123 9 32MB
English Pages 190 [97] Year 1985
ll N I W E R S Y T E T
J A G I E L L O
N
S K I
ROZPRAWY HABILITACYJNE NR 101
ANDRZEJ PISOWICZ
ORIGINS OF THE NEW · AND MIDDLE PERSIAN PHONOLOGICAL SYSTEMS
KRAK6W 1985
\J N l W E R S YT ET
J A
o· I
E L L O
N S
KJ
ROZPRAWY HABIUTACYJNE NR 101
ANDRZEJ PJSOWlCZ . ·. ·,·
ORIGINS OF THE l\1EW· 1\ND MIDDLE PERSIAN· PHONOLOGICAL SYSTEMS .
.
.
.
.
-
.
NAKLADEM UNIWERSY TETU JAGIELLO�S,KIEGO
Rec;laktor to11111 J6ze:r Reczek Reoenzenoi Leazek Bednaro:ttak AndrzeJ Czapkiewioz
Contents
Nr og.· 1206 .
:
-.
. ·
..
'.
-
.
Page
. •••••••••••••.• •:• .... !• ............ o• �.o··.__........·.�·-·�•-· ···•�-:.·• •�:•••7 ',: O h o. p t e r I. Phonolog.ical S,}'stem of Hodern ?ersian .••••••• u ..•.9, . ' � 1_. In�roducto� Re�rks •••••• o ••• • e "'.• e .... -. c11_•-�.� ••••••••-•• - ••· .• ••• 9 2. Vowels •-••••••••• ,,. e_•. ••· -e.••••·�·•••,. ._.••., • •-• •. o· • .•••.• .. ••·•.••• ., 10.. . · 2.1. Distinctive· Fea _ tures �-••••••• ·•� ••••••.• �.-� •• G. •. •-•• • •• •·• .1o . P11 11tao�
Przelo•yla na j�zyk angielski Barbara Buja-Stanek
- e�.tUres . • " ••••.•• •.•••••••••. 11_ _ _ •. ••••••· .... 2.2 ·. Redundant F 2 .3. Main Vowel Allophones �-•••• o • • • _• •--•. • ••• ._ •.•.,. � •-• •• : •••••• 13 2.4. Fluctuations in the Use of -Vowels • -'••••••• •••••�•••••·•14 . . 16 2. 5. Distribution of Vowels •••••••• .....................· .......-••••• . •••••••••••••,•••·•• .•••••••••·••••••••17 3� Corwon�nts • .-••• �•••�•••· · 3 ."1 � Introductory Assertions•• .-·" •••.•••••••••• � •••••••••••••• 17 3.2 . Semivowels /y/, /w/ ••e_o • ., .............. ��••••�• .... ••••••:--•••·• �_ _ja 3. 2 0- -1. /y / ••• 0 ............ � •••• !I .�.-••• ................ 0 ....... � .1 a· 3�2.2. Problem •Of ])ip_hthong /ey/.••••,•••••••••••••••••••22 '.5.2.3. /w/ - Problem of Diphthong /ow/ ••••••••••••••••,24 $ •.• • • •••
ISBN 83-233-0081-X ISSN 02J9-782X
Wykonano z gotowych dostarozonych oryginaL6w w Druka.nli Uniwersytetu Jagiellonskiego w Xrakowie, ul. Mani1'estu Lipcowego 13 'Wydanie·I. Nakl.ad 150 + 22 egz. Zam. 441/85 P:rzekazano do powielania w lipcu 198.5 r. .Powielanie uko6ozono w grucln;lu 1985 :r... Cena zl 185,-
Ark. wYdo 12,25 1 Ark. druk.· 11 /2 1406/85
3.3• . Liq�ids /1/, /r/ ·•••••·: .•••.•.••• �•• � ...........· ••• � .... . �•• 27. - s /ID/-, /ri.L_ ••.�::.� .. ;••.•••••••• 3.4, llasa1 · · ....•••�••••�•,e•• _ �28:_ · ·• 9! •� _ 029 3.-5. Fricatives ••••&•' ...• ..•••!•••••-•••·•• .. � ............... e:.•,. .....•• · · · lt •••• � ·.,29· ........... � .. 3.5· .. 1. Labio-d.8nta1- -1r1,- . Iv/· ······· 3.5.2. Dental /s/, /z/ and !lveoll;l.r /s/, /z/ ••••••••••30 3• 5 .3. Uvular (Y:t, [rJ and; Laryngeal /h/ ••••••••• .; •••• 30 3.6. Af�icates /'c_/�(·.,,/J/: •• - . o . •-• ...... � •·•·•-•·•, ••·• • .-. � � •.•• ·...... �1 3.7. Plosives /p/~/b/, /t/--/d/, /k/~/rJ •••••••••••••••••.32' 3.8. Distinctive Features of .Phonetically Voiced an,d vo'{Cel�Ss - Consonants ............. .·.•····�·•- ....... .. ... ·e•�,--�-••• . �---33:. · 3 .s.1. Plosives and Affricates •••• � ............... �• •••••• 33 . 3.8.2. Spirants ••••••••••9 ...................................39 3.9. Distinctive Featureri of Phoneme /q/ and Its Position · $ys_ tem of ·consonants · ............................. -��in - the
5
4 3,·9. 1. Ailo;r..-on�S of /q/
3.9.2. Place of Phoneme Consoii.9.nts
_-42
o ••••••••••••••- ..................•
/q_/ within
O • • • • • .• • • • ·• • • • • •
the Systei:n o:t;
ii .
9 •• • • •
o- e • •••
I . -• •• ct • •
44
. 3.10, Phoneme /?/ ••••o••••·•••••�-•-, .................... • - • � .........�47 . · - , �- ••.-,••...•.. .-4 .- •••••• , • • ,50 3, 10.1. 1n1't1ai / - 7/ .... , .. -.. 3.10.2. Place of/?/ in the System of ConDonanto ....... 51
3.11. /tabiea/ Consonantal Phonemes of Modern Literary Persian •••••••• 52 Acoustic Fea-tures of Phonemes of Nodern.Literary. Pers. Ian , - _,••••·, -.• •.••-••••• a • • .-• • •• " , . •�••·• • • ,,·• • . • • • , • ,-.5 _ 3
��
•. •.,•• · •,•, � •,.•.•••.•.••,•••:? + 4. The Syllabl_ e �.,-: • ·••••••� - ••,•.••. •• . ,., -. , • • • • � •·•."., �-,, • • �55 • �• • , ., • •· . • ••• • e\ .• · · • •• ,_ . 5, Stre·ss • .-� • ••••.••. 6. Distincti've Character of Collcquial :Language Vs. · Literary 7 Language in Relation to the Inventorr of Phonemes ••••••••••••5 · •••57 ' '°. •••• · •• ••• • •••• . · ••••••••·• 6.f. SynChro�ic.Data •••••!'••• .. ••• - •.•!I •••• II.� ........ e • • • • • 5� .- ·· ••••••• 6 .. 2•. DiaChronic Conclusions • •• .
.
. '
.·
"
� .\), Descriptions of Pers'ian Vocalism of the First· Half
of tho 20th c. � ....... .._ e .. . .. . .. .. . . . ,, . . . ... .. �••• .. .... . ... 4'-•• �.85 h n J' I: o r III. An Attempt at Interpretation of Causes/ _Purposes of J;ievelop:nent of ?e1•sian Vocalism
b_e"h·1e€n_13t"}?. and 20th cc. ••••"•••.•••••.,_s" ,,.- ... .. .;.87 " I; " r t o r IV. Haterials for Recon.Gtruction cf Pt.onologica.l. 1 , �IC)dern
6112.2 ,-_ Consonantism_ • � ••• .•.- �:••.•• o
,,. •• " •••••• e·• .... o•. •• - ••
C h a p t e r II. Materials for Reconstruction of l'h'onological System of Classical:·Persian /13th century/:
VocA.lism • ,,·.
. e. + •·11 ; • .-:a"� s-•·• Ii .. " ••.• •• - ." -1 ..... •·• �. 6'":.
.60
"65
. ••••••• ,65 1, General Survey of Sources .of Information ••. •••••••••• 2·. Diachronic Conclusions Emerging from Fluctuations Occurring in · Vocalism of !;fodern ·Li terarJ. Language· •••••••••••••••.••••• 6'1
3. Reconstruction of Vocalism from Which Stems the Hodern Vowel System of: Persian . Language ••••• ,••••••••••••• , •••,•••• 72
4. Absolute Chronol ogy i,f Development of Persian Vocalism in _· 13th - 20th Ceht1(ries· • .••• � •••••• -.a •. • � ••••••-� ...........•.• 73 . · 4.1•. - A� 2, D . ata .. from · Codex Ouma.nicus. • .••• . •••••••••• •••••••••73 4.3. Armenian Transcri'ptions of Liturgical Texts /ATL/ •••••• 76 4,4. Transcription Used. in the Mediaeval Armenian Dictionary _ Or PerSian Ara1)isms /APA/ � •••••••• .- •• II • • • • " . 77. 0 . .. .
• -.
Iii Iii • .• •
4.5. Persian Vowel s Recorded in "A ';fonder beyond Three Seas!' - • _ ·Nikitin · - ....... by-Athanasius - ..... � ••♦• · •••_,,c_.•••• , ...... ••., .. 79
4. 6. Data from Latin Transcription of Persian Translation _ of K
a trend towards the least. expenditure of energy/ and beyond it - in cen�
turies-old relationships, often very close, with other languages. One of the stimuli which induced the. present author to -take up
the
subject of Persian diachronic phonology was the following statement.by
G. Windfuhr coiltain�d in his surv;{ of 'rranian studies and referring to
the state of research on the New Persian period: "The. widely scattered
accumulated knowledge of this .subject urgently needs a thorough review and revision" /ivindfuhr: 144/. This postulate may and shoiild, be exte�d.;;; · ed to encompass the earlier developmental" stag'es of. Persian as well/for they pose at leaat as many problems as the New Persian period ,.
"' Nan:i( s /possibly with the year of publication/ in l'.lc>.rentheses rei'er to the bibliography at the' end of the paper. Numbers which follow them., after a colon, denote page number. If the whole of a text is meant, page numbers are not given.
8
The present author also received encouragement for the project from the late -Professor Jerzy Rurylowicz, Andrzej Pisowicz
· Chapter I
PHONOLOGICAL SYSTEM OF MODERN PERSIAN 1. INTRODUCTORY REMARKS
.·
.
.
.
.
;
.
Starting a phonological analysis of a particular language material one must fully realise exactly what language he deals with /Sommerste1u:· 16/. It is equally important in b_oth diachronic and syn�hronic studies. As- far as Per;ian /NP/ in particular is concerried 9 since the term ·1 1� sel;f has an exceptionally wide range of m�aru.ngs, an iJnmediate specifi- cation o:f the object of the research is necessary. This is no -place :for a detailed scrutiny �f•the sociolinguistic prob lems of Iranian Persian, which are largely unexplored anyway. _Let · us confine ourselves to establishing the literary Persian .of Iran, i.e. its official variant� sometimes cailed by the'· Irani�ns "fflrsi..:ye a.rftste" · /florid Persian/, as the field o:f- our research.. This style. based on the . written language, .is used for special perform�nces, and official ann6unc$� ments /e.g. TV' news/, that is, in the sjtuations when the existence of the corresponding written text is taken fC>r granted., --- .-... _ In order to make conclusions concerning diachrony, it .will be helpi'µl to use the data on the lit�riry oral style /f�rsi-ye moM:irerei/, wh..ich · · is dramatically different -from the officiaLone, and very close to the colloquial, language /fllrsi-ye>funmiya.rie/, based on Teheran dialect /Tehruili/ Finally 7 · a reservat�on '.conc!::rning the range of materia1" presented in _this chapter. Facing the - fact, that there are numerous papers on this_ sup� ject /Kr�sky 1947,. 1966, 1939; Nye; G-G; Giunasvili 1965; HO:dge; Towhid:L Provasi/ this presentation has n:o claims '.to completenes·s. is going give a broader treatment to these elements of the system which provide:-/ more data for the diachronic research,_ hence the omission of description of such suprasegmental mE;ans as intonation for which there exists no his torical verification. .
·_
-
. /
-
.
.
It
to'
11
10
following table: · 2 0 VOWELS 2.1.
Distinctive Features
On the whole, descriptions of Persian vocalism are unanimous as to the quantity of simple vowel phonemes. Generally, there are· six of them men tioned: /i/ ,/e/ ;;a/ ,/a./ ,/o/ ,/u/ � cf. /G-G: 11/. C ,T. Hodge /p. 357-8/ ai�o mentions the reduced vowel /a/, how�ver making a reservation that this phoneme does not occur in every informant's speech, and if ever, it does solely in the informal style,. The material he re.::'ers to indicates that it is an optio:,al· allo:·:;oi1e o: the phoneme /a/ which occi.:.rs in an u.n stressed p�siUon before a syllable containing /A/. A characteristic feature of Persian vocalism is the coe,icistence of two vowel phonemes of the broadest degree of opening; they give the whole sys tem a shape of a "quadrangle 11, Which :Hmpli.fied, is 1the following: front back high ii/ /u/ mid /e/. /o/ low /a/ /d/ /For technical reasons, the back!!,:,'lill be rriarked ! in the present work/. E x a m p l e s : ' sir· "satiated" - ser/r/ "secret" - sar "head" - sAr "starling" .. sor ''slippery" - sur "feast". bi "without" - be"to" - b� 11 with" - bu "smell", mehr "love" - mahr "marriage _gift" - mohr "seal". Here is a short characteriStics·o:t_ each vowel based mostly on the data provided by the sdurces mentioned i:n the Bibliogr1'1.phy /Horn 1898; Lambton; L9.zard 1957; 'rl'ye; Rastorgueva 1960; Rubin6ik 1960; Shaki; Zirkov/. The data was compiled by Gaprindal3viJ,i and Giunasvili /hence abbrev iated as G-G/ in their work Fonetika _persidskop;o _j__azyka; /i/ - front. high, unlabialized. tense, stable, long le/ - front, mid /between high and low ones/, unlabi alized, lax, un ·stable, short /a/ - frpnt, low, unlabialized, unstable, lax, short /�/ back, low; slightly labialized or unlabialized, tense, stable, long /o/ - back, mid /but closer to high/, labialized, lax, um,table, short /u/ - back, high, labialized, tense, stable, long. The above-mentioned features of articulation can be presented in the
front high · labialized tense stable lo:ng
/ii
/e/
+
:t
/a/
+
/u/
+
+
+
±.
+·
+
+
..
+
+
-
+· +
.t
+ +
+ +
Detailed experimental research carried out by �G gave them grounds for a .certain modification of the traditionally accepted desc:,:-iption o� the articulation of Persian vowels. Here is their ·table/ in parentheses transcip:tion-used by the present author/: high mid
I
2 1
central
front
I
/1/
i /e/ a
.
unlabialized
I
back.
..
a /§./
.!
COll!paCt . d.i:ffuse grave
/a/
0
0
+.
· •
+
+·
2.2/ Redundant .Features
/u/
i
+ ..
+·
/i/·
/Q/
-
+
..
+
,1
lo/
labialized.
:
T,he modification b:y the Georgian phoneticians does not cal cha:nge in the attribution •of the distinctive features vowels in acoustic terms /Jacobson, Halle - Phonology and three commonly occurring distinctive .features su.ffice for cation of Persian vowels:
/a/
..
u/u/
j ;
1
make any radi to ind:i,vidua1 . Phonetics!._ The . • the identif'i•
/el "'!
The remaining :features oharacterisi� the a.rticulaU�n Of P�rsii;µi vdw- < · els are therefor� redundant. :t,fonethel:ess, they should be anal:ysed, riot only f�r the sake of accu;-.acy _ of description; but b�cause of their pos sible import for diaohro�io �nalysis., .The features now redundan: one� . might have been rel�vant� . .·· . ·•. .· .. . . . . . _ And so, those of them worth mentioning. a.re:, labialization on the one. hao"ld, a!ld the ·set of tallying features ... stable, te�se, long.... unstable, . ·1ax, ·short on the other /cf. Krarnsky 1966: 216 and below/, The labialization is clear in case of /o/, /u/, while for /A/• G-G research pol�ted to individual fluctuations in the pronuncint,ion o:f di:f-
12
13
fer.ent informants: from a slight forward movement of the lips to their complete passivity /G-G: 13/. Contrary to the opinion of some resea;chers /Nye: 6; Rubincik 1960: 1 8/, there are no grounds for acknowledging /&/ as a labialized vowel. Moreover, grave (o/, /u/ differ from acute /e/,/i/ in respect of lip movement. Considering the groups of features conneoted with tensen ess and length the t1:_aditfonal stand-point referring to diachrony, orthog raphy, as well ,as verslficatiori data, bid.a one to perceive /i./. Ii.I. lu/ as long vowels ,contrasting short /a/, /e/, /o/ respectively /cf: Horn 1898� 21, 26, 3137; Jensen: 17; Shaki: 46 ff.; Kramsky 1939: 69/. Indeed, the origin of the latter ones from the historical short -vowels is fodicated even by t:i,e data �n loan-words, where the contemporary /e/, /o/ correspond to foreign /�/, /u/, e.g. es14m "Islam" (Arabic ?isla m; modir "manager"124; :t-Iye: II; Mattews: 3·•. G-G••. 45• aiv1·11 1965• 96/ • The above-mentioned · � G"un J. fact con. • . 0
oorna literary pronunciation, In colloquial speech, the problem is dif-. ·" t'orent, as it will be seen in i6• 1. of thls chapter /p.57/e Some researchers, for example v.s. Sokolova /publications from 1_951, ·1952/, G. Lazard /1957: 12/ speak about "stable" vowels /!1/, /1/, /u/ and "unstable" /a/, /e/, /o/. The latter can be found in G. Lazard'B· descrip-, tion: "de duree variable et de timbre souvent flottant,"11 In a rapid s-peech and a,lso in poetry, occaaio�ally they maybe completely reduced, e.g. migoz1!.ram 11 /I/ put"-- migzf1ram, bene�in --benllin "sit down", Abeie--�ble "smallpox". They assimilate and they have anaptyctic functions ·/cf. Laz11rd 1 957: 12-3/. "Stable" vowels /i.e. historically long vowela/ do not show· any of the already mentioned features of vowels /a/, /e/, /o/ • and as, Lazard notes, are.: "de timbre net et de duree rela.tivenient constarite" /Lazard_ 1957:·12/. The presented feature contrasting /a/, /e/, /o/ with /a./, /i/, /u/ cor-. responds to the classical description of. the feature tense and lax of R. �acobson and M. Halle /Tensene�s and Laxne.ssJ. Th� authors suggest that "the tense vowels are necessarily lengthened in comparison with the cor- · , · responding lax phonemes" /p •. 552/� Their further claim that "the lax. vow-- . el, notwithstanding its opener articulation, displays a shorter duration' than the corresponding tense vowel" /Ibid./, finds a good illustration in the Persian language: lax /e/, /�/�more open than tense vowels /i/, /u/. However, as it has been. pointed out, th):! feature in question is ir, · relevant to the vowel system of modern literary Persian. E. Provasi /p•.261, accepts it for spoken language� 2.3 � Main Vowel Allophcnes /1/ � s. Halm /p. XI/ and M. Shaki /pp. 48-9/ speak of short pronund.;. ation of this vo�el in the final position, and before tautosyllabic -n. ,· The same is true for th:e vowel /u/ /Shaki: 50/, and in case of the posi tion before -n --also for /!/. /e/ --is generally m-id, though aome·times it is pronounced higher.,· .· G. Lazard /1957: 11/ describes it as follows: 11»e is Shaki's description /p. 51/: "/a/ - a vowel phoneme slightly more open than the front open .vowel sound in the English words !!!§D., � - ... • Ti,P of the tongue in Persian recedes a little more and the lips are more open sideways than in.lee/." Indeed, experimental data by G-G suggest central /not front/ articulation of /a/. According to V.S. Sokolova /1952:173/, the highest allophone of-/a/ oc curs in the neighbourhood of p:ilata.l consonants as in the word kask "sour milk". ·The most-open and back allophone is reported to. occur together with U'V1ilar consonants /cf. below p� 81 /.. /a.;,.::.- similarly to /i/ and /u/ is pronounced shorter wh�n in front of -n /Shaki: 47/, and possibly at the end of a word• Furthermore, i:ri a .
particularly careful pronunciation /especially in poetry/, - one can hear o .f /fl/ with a clear lab:l.alization in the final stage of articulation /a:4d with a. c onsiderable lengtheningi. /o/ �•in G. Lazard's description /1957: 11/, it is: "moyen. ou terme; il peut �tre semblable au »� .frans:ais de »pot7/. In the word .I!.1:.£ "empty" • asp · • n iratio lasted for 40 msec •• "' �nd J n .!!!l 11leg" - 24 msec • /G-G : ·1,7 � -8 / •Asp iration occurs also in the i n . ter vorm.Uo and final positions /G-G•• 1380/ __, • It is . sim . J.la . r with /t/ and /k/ • C-G /·.P, 144/ write about the aspirati on _ which follows the ante'llocalic .inj tial /t/ that its length is six t. imes greater than the occlusion itflOJ.f,
As regards the affricate /c/, the authors of Fonetilca J?e-,.si· dsk jazyka_ speak also of the aspirati * ago on foll owing- · the ace1usi. on - and las , . . ting e.g. in the word casm "eye 11 aa tnuch as 66 msec. /G-G:1 151/, ' Althou gh in the a ffr - icat: . ;•c/-aspiration overlaps with. the dent al-- alveo1ar friction resembling /s/, it is, to a cer tain degree • Percepti-bte if e.g : Armenians, dif.ferentiating, in their phonolo g ical system, non-aspirated . . /c/ . from the _ - ·irated /cv'/ , consistently asp render •'in their • -borrowings the New Persian I cl - excl usiv e 1Y as /c, ' /, e.g. Arm • .£.;_a! "thick . ,,, -• fa+n " < 1NP-,£fill. , Ai sp rati· on characteristic for the articulation of / Pj • I tI • /k/ does not, however, occur af'ter sni • rants /s/ • /"/ 8 • cf. G-G· 143 Some asp· · ble aft · ra tion is discerni er /f/, cf. G-G: 143 , .wher- e the . . y sup • ply the - exa . i mple daftar "copy- book"• but it is ins ignific . tio ant. The lack of aspira . n in plosives occurring after /s/ • ;i; v is common, known • e g fr • ,. om Eng1·ish, and rel . ated to dissimilation in fluence of the pre ced ing spi ran ts /s/ � /v/ . h contain an element of .,. s • whic riction ana_og 1 ous to the aspirat· ion accompanyinq nlosives • Th"is P-Llen � omenon does not occur in la . -�.ages _ employi • ng asp.J_rat"'ion as a distinctive feature- of voiceless plosives . • e.g. in Sanskrit or Armenian, clusters lik e· suh sth • -skh .-. and !!.12, il, sk are equal ly possible ._ In case of langua ges-;:�h-as .-· Persian o� nglis,h, wh·ic1 � E·� 1 posseas only two series of pl e 5.ive s /S anskrit and Armen i n have ri che r systems/ • the observed neutraliz ation -o·f asn-- iration may :e int . ernreted .tn two Wa.Js, depending on the _ perceptlon of the voiceless unas"i ,_, rated /p / , ItI • /k/ occurring after spirants /s/-• s I • If . suc h plo siv es are identi��ied by the native speakers with phonetically voiceless h P one mes /p/. /t/, /k/. which haye aspirated P_ - rontinciat·ion 1n any other -Posit·Jons, a given system will -be based on . • the opposition, of voicing, - voi. celess and aspiration of the consonants, in poaitions oth er than after opi � rants, will be a redundant_ feature. However • � if the.,e " plosiv . es in • sni , te o f -thei. r obvious voicelessness, are identifi� d W1th the phonemes /b/ • /d/ • /g/ , pho netically voiced in oth. er positions . , it wil l be _ clear , tha t voi_ c·ing is not tne . tinctive feature which dis differenti. ates the pairs /p/ ~ /b/' It/ ~ /ct/• . /k/ �- /g/ • That is whe'"'. t we can. observe in Pcrrcian /Lazar - d Fr72 - .· 14"/ · It is con. f.irmed by orthogr aphy. .Na!Jlely • when· wri· tten {b} and not { pJ-p-eneraJ.lyfol _ _ lo;•:s {'slf, c&o, t jalova had at her disposal.
/c/,
/j/
/c/.
38
3.8.1.1. Drawing a conclusion from the data on distr1bution of tense and lax plosives /and affricates/, we may assume that the opposition functions chiefly in the initial position, between vowels and in the fi nal position, whereas in the ,clusters containl'ng spirant, the opposition under discussion is neutralized lax plosives being, as it seems, the only . 'J ones to appear in such position 2 • This is tested by etymologically different clusters. According to the data supplied by Lazard /1972: 147/, the Persians do not distinguish intervocalic clusters of words such as: � "to bind" and gasdan /from Arab�/ "purposely". or� /from . Arab./ "soldier", and mesgar "copper-smith", in the first pair we are dealing with the cluster /sd/, in the second - /sg/. Plosives are phone tically voiceless when preceded by /s/, e.g. [basgan], [}s:asgan], but thanks to a weak articulation tension, they are phonologically identic al with the allophones of the phonemes /d/, /g/, which are vo. iced in other positions. When following voiced apirants, lax plosives have voi· ced allophones typical for initial and intervocalic positions, e.g. in worrls: � "above..:mentioned", ezdeh§. "dragon". . The opposition of lax and tense plosives is also neutralized in front of spirant,;3. Also here, the only possible articulation, as it seems, for there is no detailed experimental data, is .lax: phonetically voiceless before voiceless spirant or - . voiced before voiced _spirant. ThllS the words containing an etymological'/d/ preceding /s/, e.g. hili!.§. 11guess 1 conjec turing" are pronounced exactly the way as word.a with etymological cluster /ts/, e.g. atse "sneezing". :Before voiced.·spirants /z/, /zl. lax plosive, as it has already been mentioned, is obviously voiced, e.g. in word� "tool". Neutralization does not occur before /v/. There is a minimal r::air: /advAr/ 11circles" N/atvf1r/ llwhims 111 in which /ct/ is voiced, and .. /t/ fullJ' voiceless, /tending to devoice Y..4 .f ii,i col, langu:'l.ge/. Remaining to be discussed are clusters plosive+ plosive. Most probab ly·, the feature tense lax does not undergo neutralization in them, though the lack of detailed ·experimental data impede the drawi,ig of cor rect conclus_ions. For beside homogeneous clusters like: tense + tense /e.g. maktub "letter 11 /, and lax+ lax /e.g. tadM_r."idea"/, there are mixed· clusters, e.g. tatbi!l 11compadng 11 , ebtekA� "initiative". The lat ter undergc, assimilation in regard of voicing /allophones of /b/ are j_n both cases voiceless: [ iat9il:), [ esteM.r]/, but since lax articulation . is retained neutraliza.tion of the opposition /b/ /p/ does not take place. . Such articu�ation is accompanied by the coming near of the vocal cords, which is characteristic of lax consonants /whenever they occur//G-G: 120, 158; Giunasvili 1965: 111/. During the articulation of tense consonants,
~
~
I.ho vocal cords are further apart, which entails aspiration· /Ibid./. . 'l'he sa.rne is true for affricates. In the word /ejteT.Ai'/ "assembly", "1nPirically analysed by G-G /p. 17>4/, r•eriodic ·vibration of the vocal oords does occur at the first stage of the occlusion of /j/, thus •the. . ' . ,., .., noriar,ateness /:1/ •" /c/ is :preserved also 1'or this nosition, which is 1inalogous to the cluster plosive + plosive. 1 rhe occurrence of lax consonant in the nosi ti,on of. neutralization. oonfirms the plausibility of the view that.. tense co'-nsona'°lts /p/, /t/, /){/, /c/ are marked .members of the.opposition. The . fa�t that a ltense consonant is a marked member is also indicated, apart from the above montioned neutralizations, by external sandhi. Namely, after the final tense /consonant/, the initial lax /co11sonant/ of the following word is oi ther _fully voiceless or only partly voiced, e.g. � d�n jfacquJre . . .· knowledge". is rendered in the recording of poetic text I made, as [ebrhga.t , .· with a voiceless lax [g). As far as the quantitative data is .concerned, which might throw some light on the nature of the investigated opposition /Kra;msky 1962: 169/, �,e find that, in NP, the picture is blurred by the mixed character 6.f the vocabulary. Namely, Arabisms constitut:!ng a large part of the voca- · bulary lack the consonants /p/, /g/, /c/, which cons_iderablychanges the quantitative pro]Jortfons /cf. Giunasvili 1965: 139, Ftn� 1/ 1 ar.d renders it impossl ble to draw any us,eful conclusions ,. 3.8.2. Snirants
Considering the occurrence of tension as a distinctive feature dif••-' ferentiat.ing plosives and affricates, it.may be expected that this fea.:. �u.r-e i,3 a:.:>o 6fr1.�/J.. o;y~elloquiai speech, but it gradually spreads in·. the.·· pronunciation regarded as literary. Quot_ed here ·are such words as: /vaqt/ "time" or /raqs/ 11 dance 11 , where the use of the plosive allophone [�] happens only in a particularly careful Ara ble-like style,. cf� the example /eqtes!s/ "riots" /0-G: 155_/. G. Lazard /1957: 22/writes about it: "la postvelaire 9., devant consonne sourde, s > as.sourdit, mais, dan:s la prononciation soign&e� elle est, au moins par.;. tiel1ement;. occlusive; elle reste ainsi distincte de la spirante sourde " !• In an i.nformal conversational style, fricative pronunciation prevails, :but it does not always mean; a� one might. expect, the voiceless spiralit . [·;rJ, ;11hich differs from /xi by its mel.low articulation. A strong vibra- · tion characteristic,f'or /x/ adds tallem ketib-j_ m1dehad "the pupil is giving the teacher a book". How ever• if two stressed syllables. were to occur side by side, the stress of predicate may be easily shifted, for euphonic reasons 9 to the last syllable: S.5.gtra be mo,'allem· javAb mideh�d /stress in midehad is 11eaker since it is the verbal part of the compound verb javib da.dan "to answer il/ "' "the pupil is answering the teacher"• Then, - only the intonation will signal the end of the sentence: a falling intonation _fo:r affirma.tive9 a rising one for interrogative. The case would be different for a predicate without pe-rsonal endings� If no prefix or negativ� _particle appear here,_ the stress falls on the. last syllable, as it happens in the.3 p. sing. of the pra.eterit. e.g. Sagerd a.mad i1 the pupil came"� For a compound verb, there - would 'oe weaker .. . . ..., ,, ,, "" . - .: � secondary stress, e.g. S§.gerd dars�ra. ylid gereft ".the pupil learned his lesson". ' Within a syntagme preceding a; 'predicate. the dynamic stress indicates the end o:f a iexical unit /a :-11ord/,- separating it from other possible grammatical elements /which are not word-forming morphemes on the lexic- al level/. Encliti cs are following: 1 / /in/definiteness suffix -i. 2/ izafet /exponent of determination/ - -/y/e, 3/ personal suffixes, 4/ particle -rll /expone�t of direct object/• 5/ conjunctions -o, -ham "a.�d" /
57 6/ short forms of CO]lllla, 7/ archaic vocative particle -A. Stressed are /except all word-forming suffixes/• the following: . fi.Xes of nouns -M., lln, _ 1 / p1u:ral suf 2/ CO'":parative suffixes of adjectives and adverbs, and superlative suffixes occurring after adjectives, }/ numeral suffixes, e.g. ordinal number suffixes -om // -om-in. Thus the place 'of stress differentiates certain phonemic sequences, .. which are lexical ·· .. units /these are oxytones/ from junctions of words · -with various grammatical morphemes, e�g Q (ket!bf "bookish" lketl!.b-1 11 a book" . fket!be �inscription" _ __ . _ \_ketAb-e /-rub/ . 'fthe/ good book" fket,bat "writing" · \ketl!.b-�t "your book" Pa,:r-ticles /e.g. bale "yes", guy4"probably"/, conjunctions /v�li. a'.mmd "but"/ and some adverbs sQmetimes- functioning as particles /e.g. albatte "of course"/ deviate from oxytone stress in favour of initial one. A noun used vocatively, e.g. /ey/ p��ar "father!" is also stressed initially. A coriimon feature of these deviations is a rather loose connection of these forms with syntagmes forming a sentence. A.stress which is different from normal. furtli�·r emphasizes this separatenes·s /usually also accompanied·. by . . . a pause/. In the above description, the role of se�ondary stresses .connected with . the occurrence of tense and.lax vowels.witbin:a syllable has been disre-· . · · · · ·garded. I lay aside the problems of intonation /cf. Towhidi; Jazayery, Paper: 32-8, '30-60; Hodge: 357/� since the la'ck of hisforical data d 1es not al. low_us ·to·draw verifiable conclusions for diachrony/ and that is the main. subject of this pa.per.
Beside the proni.:n . ciation rc1carded as lJ terary • described. ln: the .abo,re sectfons. there exists collo:::_uial PX'.Onunciation based on Teheran dialect.a It differs cons:i.derably from the literary one� not ·only in the r.ih�notac;.;;-: ri.f firi-· tic ns-pect /ni.mnli:fj. cation of consonant clusters� . disappearance . . al consonants, assimilations, �etatheses, etc.,cf. ?ejsii,;�v: •fi.-37(,' but, · more im-oortan_t, in the very 1n'lentory of phonemes. It applies' chiefly· toC· colloqujal vocalism which. contrary.to literary 1anguage, l:>egins to :_ cteristic for colloquial system, which chron?logically stems from. the: system represented . by the literary •language used simultaneously over "t�me. - . _ -_ _. ·· · It is confirmed by the data on the remaining vcwe113: · _ also spor:adically /a.he/ /eM/. /flM./, /ah�/. lit. coll. /A:/_ /Ma/,', eh coll. /e:/ - lit. /ey/, /e?/. / /coll. /o:/ - lit. /ow/, /oh/, see above, p. 57. Thus we can speak of the rise of new long vowels resultj,:ng from the · ·-. -__ _ · ·. contraction: of various. vowel_-glide combinati,ons. The situation is different, however, in case of the_ new opposition of__ length which contrasts colloquial /i/..., /1 :/, /u./ /u:/. A chronologic, al seouence may also be seen here: li�. /older/ /e/-coll. /more 'I'.ecent/ /i/ lit. /older/ /o/� coll. /more recent//u/ with a ho m.c and this development may be accounted for by the assim1lation , _ an a1�� of geneous vowel of the following syllable, or by thE! influence olar er palatal consonant exerted to narrow the articulation /e/-/i/�; Such solµtion is probable, but it is not the only one possible /it. going to be discussed on pp�95-6/. For the :picture is made complicaten by words like .coll,. nid. 11a look" /Vahidyan: -12/"' lit. negah in which it is difficult to fi.nd justification for the possible development /e/
O:Jt
and
~
i�i.
-/i/.
To state the secondariness of colloquial Persian vooalimtt in re'.!:ati�x, to the li teraJ"y om� is easy• though not banal,•- since not necessarily all elements of·c()lloquial language must be new compared- to the usuallyc�n-, servat.tve literary lar,gu.age. So it han:pens iri fa_ct, ·but certai1: arc�s•:· m,'ly- alsu appear in cqll. lanimage as they do. in any dialect: In col.L• .,, Persian based on Teheran dialect, numerous archalsms occulT1ll,!r, in morp... logy and syntax are known /e.g. the case of personal suffixes/.
61
60 Also as-regards distribution, colloquial vocali�m illustrates a more recent state than literary language. 'f'his is pointed to by numerous dif ferences which can be explained by assimilation or dissimilation /Lazard '1957: 14-8/, and also the characteristic correspondence: lit. An, Am coll. un 0 um,_ e.g. lit. An. "that", �- "house", � "naTJ1e", � ''/he/ came" - coll. :!:!Jl, xune, num, � /Vahidyfln: 12/. Yet, there exist in stances with /un/ in the literary language, e.g. xun "blood". The col loquial: uniformity, where there_ is only /un/ • but never /nn/ /except the most recent borrowings from 1iterary_ language/ in the native vocabulary /including the old Arabisms like� = lit. tamfun "whole; end", corre spond to lit. fom with /un/ or /An/. Also cf. coll. birun, })fl.run · and lit.� "outside", but bll.ra:n "rain". It indicates the secondary na ture of colloquial pronunciation. The labialization /A/:-�--/u/ before la bial /m/ has spread due to the ac.oustic relationship of nasals. · - Literacy borrowings in colloquial language r,etain /f1n/, /Am/. e.g. d§.nesga.h "university" /coll. and lit. f'orm, · c:r; coll. miduna.m i,r kn�w", where the same root J!!!:!- which occurs in word d�nesE:iih appears as dun-/, e :ter.§.m /lit. ehterf!m "reverence" from Arabic/. Thanks to the borrowinp;s • the opposition /IJ./ /u/ before /n/, /m/ suppressed in colloquial language in the past, is now being partly restored* Among the phonotactic changes in colloquial vocalism in relation to the literary language, worth menti_oning is the spreading of the distri bution of the opposition /a/~ /e/ to the final position. As it is known /see above, p.16/, in the literary language, ,this opposition does not function in the final position, and the only possible vowel there is /e/. In colloquial speech, in turn. thanks to the disa.ppearanc� of final /h/, Ii'/, also /a/ may· now o·ccur in the· final position, e.g. _v�zda� mostala, morabba /lit. yl).zdah "eleven", mostalah "idiom", morabba? "square" /cf. Lazard 1957: 25/. Also /o/ rarely occurring in literary language in the final -position may appear more frequently in this -position due to the above-meri��oned disappearance o:f /h/ • Ii'/, e.g. tava;llo, tavaaao versus lit. ta,rajjoh "attention", !avagao? "expectation".
~
6.2�2. Consonantism
Besides the alz:eagraphy advis'�,ca�tion in the drawing of 'con�lusions .. No. . . . essential differences can be stated when comparing the orthograpiy- used: in the 13th c. with that of today. It wo�+d .suggest a very slow rate .of pronunc:iation changes which, · after all, might be expected' over a i;iJlje· . span·of almost •eight c.enturies that stand between our time and the be-···· _, . ginning of the classical PE:riod. The above implies that the mere analysis of' the old .texts wiil add iittle to the reconstruction of' the 13th-c. pronunciation-. On the _other
1ahL?
.
"l
66 har..d , the basic data is provided by the material of modern .language which nay be called the internal material. �."aat I mean is the data on the fluc tuz.tfo..-, 3 occur-ring today /internal reconstruction/ , and redundant features of be :.ndividual phonemes /or their clusters/• i•rhi ch sometimes make it poss ible to draw diachronic conclusions . The analysis of orthography, the histori cal character of "'.hich provides certain information about old pronunciation, also belongs to the "internal" data . Between "int·ernal" and "ex ternal" data , • lies the material of cclloquial la:1.:;1..a.ge., basi cally staying "outside" literary larigiiage , which is. the su·b ject of our study . Still, the two languages coexis t both in time and in plac� . · · The d.a.ta ·o:f other diale ct , · related to the Iran Persian constitute s • as ·a whole , "external " la."l.gUage mate:rial /comparative· method/. In this regard. the most thoroughly investigated and ver--,1 useful point of re:fer-, ence is the Ai'ghan Persian, .so- called Dari . · · St_ill more "external!' though . by np means the least important material is �ade up of the :foreign .trans cri ptions a.>1d ·borrowings . Here , we have :primarily the· w·oras , or whole l'ersian texts noted - do1-m. 1'lith Jfebrew , Latin, and Armenian1etters • .These notations, due to separate systems of s igns , per�it the ·confrontation With the native notation done by means of .the · letters of Arabic · alphabet .. . . � �·-� 1 oan materi&l compr:i,ses both :foreign words .borrowed by Persian language ; a.nd Persian ones - occurring in other la:nguages. Tbe former · include chiefly the exceedingly numerous words o f .Arabi c origin. On the other hand ;, foreign lal".guages , which ado pted a . number of Persian l exical elements, a!'e :first of all Turkisl: /wi th Azeri/ .and Urdu · iEi:ndi/ and also . Arabic, Armenian /mainly ttialects , · not the literary languap:e/, Georgian,. and related to .Persian Iranian. languages : Kurdish and }'ashto. The analysis of the borrowi ngs · requires considering the phonoloaical d lop ent of the individual languages , and thus , it is o ccas ionally · :? � 1",---naerei: t by the f'act that the conditions in which the Jndivi.dual borrow ings were- acquired are not known to :us . It i s d i fficult to determine in w:t-,e,t .iarticular form a.. foreign word was assimilated,; whether its adoption hapnen.ned i.ffl!!lediately, or whether it took place gradually , etc. Dest,it:e t!lese difficulti es , bOrrowings are sometimes a. very valuable material for · diachroni c studies . · · . .rn addi. tio�, t:here . wi· 11 re:ma1n-· use:ful · · · dcj;! cri l)tive data concerning the · . . . ... . . . . · old pronunciation: preserved in works by Irania.'Yl.a thems elves , as •;: eli as . · · ~ t'Ce:r , of· course .. . ..- · b" des. criptions mad ., '•ur .... opeans. · ·Tlne '-1.a. , are. relatively . . . - . rec.ent /dati. .ng back. to . the 1 7t.h. c. at most/· ana' , as 1·t ma�r be exnec ted , . . tney rarely ·:fulfil our requirements for preciseness .
67 2 . DIACHRONIC COHCLUSIONS EMERGL?G FROM FLUCTUATIONS OCCUR.iUNG VOCALISM OF MODERN LITERARY LANGUAGE 2 . 1 . Fluctuatl on /a/ //
rn ·
/el_
The analysi s of the present ed fluctua tions /a/ // /e/ /pp. 1 5-i 6/ show.a that in most case s /a/ j_s older. Na.'Del7 , in fluctuations like rasidan // .. .residan , panjare // nanjere , the presence o.f the vowel · /e/ may be ex• lier: Ara'bic /1/, /e/. ng followi plained by the assimilat ing e:ffe�t .of the •· rowings also impiy that numerou s /e/ of partiei-ple _forms and of verbal .g. e .. nouns /p. 1 5 / occur in the place of original Arabi� : from stem harekat t, mo:dlje?a , e mok1lte1> ke;, motesak NP Arab. rautasakkir; mukataba/t/, murajaj,c lat1;_/ · . . . · :· .. . gua;;e. · · . _: e d 1:ly the ·old la.tin The cld :pronunciation. of l /mod. 2,./ is oonfil'Ili : . fra.nscri ntions • For instanc e . in Lat�-1'ers:iana'l'urkish ·. /fJ-�n/ diction-(': ary. Cod;x Crun�nicus fro� the iTrat half of the t4th c •. there �:re ni:.\tat1:i:i�a dil "heart". gil "mud II" xist fluglyn c�i'respondi!ig to modern :fo:rmsf dei� gel, . zest ]Bodrogligeti 197.1 : 4 i /. ·.
~
/a/. rd Shif. t forwa the > o and lowered articulation /i/ > e, 11
/.u/.
,/
._
/a/
·.e:
.
Dictionary of 1'ersia!. 4 .4. Transcri 11tion U1:3ed in the Medieval Armenian Arabisms /APA, for short/ ising s of the initial part of that dictionary compr 11 � Two manuscript in the kept y "te / are 204 entries /the lett.ers: 11 alef•i, "be", and partl the .15th c. and the other Yer·evan Hatenadaran. One of them dates back to place of its creation are to 1589. Neither the author, nor the time and
79
78 known; according to 13. C'ugaszyan /u. 150/ it was probabl:t 13th - 14th· cc� The dictionary gives Arabic words /in their·Persian pronunciation/ and their native Pernian equivalents, all in the Armenian alphabet which u. Furthermore the dictionary i, o discrimin�tes between vowels e provides equivalents in clas.sical Armenian • . This source is. particularly valua:ble for its considerable consistency of the transcription used. The latter is . best characterized by the nota tion o:f the vowel corresponding to cl. WP- /a/, by the letter denoting Armenian r�duced vowel (a) alternating, in Armenian, with /1/, /u/, but occasionally pronounced quite crpen•. Examples: ravt'an ,r: cl. NP raftan "to go". asxar.;, cl.NP aspr ( Arab. ?a�F "smaller" /C'ugaszyan: 159/;, Since in Armenian language, "in the initial cluster, this vowel has a of anaptyxis, and, then is not writte� at all, in Persian /Arab./ character . words, illthe dictionary, itwas freq-ue11tlyomitted in the onset sequence CaC, e.g. brak'at' ·/,.*barak'at/ as opposeQ.to Persian and Arabic barakat "blessing".· Also in .the. final position, /a/ was written as (a), e.g. fayi_da .. cl.N_P and Arab. fa?ida 11use'' /C'ugaszyan: 152, 164/. Only in .sparse cases, the Armenian authors rendered the cl. NP /a/ by (a) /cf. data ,of later sources, see below, p. 81 /, e.g •. in the position: 1/ before glottalstop / 0: jodau·enda = cl� xuda.wand-a "oh, my God! n,. boued= cl. bu.wad "is", cf.· Bodrogligeti 196.1.: 267, 269, 270. Long /e/, /o/ were on .the verge of coalescence, or completely - identj_ fied with /1/. /u/ respectively. Except .for an isolated case of the ret ainlng of cl. /e/ in rejten .._ cl• .rextan "to. pour", we still have (i), .e.g.' prefix /e/ini = cl. hame, ziba = cl. zeba "beautiful", deruiX = ci. darwes "dervish". In plac� of cl. /o/, there is continu.ously (u), e.g. duruj ;,. cl• duror "lie II t frv.;jt=cl. furoxt 1I/he/ sold II• Cf •. Bodrogligeti 1961: 267-9, 273-4� In LPK, diphthongs· we:i:-e represented in the following .form; cl. /ai/ · ei, cl. /au/- eu /au/, ,e.g. vei cl. wai "he, she", peiga.'nber = cl. paifampar -"prophet", seuken = cl._ saugand "oath", meut = cl. rrtaut -/from Arabic/ "death''• but in gloss : cetaur = cl. ci_;taur 11how",. Cf. Bodrogli geti 1961: 269, 270; 272�4.
/u/
-=
In the 17th c a thex·e a::�"'G2..red the firat European gra."nmars of Persian language. The ear-1.iest one known. ·.-m:o written by LudoviCUSf de DieU: /lili£1menta Lini::uae Pe:,:-Gi utor uassim in Turc.icis, ut �- in Arabicis &Persicis vodbus, ad ex,pri�. menda� nota..rn vocalem' ante 6} _]'..lost literas duriores, qua-e sunt J'J' _ lo \o,ubi debet effe5r1 clare,, £ere ut,_?! I!a�icum aut Ga�licumf J seu, ut } l'olonicum & Eun,garlcum, . sed est natura sua oreve ad .difieren'."",' tiam seouenti.s a.11 /Menirfslci: 3/. · 1 · _:::>Q7-, /_.· t?-re:f___..;_: Here are' Heninski' s Persian examp1es _/pp. •q· , . -_,z • -1· 9rJ;,: cl. taraf / a ;
represented ·oy APA, probably COll".ing
f':rom the nth c. though the man:uscript rs much later. /C ugaszyan: 150/. c
Na."llely, granted t.,"1e :faet that, i.n. the development of Persian vocalism towards the abando�nt o:f the length �ppositlo�. the Ttlrkfo adstrat ex erted its in.fluence, we may speculate that [a]/still mid in :r-espect to
short /a/ retained its probable open and mid pronunciation presumably on the whole Persian 1anp;uage area. Cf. the quotation from F. Heninski9s graJ;Jmar, above, p.81 , or the transcriptions of {a] side by side with the usual { e} in LPK - p. 80 /.
On the other hand, in certain variants Of Persian� the. art.icul�.tion not only of the short /a/, but also of the long /a//!/ may have been shifted forward. ,,uch development is traditionally described by the Ara� ic term "imala". There are numerous examples of it in CC, where, :e.,g.� ;,
.classical words: andam "member"• astI "peacett, s ilah /i'rom A.rab./ "wea�? ••. on", rjkab/from Arab./ "stir-up" are rendered as and:en, hesti; se1ee� r.icheb /Bodrogligeti l971: 46/. Also in the early 15th.;.c •. transcriptio.tis w ;f ATl, the cl. ;orm barx ast "he rose" was twice written as: bane.st,--/
1"6riil
< /P'a:p .:1zyan: 220/. Also cf. in the Zoroastr.ian di�leet /Gabri/, the correspcndir>.g to cl. lca.h "straw" /Geiger: 384/. Here, we could per- . ceive certain possible fluctuations a// a occurring occasionally i:ri the classical language, e.g. rah// _rah "way", but it seems to be contradiif
ill
ed by Arabisms in which no such n11otuat_ions are observed;, ..·.. in some variant of Persian appears to;.. The front pronunciation of be testified to also by Turkish borrowings like den /Siirt dialect/, cf� cl. YP dan, mod.· UP d�ne "seed" or yeren "friend" /dialects of Urfa and'·
/a/
Gaziantep/, cf. cl. }ff varan - /Tietze: 141, 157/� In the courseo:fi:irne, however, 1rnch pronunciation _gave way to the ,back [a]. • After the first modification. probably in the 14th c. the RP vowel system acquired the following form:
the horizontal position of the tongue/ was. as higher than /a/, identi: :fied by Turks with their native front /Ii/� higher than /as may be seen from its fiL.-ther de.,;.el.opment, in th� Ti.n:-1'.'Ssh-Osmanli,.· tm·1a.rds �I. belong
/el
ia/
ing - wi thi.n the '!'urkic vocalic harmony .,. to the group of back vowels�
Thus the impulse to shift the articulation of 'a/< a/ towards the front; .rart of the oral ca:vi'ty ig·:possible •. Cf. Turkish and ,:'lzerb. borrowJ�s
1tS and also - cif;l far as .Arabisms are conce:;:-ned - side by side 1d th the so _ called emphatics; glottal stop origi11ating from the splrant 11 ',ayn" /, the
/a/> a,
in
Tu/ Further development and fluctuations in the transcriptions Hke: classical II
II II It
" II
/a/
Ii/
/u/
i:e1·
/o/
/ai/
/au/
transcription a/e " i/e ri u/o II
11
11
"
e/i o/u
ai/ei/f'J au/ev/ /o/u cf. table�·
w_:;.·1-l become core J.uqid, i.f we ackno1,1ledg.e the_�--develO.pine stylistic di-f- ferentiation of the standard l'er�,i.an language /different fl•om old dialect,
95
94 which did not stem from the koine, more or less unifc.,rm in the 13th c./ Namely, in the official style, .fostered by the literary tTadition, preserved was, until the 19th c. a status little different from the pre sented postclassical one. The only difference visible after the 14th c. was probably the further narrowing o:f [a] towards the open[€.] transcrib . ed as {�J. It was the continuation of 'the qualitative differentiation in fa/• and the filling-in of the ca� vide the classical opposition of •t. In the eastern parts /Khorasan, Afghanistan/, the postclassical status /.front a/ is retained in all .positions to the present time. Beside the conservative of.ficiai literary language, in which vocalism was based on the length opposition, mani:festing itself' particularly strong-
ence of long /e/, /o/. In other ��ords, classical short /a/, .- /I/, had a greater possibility for the differentiation o.f their allophones ·which. in consequence, might have led to qualitative changes. And indeed9 .not ..· only /a/ began to be realized as. higher and more .front th
NP
Ii/.
/i/,
/s/, /z/, /c/,
/u/
tion of _,;,ere - for /i/: the neighbourhood of /j/ and the position before the syllable with /I/; for /u/: position before the syllable containing It is i�dicated by the retainemnt of the pro nunciation of [1] • [uJ in the modern colloquial language in words like . s"ikam "belly", kili:d ,;key", furu:a "sale" = mod. lit. NP sekam, kelid, foriis /cl, liikam, kilid, furus/ see above, pp. 58.
)u/.
The proposed vocalism .in the post-classical colloquial lane;uage has been reta.ined. in the Afghan Dari, where new /e/, /o/ are still close, as regards articulation, to the old /u/ /Dorofeeva: 13/, and the r.etained long /el, /o/ approximate the continuants of' the classical /!/,
/1/,
/u/:
/a.I
/a/ /e/
diffuse
grave
a +
a
+
/I/
+
tense
u
i
+ +
+
..
/a/,
0
0
+
/u/ >
/o/.
/I/,
..
/a/
/a/, /e/•
/e/, /o/
/u/
are retained especial
/e/, /o/
may become identical with /1/,
/u/.
. , .
There stUl remains the problem of the front, historically short ,,
/a/,
/{;. For the chronology of the transcriptions is. surprising:. in the 14th c, {a} was usual, {e} less so; later, up to the 19th c. inclusively• Je} was usual, {a} less so, s3,nd in the 20th c. in turn, again /a/, and /�/ ·., . only finally. In APA, {a} was consistently uf',ied /see p. 7£? I. but in the
/a/ but
/. Perhaps that {a} signified i
[;J -
for the Armenian authors. of the trarisc'ripticms not to ap-; :
of'
/ii./,· to :/new/ /e/ /old/ /new/ /o/ ,..,/old/ Owing to a consider able differentiation of place of articu:la tion of the pair: /front/ /a/~ /back/ /a/, permitting the employment Of u . niversal, ever distinctive fea. . ture grave,.., acute, the tensenes·s of became redundant. The system is still asyinmetricat and still sho\orn the. tendency to acquire a better .bal ance between the historically short and lax and the histo rically long and tense /I/, We may expect a gradual elimi:11ation of the vowels from the system. Such, indeed, is the course of development the Pers:tan of Iran pursues. Long gradual lY. became identical with /I/, as it can be seen from the transcripti ve and descriptive material presented in.the previou s sections. This proc ess is completed in the 18th c. /see above, p. 82 /. Since then, the sys-
-
Iii/
P• 92 / which, .from the 17th c. on, is to' have led to the. coalescence /a/ and /u/ before /n/, /m/,. there was left ,a Wide dispersion range for historical /'lr/: from the mid /and, in some areas, perhaps still complete:..
+
/e/,
cases,
ly open -
e
/I/>
~
/o/,.,
ply the letter {e} /cf., F·tn. 8/,. which,· in relation to the Armenian lan.;. guage, rendered rather the high ·"/�/ alternating with /i/. . After the probable shift of /a/ towards the back pronunciation /see
.In this system, the vowels /!/, /u/ are redundantly long and ten se, whereas /e/ are redundantly long, employing tenseness as a dis tinctive feature. As the result of the· lowered articulation e, :o·, · the use , of the feature tense "" lax was shifted froni 'the pairs /{/ ;._, /u/ ,,.,,
/o/,
II/'
ly in the minimal pairs /the role of semantic level/!. In the remaining
· such variant of Persian pronunciation in which a was ·front and sufficien�
:+
+
/o/,
mid 'as far as the horizontal tongue positior is concerned. While ATL {a} also rende,ring /a/ may have· been applied in
/ u/
e
Kabul dialect, the oppositions /e/'""'
a vowel higher than
The distinctive features of separate phonemes of the system may be presented in the following way: compact
It is worth adding here that in Dari the pronunciation of /e/,
also approximates /I/, /u/. According to L.N. Dorofeeva /p. 14/, in the
15th-c •. ATL {a}was still in use /see p. 76
/el·
/o/
tern is close to the modern one,; it has six vowels. that is. ·two vowels
fewer than in the classical period.
/ai, Je/, /o/ /o/, /u/.
/e/, /o/
.
ly low [a], through the front fa]• to [e] which initially was not replace� !'U, 2tlso evolvin!". towards "[e), but probably not strar ing for some time beyond: high [�]. in the areas-where this/!/ was still relatively hig:e It was especially ·
by tJ-:e hiotorical
·
·
·
·.
·
·
. :£.·. 11 :•
I and of a lax character like the analogous English vowel /or Polish that the historical /a/, in its tendency to amplify the-quq.!itative dif� may have reached the level of the mid ·teJ. ference in re.lation to
/al
�nd
.so it must have happened, since from the 17th c. until the second hal.f of the 19th c. Europeans wrote that vowel, as a rule /except for certain positions, cf. p.so.:.11, with_the letter {e}. It may, in .fact, have pos'"'. ·
[a:J', or [6], but the mid pronunciation /like quite probable since it is still encountered was vowel this Polish !!;./ of
sibly signified the open
in the new Isfahan dialect.
. That dialect, close to the literary language •. probably derives from l the standard koine in the variant used in Sefev:id state capital, whicl in 16th - 18th cc. \'las Isfahan. In, the mentioned dialect, /el frequently corresponds to mod. lit. /a/. Here is the comparative material according to the monograph by L,F. Smirnova /pp. 11-2/:
99 literary NP
Isfahan dialect
zadan "to hit", panj�h "fifty"
madar ''mother", doxtar "daughter", sar "head"
kas "person", sag "dof.'.", ast
11 is",
az " from"
zeden, penj§.
m�der, doxter, ser
kes, seg, es, ez
The tendency t o the highly narrowed pronunciation of th� historical /a/ was particularly strong in the final position, if in the modern lan guare, generally displaying /a/ in place of the classical /a/, there con;. sistently appears final /-e/. In this position, literary language anci in both of them, there iS [cale]
Isfahan dialect are in accord, e.g.
"hole", [gonde) "big" /Slltirnova: 70, 74/. In many cases, h owever, Is!aha11 dialect displays [a] /except for. the non-i'inal position/. similarly to the literary lan:/!Uage /i.e. histori�.g. perande 11 bird 11 /lit. narande/� galan,:ler lldervish'; /lit� .cal
/a//,
oalandar/, zahll'ietfftoil11 /lit. zabmat/, xalvet llsmaJ.l f.00m n /lit. xalvat/* ma.�varet "council" /11t. manvarat/ /Smirnov'a: 12/. It� is· dHficul t to establish any regularity, but at least some of these examples confirm .
the o ld notaticins aignal1ing the. retainment of the pronunci�tion closer to [aJin the neighbouxhood of uvulars and Arabic emphatic consonants /cf. abo ve, p. 81 /.
Other cases of the occurrence of /a/ in Isfahan could be accounted for by dissimilation or assimilation process� e.g.. per;§:nde, qali!nder. or. by the influence of literary language. tfost probably, fa] an4 [el are,
in the Isfahan dialect, aUophones of one pho�emi, as they were in the Persian of Sefevid period, but. due tp strong infl�ence iitexary lan guage, in which the opp osition /a/( ci. /a/.,-.,/e/-ent, ed materlal, the tendency to bro aden the disperiion range o:f the cl. /a./, was leading . both ph onemes towards their.identiftcation. So it partJy happened in the Isfahan dilllect, where there are,·· e.g� ��/Smirnova: 12/ ( cl. pisar "son" with the identical continuants of' two di:ffereht :_)!10. nemes. Other examples from this dialect: peder cl. j;J.i d,a:" ".fatb.er", · � ( cl, �·"bear",- i:,roru ( cl. im:?:"oz "today" /Smiinoya: 11 /. Eut the izafet.
/i/
.t!