Indo-Iranian Frontier Languages; Vol. I; Parachi and Ormuri

INTRODUCTION.

1. Parachi is first mentioned by Babur in his Memoirs 1 : "Many
other of the villages and districts (of Kabul) are occupied by Pa-
shais, Parachis, Tajiks, Berekis and Afghans." "There are eleven
or twelve different languages spoken in Kabul: Arabic, Persian,
Turki, Moghuli, Hindi, Afghani, Pashai, Parachi, Greberi, Bereki
and Lamghani."

Later in the 16th century the Turkish admiral Sidi Ali mentions
the Paras! tribe as living in the vicinity of Parwan, which is close
to Shutul 2 .

Elphinstone 3 tells us that " the Puraunchehs, another class of
Hindkees, seem to have been considered as a separate people in
Bauber's time: they are now only remarkable for being great car-
riers, and conductors of caravans."

And according to Masson 4 : "the Perancheh is spoken by a few
families of the same name, resident in or near Panjshir". Cf. p.
221: "The Peranchehs, besides the few families at Panjshir who
preserve their ancient dialect, are found over a large tract of country,
and it is well known that their conversion to Islam is of compara-
tively recent date. At the city of Kabal some of the more eminent
merchants are Peranchehs. They occupy a considerable village in
Kameh; they also inhabit Makkad on the Indus, and again are

1 Transl. Leyden and Erskine, ed. King, I, 224 f.

' Marquardt, Eranshahr, 287.

8 An Account of the Kingdom of Caubnl, new ed., I, 418.

* Narrative of Various Journeys in Balochistan etc., I, 219.

found at Atak, and the towns between it and the Jelam river. In
all situations they are a commercial people."

Till recently this was all that was known about Parachi, and,
in a note on Babur's Memoirs (1. c), Sir George Grierson explained
the name as "PrachI": "Eastern", denoting a language of Eastern
India, brought to Kabul "trough the Purbyas (of Oudh) who were
and are great travellers." 1

2. When staying in Kabul in 1924, on a linguistic mission from
the Norwegian " Institute for Comparative Research in Human Cul-
ture", I made, however, constant inquiries about Parachi, relying
on the authority of Babur and Masson, who described it as a local
dialect.

Towards the end of July a Pashai, whom I had brought to Ka-
bul from his native village of Kohnadeh (Satha) near Gulbahar, told
me that a friend of his, living in the valley of Shutul, spoke a
curious language called Pardii. He could also repeat a few senten-
ces in this language: ita'yun xureman, ^au tereman "I eat bread and
drink water"; Mnjai "from where art thou"; l usti l pareman "rise
and let us go", etc. Another Pashai, hailing from Tagau, confirmed
this information by the statement that in the dialect of Pachaghan
in Nijrau, about which he had already roused my curiosity, "bread"
was called nayiln.

This information made it clear that Parachi was an unknown
Iranian language, and ter- "to drink" reminded me of Orm. tr-.
When I sent the Kohnadeh Pashai home, I promised him a good
bakhshish, if !i«' would bring his Parachi friend to Kabul. He did
not, however, appear, and I had to send another messenger to Shu-
tul. This man duly returned with a grey-beard, who professed to
speak Parachi. But it at once appeared that his language was the
north-western dialect of Pashai, which I had recently had occasion
to study. He, of course, had imagined that Pashai and Parachi

1 Grierson also explains Babur's "Geberi" as Gabri, the dialect of the Parsees
of Yezd and Kinuau. But it seems more probable that Geberi, too, denotes an
ancient local language of Kabulistan, cf. Gawar Bati.

Introduction 5

would have the same market value with a "majnun" Firengi, and
I shall never forget the expression of his face, when I started read-
ing out to him the parable of the Prodigal Son in his own dialect.

Fortunately, the malik of Shutul, together with the other maliks
of Kohistan, had just that day come to Kabul to discuss conscrip-
tion with the government. I wrote down, in Persian letters, the
Parachi sentences which I had heard, together with a Persian trans-
lation, and sent my messenger to the malik, asking him if any such
language were spoken in Shutul. The malik confirmed this, and
mentioned the name of some speakers of Parachi.

The messenger started again, and after some nine days returned
with a genuine Parachi shepherd. When I asked him to count,
the very first numerals, zu, di, si etc., proved that Parachi was
an Iranian language, occupying a rather independent position.

3. The name of the shepherd was Mahmad GhanI (abbreviated:
M), and his home was Euidarra above Shutul, at the foot of the
Arzu Pass. He said that he was about fifty years old (being one
year old "at the time of the great earthquake"). He was very dull,
and exceedingly trying to work with; but in the course of the week
which he agreed to stay with me, I managed to get an idea of the
main features of his mother-tongue. When the Parable had been
translated into Parachi, he felt very proud and insisted upon my
writing it down in Persian letters, that he might show the inulla
of his village that he, Mahmad Grhani, had produced this piece of
Parachi literature.

In September I got into touch with two Parachi recruits, staying
in the cantonment of Sherpur. They were Ghulam Maheuddln (Gr)
and Tabakkal Shah (T), both probably about thirty years old. The
latter came from Euidarra, and was the poet laureate of the Para-
chis. The home of the other was Deh-i Kalan in Shutul (or Chutul,
which he asserted to be the correct form). He was comparatively
fair-haired. He seemed to be really interested in my linguistic work,
and was intelligent and always cheerful and reliable.

He tried with some success to apply his knowledge of Persian

6 Parachi

writing to his mother-tongue (at my suggestion he introduced some
differentiated letters from the Pashto alphabet). Of an eve-
ning in his tent he wrote down the songs which his poet friend
recited to him. The next day, when they came to see me, one of
them recited the poem, and I got the manuscript, by the help of
which I could afterwards to some extent check my own phonetical
transcription. They said that there existed in Shutul one Parachi
book, containing legends about Ali; but it was impossible for me
to get hold of it.

I worked with G and T for about one hour daily for over a
month.

All the three persons mentioned above spoke very nearly the
same dialect. l But in Kabul I also, for a short time, got hold of
a soldier from Ghujulan in Darra-i Ghush {Ddrri^us) in Nijrau. He
would not disclose his name ; but he is indicated by the letter D.
He said that his dialect was similar to that used in Pachaghan,
and this was confirmed by Pashais from the neighbouring village
of Ishpi.

In Peshawar, at the beginning of November, I had occasion to
meet Mahmaddin from Pachaghan (P), and to write down a short
vocabulary of his dialect.

The Nijrau and Pachaghan dialect has a pure a instead of Shu-
tuli a (v. 16), fi instead of o, and, in some cases, postvocalic c in-
stead of 6. The vocabulary differs considerably from that of Shutul,
and contains a still greater number of Pashai loan-words.

An account of the linguistic position and chief characteristics of
Parachi is given in my " Report on a Linguistic Mission to Afgha-
nistan", pp. 18 ff., and I have provided a short vocabulary for the
Addenda in Vol. I of the LSI.

' M's a's were less rounded than G-'s and T's, and he generally used the 1st
pi. of verbs, instead of the 1st sg. In lw.'s he usually had -a for Prs. -a. T had
a stronger tendency than G to broaden stressed u into 6 or even a; and he did
not employ the oblique case of the pronouns as a subject with past tenses of
transitive verbs. The vocabulary of all three varied slightly.

Introduction 7

4. In 1926 Panjshir was visited by the Russian scholar Professor
M. S. Andreev, who has published a preliminary report on his jour-
ney (in Russian) '. He collected some materials about Parachi
(op. cit. p. VI), but has not yet published any account of the lan-
guage.

According to him (p. 3, note) Parachi is spoken in the village
of Kuroba (Koraba of the survey maps, in Panjshir, about five
miles above Gulbahar), in Nijrau and in three of the eight villages
of the Shutul valley, viz. De-i Kalon (100 houses), Mo'ra (20—30
houses) and Andosot (40 houses). The other villages in Shutul speak
Persian. According to the traditions of the Shutulis their ancestor
Shutul migrated from Nijrau together with Kuroba and Solang
(: Salang). Solang went to the valley above Parwan, which bears
his name; but his descendants speak Persian. The dialect of Ku-
roba, Shutul and Nijrau is said to be the same (hx a3tiK o6mea).
They call themselves Tajiks; those who have preserved their own
language occasionally also Parachis.

5. My informants differed considerably regarding the number of
persons speaking Parachi. According to M they were in all 100
persons; G said that there were 2 — 300 houses in Shutul, 100 of
which were inhabited by Parachis, while the poet T held that bis
tribe occupied 400 houses in Shutul and 600 in Nijrau. Finally,
D told me that there were 100 Parachi houses in Nijrau.

Both M and G agreed that the Parachis came from Nijrau a few
generations ago (or: "6 — 700 years ago"! G), and that the two
branches of the tribe still intermarried and were one people (yah
Mum). According to G, when they saw the uninhabited valley of
Shutul they exclaimed: chu, tul "go and look (bum, sail hi)". He
said that a couple of generations ago Parachi was spoken in Panjshir,
in the villages of Feraj, Zamonkor and Dostomkhel.

It may also be noted that there is a village called Parachi in the
Paghman district, some fifteen miles west of Kabul.

1 Ou the Ethnology of Afghanistan. The Valley of the Panjshir (The Society
for Studies of Tajikistan and Iranian Peoples beyond its Borders). Tashkend 1927.

g Parachi

6. It seems doubtful whether these Iranian Parachi-speaking
peasants and hill-men are connected with the commercial commu-
nities mentioned by Blphinstone and Masson. Regarding this latter
group, cf. Pashto pard(n)ca "a mercer, draper, cloth-merchant",
Waziri paraca, paroca "a Hindu convert to Muhammedanism, n. of
a caste", Panjabi paraica "a caste of Muhammedans engaged ordi-
narily in peddling". At any rate the hill-Parachis did not know
about any such connexion.

It may be that the name is the same (cf. Skr. pardci- f. "averted,
outside of, distant", applied in the sense of "western" to the near-
est Iranian neighbours of the Pashais and other Indian tribes?),
and that it originally denoted more than one pre-Pathan Iranian
community of Kabulistan. And, very probably, the language for-
merly extended over a greater territory than at present. It is also,
however, quite possible that we have to do with two different words.

7. In the absence of any historical information regarding the
origin of the Parachis, we must depend on linguistic evidence only,
if we want to determine the position of their language within the
Iranian family.

As indicated in my Report (pp. 28 ff.), Par. agrees in several
respects with the W. Ir. dialects, and not with the eastern ones. One
important point is that Par. and W. Ir. have initial voiced stops (b, d,
g) and the palatal affricate j, unlike the eastern dialects which have
fricatives (|3 (•;;), b, y, e). This seems to be a very ancient distinction
between S. (W. and S. E.) Ir. and N.E. Ir., and I do not think
there is sufficient reason for assuming a general secondary tran-
sition of initial voiced fricatives to voiced stops in S. Ir. 1 Like
W. Ir. Par. treats initial and intervocalic Or in the same way.

With the N.W. Ir. dialects Par. (but also With.) shares the
development of dw>}> (Tedesco 12). The loss of intervocalic -d-,
■t- is also characteristic of Par. and most of the modern N.W. dialects.
We do not find this change in the N.W. Turfan texts; but the

1 The t- of NW. Ir. dialects is probably due to a secondary development of J-
parallel to that which must have taken place in the case of intervocalic 6 >]> z.

Introduction 9

i tendency may be ancient. The change of 9r > $ is found
both in Par. and in the N.W.Ir. dialect of Sangsari; but there
is not necessarily a direct connexion between the development in
the two dialects.

Moreover, we find points of resemblance to the N.W. dialects
in the formation of words and in the vocabulary. Cf. e. g. the
numerals Par. zu ] was "11 ", d(u)wds " 12 " : N.W. Turfan 'evandas (Zaza
zuendd's), duvades* (Ted. 24) p)); Par. (h)o "that": N.W.Turfan
hau (Ted. 22)); Par. nhin-, nhast "to sit down": Samn. -nin- : -nidst;
en- "to bring": Kurd., Zaza an-; dah- "to give" N.W.Turf. dah-
(Ted. 25) p)); Par. pes, pesti "behind, afterwards": Samn. pasti.

It is natural that Par. does not share any phonetical innovations
with S.W.Ir. (Prs.). The numeral sus "30" shows the same for-
mation as Prs. sih; but similar forms are found in E.Ir., too (Ted.
24) p)). Eegarding an: "I": S.Turf. 'an v. 113. The past parti-
ciples in -i (203) resemble the S.Turf. and Prs. in -id (N.W.Turf.
-ad, Ted. 26) ); but a similar type is found in Bal. (borr. from Prs.)
and in Minj. 1

8. The points of resemblance between Par. and Orm. have been
mentioned in my Eeport (p. 26 f.). As regards the phonology the
most characteristic point is the change of w > y(w) 2 . The development
of St into Par. 8, Orm. |r is not exactly parallel in the two lan-
a^iages, and the loss of intervocalic -t-, -s-, and the development of
iw- > b-, nt, nd > n is found in other dialects, too.

A number of words are more or less peculiar to these two dia-
lects (cf. Eeport p. 27):

1 If the Minj. forms in -iy-, eg. rddly- = Prs. xandid. Par. khani "laughed" is
)orr. from Prs., the borrowing must have taken place early.

a Traces of this development are found in other dialects, too. Cf. Wkh. (lw.s?)
(us "male" (* wfsan- ?), -\6r "wool" (A v. vanna-), Prs. (dial.) yeSa "forest" (Prs.
>e8a, Phi. weSak), Tsakhnr (Caucasus) yarg "lamb" (<Ir.? Cf. Prs. barm, Kurd.
vark, Shgh. warg etc.) It is not, however, possible to determine the geographical
listribution of this transition outside Par. and Orm. Possibly Greek YNAO«I>EPPOY,
[haroshthi Gudapharnasa on the coins of Gondophares point to a pronunciation
pith fricative y(u).

10 Parachi

Par. an ] darf- "to sew": Orm. tmddrsw-.

» gap-dr "fireplace": » gap "stone".

» ^mindut "apricot": » matat.

» mat "killed": » mat "withered".

» ni-: nar}yo "to go out": » nis-: nay oh.

» ner-: not "to take out": » naw a r-: naw a lak.

» plndrdt- "to sell": » prawak.

» ru "iron": » ro (but Prs. roi etc.

"copper").

» ter-: thor "to drink": » tr-: tatak.

» zdrna "winter": » ] zemak.

» zut "very": » (d)zut "very much" (?).

Some of these words have the appearance of being loan-words;
but other cases of lexical agreement between Par. and Orm. will
be found in the list of E.Ir. words in Par. given below.

On the other hand Par. and Orm. differ on several important
points of phonology and morphology (e. g. preservation of -k- in Orm.).
9. The Par. development of rt, rd > ;• reminds us of Psht., and
str > s is found in several E.Ir. dialects. Apart from these, pos-
sibly late, changes, the Par. phonetic development does not present
any special E. Ir. features. Some of the enclitic pronouns resemble
the forms found in the Pamir dialects (v. 116); but it is in the voca-
bulary that Par. shows the greatest similarity to its E. Ir. neigh-
bours, as will appear from the following list (cf. Rep. p. 31):
d l yun-: d l yust "to dress": Psht. dyusM etc.
] dmar "apple": Psht. mana etc.
dr l zun "millet": Mj. yurzdn, but Prs. arzan.
el stec "star": Shgh. s Herj <Z *starci-.
dstar- "to smear": Mj. aster-.
us~t- "to rise": Mj. wusk'-.
i] §ten, ^stdnuk "kid": Ishk. stunuk etc.
'stdwo "cold": Shgh. seHd etc. "frozen".

[ uzeh-: v)zd "to remain, be left behind": Yd. uzaiyah, Orm. ozuk
"left behind" etc.

Introduction 11

bhdm "smell": Sak. bvdma-ta- "intelligence".

bhin "tree": Shgh. veyzn etc. "birch"?

banapdi "pillow": Mj. vezni etc.?

berhh- "to fear": Psht. wera, yera (Waz. wydra) "fear"?

ba& "rope": Psht. was, Orm. bes etc.

] bi§to "long": Psht. uzd etc.

id "how many": Psht. co.

due- "to milk": Shgh. biij-, Mj. luz- etc.

dhor "saw": Mj. IdSlcy.

derz- "to take on one's back": Psht. Uzal "to load", Orm. da? - etc.

^dusara "kid, two years old": Psht. dosaral.

dos "hair": Wkh. burs "goat's hair" etc.

gi ] no "hair": Psht. yuna etc.

gir "stone": Shgh. zir etc.

yuh-: yust "to throw": Psht. wistel etc.

yan "oak": Psht. ivana "tree" etc.

^yanuho "short": Sak. vanda-, but also Bal. gwand.

ya)nir "field": Wkh. wundr.

yarw- "to be boiling": Mj. wurv-, Sar. wdrav- etc.

yu ] rok "child": Psht. wur "small", wdflcai "boy".

ydzd "fat": Psht. wazda.

hi "bridge": Shgh. yed etc.

hariv- "to hear": Psht. arwedsl.

)ar- "to say": Psht. zarsl "to cry".

ho "roof": Wkh. hut etc.

hhuf- "to cough": Yd. kofah etc.

Mma "throat": Psht. dial. humai "Adam's apple".

me l hi "month": Psht. -mai, Orm. mat.

mux "face": Psht. max, Orm. mux.

na'yon "bread": Psht. nayan, Orm. txan etc.

norh "nail": Sangl. narxah etc.

pe "milk": Psht. pai etc.

p(h)i "spade": Shgh. fe, Wkh. pay etc.

pen "with, by help of": Psht. bande "on" etc.?

12 Parachv

pa l ndn "road": Shgh. pdnd etc.

pa l rl6- "to shake a sieve ": Shgh. parwiz-.

pa'tdsur, pdrasur "last year": Wkh. pard, or Psht. paros.

rue "flea": Mj. foriga, Orm. s r ak etc.

rlidydm "spring": Psht. waryumai "male kid"?

l-rhine "fire": Zeb. roSni, Orm. run.

rhdz- "to fly": Shgh. rewaz- etc.

my "word, affair": Sar. saug "proverb, tale"?

sa ] fwk "hare": Psht. soe, Orm. sikak.

m [ ni-: su ] na "to wash": Shgh. ze ] ne-: ze [ ndd etc.

si "horn": Sar. "shatv" etc.

weak "female": Zeb. See etc.

toy "male mountain goat": Wkh. tuy "goat".

thi- "to burn": Wkh. \fau-, §i-iiw- etc.

tar- "in, to, from": Ishk. tar "into" etc.

xer "hay": Oss. xor "fodder".

] xera "mill": Mj. xlrga etc.

l xdra "summer": Psht. worai etc.

zdy "son": Psht. ede, Soghd. zdk.

zdm "son-in-law": Psht. zum.

zox "firewood": Wkh. yax "twig".

With Minj., its nearest neighbour towards the north-east, Par.
shares the treatment of the secondary group of consonants *rt, cf.
Par. zito, Minj. sit "yellow" < Av- zairita-. The present stem kan-
from the root kar- "to do", which is alien to N.W. Ir., is found
in Par. and Minj., but also in other E. Ir. dialects. Further, the re-
partition of the stems in Par. zb-: dya "to come" (*aydya-: dgata-) is
found in Minj., too, and not only in N.W Ir., as stated by Ted.
(p. 231).

10. This survey shows that Pap. occupies a rather central position
among the Ir. languages. The points of agreement with E. Ir., which
we found especially in the vocabulary, may be due to a protracted
contact with these dialects. But the agreements with the W.
(especially N.W.) Ir. dialects, probably date from a very early time,

Introduction 13

when an unbroken series of dialeats (in Hazarajat, Herat etc.) con-
nected Parachi with Western Iran.

If, as I think probable, both Par. and Orm. are the remains
and descendants of the original Ir\ languages of Afghanistan i , the
points of resemblance between the: two languages can easily be ex-
plained. The profound divergences in some respects indicate that
the connexion between the two was severed at an early date, prob-
ably on account of an Indian expansion towards the West, which
partly embedded the Par. tribe in Nijrau and Tagau, and perhaps
created a Parachi diaspora of pedllars (cf. 6).

11. The very strong Ind. influence to which Par. has been sub-
ject shows also that this dialect has been spoken for a long time
in, or near, its present home.

A great number of Ind. loan-words of Hind, and Lhd. origin
have reached Par., probably to a (great extent through the medium
of Prs. and Psht. Cf. Voc. s.v.v.. bham l bur, 6u l ri, dzWdk, l ghana,
ga'ri (ga'ri), gat, hadd, l jumki, }6 r , ftwdri, ka l cdri, ka6era->phor, Ml,
kancatni, kut, luy'gi, laygd'fa, Hundu, lur-, lataH, ■maila, maina,
l matta, na'ti, plvund'6, pher-, paildn, paisa, pet, put, rupa'i, sir*, til,
tol, tok, weh-.

The Engl, words in Par. have also come through Ind. (e. g.
kar [ nail, l ardali, palHun, [ waskai).

Ind., but borrowed from sources not easy to define in each case,
are such words as ] dyes, budh-, humbu^-u, char-, ihatHo, 6i l no, dha)ram,
dhew-, daHew-, ghdnd, ^hanu, jal, ka)io, 'khandi, ^katt'6, lam, Hdwar,
[ mdneS, pen [ di, pendar, pdr'bn, patt, ra [ ho, se l ya, sel, si l na, sum], tekku,
ml, wdl Prob. of Ind. origin are also : ba l lur, dumb, l d'6nas, khdnas,
*kutur[ika), landd, Wre, luta l ki, paddo, pal-.

Words such as l kano and l wdyar (cf. Eep. p. 25), which have
their nearest parallels in the Kafir languages, render it possible
that some of the words mentioned above may be of Kafiri origin.
And it is very probable that many of them have been borrowed
i introduced from the West will be discussed

14 Parachi

rrom Pash., even if they do not at present exist in, or have not
been traced in that language. We know that e. g. Hiwon and phor
have been borrowed from an earlier stage of Pash. and the same
is probably the case with dha l man etc. (v. Voc).

But the number of words certainly borrowed from Pash. and still
found in that language is quite large. Cf. the voc. s.v.v. a ] i, u6,
andi ] wdl, ira, bhdr, halo, l bilru, cur, £a)tak-, l dddd, dhar-, dhdr, dhd)ri,
dhdw-, dale, dak, dmu'ruk, dak-, dal, gan l da, ga ] rat, M, hey l gas, huper-,
jut, khuj-, khen, khur, khu l ri, bun, kopdn, kor, ka l sayan (Pash. < ?),
kausdra (Pash. < Prs.), kat, ka l war, kiza)re, lejj-, ma [ 6i, l mdma,
man-do, munda l ran, hnaruk (?), hnawul, muz, po6 (?), paedgi, pa l po,
par-, par-, si, sat, sap, sor, ta'p'6, taWdnl, tag, toygok, w{i) l ydr.

13. But the Pash. influence upon Par. is by no means restricted
to the vocabulary. Like most Pash. dialects Par. changes y- into i-
and the distribution of this development (cf. G-awar-Bati y- >],
Traieguma Waigeli y- > z-) shows that it originates from Pash.

The treatment of intervocab'c -t- and -c- is identical in the two
languages (v. 50, 55) (t is elided, but c retained). The development
of -p-, however, differs (Par. -w-, Pash. -p-), while the fate of -k- in
Pash. is uncertain, and the two languages have probably developed
on independent lines as regards intervocalic stops. l

But the most important result of Ind. (prob. Pash. or pre-Pash.)
influence on the phonetical system of Par., is the introduction (or
preservation) of aspirates which will be discussed below (73). Even
if we admit that the aspirates may in some cases have been in-
herited directly from a pre-Ir. stage, the preservation and extended
use of them is undoubtedly due to a strong Ind. influence.

13. If we turn to the morphology of Par., we find that this
language has borr. from Pash. the pi. suffix dal and the case suf-

1 Pash. is the only Ind. language in which intervocalic t is preserved in some
cases, owing to an early vowel contraction. E.g. Pash. L vdst "20" <*vi&ati (bnt
e.g. traivya "60" < *tray6vi£dti\ \hanata "yon shall beat" < 'handthana (Turner,
JKAS, 1927, 237); but e.g. Waig. wtii, -i (< -athana). The contraction of an unstressed
vowel in similar position is common to most Ir. dialects, and it is possible that
Pash. may have been influenced by the Ir. system of accentuation.

Introduction 15

fixes Jcun and wano (v. 82, 97, 103). Probably forms such as Par.
bawdha "father and son" etc. (v. 83) are of Pash. origin (cf. Pash.
L. batoya).

In some cases we find strikingly similar forms, in Par. and Pash.
Cf. Par. an "I": Pash. a (v. 113); the enclitic pronouns (v. 116);
Par. and Pash. S a, a "he, it is" (v. 194); Par. and Pash. bin "he
was" (v. 182); poss. Par. 1 pi. -man: E. Pash. 1 pi. konj. -man
(v- 193). In each of these cases, viewed separately, the similarity
might be due to a parallel but independent development; but the
number of these morphological points of resemblance peculiar to
Par. and Pash. renders it probable that they are not all of them
accidental. No borrowing of forms need have taken place; but
forms which resembled those of the neighbouring language may
have had a better chance of being retained. Considering the numeri-
cal superiority of the Pashais and the great number of Pash. loan-
words in Par., we may safely assume that the source has been Pash. l

At any rate, Par. si "it exists" and the suffix of the present in
-ton (v. 176, 198) have been borrowed from Pash. In particular the
latter instance, the borrowing of a purely inflexional element, shows how
deeply Pash. has influenced Par. This is certainly not due to the
existence of a Pash. substratum in Par., seeing that Par. is the lan-
guage which has lost ground ; but it indicates a former social or cultural
superiority on the part of Pash. and a close connexion between
the two languages during on extended period.

14. The influence of Prs. dominates the Par. vocabulary completely.
Among the words which I collected, the words of Prs. origin by
far outnumbered the genuine Par. ones. A certain number of Prs.
words, however, occurred only in the poems, and were probably not
much used in the ordinary language.

But Par. seems to be approaching a state where the Prs. influence
becomes altogether overwhelming. Most, if not all, of the men are
bilingual, and any word wanting in Par. may be supplied from Prs.

1 i;» yarding a few possible Par. loan-words in Pash. v. Rep. p. 26 (par- is prob-
ably it. into Par. from Pash.).

16 Parachi

It is frequently impossible to decide whether we should consider
a word to be naturalized as a loan-word in Par., or not. In many
cases a Prs. word is used even where a Par. word with the same
meaning exists. E.g. l ddam "man" is used along with l tn&ne§,
and sometimes x sar-i<xud "his own head" is said instead of xu ] sor,
or l xdna "house" instead of yus. This shows that Par. has reached
a state when the resistance to the introduction of Prs. elements
is very weak. In some cases, however, the Prs. words have been
"translated" into Par., v. Voc. s.v.v. i aiv9-i didattka, kho^buj,
khor9 l gu.

The syntactical system, the "innere Sprachform" of Par. has a
strong Prs. stamp, and most idioms are formed after Prs. patterns,
as will be apparent from the Prs. translations of Par. sentences
given in the Texts and in the Vocabulary.

Through the influence of Prs. the Ar. qaf has been introduced
into Par., although k is frequently heard in unguarded speech, and
even the 'ain occurs occasionally. The pronunciation of d in Shutul
(G and T; M, who was some 25 years older, had a less rounded
pronunciation), compared with a in Nijrau, is certainly due to Prs.
In Shutul the Parachis are surrounded by a Prs. population, and
the men probably often speak nearly as much Prs. as Par., and
easily adapt their pronunciation to that of Prs.

The morphological system of Par. is nearly intact. But the use
of the izafat, which was probably introduced in loan-words such as
band-e , dest "wrist" (also band-e 1 dost, band-e dosti l ka), has been fully
established. It is possible that the pi. in an is of Prs. origin, the
genuine pi. being found in phor, pi. phar etc. (v. 83). T's tendency
to employ the nom. (e.g. an "1" instead of mun) with past tenses
of transitive verbs is certainly a Persianism.

15. A language like Par. — this applies also to Orm., some dia-
lects of Pash. etc. — is therefore not a comparatively self-sufficient
and complete unity like the great languages of civilization. Nor
does it occupy the independent position of e.g. the Kafir languages,
which till recently had a separate civilization of their own, and

Introduction 17

were, in the main, able to express the ideas known to them by
means of indigenous, or completely assimilated, linguistic elements.
It is, of course, impossible to consider Par. as a dialect of Prs.
in the ordinary, genealogical sense, since it has retained a morpho-
logical system and, to a considerable extent, a vocabulary which
differs very much from Prs. But it might, in a certain sense of
the word, be called a "cultural" dialect of Prs.

INSTITUTTET
FOR SAMMENL1GNENDE KULTURFORSKNING

INDO-IRANIAN FRONTIER
LANGUAGES

BY

GEORG MORGENSTIERNE

VOL. I.
PARACHI AND ORMURI

OSLO 1929

H. ASCHEHOUG & CO. (W. NYGAARD)

LEIPZIG PARIS LONDON

OTTO HARRASSOWITZ HONORS CHAMPION WILLIAMS & NORGATE, LTD.

CAMBRIDGE, MASS.

HARVARD UNIVERSITY PRESS