Friday, 17 August 2012

an Arabic mystery

Most of us know, I think, that Standard Arabic has no p in its consonant system; nor, as far as I know, do any of the mostly widely spoken versions of colloquial Arabic (though apparently Iraqi Arabic is an exception).

So how come, you may ask, that the largest city in (Arabic-speaking) Syria is called Aleppo, while both in Syria and in Libya there are cities called Tripoli? How can the locals pronounce these names?

The answer, of course, is that these are not the Arabic names of the places in question.

The Arabic name of Aleppo is حلب Ḥalab ˈħalab, with the expected b rather than p. Aleppo is the Italian name.

Tripoli, likewise, is Italian, namely the Italian version of the Greek Τρίπολις Trípolis ‘three cities’. The Arabic name is طرابلس Ṭarābulus, which in Libyan Arabic gets reduced to tˁraːbləs. Again, no p. Simples!

55 comments:

  1. Yes. On the other hand:

    Protosemitic /p/ appears to have got /f/ in Arabic,they say. And also, though there might be no connection, in th'olden times they would make an /f/ out of the Greek /p/, for instance:

    أفلاطون, 'aflat.u-n (emphatic /t/, long /u/), Plato, probably the latter as a reflection of the omega in the Greek name.

    أفلوطين, 'aflu-t.i-n, Plotinus, long /i/, long /u/

    strangely, they made an emphatic /t/ out of the Greek tau, anyone know an explanation?

    أرسطو, Arist.u-, Aristotle.

    Why did they make an /f/ out of /p/? Aramaic, Persian perhaps, the intermediary of? And why did they substitute emphatic /t/s for the Greek /t/s? Another Arabic mystery.

    ReplyDelete
    Replies
    1. Wojciech

      Aramaic, Persian perhaps, the intermediary of?

      I believe the intermediary was Syriac. Quite how different this was from Aramaic I have no idea.

      Delete
    2. If Wikipedia is to be believed,
      • Syriac was eastern Aramaic
      • [p] and [f] was one of six pairs of consonants which could vary for the same morpheme in different environments
      • Syriac letter pē ( ܦ ) represents either value
      • Arabic letter feh ( ف ) represents a fricative when consonantal
      • pe and feh occur at the same place in their respective alphabets

      Delete
    3. yes, thank you, I have known that. The situation in Aramaic (called 'Syriac' in some later versions) is similar to that in Hebrew, if this be of any help to you. May be an explanation of the vicissitudes of the Greek pi in Arabic. Re Arabic, I seem to know that ف always represents a fricative in't, as it always is consonantal and there is no /p/ in that language...

      Delete
    4. Wojciech

      I have known that

      I suspect you mean I knew that already. If so, I feel cheated that I could have got it from a reliable source like you and not from the usually-but-not-invariably reliable source that is Wikipedia.

      I'm sure it was from a reliable (though alas forgotten) source that I heard of the debt that Islamic scholarship owed to Christian scholars who knew both Greek and Syriac. There was, of course, another debt to Jewish scholars writing in Hebrew, but i don't think there was any tradition there of extensive translation of Aristotle etc.

      Yes, I realised too late that it was silly to write of fen when consonantal. I was thinking of another letter entirely.

      The Syriac practice is an interesting mode of spelling. It's as if Welsh, say, ignored all the mutations of initial consonants, because Welsh readers could be relied on to recognise the morphemes and apply the changes automatically. Even English has an element of this — as in the morphemic spelling electric whether the final consonant sound happens to be s or ʃ or the expected k.

      All that remains to be explained is whether the Arabic spelling is the result of transliteration from Syriac or whether there was a sound change in Syriac itself for these particular words in contexts, and that Arabic chose the f variant as citation form.

      Delete
    5. Hi,

      no, come on, I am not a reliable source, on THAT at least, no. I meant 'I have always known that [since times immemorial, i. e. the times of my univ. studies]'. The thing with the debt and so on is correct, after the closing-down of Plato's Academy by Justinian AD 529 many scholars emigrated to Persia, from where they started disseminating Greek wisdom in what was to become the Islamic Middle East... to cut a long story extremely short. On Syriac too you're right, at least as far as I can judge. I'd suppose, given that in Syriac /p/ and /f/ were allophones (or morphallophones, or something like that) and that in Arabic there was no /p/, the Arabs simply mapped all Syriac /p/s and /f/s to their /f/. But that's probably again an extremely foreshortened perspective.

      Which (non-consonantal? in a Semitic alphabet?) letter DID you have in mind while mentioning 'ف'? I am puzzled.

      Delete
    6. Wojciech

      I meant 'I have always known that [since times immemorial, i. e. the times of my univ. studies]'.

      Only teasing. Without that always, the sentence means 'I used to know that but I don't know it now'.

      You remind me of someone who taught in Leningrad a year before me. Because of his seniority, colleagues in the English 'cathedra' asked him, as they didn't ask me, whether he was married. His answer caused consternation:

      I ↘have ↗been ↘married.

      It just wasn't in the grammar they taught their students.

      What differentiated the Syriac scholars from the Persians, I believe, is that a number of them regularly operated in three languages.

      Delete
    7. No, hang on, I mean, with 'I have known' I meant something like: I had known it since times immemorial, then forgot it, then it was brought to my awareness by your WP stuff... . Probably too much for just this simple sentence.

      In th'olde dayes of that kynge Arthure I had a very thorough course in the history of philosophy where they explained such things. In yon dayes, they treated history seriously, as a chain of events with causal connection and 'wie ist es eigentlich gewesen'. Not so today, where history is an Olde Curiositie Shoppe at best, or a storage room for things weird and abstruse. This is at least my perception of today's historical awareness. Small wonder they have regular recourse to WP, where information is served fragmented and connection-free.

      Syriac scholars, whether in the Persian or Byzantine Empire, were usually fluent in Syriac (a version of Middle Aramaic), Greek und oftentimes a third language, depending on the circumstances. After the Islamic conquest, in Arabic.

      btw, it would interest me if you see a difference between 'he has gone' and 'he is gone' and what, to your mind, this difference be, if any.

      Delete
    8. Wojciech

      I meant something like: I had known it since times immemorial, then forgot it, then it was brought to my awareness by your WP stuff... .

      This is what is meant by I used to know that.

      As for he has gone vs he is gone, it's complicated. There are (at least) three sets of norms: historical, geographical and personal. Plus, in most styles speech and in informal writing the difference is obliterated by reduction: He's gone.

      For me personally, He has gone to London may to refer to a brief visit or a permanent move, while He is gone to London implies a long-term if not permanent change of address.

      The presence or absence of a place adjunct also makes a difference. Bare He has gone is less likely than He has gone away or He has gone to Leeds. Bare He is gone is less likely than He has left but, even so, more likely than bare He has gone.

      It's even possible to say

      He has gone by train, but only once or twice when he was young and even then he hated it. Now he always flies there.

      As for the multilingual scholars, it's not so much a question of whether they could translate things as whether they did have patrons who in effect commissioned them.

      If you can read a language easily enough to be capable of translating it, then you don't yourself need a translation.

      At certain times in Islamic history there were rulers who fancied having courts that were centres of transcultural scholarship. Unless I've misremembered my source, it was the Caliphs of Baghdad that got the Christian monks to translate from Greek to Syriac so that they or others could translate into Arabic.

      Delete
    9. thank you for your explanations.

      You can translate things from a language you read fluently in order to make them accessible to others. Or, if you don't read the language fluently, you can still translate them for yourself, in order to be able to read them fluently. I have done such thing myself in my school-days.

      'Unless I've misremembered my source, it was the Caliphs of Baghdad that got the Christian monks to translate from Greek to Syriac so that they or others could translate into Arabic.'

      the translators were for the most part native Syriac speakers, Christian (Nestorian or monophysite), or 'Sabian', whatever that be (e.g. the famous Thabit bin Qurra), in any case non-Muslims. Muslims were so convinced of the superiority of Arabic that they simply did not stoop down to learning foreign languages, least of all Syriac. Whether they (the translators) mostly be monks I don't know.

      Delete
  2. Alveolars in loanwords are regularly mapped to the emphatic consonants, and have always been - even early Latin loans like Siraat (sorry, no Arabic on this keyboard).

    In modern loans, p is regularly mapped to b, but you're right, in older loans it seems to have been f. Presumably these words were borrowed before the Arabic p became fricativised.

    ReplyDelete
    Replies
    1. 'Alveolars in loanwords are regularly mapped to the emphatic consonants, and have always been'

      yes, thanks, this I know having studied various old philosophical texts, but d'you know of any explanation thereof? Did the Arabs actually hear the Greek consonants as their emphatics, like?

      Delete
    2. It may just be me, but I hear the "non-emphatic" tā’ as distinctly more aspirated than the "emphatic" ṭā’. If the same is true for the variety of Arabic that borrowed these words from Greek, that may have influenced the choice of ṭā’ over tā’ (and qāf over kāf).

      Delete
    3. 'I hear the "non-emphatic" tā’ as distinctly more aspirated than the "emphatic" ṭā’.'

      same is true for me. Maybe the non-emphatic is fortis, whereas the emphatic is lenis (as a concomitant feature)? Appears a good explanation of this 'Arabic mystery'.

      Delete
    4. Georgian borrows Greek and Russian /ptk/ as ejective/glottalised /p't'k'/, and /f/ as /ph/, so this phenomenon doesn't seem to be limited to Arabic.

      Delete
    5. Borrowing /f/ as /pʰ/ is not unique to Georgian; my native Bengali (and other Indo-Aryan languages, except Urdu, maybe) borrowed English, Portuguese, Greek, Persian, Arabic and all other f's as /pʰ/ - well at least until /f/ started encroaching on /pʰ/'s territory and started replacing it even in native words.

      Delete
    6. Lithuanian and Latvian used to borrow foreign /f/s as their /p/s, e.g. 'prancūzas', Frenchman, Polish 'Francuz'. The simple reason is that /f/ does not exist in these languages.

      Delete
  3. The answer, of course, is that these are not the Arabic names of the places in question.

    Of course.

    It is either that or that, in some languages, people perceive voiced plosives as voiceless.

    Do you speak Arabic?

    ReplyDelete
  4. If /a, aː/ is “retracted to [ɑ] in the environment of a neighboring /r/, /q/ or an emphatic (pharyngealized) consonant”, then I guess Ṭarābulus must be tˤɑˈɾɑːbulus.

    ReplyDelete
  5. If that's right, they should have used unemphatic T for Ancient Greek theta (IIRC, delta, tau, and theta were a VOT triplet); did they?

    ReplyDelete
    Replies
    1. well, at that time the Greek theta had been the 'th' in 'thin' since long.

      Delete
    2. "well, at that time the Greek theta had been the 'th' in 'thin' since long."

      Really? I could say that this would have been the case for most of history; but then, the Greeks and Arabs have been interacting for way too long and I doubt that Θθ was pronounced as a fricative in the early days of association.

      Delete
    3. in the early days, that is before ca. the III c. b. C., certainly not, but at that time the contacts between the Arabs and Greeks were infrequent enough for no words (at least no known to me) to be borrowed.

      Delete
  6. And this /p/-/f/ confusion made فارسى Fārsī /fɒːɾˈsiː/ out of پارسى Pārsī /pʰɒːɾˈsiː/. But why would the Iranians morph their own language, just 'cause some Arabs said so, ideas?

    ReplyDelete
  7. Is my impression right that b undergoes a bit of devoicing at the end of Arabic words? So Halab would sound almost like "Halap"?

    ReplyDelete
  8. What foreign language /p/ becomes in Arabic might depend on the sort of phonemic distinction (in terms of the labial consonants), the foreign language (Greek or Latin or Italian) has. Exempli gratia, ancient Greek had /p/ vs. /b/ (vs. /pʰ/, but that's irrelevant) but no /f/; so Greek /b/ was translated into Arabic as /b/ and /p/ became /f/: on the other hand modern Greek and most other languages today have /f/, so, /p/ becomes /b/ in Arabic (as /b/ is closer to /p/ than /f/) - the fundamental point being the sort of phonemic distinction the parent language has.
    That's my theory anyway.
    As for Aleppo حلب, it has to be the Italians who translated the word into Arabic and not the other way around, so, what حلب becomes is basically an Italian transliteration problem not concerned with how Arabs perceive sounds.

    ReplyDelete
  9. '- the fundamental point being the sort of phonemic distinction the parent language has.
    That's my theory anyway.'

    this might depend on whether the speakers of the borrowing language care much for the 'phonemic distinctions' that the lending language has, or at least care more than for what they hear.

    Germans are notoriously unable to pronounce the voiced 'j', their 'John' is 'chohn' and so on. They obliterate a distinction. They can't, either, pronounce the initial voiceless /s/, so for instance the Polish name 'Sopot' comes out, with them, as 'Zoppot', notwithstanding the fact that their 'z' exists in Polish too, spelt 'c'. They again obliterate a distinction. But they could hardly care less.

    ReplyDelete
    Replies
    1. I believe foreign language speakers do care about phonemic distinctions of another language, that is if the phonotactics of the mother tongue allow it. For example Polish distinguishes Ś /ɕ/ and Sz /ʂ/ (not actually retroflex, rather retracted, but I am going to stick to Wikipedia conventions), and Ć /tɕ/ and Cz /ʈʂ/. But, English just has Sh /ʃ/ and Ch /tʃ/. Now, both Polish Ś and Sz would seem to English ears as Sh as English doesn't distinguish between 2 post-alveolar fricatives, but no one would confuse the 2 affricates with them; on the other hand a Frenchman might confuse all 4 as French doen't have affricates.
      Your German example holds to a certain extent, but not entirely, as to the best of my knowledge, German doesn't allow /s/ at word-initial positions, so it basically is a toss-up between /z/ and /ts/, and hence your observation.

      Delete
    2. ' Now, both Polish Ś and Sz would seem to English ears as Sh as English doesn't distinguish between 2 post-alveolar fricatives, but no one would confuse the 2 affricates with them;'

      The Anglophone (and not just them) confuse (pairwise) the Polish 'sz' and 'ś', "cz" and "ć", 'ż' and 'ź', 'dż' and 'dź'. Many of them say not to hear any difference between them (again, pairwise) at all. Maybe I haven't got you.

      What you say about German is of course right and bears out my observation that people most often don't care about the phonemic differences in foreign languages (those they learn, let alone those they occasionally borrow words from) but only about that which they hear (which, of course, is filtered through the phonemic system of their own language(s)). To a germanophone (here the best example is my own son, predominantly germanophone, whose linguistic development in the different languages he has been learning I have been following since he was born) the Polish initial /s/ sounds like their 'z' (i. e. /ts/), hence Sopot-->Zoppot.

      On the other hand, I don't know if you have ever taught languages, I have to some extent on the margin of my major occupations, and hence I know how difficult it is to make learners aware of phonemic differences in languages they sincerely want to learn, not just borrow a word from every now and again. For instance, Polish learners get most English vowels wrong, so different they are from theirs, and it takes hard work to teach them that 'man' should sound different from 'men', or 'flash' from 'flesh' and 'flush'. They at first won't believe that there be any 'phonemic difference' adall...


      Delete
    3. Oh, now I see, sorry, you meant that no Anglophone would ever confuse the Polish 'sz' and the Polish 'cz'? No, you're right. But the point was: even though the difference between Polish 'ś' and 'sz' or 'ć' and 'cz' be (in that language) phonemic, most non-polonophones would disregard it and not even hear the difference. Witness numerous germanisations of Polish names as Śródka-->Schruttke, Wiśniewski-->Wischnewski, and so on. This bears out my point. The ancient Arabs heard the Greek 'pi' in a way (we can only guess which) which made them transcribe it with their letter that is nowadays pronounced /f/ in Arabic, hence Aflatun for Plato and so on.

      Delete
    4. Could is just be a general befuddlement with foreign words that is hard to explain by reference to individual sounds? I was listening to a couple of English sports commentators' attempts at Szczęsny: one said ˈʃɛzni, the other ʃəˈt͡ʃɛzni. The difficulties with szcz, ę and y are all understandable, but why the z? A recording from a Polish native speaker contained what to me very clearly mapped onto English /s/, and given English words ˈpɑːsnɪp and ˈt͡ʃɛsnʌt, it's not as if there's any difficulty in pronouncing it in that position.

      Delete
    5. could be Italian influence. In Italian, "sn" (occurs mosttly word-initially, though) is always /zn/.

      Delete
  10. Alan

    given English words ˈpɑːsnɪp and ˈt͡ʃɛsnʌt, it's not as if there's any difficulty in pronouncing it in that position.

    When average punters read names aloud, it's not the phonology that informs them so much as the symbol-to-sound spelling norms. Both parsnip and chestnut have more than one one consonant letter before the N. The expected norm for -sn- would be -SSN-.

    If we English never saw Glasgow written down, we'd stop annoying the Scots with ˈgla:zgəʊ

    ReplyDelete
    Replies
    1. OK David, but this still leaves Alan's question unanswered, nay, adds another one of the same kind: why do you English annoy us Poles with 'shchezni' and your Hibernian compatriots with 'glazgow'?

      Delete
    2. Wojciech

      I thought the answer was obvious. There were phonological constraints at the time that the spelling conventions developed. Those constraints no longer exist, but spelling norms take much longer to change.

      Another -SN- name from a Slavic-language region: Bosnia ˈbɒzniə. In Cyrillic Босна, not Бозна.

      Delete
    3. sorry, I am not geddin' it. You mean to say: in th'olden dayes 'sn' or 'sg' had to be pronounced with a /z/, now no longer? (tho' the spelling persist)?

      The pronunciation [bozniə], with a [z] has always puzzled me (and continues so doing up to, and including, the present day). In Polish it's "Bośnia", with a palatal voiceless "ś". Whence the [z]? Italian influence?

      Delete
    4. This comment has been removed by the author.

      Delete
    5. It could be that the voicing of the /n/ and /g/ sound causes some sort of assimilation that makes /z/ instead of /s/ sound more natural in that context. Parsnip and chestnut could be explicable because of the environment of the /n/; chestnut contains the voiceless /t/ (which probably was historically pronounced) and parsnip has /ɹ/ (or historically did for non-rhotic accents) which does allow voiceless consonants to appear adjacent to it (compare, bag[z] and pin[z] versus pur[s]e). Even as a non-anglophone I feel /z/ and not /s/ is a natural choice for all of Szczęsny, Bosnia and Glasgow.

      Delete
    6. Even word-initially, like in Italian 'snob' [znob]?

      I find /z/ natural in Glasgow, but not in front of anything like /r/, /l/, /m/ or /n/. But I am conditioned by my native language, Polish, in which there is no /sg/ (at least I can't think of any) but a lot of /sn/ etc.

      Delete
    7. Wojciech

      I think the old phonological rule was that S became [z] before a voiced consonant by backward assimilation. (Either that or consonants after S and Z became devoiced or voiced by forward assimilation.) Your snob example shows that this was not independent of context. My impression is that the rule applies typically in stressed syllables.

      Letter Z is very much a Johnny-come-lately in English spelling. My impression is that it was no great loss in earlier periods of English because [s] and [z] were to such an extent in complementary distribution.

      The association of spelling and sound persists even with recent coinings including the acronyms asbo (antisocial behaviour order), Asda (Associated Dairies & Farm Stores), Usdaw (Union of Shop, Distributive and Allied Workers).

      I find /z/ natural in Glasgow, but not in front of anything like /r/, /l/, /m/ or /n/.

      That's the difference between Polish and English. We have Israel, Islam, Kismet, Disney all with z.

      Delete
    8. Parsnip seems always to have been a phonological oddity. According to the OED, most Middle English spellings used an extra consonant even before the R was introduced, for example passenep, passenepe, passnepe, pastenepe, pastnep, pastnepe, pastyrnepe.

      The T in some spellings reflects the Latin source pastināca — just as the T in chestnut reflects the etymology chesten nut.

      Delete
    9. Ah, OK. Polish "pasternak". We have 'Izrael', with a /z/, it could have been 'Israel' but is not. Whence the /z/ I dunno, perhaps from French? Otherwise a /s/ in the words you mentioned, including 'Disney' /disnej/. /dizni/ or such-like are considered kind of snobby. But we also say /glazgou/, annoying the poor Glasgovians...

      Delete
    10. I, an American (not sure if that's relevant), would never voice the S in Glasgow, but I might de-voice the G, glæskou. Contrarily, Wikipedia and Wiktionary don't list any proununciation with an S sound.

      Delete
    11. I know ONLY this pronunciation from the US. That is why I kept marvelling at the English's habit of annoying their fellows in kingdom with a voiced 's'.

      Delete
    12. I've just bought David Crusytal's delightful new book on English spelling. It confirms and illuminates what I thought.

      Old English had no use for letter Z. The single sibilant phoneme was voiceless in voiceless environments and voiced in voiced environments.

      After the Norman Conquest, our scribal tradition was French, so a few borrowings spelled with Z in French were copied into English. But we were strangely averse to the letter — right up to Early Modern English times.

      • A choice Shakespearian insult:
      Thou whoreson zed! thou unnecessary letter!

      • In 1582, Mulcaster, a writer on spelling (quoted by Crystal with modernised spelling & punctuation):
      Z is a consonant much heard among us and seldom seen. I think by reason it is not so ready to the pen as s is, which is become lieutenant general to z.
      Crystal's interpretation: Z is more difficult to write than S, which therefore deputises.

      Delete
    13. Yet it can be useful for mimicking Zumerzet dialect/accent, perhaps other Zouthern 'uns?

      Delete
    14. That's the exception that proves the rule. Z is so useless for representing familiar accents of spoken English that it's particular suited for signalling the unfamiliar

      Delete
    15. Then it has replaced the old English letter 'yogh' or some such in McKenzie, Menzies, und such....

      Funny that if for the southern f->v there are a few loanwords (from the dialects) in Standard English, eg 'vixen', 'vane', there are none (for aught I know) for the analogous s-->z shift. Or are there?.. That'd be a sur- not to zay a zurprise to me... . Zounds!

      One contrast, perhaps not very worth-while dwelling on: seal vs. zeal. Also, it wouldn't be bad for foreign learners to have the multifunctional ending -z spelt correctly, wolvez (why don't you English spell 'wolfes'?) learnz, housez. But your spelling is (iz) such a mess anyway...

      Delete
    16. No. Zed replaced yogh in Scottish use. (In English use it was dropped.)

      I'm not sure about the history of [z] in Scots. One current spelling of Glasgow is Glesca.

      As well as the spelling vixen, Crystal cites vat, vane and van alongside fat, fane and fan. Initially, English was resistant to French V — vers was spelled ferrs. But then V became acceptable and took over as the symbol for [v]. It was useful to distinguish view from few.

      Zeal was a French word (of Greek origin with the zeta preserved in Latin) introduced in 1382 by John Wycliffe in his translation of the Bible. By then the spelling seal was already in use.

      For [s] in some positions, there was some use of -SE — for example to distinguish the words now spelled grass and graze, but it didn't catch on. But we still have a lot of words from French with CE for [s] — advice, price etc.

      The spelling wolves is down to those pesky French, who insisted on hearing a difference between [f] and [v]. At least, that's where the V comes from. The O is one of those strange substitutions for U — to make the word more legible in Gothic script.

      Delete
    17. No. Zed replaced yogh in Scottish use.

      Sorry, I misread you post, Wojciech. I should have said:

      Yes. Zed replaced yogh in Scottish use.

      Delete
  11. Same for English snap /snæp/, not /znæp/.

    ReplyDelete
    Replies
    1. But your micro-theory, set forth above,

      'It could be that the voicing of the /n/ and /g/ sound causes some sort of assimilation that makes /z/ instead of /s/ sound more natural in that context.'


      does not (seem to) provide for this limitation. It would, 'as is', work perfectly for Italian.

      That the English should say 'glazgow' amazes me, though. I thought they did not have this kind of forward-looking assimilation. We said 'wezd germany', because Polish (and Dutch to an extent) works that way, they (I thought) said 'west g...'. Did they not?

      Delete
    2. Re Dutch: they say [ig ben] for 'ik ben', I am, whereas the Germans pronounce their 'ch' in 'ich bin' voiceless.

      Delete
    3. I think morpheme boundaries and word boundaries as in West Germany function differently from words such as snap or Szczęsny. But, I seem to have found a few contradictions myself, as in mince or pounce. So, I guess my micro-theory applies to non-word initial /sn/ and not /ns/ as in pounce.

      Delete
  12. I'm Arabic, and I'm studying English Linguistics.
    standard Arabic has a nickname which is "لغة الضاد" the English translation to this is " The language of Dhad".

    Dhad (ض)is a sound in Arabic Alphabet (similar to [d] sound.
    They gave Arabic Language this name assuming that
    there is no other language consists of this sound in its phonetic alphabet!
    Is that true?
    and what is the phonetic symbol and description for Dhad(ض)sound?

    ReplyDelete