THE CHARACTER OF THE INDIAN KHAROSTHI SCRIPT

AND THE "SANSKRIT REVOLUTION": A WRITING SYSTEM

BETWEEN IDENTITY AND ASSIMILATION
Ingo Strauch
Introduction

The history of writing in India is extraordinary from different points of

view. First, there is the fact that writing as a cultural technique was introduced into Indian culture at a rather late point in its history. It was long
after large amounts of its literature had already been created. Not only
the Veda but also the late Vedic literature and even considerable portions
of the Buddhist and Jain canons were already in existence when India
saw its first attempts to fix language to a writing system.1 Remarkably,
this introduction did also postdate the early Indian state formation processes in the sth century BC, and consequently it was also later than the
urbanization which accompanied this state formation in the Ganges valley (cf. Strauch 2005).
The focus of my paper will be dedicated to another peculiar feature
of the Indian history of writing. From its very beginning two rather different scripts were in use: the Kharo~thi, a script which prevailed in the
Indian North-West, i.e. modem North-West Pakistan, and spread from
there to Northem Bactria and Central Asia, and the Brahmi which began

' The question of the introduction of writing in India has been controversially discussed. An excellent survey of this academic debate and a plausible scenario of this process was presented by Falk (1993). Other valuable contributions concerning this problem
are von Hiniiber (1990) and Falk (1996). Although it is probable that the introduction of
Kharo~thi preceded that of Brahmi for some decades there is no positive evidence for the
use of either of these scripts before the time of Asoka. Any suggestions about a gradual
development of Kharo~thi in the centuries before ASoka and a reconstruction of this process (e.g. Glass 2ooo: n-20) must remain highly speculative. For the suggested very small
time gap between the emergence ofKharo~thi andAsoka see Falk (1993: 103-105). According to Salomon, "there is no clear evidence to allow us to specify the date of the origin
of Kharo~thi with any more precision than sometime in the fourth, or possibly the fifth,
century B.C." (1998a: 46). Recent discoveries of presumably earlier Briihmi texts from Sri
Lailka still wait for further confirmation of their suggested dating between the 6th and
4th centuries BC (see Coningham et al. 1996). Since this article will concentrate on the
final phase of Kharo~thi, the discussion on its introduction is of minor relevance.

INGO STRAUCH

its career in the East of the Indian subcontinent and conquered almost all
of India until it replaced even the Kharo~thi in its mother-land after the
3rd c. AD. It is the Brahmi script which became the "mother" of all modem
South-Asian scripts and of many scripts in South-East Asia. The Kharo~thi
remained a footnote and left no further traces in the writing systems of
the region. Nonetheless, it was a rather influential script which not only
left hundreds of epigraphs but also a considerable corpus of Buddhist
manuscripts which belong to the earliest witnesses of Buddhist literature
in generatz In recent years our knowledge about this script and its use
increased considerably due to the discovery of a series of new Kharo~thi
manuscripts which are presently being studied in Seattle and Berlin.3
Why this script died out, is still a matter of academic debate. Only
recently, Richard Salomon, one of the leading experts in Kharo~thi studies, expressed the view that "the decline of the Kharo~thi in its homeland was closely connected with, if not directly caused by, the collapse
of the Ku~aJ).a dynasty." He attributes the Kharo~t}}.i case to those "cases
of script disappearance that are directly attributable to dynastic changes
or declines (Salomon 2oo8b: 149 )." According to Salomon, the Kharo~thi
"could in theory have fulfilled the role of a Pan-Indian script, that, as
the accidents of history had it, actually fell to Brahmi and its derivatives
(2oo8b: 144)."
It is the main aim of this paper to demonstrate a different approach to
this phenomenon in the Indian history of writing and to show that the
decline of Kharo~thi can alternatively be interpreted as the direct result
of a cultural and linguistic shift in the communities where it was used,
and not-or at least not mainly-as the consequence of a political event.
Taking up the general title of this book this shift could be characterized
as a forced and intentional, but eventually unsuccessful border-crossing
by which the Kharo~thi script was supposed to adjust itself to another
language, namely Sanskrit.
The paper will be divided into two parts. The first section describes the
character of the script, particularly with regard to its suitability to write
Sanskrit. Since the missing success of the Kharo~thi is also the result of a

An easy survey about all material available in this script can be found in the "Catalog
of Kharo~thi Inscriptions" (CKI) and the "Catalog of Kharo~thi Manuscripts" (CKM) which
are both available on the homepage of the "Dictionary of Gandhari", a project by Stefan Baums and Andrew Glass (www. Gandhari.org). The page also provides an extensive
bibliography.
3 See for more information about the manuscripts studied in both projects Salomon
1999, Allon 2007, Strauch 2008.
2

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

133

competitive situation between both Indian scripts and their cultural contexts this section will also shortly refer to the characteristics of Brahml.
The second part will illustrate different attempts of Kharo~thi to adjust
itself to the introduction of Sanskrit as a new literary language and lingua Jranca in the Gandhara area and its relationship to the Brahmi script
which accompanied the advent of Sanskrit.
It should be stressed that this paper does not aim at a comprehensive
study on the subject. It tries to introduce a new perspective on the problem
in a more essayistic form, including some new data which are based on
the author's recent research on a collection ofKharo~thi birch-bark manuscripts (Strauch 2oo8). Since the field of Kharo~thi studies has recently
been much in transition due to the ongoing discovery of new texts and
inscriptions, the general and comprehensive history of Kharo~thi remains
to be written.
The Character of the Kharo$(hi Script
The beginnings: Brahmi and Kharo$(hi under Asoka

The first safely datable monuments of the Kharo~thi are associated with the
reign of the Indian ruler Asoka (3rd c. BC). On behalf of his imperial order,
rocks and pillars all over the South-Asian subcontinent were inscribed
with what one might call a kind of state ideology which is partially based
on Buddhist ethical conceptions. 4 While for the majority of texts the probably newly developed Brahmi script was used, those inscriptions which
are located in the extreme North-West are written in Kharo~thl. Contemporary to the Asokan Kharo~thi epigraphs are some texts which continue
the use of the earlier Greek and Aramaic alphabets and languages in the
Indian North West (Falk 2006: 241-253). These contemporary writing cultures also indicate the most probable candidates for the stimulus to create
a script which is capable of fixing an Indian language. According to the
prevailing theories, the Kharo~thi script was developed on the base of the
Aramaic alphabet as used during the Mauryan period in North-Western
India. 5 This presumable prototype is not only responsible for the shape of

4 A comprehensive survey of the A.Sokan epigraphs and a bibliography are now available with Falk zoo6.
5 The complex relationship between Asokan Kharo~thi and Aramaic has been dealt
with by Falk (1996). The Semitist's point of view was expressed by Voigt (zoos), whose
remarks provide valuable additional data but suffer from the author's lacking familiarity
with the Kharo~thi script and Gandhari language.

134

INGO STRAUCH

many individual letters but also for the writing direction of the Kharo~thl
from right to left. Despite these parallels, the Kharo~thl has to be defined
as a newly created script with a distinctively different character.
This first phase of Indian writing is not only distinguished by the multitude of scripts but also by the fact that its texts were composed in different dialects. Consequently, beside their doubtless ritual function to
mark the extant of the ruler's actual or asserted influence, these edicts
were clearly meant to be understood by the local communities. For this
purpose Asoka ordered them to be translated into various Middle Indian
dialects. Among these translations only those in the North-Western language were written down in Kharo~thl. This speaks not only in favour of
the assumption that this script was already in use at the time of Asoka,
but also that it had been intentionally designed for the local language of
the region, the so-called Gandharl.
The Brahml, however, fulfilled from its very beginning a much more
universal function and was designed to express different local dialects of
Northern India. What both scripts shared, however, is the fact that none of
them was created to write down a Sanskrit text. This had a direct impact
on their inventory of signs.
Kharo!Jthi and the Sanskrit var:r:tamala

Usually, Indian alphabets are arranged according to the so-called

van:zamiilii sequence: an inventory of sounds which was created on the
basis of Sanskrit phonology as early as in the 4th c. BC, i.e. before the
introduction of writing. It classifies the sounds of the language according
to their physiological features into different groups (cf. table 1).
A script which could claim to match these Sanskritic phonemes would
be expected to provide graphemes for all these sounds. Since not all of
them are needed for Middle Indian, the earliest Brahml and Kharo~thl
have only a restricted inventory of signs. The table below contrasts the
traditional van:zamiilii alphabet and those sounds which were not represented in the scripts in their initial stages. It should be stressed that in the
case of Kharo~thl this survey mainly relies on the conventional phonetic
evaluation of the respective graphemes, which is in many cases, however,
far from certain. 6

6 For a promising new approach to the phonology of Gandhan including a discussion of

the phonetic value of Kharo~thi signs see now Stefan Baums (2oog: no-2oo ). The later use
of the Kharo~thi graphemes for writing Sanskrit, however, shows that the Kharo~thi signs
could indeed carry the phonetic value implied by the van:zamala (cf. below, table 3).

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTIO~"

Table

1.

135

The Sanskrit vall).amala alphabet and its relation to early Brahmi: and
Kharo~thi

L -_

__.1

not in Kharo~thi
not in Brahmi: and Kharo~thi

Vowels
Simple vowels

short

long

ii

short

long

ai

au

Diphthongs

Consonants
Explosives

voiceless

voiced

Nasals

non-aspirated

aspirated

non-aspirated

aspirated

Velars

ka

kha

ga

gha

Iia

Palatals

ea

eha

ja

Ijha

fta

Retroflexes
Dentals
Labials

ta
ta
pa

tha
tha
pha

qa
da
ba

Qha
dha
bha

Semisonants

ya

ra

la

va

Sibilants

sa

~a

sa

Spirant

Vowel
modificators

:q1

Anusvara

Visarga

Consonant
modificators

Virama

I).a

na
ma

INGO STRAUCH

Relying on this van:zamiila based comparison one can easily get the
impression that both alphabets contain by and large the same inventory
of signs. Their coincidences can be explained on the basis of the common
Middle Indian phonology: both lack the velar na and the special aspiration sign Visarga, and both lack the signs for sonantic .r/l and the diphthongs ai and au.
Beside this seemingly common inventory both alphabets also share
their abugida character. Every consonantal sign of the script includes a
subsequent vowel (cf. Daniels & Bright 1996: 4), in the case of the Indian
scripts a short a. To change the quality or quantity of this basic vowel
one has to add another diacritical sign. The inherent vowel expressed by
a sign makes it also necessary to signify vowelless consonants. This had
to be done either by the graphical instrument of sign combinations, the
so-called ligatures, or-if the consonant appeared at the end of a wordby a further diacritical sign. Since the Middle-Indian languages are characterized by a strong tendency to assimilate different and to omit final
consonants, both early variants of Kharo~th1 and Brahmi contained a
very limited choice of ligatures and no special diacritic mark for a final
consonant. In addition, both scripts generally refrain from designating
geminated consonants, in which the Middle Indian languages are extraordinarily rich due to the described tendency of consonant assimilation. 7
Beside these shortcomings which had to be overcome to make both
alphabets fit for Sanskrit, the Kharo~th1 shows a further significant difference: It is unable to signify the vowel quantity of d, l and ii. It is rather
improbable that the Gandhar1 speakers did not know about this distinction. Thus this fact can be best explained as a structural inconsistency
which was probably inherited from the Semitic model of the script. It is
mainly this difference, which marks the Brahm1 as the more developed
of both scripts, which is better adjusted not only to the requirements of
Middle Indian phonology but also to those of Sanskrit.

7 A special position among the consonant clusters is held by the combination of consonants with the semivowel r. Since the local dialect ofthe Indian North-West obviously did
not assimilate this sound in post- and preconsonantal position, the Kharo~tl.Ii knew special
diacritics for these combinations from the very beginning. Although this sound combination was obviously also known in other Middle Indian dialects, its adequate realization in
the Asokan Brahmi caused a few problems which were, however, quickly overcome in the
subsequent century (cf. Dani 1963: 54f. ).

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

137

Kharo$thi and the Arapacana

As useful as the van:tamiilii arrangement might be for a direct comparison

of both scripts, it hides some of the characteristic features of Kharo~thi
and gives the impression that with some exceptions both alphabets share
a common character. This impression quickly disintegrates when we consider the sign inventory of the Kharo~thi on the base of its own alphabet,
which is known today by the name Arapacana because it starts with the
five letters a-ra-pa-ca-na. In recent East-Asian Buddhism this alphabet
is associated with the Buddha Maiijusri, and its name--Arapacana-is
regarded as a kind of mantra, i.e. a magical spelling (Salomon 1990: 255).
In a much earlier period the alphabet was also used as a kind of memorizing device: important categories of Buddhist doctrine were memorized
in the sequence of the alphabet. Remains of this use can be observed in
quite a large number of Buddhist texts, where various lists of Buddhist
terms contain passages which can be traced back to the sequence of the
Arapacana (Pagel2007: 18-38).
We do not know how old this alphabet is and since when it had been
used for the arrangement of the Kharo~thi sign system. The fact that it
contains some signs which are obsolete in the mature phase of Kharo~thi
(no. 28, no. 40) seems, however, to speak in favour of a quite early date
of the Arapacana.
Until most recent time the inventory of the alphabet had to be reconstructed on the base of few Sanskrit texts, like the Satasahasrika and
Paiicavi~p.satika Prajiiaparamita or the Lalitavistara, which contain an
almost complete but sanskritized version of this alphabet in Brahmi
writing (Brough 1977, Salomon 1990). In many cases it was not possible
to determine which concrete Kharo~thi sign hid behind the respective
Sanskrit-Brahmi "transcription". Only recently, several incomplete variants of the Arapacana alphabet written in Kharo~thi could be identified. The oldest of them was discovered on an ostracon in the course of
excavations of a Buddhist monastic complex at Kara-Tepe, near Termez,
Uzbekistan (Salomon 2004, Fussman 2on: 87, 107 KT). According to its
archaeological context its date can be fixed into "the late first or early
second centuries A.D." (Salomon 2004: 45). Another part of the Arapacana
alphabet was found among the inscribed wooden tablets from Niya in
Chinese Turkestan, which can be dated into the 3rd, early 4th centuries
AD (Salomon 1990 ). This rather limited evidence is supplemented by the
widespread use of Arapacana letters as location markers in Gandharan
art and architecture. Many objects-reliefs, sculptures, etc.-bear small

INGO STRAUCH

Kharo~thl

signs which indicate their arrangement in a certain complex

(Salomon 2006, Koizumi 2007). All this evidence allowed a rather reliable
reconstruction of the alphabet up to its 35th letter (cf. table 2). Fortunately, among a collection of manuscripts which was discovered in the
Bajaur district, not far from the Afghanistan-Pakistan border, and is currently being studied at Freie Universitiit Berlin, a text could be identified
which contains a compilation of verses which are arranged in the order of
this alphabet (= BajC s, cf. Strauch 2oo8: 121-123). Each half of every verse
begins with the same letter. On the basis of this text and the incomplete
versions which were previously known it is now possible to establish the
complete inventory of signs as listed in the Arapacana alphabet.
The following table (table 2) will demonstrate the inventory of signs
as given in the Bajaur fragment 9 and its relation to other versions of the
Arapacana alphabet. Arapacana signs which are not part of the vafl}.amala
alphabet are left unshaded in columns 1 and 6. (Section 1). Moreover, the
table will show the varieties of secondary modifications attested in various sources ofKharo~thl writing (Section 2). A third section will illustrate
the use of the basic and modified signs in a Kharo~thl text from Bajaur
which is composed in Sanskrit (= BajC 9 ).
The Arapacana alphabets in the table are reconstructed according to the
following publications:
- Brahmi representatives from Skt. texts: Salomon 2004: 47
- Niya tablet: Salomon 1ggo: 265-268, fig. 7 (with the exception of no. 21
(spa), no. 28 {fa) and no. 35 {tha) which were corrected on the basis of the
digital image of the plate (available at http:/ /idp.bl.uk/database/large.a4d
?recnum=16781&imageRecnum=zo818) for spa, (tha), and [tha]
- Kara-Tepe ostracon: Salomon 2004
- Location markers: 1-17 (a-ma): Zar Dheri (Koizumi 2007), except 13 (ya)
(Salomon zoo6: 199), 18 (ga): Nimogram (Salomon zoo6: zoo), 24 (kha):
Butkara (ibid.), 26 (sta): "Chorasan, Gandhara" (ibid.).
The Bajaur signs are extracted from the verse initials of Bajaur fragment 5,
except when preceded by *. In these cases the signs had to be taken from
other portions of the text.
Most of the modified signs are extracted from Glass zooo. Their exact
provenance is indicated in the following annotated survey. This table does
not aim at a complete inventory of modified signs and is meant only for
illustration purposes. The actual-recently attested-number of modifications is higher than shown here. A more complete survey has to be undertaken after most of the recently studied Kharo~th1 manuscripts are edited. It
is possible that some of the signs identified by Glass as phonetically relevant
graphemes are mere ornamental varieties of the basic sign. If the character
of a modification is unclear, this is indicated by*.

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

139

- *pa: Glass 2ooo: 84 (Gardner plate 30, no.3). Although Glass mentions
the variant pa (transliterated as pa) among the letters composed with
the help of the respective syllable modificator ("Cauda") (2ooo: 136), his
chapter on the variants of the sign pa contains only one letter of this type
which is considered by him as a footmark (2ooo: 84). It is generally difficult to distinguish the phonetically relevant substroke from the merely
graphical footmark of almost identical shape.
- ea: Glass 2ooo: 62 (Niya documents)
- na: Glass 2ooo: 82 (Khotan Dharmapada)
- da: Glass 2ooo: 79 {British Library scribe 21)
- 9-a: Glass 2ooo: 72 (Niya documents). Although Glass is referring to this
variety in his chapter on modifiying signs (2ooo: 136), he refuses this identification in his chapter on the sign if-a and suggests considering it as one
of the phonetically irrelevant footmark variants (2ooo: 73). For sake of
completeness we will, however, pertain to distinguish this variant from its
basic sign. Rapson et al. (1920-29) transliterate this letter as f!.a.
- $a: Glass 2ooo: 103 (British Library scribe g)
- a: Bajaur Collection scribe 4
- ~a: Glass 2ooo: g8 (Niya documents). As Glass points out, this letter is
already found on the coins of Vima Kadphises, where his name is spelt
~ima.

- ti: Glass 2ooo: 77 (Bajaur casket)

- *yi: Glass 2ooo: 94 (Bajaur casket). It is not clear whether this variant
really represents a modified consonant. Again Glass mentions this letter
among the examples for the consonant modifier (2ooo: 136), he considers
it in another chapter as a footmark variety of the letter ya (2ooo: 94).
- fs.a: Glass 2007: 95 (Senior scribe)
- k: Bajaur Collection scribe 3
- se: Glass 2ooo: 109 {British Library scribe 14)
- a: Glass 2ooo: 107 {British Library scribe g). It is not clear whether the

variants (Wardak vase) and

of some British Library scribes represent the same sign. They are regarded by Glass as early variants of the
letter a. In one of the BajC manuscripts, however, both variants were
used side by side, indicating perhaps that they were perceived as different
graphemes. The same graphical distinction between the ordinary a and
this under-bent sa is also found in the Arthapada manuscript of the "Split
Collection" where both signs perhaps even designate different phonemes.
C Falk 2onb: 14f.
ma: Glass 2ooo: 93 {British Library scribe 8)
*mi: Gl~ss 2ooo: 92 (Wardak vase). Again the status of this sign is unclear.
Like the pa, the 1-a and the )!i it was listed by Glass among the modified
signs (2ooo: 136), but considered in the description ofletters as a footmark
variety (2ooo: 92).
gi: Glass 2000: 59 (Khotan Dharmapada)
ga: Glass 2ooo: 58 (Khotan Dharmapada)
tfia: Bajaur Collection scribe g

140

INGO STRAUCH

- ja: Glass 2ooo: 65 (British Library scribe g)

- ja: Glass 2ooo: 66 (Bajaur casket)
- -sa: Glass 2ooo: 99 (British Library scribe 10 ). According to Glass, "the
superscript line with sa is not thought to signify a different phoneme,
but rather to distinguish sa fromya which can be almost identical" (2000:
99 ). Although this interpretation is quite plausible, it seems advisable to
consider this sign for the time being as a separate and phonetically relevant variant of the basic grapheme sa.
- a: Glass 2ooo: 100 (Bajaur casket)
- k$a: Glass 2ooo: n6 (Niya documents)
- n: Bajaur Collection scribe g
- fie: Glass 2ooo: 113 (Niya documents)

Table 2. The basic and modified signs of the

1

Kharo~thi

Arapacana alphabet and their Sanskrit reflexes in BajC 5

Arapacana alphabet- Basic signs

No.

Sanskrit-Brahmi Niya tablet

Kara Tepe

Location

Bajaur fragment

representatives . _

ostrakon

markers

BajC 5

Modifications

I Bar above

3 Sanskrit reflexes (BajC g)

Extension below

I Basic signs

Modified signs

I::)

~
~

ra

ra

ra

pa

pa

pa

[pa]

.!}

ea

ea

]'

ea

na

[na]

na

na

iia

-1

na/T).a

la

la

la

da

da

da

darp.

.s
J

ba

ba

ba

ba

c,l.a

[c,l.a]

c,l.a

c,l.a

~a

~a

va

va

ya

ti

12

ta

ta

ta

ta

la
c;Ja

da
ba

'I

<Ja
~a

'fi

sea

t;:j
~

I::)

la

[va]

::t:
Cll

ea

va

l!

pa

C')

[ea]

ll

Cil

*pa

ea

~a

ra

10

~a

'f..

c,l.a

~a

z
~

::t:
t%j
::

Cll

~
~

t"'

va
ta

~
~

0
Z:::
1-1

-e

Table

1-'

(cont.)

~
~

1 Arapacana alphabet- Basic signs

2 Modifications

representatives

Kara Tepe
ostrakon

Location
markers

BajC 5

13

ya

ya

ya

ya

.11

14

~ta

[tha]

tha

tha

No.

Sanskrit-Brahmi Niya tablet

15

ka

16

sa

17
18

ma
ga

[sa]
(ma)
( ga)

Bajaur fragment

ka

ka

ka

sa

sa

sa

ma
ga

ma
ga

ma
ga

19

tha

tha

tha

tha

20

ja

(ja)

ja

ja

21

8va

spa

[s.. ]

spa

22

dha

dha

dha

dha

23

sa

sa

sa

24

kha

kha

kha

kha

.b

y
fb

Bar above

JJ

kha

Extension below

*yi

Basic signs

Modified signs

,.fL.. I ya
~tal
~tha

ka

;,-

ka

ka

se

sa

ma

ga

ma
gi

tha
ja

sa

3 Sanskrit reflexes (BajC 9)

~a

If
1y
_.,

*mi
ga

C:l
0

t;f.)

t-3

C":l

ja

tha/
stha
ja

I dha
~a

(L

I sa
I kha

stha

:I:

Table

(cont.)

1 Arapacana alphabet- Basic signs

2 Modifications

3 Sanskrit reflexes (BajC g)

1-<

No.

Sanskrit-Brahmi Niya tablet

representatives

Kara Tepe
ostrakon

25

~a

~a

~a

26

sta

sta

[sta]

27
28

jiia
rtha

iia
ta

[iia]
ta

Location
markers

Bajaur fragment
BajC 5
~a

sta

sta
iia

ta

29

bha

bha

bha

bha

30

cha

cha

cha

cha

31

sma

spa

_p

rjS
7'

Bar above

~a

Extension below

y-

Basic signs

~a

jiia

.P
!f,

bha
cha

32

33

tsa

tsa

tsa

j3

tsa

c:e

gha

35

tha

[tha]

{ha

36

I).a

37

pha

I).a

pha

1:1

-p

(.j

""d

gha

jiia

1-<1

vha

[gha]

Cf.)

vha

gha

Cf.l

hva

34

~a

sta
ii

ztj

spa

Modified signs

tj
~

=
l:;lj
:::

Cf.)

1-<

t"'

l).a/na

c:~
1-<

pha

.z:::
~

-+:>.

Table 2 (cont.)
1

Arapacana alphabet- Basic signs

No.

Sanskrit-Brahmi Niya tablet

representatives

Kara Tepe
ostrakon

Location
markers

Bajaur fragment

Modifications

Bar above

3 Sanskrit reflexes (BajC g)

Extension below

Basic signs

Modified signs

BajC 5

38

ska

ka

39

ysa

za

:1

40

sea

ea

41

ta

ta

1'
i-

42

Qha

Qha

ha

*ha

T
2

ska

iie

2::

....
z

ha

Cf.)

C')

:I:

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

145

It might be worth noticing that the versions of this alphabet represented

here-although they originate from quite different regions and cultural
contexts of the Kharo~thi using area-are remarkably consistent. This
clearly speaks in favour of a coherent use and commonly accepted shape
of the Arapacana alphabet throughout the attested history of its use.
If we compare this repertoire with the van:zamala, we observe:
1. Nearly all phonemes of the reduced Middle Indian variant of the
van:zamala as represented in the Asokan scripts are part of the Arapacana
alphabet. They are indicated here by shaded cells. The exceptions are the
phonemes }ha and ha which are not part of the alphabet. In the case of
the aspirated }ha it is quite possible that it was unknown to the Gandhari
language. 8
The reason for the absence of ha is however a mystery. It is clearly
present in the earliest varieties of Kharo~thi. A not very convincing explanation could point to the fact that some dialectical variants of Gandhari
seem to reduce the Old-Indian phoneme h to yor hiatus (cf. Brough 1962:
92f.). Accordingly, the Arapacana could have been developed in a region
which did not know the voiced glottal sound h.
2. In comparison to the van:zamala arrangement the Arapacana alphabet
contains a considerable number of additional letters (here left unshaded ).
These additional signs can be divided into different categories:
A-Some of these signs can be explained as representations of sounds
which are only found in the Gandhari speaking area for which the
Kharo~thi was designed. To this category belong the signs for vha (no. 32),
which most probably determines a voiced or voiceless spirant like vlf, and
za (no. 39), a voiced sibilant which is absent from other Indian dialects.
B-The phonetic value of the remaining signs is less clear. Although
the graphical shape of some of them seems to suggest that they represent either sound combinations (no. 31 spa, no. 33 tsa) or modifications of
basic phonemes (no. 26 sta, no. 28 ta, no. 35 tha, no. 38 ka, no. 40 ea), the
respective transliterations are mere reconstructions and do not reproduce
their actual phonetic value.

8 In earlier transliteration systems the Kharof?thi sign za was transliterated as jha. For
its now commonly accepted transliteration cf. Glass 2ooo: no.

INGO STRAUCH

Without entering the discussion of the complicated and not yet completely understood phonology of the Gandhari language and its relation
to the orthography of the Kharo~thl script, we may restrict ourselves here
to the observation that the Arapacana alphabet contains eleven letters
which are not part of the Sanskrit phonological system as displayed in
the van:zamala and are thus strictly taken not necessary to write Sanskrit.
At the same time these "superfluous" letters form an indispensable part
of the Kharo~thl alphabet.
Regardless of their original phonetic value all of the signs of the Arapacana alphabet are perceived as basic letters which can be subject to further
modifications, like the designation of the vowel value and its nasalization
by a set of diacritics. Another important category are special diacritics
for the pre- and postconsonantal semivowels -y-, -r-, -v-. Obviously, consonant clusters with these semivowels occurred in at least some varieties
of Gandhari and are therefore part of the Kharo~thl script from its very
beginning (cf. table 3).9
In addition to this initial inventory the Kharo~thl scribes developed a
method to further modify the original value of the basic signs by attaching
Table 3 The vowel and semivowel diacritics of conventional Kharo~thi
Vowel diacritics
ka

ki

ku

ke

vya

sva

ko

paJ11

Semivowel diacritics
rka

kra

1
All sample letters are extracted from the Bajaur manuscript BajC 2, with the exception
of rya (BajC gr).
9

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

147

different diacritics. The structural basis for this approach can already be
observed in the fundamental sign inventory as displayed in the Arapacana. Here we can discern sign groups like pa: pha (no. 3: no. 37), ea: ea
(4: 40), va: vha (n: 32), t;la: tha: tha: sta: fa: fha: ta (g: 14: 19 : 26: 28 :
35: 41), ka: ka (15: 38) or ga: gha (18: 34), where phonetic relationship is
clearly expressed by the graphical modification of basic signs.
It is remarkable that in cases where we can clearly distinguish the basic
sign from its modified variant the basic form regularly precedes the modified one in the sequence of the alphabet. It is therefore possible that the
order of the Arapacana letters reflects the historical sequence of their creation. Thus it can hardly be seen as a mere coincidence that most of the
signs for aspirated sounds, which were probably originally unknown to
the Gandhari language, are in the second half of the alphabet. I call this
process which took place until the alphabet as such was closed for further
additions of signs "primary modification". It is difficult to say, when this
process came to an end, i.e. when the Arapacana alphabet had reached
its complete shape. It cannot be excluded that the alphabet was not yet
complete by the time of Asoka. At least some of its basic signs (no. 21:
spa, no. 32: vha, no. 33= tsa, no. 38: ka, no. 40 : ea) are obviously absent
from the inventory of the Asokan Kharo~thi as represented in the rock
edicts at Shahbazgarhi and Mansehra. However, like any argumentum ex
silentio this absence cannot prove that other unattested varieties of the
script did not contain these missing signs. It is therefore possible-though
hypothetical-that Glass' statement according to which "(b )y the time of
the Asokan inscriptions, the Kharo~?thi alphabet was complete" (2ooo: 20)
is correct.
In the centuries following Asoka this primary modification was supplemented by two types of secondary modifiers-a horizontal bar above the
letter and an extension added to the foot of the letter (cf. above table 2).
The phonetic value of these modifications is not completely clear, and
according to the present state of research their use was by no means consistent. Concerning the first variety, however, there now seems to be good
evidence that it mostly indicates the duplication-or prolongation-of a
consonantal sound. It must be stressed that it shares this function with
other graphical devices. Thus a preconsonantal rand a post-consonantal
v was also occasionally used to designate consonant clusters, including
geminated consonants (cf. Baums 2oog: 198, table 45). Less clear is the
function of the underlining variety which is sometimes difficult to distinguish from an ornamental foot-mark without phonetic meaning. In many

INGO STRAUCH

cases it can be interpreted as a marker of a fricative pronunciation, but

other types of modification are not excluded.
The number of true combined letters which use two or more basic
signs for designating consonant clusters is very limited. According to
the hypothesis suggested here they are originally alien to this script and
belong to the later phase of Kharo~thi writing which experienced strong
influence from the Brahmi writing system (c below).
This detailed description of the alphabet and its use was intended
to show that the Kharo~thi by its nature was not phonological like the
Brahmi but phonetic. Throughout its history it showed a strong tendency to depict different allophones by different or modified letters. This
resulted in a rather large diversity of signs which hardly corresponded to
the phonological structure of the language the script designated.
This phonetic character is closely connected with and still complicated
by the fact that the Kharo~thi shows very weak tendencies towards a standardization of its orthography. Thus we easily find for one lexeme a multitude of different spellings. Although it is not excluded that these various
spellings reflect different phonetic realizations of the same word, we have
to consider that they might simply be caused by different orthographical
conventions, including historical spellings. This results in considerable
synchronic and diachronic diversity of orthographical usages.
As an example for this feature one may cite the frequently attested
word derived from Old Indian prati$thiipita "erected". Its orthographical reproductions in the Kharo~thi recordsn include pratistapita (CKI
155), pratithapita (CKI 328), pratithavita (CKI 257), pratithavito (CKI 48),
pratithavida (CKI 243), pratithavida (CKI 32), pratithavidu (CKI 405),
pratithavite (CKI 176), pratithavito (CKI 65), pratistavitae (CKI 403),
pradithavida (CKI 464), pradistavita (CKI 6o ).
Although such orthographical variants might provide valuable information about the dialectical variance in the Gandhari speaking area, they
are rather obstructive in terms of standardization which is, however,
10

1 For its assumed function as fricative marker cf. Glass 2ooo: 136f. More differentiated
is the interpretation of this graphical device as suggested by Baums (2009 ). While k. "indicates UJ derived from a velar" (140 ), 9- as used in the Niya documents and by British Library
scribe 14 indicates the sound [r] (141), r1 could designate a fricative (141) and is said to
indicate "the change from [s] to [z]" (150 ).
n The inscriptions are cited according to the "Catalog of Kharo~thi Inscriptions" (CKI),
see above fn. 2. Cf. there for detailed bibliographical references.

INDIAN KHARO~'.fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

149

necessary if a script is intended to be used for other languages than that

it was created for.
The ''Sanskrit Revolution"

The Asokan inscriptions and all other epigraphical documents in the subsequent two centuries were composed in Middle Indian languages. Not a
single written Sanskrit text of this period has been found so far. This fact
can at least partially be explained by the adverse attitude of the Brahmins
towards the newly introduced writing culture. Thus still the Mahabharata,
a work which was composed in the three centuries before and after Christ
(13.24,70) states:
vedavikrayi1J.as caiva vediinii:rrz caiva dil{;akii/:1.
vediinii'lJ lekhakiiS caiva te vai nirayagiiminal}
Those who sell the Veda, who spoil the Veda,
who write down the Veda will certainly go to hell

Writing was regarded by the Brahmins an inappropriate means to preserve and transmit their religious texts. This attitude was mainly due to
their monopolistic position in the sphere of religious literature. It was one
of their main tasks to preserve the textual tradition and to apply it in
ritual contexts. Any attempt to democratize this sphere-and script is of
course an instrument in this direction-would challenge this position and
was consequently rejected. It is therefore not surprising that the growing
writing culture was first adapted by other social and religious groups. The
most important among them were the Buddhists. The majority of inscriptions which are attested between Asoka and the 1st c. AD consequently
belong to Buddhist monasteries and record the dedication of buildings
or sculptures. The persons who left these epigraphs help to determine
the social strata which used writing. They belonged either to the ruling
aristocratic class or to the class of merchants and artisans.
The language policy of the Buddhists promoted the development of
Middle Indian dialects as literary languages. They translated their literature into the 'dialects spoken in the areas of their activities. Although
in the first centuries following Asoka the texts were mainly orally transmitted the Buddhists soon started to make use of the new cultural technique of writing. The exact date of this event is in most cases unknown,
but the Pali canon e.g. is said to have been written down in the middle
of the 1st century BC (von Hiniiber 1ggo: 63-66). The newly discovered

INGO STRAUCH

manuscripts from Gandhara can be dated slightly later, into the period
between the 1st c. BC and the 2nd c. AD, and thus provide another reliable
date for this process.
In the same time, however, the Brahmins tried to gain ground again.
Especially those among them who specialized in legal affairs and political
sciences and were closely related to the ruling elites began to adopt writing as one of their traditional disciplines of learning (kala). It is possible
to trace this gradual process in the literature of this time, e.g. the introduction of written documents into legal procedure as witnessed in the
Dharmasastra literature (cf. Strauch 2002: 19-52). This development was
accompanied by the growth of a new, mighty religious movement which
is nowadays subsumed under the term Hinduism. The texts of this movement such as the great epos Mahabharata with its influential Bhagavadgita
were no longer the secret knowledge of a small group of specialists but
were widely propagated. Hinduism provided the ideological base for the
spread of the Brahmanical culture over the entire Indian subcontinent
(cf. now Malinar 2009: so-66). The basic means of communication in this
process was Sanskrit, which not only served as lingua franca, but also
regained its status of a religiously legitimized literary language. It is significant that the earliest Indian inscriptions which show clear traces of
Sanskrit phonology are not originating from a Buddhist background, but
can be ascribed to the newly arising Hinduist culture.12 The status of Sanskrit gradually became so strong and influential that even the Buddhists
in India started to sanskritize their texts. 13 Traces of this sanskritization
are felt throughout: not only in early Buddhist literature but also in the
inscriptions of that time (Damsteegt 1978). The Kharo~thi epigraphs show
that also the Gandhar1 speaking area was subject to this process from
about the late 2nd/early 3rd c. AD onwards (Salomon 2001: 141).
In India proper the specialists from the briihmar.za circles had developed the Brahmi script into an adequate instrument for a phonologically
correct reflection of the Sanskrit language. By the end of the 2nd century
AD the Brahmi disposed of a complete inventory of signs and sign combi-

Cf. e.g. the Brahmi epigraphs from Ghosul).Q.i and HathibaQ.a (Salomon 1gg8a: 86f.).
Cf. for the general conditions of this process and the influence from the side of ~rah
manical culture Salomon 2001: 248-251. Controversially discussed is the significance of
writing for the emergence of a new Buddhist movement, the Mahayana, which is said to
have a special attachment to writing and written artifacts like books (for a summary of
the recent discussion and further references cf. Drewes 2oog). As recent research shows,
Mahayana was also influential among Gandharan Buddhism (Strauch 2010 ).
12

'3

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

151

nations in order to express Sanskrit in a non-ambiguous way. The general

consistency of orthography and the graphical solutions which were found
speak in favour of a somehow controlled or at least harmonized process
based on a sound knowledge of Sanskrit phonology and grammar.
The ('Internal Sanskritization" ofKharo$thi

But how did the proponents of the Kharo~thi script react to this Pan-Indian
challenge? The number of manuscripts or inscriptions which belong to
the group of sanskritized or Sanskrit texts is rather limited. Nearly all of
them are quite late and are datable to the 3rd c. AD or even later, i.e. into
the final phase of Kharo~thl (cf. Salomon 2001). Fortunately, among the
texts of the Bajaur Collection there is one manuscript which seems to
belong to an earlier phase of Kharo~thl Sanskrit writing (= BajC 9 recto).
According to its palaeographical features and the evidence of the collection as a whole it should not be later than the 2nd c. AD (Strauch 2oo8:
108-111). The manuscript contains a collection of verses which belong to
the Brahmanical genre of Nlti literature, i.e. political science.14 This text
shows that the process of sanskritization did not only involve a linguistic
shift within the boundaries of Buddhist literature but did also include a
cultural change which implied a more intensive confrontation with new
branches of non-Buddhist literature composed in Sanskrit.
The Bajaur manuscript BajC 9 is written in a conventional Kharo~thl
using the typical Arapacana signs and its modification markers. There is
no extensive use of newly created conjunct signs or special indicators for
vowel quantity which are typical for later varieties of Sanskrit in Kharo~thl
script. Therefore, at the first glance the text looks like ordinary Gandhari.
But phonetic features like the use of external and internal sandhi and
morphological forms which are characteristic only for Sanskrit reveal its
true language. Moreover, the text is composed in the Arya meter.
In most cases it is possible to reconstruct from the defective Kharo~thl
spelling the correct Sanskrit text. The principles which were used for writing Sanskrit can be demonstrated on the base of one of the verses which
lists the components of the royal income (BajC 9 recto, verse 8):

14

For more information about this text see Strauch 2oo8: 125-127.

INGO STRAUCH

152

Fig. 1. Extract from the manuscript BajC g, recto.

Kharo~thi

spelling

dhal).a-dhanya-kupya-yavase1J'ldhal).{e} J).i
yatrayudhani ea rathaca
upakaral).ani ea koso
naravahana-sipi-yodhaca

Sanskrit reconstruction

dhana-dhanya-kupya-yavasendhanimi
yantrayudhimi ea rathasca
upakaral).imi ea koso
naravahana-silpi-yodhasca

Translation: The treasure (are) money, grain, forest produce, 15 grass,

machines and weapons, chariots and instruments, men, vehicles, craftsmen and soldiers.
For giving a more complete picture of the overall appearance of the manuscript, the following survey will be complemented by selected evidence
from other portions of the same text.
Vowel quantity (above= bold print)
Throughout the manuscript the quantities of the vowels are not indicated. Other text portions show that also diphthongs are not indicated
but represented by their respective monophthongs (ceva = Skt. caiva ).
The sonant r is expressed either by ri (nripati = Skt. nrpati) or ir (hirdaya = Skt. hrdaya ), or-in a labial environment-by ur (vurdhi = Skt.
vrddhi).

1.

Simple consonants
With the exception of the confusion of dental and retroflex nasal (n/7J-),
which is typical for most of the varieties of Kharo~thi and reflects the supposed indistinctiveness of the pronunciation of both sounds all simple
(short) consonants are represented by their expected Kharo~thi equivalent (cf. above table 2 ).
2.

This connotation of kupya is obvious from the chapter 2.17 of the Kautillya
Arthasastra which is exclusively devoted to the duties of the "Director of forest produce"
(kupyiidhyak$a) . See also Kangle (1969: Glossary, s.v.).
'5

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

3 Consonant clusters (above

153

= cursive print)

Consonant clusters are indicated generally only in cases when the conventional inventory of Kharo~thi offers a possible grapheme. Geminated
consonants are not indicated (t = Skt. tt).
Semivowel diacritics (-y, -r, r-, v-)
Clusters containing the semivowels y and rare usually indicated by means
of the primary modifiers of the Kharo~thi, like above nya, pya, tra. As an
example for preconsonantal r atmartha-: Skt. iitmiirtha- and durge: Skt.
durge can be cited, postconsonantal v is found in kritva = Skt /qtvii.
3.1

Basic Arapacana signs expressing sound combinations

Although the phonetic realization of these signs (Arapacana no. 25, 26, 38)
is disputed, the scribe of BajC 9 apparently perceived them as representatives of Old Indian sound clusters. Thus we find in the text k$a (no. 25) for
Skt. k$a (kosak$aya = Skt. kosak$aya), sta (no. 26) for Skt. sta (sastre$U =
Skt. siistre$U), and ka (no. 38) for Skt. ska (sarrzkrita = Skt. sarrzskr;ta) (cf.
above table 2 ).
3.2

Secondarily modified Arapacana signs

In other cases the Kharo~thi either uses the possibility of additional modifying signs, e.g. a bar above the letter (ea= sea) or an anusvara indicating the class nasal (rrzdha = ndha ). The horizontal bar is also used for the

33

r ),

a
sound combination jfia which is depicted as fia with bar above (
sign which is so far unattested in Kharo~thi palaeography. Another unat-

1-) which is used beside unmodified

tested combination is found in tfia (

f)

tha (= Arapacana, no. 19,

occasionally for Skt. stha. A different internal
derivation for this sound combination which is based on the Arapacana
sign sta (no.

26)

is found in later Central Asian Kharo~thi

(f., sthi, Glass

2000: 133)

Another special modification is represented by clusters containing a

pre- or postcorisonantal l, where the vertical lower end of the basic sign
J

j. ),

is crossed by a stroke: klarrz ( ), kle (

lkarrz ( ). This rarely attested
device can be associated with the ligature lpi found in the Kharo~th'i legends of Vima Kadphises (I') and in later Central Asian
(cf. Glass 2ooo: 132 ).

Kharo~thi (t')

154

INGO STRAUCH

No indication
Occasionally consonant clusters are not indicated at all, even if the conventional Kharo~?thi would provide a solution (tr = ntr, p = lp).
34

35 Combined signs
Only in very rare cases the scribe makes use of combined signs which join
basic letters. Most of them are also attested in contemporary Kharo~?thi
records, like e.g. tma :Z. (atmartha = Skt. iitmiirtha ).
Furthermore, the Kharo~?thi of the Bajaur manuscript does not indicate
Visarga or final consonant.
It is quite obvious that this Bajaur text represents an early attempt to
cross the language border towards Sanskrit on the basis of the instrumentarium developed within the Kharo~?thi script. I would like to call this
process 11 Intemal Sanskritization of Kharo~?thi". It has to be distinguished
from the later phase which makes extensive use of the tools which have
been developed and successfully used by the Brahmi alphabet. Due to this
indebtedness to an external source this later type can be named ~~External
Sanskritization of Kharo~?thi".
The interaction ofBriihml and Kharo$thi

While the influence of Brahmi might have been rather weak in the Asokan
period16 it became more substantial in the subsequent centuries when
Indo-Greek and Indo-Scythian rulers held sway over the North-West. The
peak of this development was certainly reached under the dynasty of the
Kul?al)as. From the end of the 1st c. AD onwards they started to establish an
empire which united Bactria, the Hindukush area, the Indian North-West
and the north of India up to Pataliputra, i.e. modem Patna, thus covering
a huge area where both Kharo~?thi and Brahmi were in use. Although the
Kul?~as did not actively promote any of both scripts, the social and economic mobility within the boundaries of their empire and with its direct
neighbours, the Western K~?atrapas and Satavahanas in Gujarat and on the
Deccan, resulted also in an interaction of both writing systems. We do not
only find the occasional use of Brahmi and Kharo~?thi in areas which were

' 6 That biscriptuality was an old phenomenon, is shown by the case of the scribe Capac.la
who produced parts of the Asokan edicts at Brahmagiri, Siddapur and Japilga Ramesvara.
He added to the otherwise Brahmi texts his title lipikara "scribe" in Kharo~thi letters (Salomon 1998: 136).

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

155

originally remote to them17 but also the increasing occurrence of BrahmiKharo~thi biscriptuallegends on seals and coins as well in Northern India
as in the Kharo~thi using area in the North-West. It might be interesting to notice that the Ku~al).as themselves did not actively participate in
this process. Instead, under Kani~ka I (after 127 AD) they replaced their
originally biscriptual-bilingual Greek-Kharo~thilegends, which they had
inherited from their predecessors, by a single Bactrian legend.
Other dynasties, however, as the neighbouring K~aharatas who shifted
from the North-West to Gujarat, i.e. into a Brahmi using area, continued the
older system and added a Brahmilegend, a tradition which was continued
from Bhiimaka via Nahapana up to the ~atrapa ruler Damazada (Senior
2001: 194-200 ). The use of Kharo~thi was abandoned when the Middle
Indian legends were replaced by Sanskrit under the ruler Rudradaman.
The opposite development is attested by the coinage of the Parataraja
dynasty who ruled from the 2nd till the 4th c. AD in Baluchistan. The earliest rulers in the middle of the 2nd c. AD used exclusively Brahmilegends
before the dynasty shifted to the use of Kharo~thi which can be safely
dated up to the end of the 3rd c. AD.18 This evidence also shows the sphere
of influence the Brahmi reached during the first centuries AD.
Although it cannot be excluded that some examples of this early
Kharo~thi-Brahmi biscriptuality were accompanied by bilinguality/9 most

For Kharo~thi scribes in the Brahmi area see e.g. the famous Mathura Lion Capital
inscription of the 1st c. AD (CKI 48, see now Falk 2onb) and the Mathura bilingual inscription dated (Ku~iil}.a) year 40 = 167 AD (CKI 440, Chattopadhyaya 1980-82, Bhattacharya
1984). Even more eastern examples are found at Bharhut (Cunningham 1879: 8, pl. VIII)
and Patna (CKI 166, Konow 1929: 177f., plate XXXVI). For a probably imported Ku~iil}.a
period Brahmi inscription found in the Peshawar valley see Falk 2004: 13gf. Earlier examples for the use of Kharo~thi outside its original territory are the 3rd-2nd c. BC biscriptual
inscriptions from the Kangra valley in Himachal Pradesh (Vogel1902-03, CKI 167+168).
18 For more information about the Paratarajas and their coinage see Tandon 2oo6, 2009
and Falk 2007. The chronology used here follows Tandon 2009, the terminus post quem of
the latest Kharo~thi issues can be reliably fixed on the base of overstruck coins (Tandon
2009: 154-156). Other contemporary biscriptual coinages like those of the Audumbaras (ea.
1st c. AD) and the Kunil}.<;l.as show that the introduction of Brahmi cum Kharo~thi legends
was initially mainly restricted to territories and by dynasties which were peripheral to the
core-land of Gandhara and its ruling elites, but located at important trade routes which
connected this area with Brahmi using territories (cf. Chattopadhyaya 2003: 59-60 ).The
much earlier attempt by Agathocles and Pantaleon (190-180 BC) to introduce Brahmi on
their coins is not considered here (cf. Bopearachchi 1991: 175, pl. 7, series 9 HO, 182, pl.
9, serie 6). It remained without further consequences for the development of writing in
the region.
19 A clear case of bilinguality seems to be the Kharo~thi-Brahmi Kanhiiira inscription
(Vogel1902-03, Konow 1929: 178, CKI 168). Here the Sanskrit text /a:$1J.aya.Sasya iiriima
miidagisya written in Brahmi is accompanied by the Gandhari krijayasasa aramo in
17

INGO STRAUCH

biscriptual epigraphs seem to represent the same Middle Indian language.

If a more accurate definition of this language is possible at all, it can be
attributed to the North-Western, i.e. Gandhari speaking region. Thus not
only many of the biscriptual seals (e.g. Konow 1929: 100, 102) but also
the biscriptual Ku~al}.a period Mathura epigraph (Bhattacharya 1984) are
nothing more than Gandhari written in Brahmi script. In many cases the
scribes simply transferred Kharo~tl}.l orthographical usages into the new
script. 20 This clearly shows the direction of assimilation and its agents.
Obviously Kharo~thl scribes tried to use Brahml for their own language,
which might affect the orthography of the Brahmi written by them. As
far as we can observe on the basis of the available evidence, there was no
remarkable tendency to an opposite influence. It seems that as long as
the media of communication was a Middle Indian language there was no
need for a Kharo~thl scribe to seriously revise his script and to leave the
above described internal limits of its instrumentarium. 21
With the beginning of the 3rd c. AD the picture changed. The Middle
Indian language of Gandhara had been reshaped into a kind of sanskritized
lingua franc a which was not very different from the varieties of Sanskrit
which were in use in other parts of the Indian subcontinent. This sanskritized language affected all spheres of writing, i.e. not only epigraphical
records but also the transmission of Buddhist literature. In the beginning

Kharo~thi

(my readings). The inscription which was found in the Kangra valley in modem
Himachal Pradesh seems to be very early and can probably be dated into the 2nd c. BC. It
supports the idea that the preferable medium for writing Sanskrit was Brahmi while the
use of Kharo~thi was at that period confined to Gandhari text.
It is not clear, whether the bronze die with the text (Kharo~thi) sidhatasa, (Brahmi)
sidhata.Sa is also such a case of bilinguality where the Brahmi is meant to designate an
East-Indian dialect-(Ardha-)magadhi-which replaces dental s by palatal s (Archaeological Survey ofIndia, Annual Report 1936/37: 39, pl. X, f,g). It is also possible that it represents
an example for the incorrect use of the sibilant which is typical for the earliest Brahmi as
used by Asoka. See for this seal now Aman ur Rahman & Falk 2on: 186, TM 07.07.01.
20 A possible example for such a transfer of Kharo~thi orthography is the use of preconsontal r to designate a geminated consonant in a Mathura epigraph of the Ku~aQ.a period
(Liiders 1961: 82f., 46): derya = deyya. Such a usage of Kharo~thi orthography can also be
verified in much later documents. Thus we find in the frequently attested spelling seryathii
= seyyathii in Buddhist Sanskrit texts (ibid.: 83) and in a peculiar Pratimo~asiitra manuscript from Qizil of the 6th/7th c. (von Simson 1997) which reads e.g. adima for adinna
(583). For this device in Kharo~thi texts see now Salomon 2oo8a: 97
An interesting case how a "foreign" sound is designated in Kharo~thi are the
Parataraja names Kozana and Koziya. The Kharo~thi coins mentioning them introduced
a new Kharo~thi sign by modifying the letter z with the help of an additional hook added
to its lower right and a circle on its top. The Brahmi parallel seems to be based on the
established ligature ysa /za/ which was, however, not used by the Kharo~thi scribes as a
model for their new sign (Falk 2007).
21

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

157

of this sanskritization Kharo~thi continued to be in use. In the sphere of

Buddhist literature this is impressively proven by the sanskritized Buddhist texts in Kharo~thi script which are part of the British Library, Pelliot
and Scheyen CollectionS.22 According to recent radio-carbon dating the
relevant Scheyen fragments can most probably be dated into the 3rd c.
AD (Allon et al. 2oo6: 288-291).
The inscriptional evidence of this period shows that the use of
Kharo~thi for this sanskritized Gandhari became gradually replaced by
that of Brahml. In many cases the transition is connected with the advent
of foreigners originating from Brahmi using areas (Falk 2004). As far as we
can judge from the available material, this process first concentrated on
two distinctive environments:
A-Non-Buddhist establishments which obviously preferred the use
of Brahmi as a script which was not linked to the Buddhist culture of
Gandhara. As an example for this category the numerous epigraphs discovered in the early Hindu site Kashmir Smast can be cited. They show
how the peculiar-still heavily Gandhari based-sanskritized language
of the region was tried to be reproduced in the newly introduced Brahmi
script (cf. Falk 2001 and Falk 2oo8b: 138-143).23
B-Peripheral places of the Kharo~thi using area with a high factor of
social mobility which were less reluctant to accept a new script. The Buddhist monastery complex Kara-Tepe near Termez (Uzbekistan) can serve
as a paradigmatic place for this category. Here we find the contemporary
use of Brahmi and Kharo~thi probably datable into the 3rd-4th c. AD.24

The Sanskritic features of the respective manuscripts are described in Salomon 1998b
(Pelliot), Salomon 1999: 123, Salomon 2001: 243 (British Library), Allon & Salomon 2000:
266-271, Salomon 2001: 243-247, and Allon et al. 2006:288-290 (Schey-en). The same category of relatively late sanskritized Kharo~?thi texts is represented by a hitherto unpublished
palm-leaf folio from the Kabul Museum which probably hails from a cave at Bamiyan (Falk
& Strauch, forthcoming).
23 Forms such as the ending -sya for the genitive sg. masc. clearly show the Skt. character.
The inconsistency in the designation of vowel quantity and prakritisms or hyper-sanskritisms such as piiJJ.tka and pii1J.tii (for Skt. pii1J.'iya) and pratitthiivita (for Skt. prati!;thiipita-)
indicate the underlying Gandhan phonology.
" 4 The editor of these inscriptions suggests a much later date for some of the bilingual
epigraphs comparing the records of the monk Buddhasira(s) with the Brahmi inscriptions
of Central India and the Deccan of the sth/6th c. AD (Vertogradova 1995: 32). On the base
of the palaeographical arguments presented by her it seems, however, more plausible to
presume a date in the late Ku~?ii:r;ta/early Gupta period, i.e. around the end of the 3rd,
beginning of the 4th c. AD. A still earlier date is now suggested by Gerard Fussman (2on:
22

INGO STRAUCH

A series of biscriptual inscriptions on ceramic vessels25 shows that Brahml

and Kharo~thl virtually reproduced the same sanskritized language.
That the gradual introduction of Brahml did also affect places in the
Gandhan core-land can be shown by a bilingual epigraph from a Buddhist site near Peshawar (Falk 2004: 146f., plate V, d) and by a vessel
with a Brahml inscription from a Buddhist monastery near HaQ.Q.a called
mahaseniiralJ.ya (pot 1: Sadakata 1999: 65-66, plate 11). Interestingly, the
same monastery name is attested on another ceramic pot inscribed in
Kharo~thl (pot 2: CKI 360, Salomon 1gg6: 238-242, plates 8-14). Here we
read in sanskritized orthography: saghe cadurdise masenaranye26 To the
community of the four directions at the Mahasena monastery." Obviously the Mahasenara:Q.ya has completed the shift from Kharo~thl to
Brahml in the period between the production of both vessels. It seems
that this shift was preceded or accompanied by a change of Kharo~thl
orthography. The initial phase of this development can be illustrated
with the help of another pair of pots. The same sanskritized spelling as
found on the pot from the Ma(ha )senara:r:tya is seen on a yet unpublished
vessel from the Mahapriyara:Q.ya monastery near HaQ.Q.a. Here the same
unusual ligature 1J.Ya and additionally even a long ii occur: mahiipriaralJ.ye
(pot 3: Strauch, forthcoming). This evidence can be contrasted with
the variant mahapriyaraJTlfie in conventional Kharo~thl orthography
as found on another vessel from the same monastery (pot 4: CKI 223,
Fussman 1g6g).27
All these vessel inscriptions represent different stages in the development from conventional Kharo~thl to Brahml, caused by the sanskritization of the North-Western language:

41-45) who dates the biscriptual texts of Buddhasira(s) between so and 200 AD while
jivananda is placed by him between 150 and 250 AD.
zs The multiscriptual inscriptions of the Kara-Tepe monks Jivananda and Buddha8ira( s)
are published by V.V. Vertogradova (1995: 106-113, 2004: 69-72) and Fussman 2011: 63-88
(Buddhasira(s) alias Buddhamitra) and 67 (Jivananda). Another triscriptual text of
Jivananda which was not recognized as such seems to be 51 KT (Fussman 2011: 75f.). The
Kharo~thi text on sherd 51 KT b can clearly be read as [i]i[va]na[J?'l]das[y]a. The script is
identical with that used on the other pots of this person.
zs What Salomon interpretes as "long, decorative extensions to the last syllable of a
word" (1gg6: 239) can now be safely identified as subscribed -ya. Hence his readings masenarane and budhagho$asa have to be corrected to masenaranye and budhaghot;asya. For
masena as contracted form of mahiisena cf. Salomon 1gg6: 242.
z7 As Salomon (1ggg: 243) rightly points out, Fussman's reading mahapriasarrrfie has to
be corrected.

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

Language

Gandhari

Script

Conventional Kharo~t}ll Modified Kharo~thi Brahmi

159

Sanskritized language

MahasenaraiJ.ya
Mahapriyara:Q.ya mahapriyararpfte
(pot 4)

masenaranye
(pot 1)
mahtipriaral)..ye
(pots)

mahiisentiral)..ye
(pot 2)

As this evidence shows, the process of sanskritizing did not only affect the

literary and administrative language of the region but also had a considerable impact on the orthography of the Kharo~thi. As mentioned before, of
both scripts it was the Brahmi which had already found the basic answers
to the challenge of sanskritization. What would be more natural than the
attempt to use these answers for the adjustment of Kharo~thi?

The {{External Sanskritization ofKharo$th"i

11

The following short survey will list the most characteristic changes
of Kharo~thi in the late phase of its adaption to Sanskrit phonology as
witnessed by the few examples of sanskritized Kharo~thi texts. 28 Beside
the few texts of the Scheyen, British Library and Pelliot collections the
most extensive repertoire for this late type of Kharo~thi is represented by
the documents on wood and leather discovered in the beginning of the
2oth c. by Aurel Stein in Niya and Endere (Boyer et al. 1920-29 ). Most of
these documents are written in the administrative language of the Kroraina kingdom which is based on Gandhari but largely influenced by an
underlying dialect related to Tokharian (cf. Burrow 1937 ). Some of them,
however, contain Buddhist verses which are composed in the Buddhist
Sanskrit of the period (Iwamatsu 1998, 2000, 2001, 2002 ). The entire corpus
can be dated into the 3rd/4th c. AD (Brough 1965). Both the foreign elements of the language as well as the desire to express Sanskrit affected the
development of the Niya Kharo~thi which is abundant in newly created
sign modifications and combinations (Boyer et al. 1920-29: 295-322 ).
In the late Sanskrit or. sanskritized Kharo~thi texts consonant clusters
are usually formed by writing the respective basic signs one above the

Since the number of these texts can be expected to increase and only some of the
known texts are published so far, the survey is just a cursory overview to indicate the general characteristics. The major texts which are available at present are listed in Salomon
2001 and Allon et al. 2006.
28

INGO STRAUCH

160

other. This method clearly adopts the way which is the standard one for
Brahmi. As opposed the "internal" way of expressing such combinations
by non-standardized diacritics, this approach guarantees a phonologically adequate reading. 29 The necessary unambiguity is, however, never
reached, since all Kharo~thi texts attested so far simultaneously continue
the less distinct ''internal method". Thus despite its otherwise sanskritized
or-better-brahmiized character the Scheyen manuscript applies s for
Skt. sna and] for Skt.jha (Allon & Salomon 2ooo: 267).
Table 4.30 Selected consonant clusters in late Kharo~thi as opposed to the
orthography of the manuscript BajC 9
External Sanskritization
Skt.

Scheyen
Collection

Pelliot
Collection

Niya
documents

Internal Sanskritization
Bajaur fragment 9

kta
jfia

ta

fia
ta

tta

tha
stha

tiia

lpa

lkarp.

29

An interesting and rather early example for this device is found with the ligature

(Ku~fu}.a)

$thu (t) on the two Wardak vase inscriptions dated into the
year 51(= 178 AD)
(CKI 159 = Konow 1929: 165-179, CKI 509 = Falk 2oo8a). Remarkably, it occurs only in the
foreign, probably Iranian, name Ha~thuna. The two dots above the final letter (nii) seem to
indicate another orthographical peculiarity which was probably introduced into Kharo~thi
from a foreign writing system. Konow (1929: 166) compares this device with the vowel designation -ii of Saka Brahmi texts. Usually such double dot above a consonant sign is used
to designate the Skt. Visarga and probably derived from the corresponding Brahmi sign.
Cf. also the discussion in Salomon 1998: 131, 143 and Glass 2ooo: 137f.
30 With the exception of stha (Allon & Salomon 2000: pl. X,1) and the BajC signs, all
Kharo~thi signs are extracted from Glass 2000.

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

Table 5.31 Additional signs of late

Kharo~thi

Vowel quantity

and their Brahmi models

Virama

Visarga

Brahmi

'(

yal).

Kharo~thi

tal).

161

W
#

tarn

dhik

]>6

Other orthographical needs for writing Sanskrit were only sporadically

satisfied. Thus the Visarga is found only in the few Kharo~thl manuscripts
from the Northern Silk Road which are now part of the Pelliot Collection
(Salomon 1998b) and in few of the later Kharo~thl tablets from Central
Asia (cf. ibid.: 131). In Ku~a:r:ta times the Brahml had started to write this
sound as double dot on the right side of a letter. The new Kharo~thl sign
which arranges these dots on the top of the basic sign can be interpreted
as an adaptation of this method. Its occurrence in Central Asia is clearly
based on the prominent presence of Brahml in this area.
Another feature which was introduced rather late into Kharo~thl is the
indication of vowel quantity. The earliest attempt seems to be represented
in the Avadana texts of the British Library Collection scribe 2 from about
the 2nd century which is restricted, however, to the initial a (Lenz 2010:
17). The long quality of the vowel is indicated by a stroke to the lower right
of the letter which can be associated to the shape of the -a diacritic as
found in some varieties of Ku~a:r:ta Brahml (cf. table 5). The same practice
is also occasionally attested in epigraphical records, generally in the case
of inscriptions which show Sanskrit influence and were written in a biscriptual environment (e.g. Kara-Tepe, Vertogradova 1995: 19f. ). It remained
in use in Central Asia where we find the only text of Kharo~thl literature
which shows a consequent approach to the indication of vowel quantity
(Niya document no. 523, Boyer et al. 1920-29: 191, tr. Burrow 1937: 103).
Remarkably, this tablet contains also one of the few Kharo~thl evidences
for the indication of.final consonants. In Ku~a:r:ta Brahml this problem was
solved by adding
a consonant sign in a smaller size than the preceding.
'
The same device was used by the scribe of Niya document 523, obviously
under the influence of the Brahml orthography (cf. Glass 2ooo: 138). This
.

31

The Kharot?thi signs are extracted from Glass 2ooo, the Kut?aJ;Ia Brahmi signs

a and

yal;t from the Indoskript database (http://userpage.fu-berlin.de/-falk/), the tam from

Sander 1968: table 11.

INGO STRAUCH

influence is also traceable in the shape of the numerals 1-3 written as

horizontal strokes instead of the usual verticals. The same brahmiized
numerals are found in some of the Sch0yen manuscripts (Glass 2ooo: 13gf. ).
These few observations might be sufficient to characterize the main
features of the external sanskritization of Kharol?thi by the end of the 3rd
c. AD. The technical realization of Sanskrit phonology as well as the contexts of the inscriptions and manuscripts indicate that this process took
place in a distinctively biscriptual Brahmi-Kharol?thi environment where
the Kharol?thi was exposed to the principles of Brahmi orthography and
tried to adapt them for the reproduction of the newly introduced Sanskrit
language.
The influence of Brahmi on the Kharol?thi orthography, however, did
never result in a consistent adjustment ofKharol?thi to the Sanskrit phonology. Besides adapting Brahmi orthographical devices the Kharol?thi scribes
continued to use a multitude of diacritical modifying signs and never tried
to reduce the inventory of letters down to a reasonable number.
Conclusions

Kharol?thi was designed for the specific needs of the North-Western language Gandhari and remained closely linked to this region and language
throughout its use. As shown by its own inventory of signs-as represented
in the Arapacana alphabet-it is a distinctly phonetic script which tries to
reproduce phonetic variants by a multitude of basic signs and supplementary modifiers. An attempt has never been made to transform the script
into a phonological writing system which confines itself to the consequent
tagging of phonemes. Moreover, the phonetic character of the script was
never regulated by commonly accepted orthographical standards.
The attested attempts to use Kharol?thi for writing Sanskrit or a sanskritized language can be attributed to two different and clearly distinguishable
approaches. While the first one-which I call"Internal Sanskritization"uses the instruments developed within the limits of Kharol?thi writing
by applying the basic Arapacana signs and their primary and secondary
modifications, the second approach is clearly influenced by the methods
which are peculiar for the Brahmi. This "External Sanskritization", which
could also be labeled as "Brahmiization", is characterized by the incieasing use of combined signs (ligatures) and graphical devices which are
otherwise confined to the Brahmi script (Visarga, Virama). Although the
((internally sanskritized" Kharol?thi seems to be historically older, it did not

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTIOI'l"

163

become obsolete after the application of the external (Brahm1) devices.

Both methods continued to be used within the same text up to the very
end of the Kharo!?th1 period.
To enable Kharo!?th1 to cross its linguistic limits and to be used for
Sanskrit would have required a serious reform of the writing system
which would have consequently adjusted its inventory of signs to the
needs of Sanskrit phonology. Such a reform would have presupposed
strong cultural or political agent with a profound interest in maintaining
Kharo!?th1 as a writing system. It seems that this interest did not arise in
Gandhara proper when by the end of the 3rd c. AD the use of Gandhar1
as Lingua franca and literary language came to an end and had to give
way to Sanskrit. Neither the political elites nor the Buddhist institutions
resisted when this linguistic shift was accompanied by the introduction
of a new script which was much better adapted to Sanskrit phonology
and in this regard much superior to Kharo!?thl. Although the Buddhists
had made a certain attempt to sacralize the Kharo!?th1 script by ascribing
the Arapacana alphabet a magic and ritual power, this did not result in a
status which would have prevented their shift to the new alphabet. The
different versions of the Lalitavistara story of the young Bodhisattva as a
school-boy show impressively how pragmatically the Buddhists handled
this problem. In this story the future Buddha is supposed to learn the
alphabet. Of course, due to his outstanding capacities his knowledge of
scripts supersedes that of his teacher by many times. Altogether 64 different scripts are enumerated by the future Buddha, the first of them being
Brahm1 and Kharo!?thl. Moreover, when the school-master starts to teach
the alphabet, the present young boys and the Bodhisattva do not only
learn the respective letters but also a great number of religious concepts
which are associated to each of them. While in the older version, preserved in Dharmarak!?a's Chinese translation of this text (308 AD), it is the
Arapacana order which the teacher is reciting (Brough 1977 ), its younger
Sanskrit parallel rearranges the whole passage according to the sequence
of the Sanskrit van:zamiilii.
The continuing use of Kharo!?th1 in some other areas was mainly caused
by the status ,the Gandhari language pertained in these peripheral communities. As long as Gandhari or a related dialect was in use as administrative or literary language, Kharo!?th1 continued to be used as well. When
the Gandhar1 lost this status and began to be replaced by Sanskrit the
use of Kharo!?th1 came to an end as well. The gradual displacement of
Kharo!?th1 was consequently the direct result of the gradual sanskritization of its cultural environment.

INGO STRAUCH

Accordingly, the mysterious decline of Kharo~thi is not so much due

to a dynastical or political change, as assumed by Richard Salomon,32 but
to a cultural and linguistic shift which is known under the term sanskritization. The new language brought along a well-adjusted writing system
which was readily adapted in the writing culture of the area.
As shown above, this pragmatical shift from Kharo~thi to Brahmi had
both internal and external reasons. While the phonetic character of the
Kharo~thi script and its missing orthographical standardization were serious internal barriers, it was mainly the existence of the superior Brahmi
alphabet which externally prevented the Kharo~thi to cross the border
which was erected by the advent of the Sanskrit culture.
Acknowledgements

I am very grateful to Harry Falk with whom I discussed many aspects of

the problems raised in this paper. He also read the final draft and made
a series of valuable suggestions. The Kharo~thi signs from manuscripts of
the Bajaur Collection were extracted by Andrea Schlosser. Many thanks
go to my friend Gordon Adie and to Heidrun Wenisch, who checked the
English of one of the "final" versions. All remaining mistakes are, of course,
exclusively my fault.
References
Allon, Mark 2007. The Senior manuscripts. In Andrew Glass: Four Gandhar"i Sarrzyuktagama
Siltras: Senior Kharo$thl Fragment 5 Gandhiiran Buddhist texts 4, Seattle: University of
Washington Press: 3-25.
Allon, Mark & Richard Salomon 2000. Kharo:?thi fragments of a Giindhiiri version of the
Mahiiparinirviil}.asiitra. In Jens Braarvig (ed.): Buddhist Manuscripts, volume I. Manuscripts in the Scheyen Collection I. Oslo: Hermes Academic Publishing: 243-73.
Allon, M., Salomon, R. Jacobsen, G. & U. Zoppi 2006. Radiocarbon dating of Kharo:?thi
fragments from the Sch~yen and Senior manuscript collections. In jens Braarvig (ed.):
Buddhist Manuscripts, volume Ill. Manuscripts in the Sch~yen Collection. Oslo: Hermes
Academic Publishing: 279-91.
Aman ur Rahman & Harry Falk 2011. Seals, sealings and tokens from Gandhara. Monographien zur indischen Archaologie, Kunst und Philologie, Band 20 = Studies in the Aman
ur Rahman Collection, 1. Wiesbaden: Reichert Verlag.

It is hardly possible to qualify Kharo:?thi as "bureaucratic script of the Ku:?iiQ.a empire"

(Salomon 2oo8b: 149). The only "official" epigraphical records of this dynasty are written
in Bactrian, which was certainly the preferred language. What we know of the language
policy seems to indicate that the language and script of the respective region was used in
administrative affairs.
32

INDIAN KHARO~'fHI SCRIPT AND THE "SANSKRIT REVOLUTION"

165

Baums, Stefan 2009. A Giindhiiri: Commentary on Early Buddhist Verses: British Library
Kharo$thl Fragments 7, 9, 13 and 18. PhD Dissertation. University of Washington.
Bhattacharya, Gouriswar 1984. On the fragmentary, bi-scriptual pedestal inscription from
Mathura. Indian Museum Bulletin 19: 27-30.
Bopearachchi, Osmund 1991. Monnaies greco-bactriennes et indo-grecques. Catalogue raisonne. Paris: Bibliotheque Nationale.
Boyer, A.M., E.J. Rapson, E. Senart, & J. Noble 1920-9. Kharo$thl Inscriptions discovered
by Sir Aurel Stein in Chinese Turkestan. 3 parts (part 3 by E. J. Rapson and P. S. Noble).
Oxford: Clarendon Press.
Brough, John 1962. The Giindhiiri: Dharmapada. London Oriental Series 7 London: Oxford
University Press.
- - 1965. Comments on third-century Shan-shan and the history of Buddhism. Bulletin
of the School of Oriental and African Studies 28: 582-612.
- - 1977. The Arapacana syllabary in the old Lalita-vistara. Bulletin of the School of Oriental and African Studies 40: 85-95.
Burrow, Thomas 1937. The Language of the Kharo$thi Documents from Chinese Turkestan.
Cambridge: Cambridge University Press.
- - 1940. A Translation of the Kharo$thi Documents from Chinese Turkestan. James G.
Forlong Fund 20. London: Royal Asiatic Society.
Chattopadhyaya, B.D. 1980-2. On a hi-scriptural epigraph of the Ku~fu.J.a period from
Mathura (with a note on eghatta and ehat;la by K. Meenakshi).joumal ofAncient Indian
History 13: 277-84.
- - 2003. Studying Early India: Archaeology, Texts and Historical Issues. Delhi.
Coningham, R.A.E., Allchin, R.R., Batt C.M. & D. Lucy. 1996. Passage to India? Anuradhapura and the Early Use of the Brahmi Script. Cambridge Archaeological journal 6:
73-97
Cunningham, Alexander 1879. The Stftpa of Bharhut. London: W.H. Alien and Co.
Damsteegt, Th. 1978. Epigraphical Hybrid Sanskrit: Its Rise, Spread, Characteristics and Relationship to Buddhist Hybrid Sanskrit. Orientalia Rheno-Traiectina, volumen vicesimum
tertium. Leiden: E. J. Brill.
Dani, Ahmad Hasan 1963. Indian palaeography. Oxford: Clarendon Press.
Daniels, Peter T. & William Bright (eds.): The World's Writing Systems. New York: Oxford
University Press.
Drewes, David 2009. Early Indian Mahayana Buddhism I: Recent Scholarship. Religion
Compass 4, 2: 55-65. (http:/ /www3.interscience.wiley.com/joumal/n7982875/home,
accessed 8.3.2010 ).
Falk, Harry 1993. Schrift im alten Indien: ein Forschungsbericht mit Anmerkungen. ScriptOralia, 56. Tiibingen: Gunter Narr Verlag.
- - 1996. Aramaic script and the Kharo~thi-a comparison. Berliner Indologische Studien 9-10: 151-6.
- - 2003. A copper plate donation record and some seals from the Kashmir Smast.
Beitriige zur allgemeinen und vergleichenden Archiiologie 23: 1-19.
- - 2004. Six early Brahmi inscriptions from Gandhara. Universita degli Studi di Napoli
"L'Orientale~ Annali 64:139-55.
- - 2006. Asokan Sites and Artefacts: a Source-Book with Bibliography. Monographien zur
indischen Archaologie, Kunst und Philologie, Band 18. Mainz am Rhein: Verlag Philipp
von Zabem. '.
- - 2007. The names of the Paratarajas issuing coins with Kharo~thi legends. The Numismatic Chronicle: 171-8.
- - 2oo8a. Another reliquary vase from Wardak and consecrating fire rites in Gandhara.
In Claudine Bautze-Picron (ed. ): Religion and art: new issues in Indian iconography and
iconology. London: The British Association for South Asian Studies: 63-80.
- - 2oo8b. Money can buy me heaven: religious donations in late and post-Kushan
India. Archiiologische Mitteilungen aus Iran und Turan 40: 137-48.

166

INGO STRAUCH

- - 2oua. The 'Split' Collection of Kharo~thi texts. Annual Report of The International
Research Institute for Advanced Buddhology at Soka University for the Academic Year 2010
[ = ARIRL4B] 14: 13-23.
- - 2011b. Ten thoughts on the Mathura Lion capital reliquary. In Shailendra Bhandare
& Sanjay Garg (eds. ). Felicitas: Essays in Numismatics, Epigraphy and History in Honour
of]oe Cribb. Mumbai: Reesha Books International: 121-141.
Falk, Harry & Ingo Strauch. Forthcoming. The Bajaur and Split Collections of Kharo~thi
manuscripts within the context of Buddhist Gandhari literature. In Paul Harrison &
Jens-Uwe Hartmann (eds.): Proceedings of the conference "Indic Buddhist Manuscripts:
The State of the Field'~ Stanford. June 15-19, 2009.
Fussman, Gerard 1969. Une inscription kharo~thi a Hac;lc;la. Bulletin de l':Ecole .franfaise
d'Extreme-Orient 56: 5-9.
- - 20n. Monuments bouddhiques de Termez. Termez Buddhist Monuments. L Catalogue
des inscriptions sur poteries (avec une contribution de Nicholas Sims-Williams et la collaboration d'Eric Ollivier). College de France. Publications de l'Institut de civilisation
indienne, 79 Paris: Diffusion de Boccard.
Glass, Andrew 2000. A Preliminary Study ofKharo$th Manuscript Paleography. MA thesis.
Department of Asian Languages and Literature, University of Washington.
- - 2007. Four Giindhiir[ SaTflyuktiigama Sutras: Senior Kharo$th Fragment 5 Gandharan
Buddhist Texts, 4 Seattle: University ofWashington Press.
von Hiniiber, Oskar 1990. Der Beginn der Schrift und.friihe Schriftlichkeit in Indien. Abhandlungen der geistes- und sozialwissenschaftlichen Klasse, Jahrgang 1989, Nr. 11. Mainz:
Akademie der Wissenschaften und der Literatur.
Iwamatsu, Asao (::!i5-f'~~X) 1998. Karoshuti bunjochii no Bukkyo bongo no geju ni tsuite
(1) 1J P -v ::z. 7--1 -Jt:ff9='0){.b.~1t~0){~~~;:0v\-c (1) (On some Buddhist
Hybrid Sanskrit verses in the Kharo~thi inscriptions mainly discovered by Sir Aurel
Stein in Chinese Turkestan).jinbun-Ronshu AJt~iiiJ~ (Studies in Humanities) 10:1-16.
- - 2ooo. Karoshuti bunjochii no Bukkyo bongo no geju ni tsuite (2) 1J P -v ::z. 7--1
- Jt:ff 9=' O){.b.~1t~0){~1.l ~;: 0 v \ -c (2) (On some Buddhist Hybrid Sanskrit verses
in the Kharo~thi inscriptions mainly discovered by Sir Aurel Stein in Chinese Turkestan). Siruku Rodo kenkyu v /v:7 P - ]"1i}f~-e :/' ~- (Silk Road Studies) 2: 21-35.
- - 2001. Karoshuti bunjo daiban 5u ni tsuite (1) 1J P - v ::z. T-{ - Jt:J::~Jlm:5u~;:
0 v\ -c (1) (On the no. sn text of the Kharo~thi inscriptions). ]inbun-Ronshu AJtfnil~
(Studies in Humanities) 13: 157-91.
- - 2002. "Karoshuti bunjo daiban 5n ni tsuite (2) 1J P -v ::z. 7--1 -Jt:ff~ii511~;:
0 v\-c (2) (On the no. 511 text of the Kharo~thi inscriptions ).jinbun-Ronshu AJtfnil~
(Studies in Humanities) 14: 183-204.
Kangle, R.P. (ed.) 1969. The KautilTya Artha.Siistra. Part 1. A critical edition with a glossary.
Bombay.
Koizumi, Yoshihid.e (;J\* 1l~) 2007. "Zaru Deri iseki shutsudo ishi chogun no fukugenteki kosatsu -!f-;v7'!) -ii:H1J);tf:l~~~O)i;l7CiJ~~." MUSEUM: Tokyo
kokuritsu hakubutsukan kenkyu shi MUSEUM: JR~ 00 .lz:tf!JW!g1i]f~w 6o6: 5-22.
Konow, Sten 1929. Kharoshth Inscriptions with the Exception of Those of Asoka. Corpus Inscriptionum Indicarum 2.1. Calcutta: Government of India, Central Publication
Branch.
Lenz, Timothy 2010. Gandhiiran Avadiinas: British Library Kharo$tht Fragments 1-3 and 21
and Supplementary Fragments A-C. Gandharan Buddhist texts 6. Seattle: University of
Washington Press.
Liiders, Heinrich 1961. Mathurii inscriptions. Unpublished papers. Abhandlungen der Akademie der Wissenschaften in Gottingen. Phil.-hist. Kl~ Dritte Folge 47 Gottingen.
Malinar, Angelika 2009. Hinduismus. Studium Religionen. Gottingen: Vandenhoeck &
Ruprecht.
Pagel, Ulrich 2007. Mapping the Path: Vajrapadas in Mahiiyiina Literature. Studia Philologica Buddhica. Monograph Series, XXI. Tokyo: The International Institute for Buddhist
Studies of The International College for Postgraduate Buddhist Studies.

INDIAN KHARO~THI SCRIPT AND THE ''SANSKRIT REVOLUTION"

167

Sadakata, Akira 1999. Girugitto shutsudo oyobi Bamiyan shutsudo no Bukkyo kankei no
Moji shinyo (Buddhist Manuscripts and Inscriptions Found in Gilgit and in Bamiyan).
Tokai Daigaku Kiyo Bungakubu (Bulletin of the Faculty of Letters, Tokai University) 71:
55-74
Salomon, Richard 1990. New evidence for a Gandhari origin of the Arapacana syllabary.
journal of the American Oriental Society no: 255-73.
- - 1996. Five Kharol?thi inscriptions. Bulletin of the Asia Institute 10: 233-46.
- - 1998a. Indian Epigraphy: A Guide to the Study ofInscriptions in Sanskrit, Prakrit, and
the Other Indo-Aryan Languages. South Asia Research. New York: Oxford University
Press.
- - 1998b. Kharol?th1 manuscript fragments in the Pelliot collection, Bibliotheque
nationale de France. Bulletin d'etudes indiennes 16: 123-60.
- - 1999. Ancient Buddhist Scrolls from Gandhiira: the British Library Kharo$thi Fragments. Seattle: University ofWashington Press.
- - 2001. 'Gandhari Hybrid Sanskrit': new sources for the study of the Sanskritization of
Buddhist literature. Indo-Iranianjournal 44: 241-52.
- - 2004. An Arapacana abecedary from Kara Tepe (Termez, Uzbekistan). Bulletin of the
Asia Institute 18: 43-51.
- - 2006. Kharol?thi syllables used as location markers in Gandharan stiipa architecture.
In Pierfrancesco Callieri (ed. ): Architetti, capomastri, artigiani: l'organizzazione dei cantieri e delta produzione artistica nell'Asia ellenistica: studi offerti a Domenico Faccenna nel
suo ottantesimo compleanno. Serie orientale Roma, vol. C. Roma: Istituto Italiano per
!'Africa e l'Oriente: 181-224.
- - 2oo8a. Two GiindhiiriManuscripts ofthe Songs ofLakeAnavatapta (Anavatapta-giithii):
British Library Kharo$thi Fragment 1 and Senior Scroll14. Gandharan Buddhist Texts, 5
Seattle: University ofWashington Press.
- - 2oo8b. Whatever happened to Kharol?thi? The fate of a forgotten Indic script. In John
Baines, John Bennet and Stephen Houston (eds. ): The Disappearance ofWriting Systems:
Perspectives on Literacy and Communication. London: Equinox Publishing: 139-55.
Sander, Lore 1968. Paliiographisches zu den Sanskrithandschriften der Berliner Tuifansammlung. Verzeichnis der Orientalischen Handschriften in Deutschland, Supplementband 8.
Wiesbaden: Steiner.
Senior, Robert 2001. Indo-Scythian Coins and History. Volume 11. Lancaster, Pennsylvania:
Classical Numismatic Group.
von Simson, Georg 1997. Eine Pratimokl?asiitra-Handschrift in hybrider Sprache. In Petra
Kieffer-Piilz & Jens-Uwe Hartmann (eds. ): Bauddhavidyiisudhiikaral;.. Studies in Honour
of Heinz Bechert. On the Occasion of His 6sth Birthday. Indica et Tibetica 30, SwisttalOdendorf: Indica et Tibetica Verlag: 583-604.
Strauch, Ingo 2002. Die Lekhapaddhati-Lekhapaficasikii. Briefe und Urkunden im mittelalterlichen Gujarat. Text, Ubersetzung, Kommentar, Glossar. Monographien zur Indischen Archaologie, Kunst und Philologie, Band 16. Berlin: Dietrich Reimer Verlag.
- - 2005. Urbanisierung, Antiurbanismus und Deurbanisierung: Wege zur Stadt im alten
Indien. In Harry Falk (ed.): Wege zur Stadt. Entwicklung und Formen urbanen Lebens in
der alten Welt. Vergleichende Studien zu Antike und Orient 2. Bremen: Ute Hempen
Verlag: 121-157.
- - 2008. The Bajaur collection of Kharol?th1 manuscripts-a preliminary survey. Studien zur Indologie und Iranistik 25: 103-36.
- - 2010. More missing pieces of Early Pure Land Buddhism: New evidence for Akl?obhya
and Abhirati in an early Mahayana siitra from Gandhara. The Eastern Buddhist 41:
23-66.
- - forthcoming. Inscribed vessels and the geography of ancient Hac;lc;la (working
title).
Tandon, Pankaj 2006. New light on the Paratarajas. Numismatic Chronicle 166: 173-209.
- - 2009. Further Light on the Paratarajas: an Absolute Chronology of the Brahmi and
Kharol?th1 Series. Numismatic Chronicle 168: 137-171.

168

INGO STRAUCH

Vertogradova, V.V.199S Inditska7:a epigrajika iz Kara-tepe v Starom Temeze: problemy

deshifrovki i interpretafSii. HniJuucKWI. 3nuzpaffiuKa U3 Kapa-mene B CmapoM. TepM.e3e:

npo6AeM.bt iJeutuffipoBKU u unmepnpemalJUU. Moskva: Izdatel'skaia firma "Vostochnaia
literatura" RAN. MocKBa H3,n;aTeJibCKaH I}>HpMa "BocTO'IHaH JIHTeparypa" PAH.
- - 2004. Indian epigraphy and inscriptions in unknown writing from ancient Termez
(Kara-tepe, 1990-1994). International Association for the Study of the Cultures of Central
Asia, Information Bulletin 24: so-Ss.
Vogel, J. Ph. 1902-03. Two Brahmi and Kharoshthi Rock-Inscriptions in the Kangra Valley.
Epigraphia Indica 7: n6-n9.
Voigt, Rainer 2oos. Die Entwicklung der aramaischen zur Kharo~thi- und Brahmi-Schrift.
Zeitschrift der Deutschen Morgenliindischen Gesellschaft 1ss: 2s-so.

The Idea of Writing

Writing Across Borders

Edited by

Alex de Voogt and Joachim Friedrich Quack

BRILL

LEIDEN BOSTON
2012