation.

  36. Trond Engen says:

    It might be a hint that the nominative is defined as “name form”… But I won’t let it go quite yet, so I’ll reformultate the question. How would people say “My name is …” or “Call me …”?

    The reason I think of it is that Cenral Scandinavian developed a rudimentary case system for personal names in the aftermath of losing the original one. It goes something like this:

    Nom. han Ola / ho Kari
    Acc./dat. hono Ola / hinna Kari
    Gen. hans Ola / hinnar Kari
    Voc. Ola! / Kari!

    (Or maybe nom./acc. vs. dat. in some dialects? Don’t remember.)

    In this paradigm the case marker is a separate word (or clitic), and the bare stems Ola and Kari are clearly felt as the names. This is a shallow and transparent construction and may well be a special case, but since most case markers start out as clitics, I imagine from similar pronominal constructions, it could also be common.

  37. David Marjanović says:

    The reason I think of it is that Cenral Scandinavian developed a rudimentary case system for personal names in the aftermath of losing the original one. It goes something like this:

    Very similar to all the southern and central German dialects (like mine) where articles are obligatory with personal names except in direct address. If you ask little children for their names, the answer will often begin with “I am the”, and this is also often used by people more generally to introduce themselves. Yet, the article is not used with the special verb heißen “to be named”, and placenames practically never take articles.

    Some have wondered if the PIE animate nom. sg. ending *-s is derived from the nom. sg. animate demonstrative pronoun *so via an unattested function of the latter as a definite article. If so, that would neatly explain why the nominative was marked and the vocative more or less wasn’t.

  38. Very interesting!

  39. January First-of-May says:

    How would people say “My name is …” or “Call me …”?

    In literary Russian, in fact, Меня зовут… would take the instrumentative case (as in Катей) – perhaps by extension from something like Зовите меня… (which would probably be in the instrumentative even in regular Russian).

  40. Trond Engen says:

    David M.: Very similar to all the southern and central German dialects (like mine) where articles are obligatory with personal names except in direct address.

    (I should have said Northern rather than Central Scandinavian.)

    It’s an oddly shaped area for a shared innovation. Didn’t we know better we might have concluded with a shared retainment. And maybe that’s not completely wrong. My theory of the name article is that it developed in conservative dialects as a way to convey meaningful case marking in a large number of newly imported indeclineable names.

    Some have wondered if the PIE animate nom. sg. ending *-s is derived from the nom. sg. animate demonstrative pronoun *so via an unattested function of the latter as a definite article. If so, that would neatly explain why the nominative was marked and the vocative more or less wasn’t.

    That’s how I think about the unmarked vocative. A good theory should explain the whole paradigm, though.

  41. David Marjanović says:

    My theory of the name article is that it developed in conservative dialects as a way to convey meaningful case marking in a large number of newly imported indeclineable names.

    That makes sense.

    A good theory should explain the whole paradigm, though.

    Oh no. Case systems don’t form all at once, and PIE is a good example of a case system that was growing at the edges, with various case endings and confusingly case-like endings also occurring as free-standing adpositions. On the one hand, the accusative *-m is shared with Uralic; on the other, whether the Old Hittite allative and its ending is retained from PIE or some kind of strange innovation seems to be pretty much guesswork.

  42. David Marjanović says:

    Latin mihi nomen est came with the nominative or, in analogy to mihi, with the dative.

  43. January First-of-May says:

    Modern Russian actually has a (colloquial) vocative form for names that basically involves dropping the nominative ending, and effectively is the bare root (for 1st declension, i.e. names ending with -а, -я, which includes many common names and most nicknames; not sure what it comes out to in other declensions, if anything).
    [Which is to say, the colloquial vocative of Петя / Маша / Лена / Гоша is Петь! / Маш! / Лен! / Гош!]

    I wonder if the “unmarked vocative” was something similar.

  44. But it is only for familiar names, which traditionally always have -а, -я endings in Russian. For example, I had known a guy whose formal name was Pavel, but who was called by everyone Paul (pronounced like a German name). Obviously, nominative and vocative coincided in this case. A number of Ivans are called Vano (yes, Russian Ivans, not Georgian) or Alexanders go by Sandro. No one would ever think about dropping endings in these cases.

  45. marie-lucie says:

    Hans: French fils ; here the nominative “s” has survived even in pronunciation

    Yes, this is interesting! I wonder if that was in order to avoid homophony with fil ‘thread’, from Latin filum.

    More generally I wonder if those cases of s preservation (Charles, etc) were influenced by their Latin equivalents, which would have been heard spoken by clergy: names at baptisms, weddings and funerals (Carolus, Iacobus, etc), and filius ‘son’ in various prayers recited in Latin, especially in the common phrase Pater et Filius. But perhaps there is another, commonly accepted explanation.

  46. Iberian dios (etc.) is another prominent nominative survival.

  47. a further example where the current French form if the noun goes back to the nominative is fils ; here the nominative “s” has survived even in pronunciation.

    In fact (per Rickard, A History of the French Language), it was lost at the end of the Old French period with all other final /-s/, but restored from liaison and the written form in the 18C, along with the final consonants of net, sens, and many other monosyllables.

  48. I’ll be damned. I love that sort of thing!

  49. marie-lucie says:

    JC: it was lost at the end of the Old French period with all other final /-s/, but restored from liaison and the written form

    In that case, it was never completely lost! if for example it was always sounded in fils unique ‘only son’. Here, unlike in other cases of liaison, the final sound is [s] not [z] as in les enfants. Another similar case is sens as in sens unique ‘one way’ or sens interdit ‘no entry’ (as in a one-way street).

  50. Rickard doesn’t say which words had liaison and which did not. In all probability it was situational back then, as indeed it is today.

  51. marie-lucie says:

    JC, it is not always situational. You cannot NOT do it with les enfants, for instance. But it is very hard to explain, as there are often many factors involved.

  52. Rodger C says:

    Iberian dios (etc.) is another prominent nominative survival.

    Certainly from the influence of the Mass; the Ladino is El Dío.

  53. List of French doublets derived from the nominative and oblique cases of Old French nouns.

  54. marie-lucie says:

    JC: which words had liaison and which did not

    It is not a matter of listing words, which are vocabulary items, while liaison (‘linking”) links together a final consonant (usually a grammatical affix) which is not pronounced final position, with a following word beginning with a vowel. So for instance un enfant, un petit enfant, les enfants, les petits enfants.

    me: about the final [s] in fils [fis] ‘son’: I wonder if that was in order to avoid homophony with fil ‘thread’, from Latin filum.

    Since I wrote this I researched a paper about the pronunciation of final il as just i in a smallish number of words, such as persil ‘parsley’ traditionally pronounced [persi] but which more and more people pronounce [persil], while fil ‘thread’ has apparently always preserved the final l.

    In the vast majority of cases like persil, the Latin ancestor had stem-final ili, as in petrosilium. For words which have both masculine and feminine forms, as in Latin filia ‘daughter’, the French equivalent has ill, pronounced [ij], thus fille ‘daughter, girl’. Final il pronounced [j] also occurs in a number of words where il is preceded by another vowel, as in travail ‘work’, réveil ‘awakening’ and fenouil ‘fennel’. Other forms of the same words which add vowel-initial affixes also have [j|, as in the verbs travailler and réveiller. Similarly persil has a derivative persiller ‘to add (to a food) tiny pieces of a condiment, such as cut up parley leaves, which will remain visible in the completed dish’, fusil ‘long gun’ the derivative fusiller ‘to execute (someone) by firing squad’, among others.

    Back to fils ‘son’ from Latin filius, it must have been pronounced [fijs], corresponding to the feminine fille from Latin fille. The Latin stem fili- is also preserved in filleul(e) ‘godson/daughter’, where the ill is [ij].

    My conclusion is that words like persil have not lost the sound [l] which is now being restored from the written form, but the sound [j] (something also attested from occasional instances), which must have blended with the preceding vowel [i].

  55. According to Mildred Pope, the [s] in fils (OFr. filz [fiʎts] > [fits] > [fis] > [fi]) was added back in 18th-century Parisian French. A century earlier, a final s was mute especially after [i]. In particular, as 17th-century sources tell us explicitly, fils was pronounced [fi] even before a vowel.

    Old French z [ts] was the regular development of s after a palatal lateral or nasal. It merged with [s] in the 13th century.

  56. I think the list of doublets leaves out empereur / Lemprière.

  57. Thanks, yes. It’s definitely incomplete as to names, which often preserve the nominative.

  58. marie-lucie says:

    Piotr: According to Mildred Pope, the [s] in fils (OFr. filz [fiʎts] > [fits] > [fis] > [fi]) was added back in 18th-century Parisian French. A century earlier, a final s was mute especially after [i]. In particular, as 17th-century sources tell us explicitly, fils was pronounced [fi] even before a vowel.

    Thank you for the reference. I think that in the sequence OFr. filz [fiʎts] > [fits] > [fis] > [fi] there may have been an intermediate step, so [fiʎts] > [fijts] > [fits]. Similarly a step [fiʎ∂] between filia and fille [fij]. (cf OProv filha with lh for [ʎ]).

    Old French z [ts] was the regular development of s after a palatal lateral or nasal. It merged with [s] in the 13th century.

    That must be the reason for Fitz in Fitzgerald, Fitzhugh, Fitzwilliam and similar hybrid names. I knew that it came from French ‘son’ but wondered about the tz.

    Roger: I think the list of doublets leaves out empereur / Lemprière.

    I know Lempérière, which is on the map of Paris (a street or plaza). Same word, or doublets? In any case a feminine word. Too bad it was not kept, instead of impératrice.

  59. I know the name from Lemprière’s Classical Dictionary.

  60. marie-lucie says:

    I did not know him or his dictionary, but there are many people called Lempérière too.

  61. I know Lempérière, which is on the map of Paris (a street or plaza). Same word, or doublets? In any case a feminine word. Too bad it was not kept, instead of impératrice.

    It is a case doublet like the others I linked to. The Old French word, which was masculine, was nom. emperesdre, obl. empereor. Both are pronounced as if Spanish or Italian, with d representing /ð/. The latter gaves us Modern French empereur, the former was lost except in this name (L)empérière.

    The first couplet of the Chanson de Roland shows us the nominative in use: “Charles li reis, nostre emperesdre magnes / Set anz totz pleinz ad ested in Espagnes.” As I said at the other post, a knowledge of Modern French is often a positive handicap in learning Old French: too many of one’s intuitions are wrong.

    (The couplet was the first email signature I ever used, and still pops up occasionally in rotation. I think I first chose it in imitation of “Ki semenat ispinaza, non andet iskultsu!”, the signature of J.A. Rea at the University of Kentucky. This was a twofold challenge: what does it mean, and what language is it? —Etienne, you’re out of this game.)

  62. Who sows spinach won’t reap somethingorother.

  63. marie-lucie: I think that in the sequence OFr. filz [fiʎts] > [fits] > [fis] > [fi] there may have been an intermediate step, so [fiʎts] > [fijts] > [fits]. Similarly a step [fiʎ∂] between filia and fille [fij]. (cf OProv filha with lh for [ʎ])

    These developments are not fully parallel, however. Preconsonantally, as in filz, [ʎ] (or whatever came from it) dropped out already by the mid-twelfth century (hence Fitz in Norman names). Intervocalically, in words like [fiʎə], the l mouillé survived as [ʎ] till the 18-th century in mainstream French, and this pronunciation can still be found in some regional accents. Some orthoepists insisted (in vain) on its use in Standard French even as late as 1870 or so.

  64. “Ki semenat ispinaza, non andet iskultsu!”

    Sardinian, I suppose, but give me a little time to decipher it.

  65. OK, I’ll chance it. ‘Who sows thorns does not walk barefoot.’

  66. What mood is andet?

  67. Trond Engen says:

    Sardinian, I suppose

    Of course. I was trying to remember which non-creole Romance language is written with k- for qu-.

    If ispinaza means what I think it means, I can’t make any sense of iskultsu. It looks like a Germanic loan, too, but that doesn’t make any more sense.

  68. Trond Engen says:

    Oh, right.

  69. Aha. Subjunctive or imperative of andare. The indicative would be andat, right? So ‘… let him not walk barefoot’.

  70. iskultsu — cf. Lat. ex-calceō ‘take off the shoes’.

  71. The thorns on Mediterranean island bush varieties can be enormous.

  72. Trond Engen says:

    Well, at least I was right about ispinaza. “Shoeless” makes sense, but I had no idea how to parse iskultsu, The best I had come up with on my own was “Who sows thorns needs no shield”.

  73. As for the form of iscultsu, cf. Romanian desculț ‘barefoot’, Venetian descolz from VLat. *disculceus, var. or discalceātus (the latter gave French déchaussé). There was also another variant, *discalceus, as in French déchaux ‘discalced’ (like a Carmelite monk). Also Italian scalzo < *excalceus.

  74. Benjamin Franklin alias “Richard Saunders” wrote in Poor Richard’s Almanack, 1743: “He that sows Thorns should never go barefoot.” It’s one of those wingèd proverbs: there are at least French, Spanish, and Italian versions, though I don’t know them.

    Someone (not me) provided essentially the same translation as Piotr, and Rea answered with the same rhetorical question that I asked.

  75. David Marjanović says:

    So, is that a continental loanword in Sardinian (-ce- > -ts-)?

  76. Whoever heard of sowing thorns?

  77. David, I don’t think so. Rather than that, it’s /ke.u/ > /kju/ > /tsu/. The suffixes -eus, -ea and -ius, -ia were confused in Vulgar Latin.

  78. Y: it’s a metaphor; cf. Hosea 8:7.

  79. marie-lucie says:

    JC: thanks for the quotation from la Chanson de Roland. I had no idea of these forms!

    Piotr: discalceātus (the latter gave French déchaussé). There was also another variant, *discalceus, as in French déchaux ‘discalced’ (like a Carmelite monk

    I had to go to the TLFI to verify the existence of French déchaux. It is listed, but marked as old. The barefoot monks I know as Les carmes déchaussés not déchaux. (Actually not quite barefoot, but wearing sandals, without socks).

  80. This form is dated, no doubt about it; nevertheless, L’Ordre des Carmes Déchaux is what the “Discalced Carmelites” officially call themselves in French.

    http://www.carmes-paris.org/

  81. marie-lucie says:

    Thank you for finding this, Piotr.

    It looks like both les Carmes Déchaux and les Carmes Déchaussés exist, and it is difficult for the uninitiated to keep them apart, or to find out which name came first in what started as a single order.

    The Carmes Déchaux (a tiny order, even if it is international, with one convent as far as Baghdad) aim to go back to the teachings of their founders, including the archaic name, but the link with the word déchaussé is not obvious, unless you know that the female branch of the order is les Moniales Déchaussées. Déchaux itself, a masculine plural form, is treated as invariable: the text refers to a single monk as a Carme Déchaux, where one would expect déchal which would be etymological (as a shortened form of the Latin original, rather than a derived one) but even less obvious to relate to déchaussé.

  82. David Marjanović says:

    Ah, so Sardinian did palatalize [kj]… that had escaped me.

  83. The development of Latin [kj] and [tj] in Sardinian was quite complicated. They were palatalised everywhere (though not at the same time) in Romance and usually merged both with each other (thus in Sardinian) and with the outcome of Lat. [k] before [ɛ, e, i] (exceptions include some varieties of Sardinian, which keep a velar if not followed by yod, and Romanian, where the affricate reflexes are different). It also seems that the early Romance outcome of [kj] (and less regularly that of [tj]) was a lengthened affricate like [tːs], hence its resistance to intervocalic lenition.

    In mediaeval Sardinian this affricate may have developed into [tθ], if the spelling th is anything to go by. The modern reflexes vary from dialect to dialect: Nuorese [θ], Logudorese [t], Campidanese [ts] (they may all be phonetically long, but this length is non-distinctive). The preferred conventional spelling today is tz. The prothetic vowel in iskultsu (= iscultzu) betrays a Logudorese/Nuorese word.

  84. Reducing language to rhythm: Amazonian Bora drummed language exploits speech rhythm for long-distance communication

    http://rsos.royalsocietypublishing.org/content/5/4/170354

Speak Your Mind

*