Review: On the Origin of the Early Indian Scripts

Reviewed Work(s):
Der Beginn der Schrift und frühe Schriftlichkeit in Indien
by Oskar von Hinüber: Schrift im alten Indien: Ein Forschungsbericht mit
Anmerkungen by Harry Falk
Review by: Richard Salomon
Source: Journal of the American Oriental Society , Apr. - Jun., 1995, Vol. 115, No. 2
(Apr. - Jun., 1995), pp. 271-279
Published by: American Oriental Society

Stable URL: https://www.jstor.org/stable/604670

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Oriental Society is collaborating with JSTOR to digitize, preserve and extend access
to Journal of the American Oriental Society

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms
 REVIEW ARTICLE

ON THE ORIGIN OF THE EARLY INDIAN SCRIPTS1

RICHARD SALOMON

UNIVERSITY OF WASHINGTON

Several recent publications have questioned prevailing doctrines and offered new views on the an-
tiquity of writing in early India and on the source and early development of the Indian scripts (Brahmi
and Kharosthi). Most of the new studies agree in assigning the origin of these scripts to a later period,
i.e., the early Mauryan era (late fourth to mid third centuries B.C.), than has generally been done in
the past, and in deriving them from prototypes in Semitic or Semitic-derived scripts. The main works
to be evaluated here are Oskar von Hiniiber's Der Beginn der Schrift undfriihe Schriftlichkeit in In-
dien and Harry Falk's Schrift im alten Indien: Ein Forschungsbericht mit Anmerkungen.2 Also dis-
cussed are two recent articles on similar topics, Gerard Fussman's "Les premiers syst6mes d'ecriture
en Inde"3 and Kenneth R. Norman's "The Development of Writing in India and its Effect upon the
Pali Canon,"4 as well as some other relevant publications. The authority and significance of this new
trend toward assigning a later date of origin for the Indian scripts is evaluated and placed in the con-
text of broader historical and cultural issues.

THE DEVELOPMENT AND EARLY HISTORY of writing inThe relationship, if any, of the historical scripts to
India of the historical period (that is, after the timethe
of writing of the proto-historic Indus Valley civili-
the Indus Valley civilization) has long been a controver-
zation and the explanation of the long interval during
sial problem. Basically, arguments turn around three which writing appears to have fallen out of use in
main issues: India.

The sources and origins of the Indian scripts of theThe principal reasons that these issues, particularly
historical period, namely, Kharosthi and especiallythe second, are so problematic are:
Brahmi.

The date at which these scripts, or their proto- There are no securely datable specimens of writing
types, first came into use. from the historical period earlier than the inscriptions
of Asoka from the mid-third century B.C. Other early
inscriptions which have been proposed by various au-
1 This is a review article of: thors as examples of pre-Asokan writing are of un-
Der Beginn der Schrift und friihe Schriftlichkeit in Indien. certain date at best.
By OSKAR VON HINUBER, Akademie der Wissenschaften und The external testimony from literary and other
der Literatur, Abhandlungen der Geistes- und Sozialwissen- sources on the use of writing in pre-Asokan India is
schaftlichen Klasse, Jahrgang 1989, Nr. 11. Mainz: AKADEMIE vague and inconclusive. Alleged evidence of pre-
DER WISSENSCHAFTEN UND DER LITERATUR; Stuttgart: FRANZ Mauryan writing has in the past been found by various
STEINER VERLAG, 1990. Pp. 75; and scholars in such sources as later Vedic literature, the
Schrift im alten Indien: Ein Forschungsbericht mit Anmer- Pali canon, the early Sanskrit grammatical treatises of
kungen. By HARRY FALK. ScriptOralia 56. Tubingen: GUNTER Panini and his successors, and the works of European
NARR VERLAG, 1993. Pp. 355. DM 136. classical historians. But all of these references are
2 See note 1. subject in varying degrees to chronological or inter-
3 Annuaire du College de France 1988-1989: Resume des pretive problems.
cours et travaux, 507-14.
4 Wiener Zeitschrift fur die Kunde Sudasiens 36 (Supple- Until recently, the received opinions on these issues,
mentband) (1993): 239-49. in the West at least, have mainly been based on, or at

271

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms
272 Journal of the American Oriental Society 115.2 (1995)

least strongly affected by, their explication by Georg now at our disposal. Major discoveries since Biihler's
Biihler fully one century ago in his highly influential, if day include the Aramaic and Greek inscriptions of the
somewhat controversial, monograph, On the Origin of time of Asoka from Afghanistan and Pakistan, and the
the Indian Brahma Alphabet (Indian Studies No. 111).5 very existence of the Indus Valley civilization. Among
Biihler argued for an early origin of writing in India and the improved analyses of earlier materials are a better
posited an extensive prehistory, going as far back as the (though still far from complete) understanding of the
eighth century B.C. for the Brahmi script, which he de- chronological development of textual corpora such as
rived from the Phoenician script. Although more recent the Veda and the Pali canon, and a clearer and more
writers such as David Diringer6 have tended to doubt cautious methodology for the paleographic dating of
such an early date for Brahmi and have looked to the inscriptions.
Aramaic rather than the Phoenician script as its prob- Previous discussions have also been hampered by a
able source, Biihler's materials and arguments have con- chronic lack of communication and understanding be-
tinued to guide the discussion long after many of them tween Indologists on the one hand and Semiticists and
have become outdated (Falk, p. 11). The arguments of other scholars of the history of writing on the other. The
specialists have largely focused on evaluations, criti- former, for the most part, had little or no knowledge of
cisms, and modifications of Buhler, while presentations the relevant branches of Semitic epigraphy, while the
by non-Indologists such as Diringer and Hans Jensen7 in latter typically had even less awareness of matters
their general works on the history of writing have relied Indian (whence their often uncritical reliance, alluded to
heavily and often uncritically and inaccurately on him above, on Buhler). Here too a major step forward has
(see, e.g., Falk, pp. 96, 123). In general, some form or been achieved in that our new authors, and once again
other of Biihler's essential thesis that Brahmi was devel- Falk in particular, have taken the trouble to familiarize
oped out of a Semitic prototype in pre-Mauryan India themselves with Semitic scripts, especially Aramaic, in
has been accepted by most scholars in the West, but re- such a way that the possible connections can at last be
jected by the majority of South Asian experts, who gen- discussed in an intelligent and objective manner.
erally argue for a separate and indigenous origin for theThe major conclusion shared by the studies of Fuss-
man, von Hiniiber, and Falk is that at least the Brahmi
Indic scripts, often by way of derivation, direct or indi-
rect, from the Indus script. script, and possibly also Kharosthi, originated in the
But what virtually all of these voices have in com- Mauryan period and not earlier. Although they disagree
mon is a focus, whether favorable or critical, on the ar-
on specifics, especially with regard to the date of the
guments presented by Buhler a century ago. It is thus development of Brahmi, all three agree that Kharosthi,
which was a regional script of the far northwest, was
appropriate and important that the authors of the publi-
older than the pan-Indian Brahmi and influenced its
cations under discussion here, in particular Harry Falk,
have finally freed themselves from the shackles of the formation. The three authors share a sharp skepticism
tired old arguments and undertaken an entirely new about alleged literary evidence for writing in pre-Aso-
look at these issues in light of what we now know kan of India, and are inclined to interpret the situation em-
Indian chronology, epigraphy, numismatics, and lin- pirically, on the grounds of what we definitely know,
rather than speculating on what might have been. They
guistic and literary history. Although no single decisive
document, such as the long-awaited certifiably pre- are inclined to take the absence of incontrovertible evi-
Mauryan Brahmi inscription, has come to light since dence for early writing as an indication that it did not ex-
Biihler's time, a vast amount of new material and a farist, rather than, as have earlier writers, adding up the bits
better understanding of what was previously known are of inconclusive hints and theoretical possibilities to re-
construct a hypothetical pre-history for the early scripts.
Among the four studies discussed here, only Kenneth
R. Norman's article, "The Development of Writing in In-
5 Sitzungsberichte der kaiserlichen Akademie der Wissen-
dia and its Effect upon the Pali Canon," follows a more
schaften, Wien, phil.-hist. K1., 132, no. 5, 1895. 2nd revised ed.:
traditional path. He analyzes certain patterns of textual
Strassburg: Karl J. Tribner, 1898. Reprint ed.: Chowkhamba
variation in Pali texts (e.g., hatthivattika/hattivatika,
Sanskrit Studies, vol. 33; Varanasi: Chowkhamba Sanskrit Series
Office, 1963. pp. 239-40, and samaya/samaja, p. 241) which seem to
reflect an early redaction in a script which did not rep-
6 The Alphabet. A Key to the History of Mankind (2nd ed.;
New York: Philosophical Library, 1953), 336. resent geminate consonants or differentiate vowel length,
and identifies this script as an early prototype of Brahmi
7 Die Schrift in Vergangenheit und Gegenwart (Gliickstadt
and Hamburg: J. J. Augustin, 1935). used in Magadha in pre-Mauryan times (p. 243). Norman

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms
 SALOMON: Origin of the Early Indian Scripts 273

finds it "difficult to accept that Brahmi was devised as a such a position, like those of others to be discussed be-
single complete writing system at one and the same time low, are cast into doubt by the three other new studies.
during the reign of Candragupta" (p. 245),8 and considers Though developed most cogently and completely in
it "even less likely that Brahmi was invented at the time these three new publications, the theory of a relatively
of Asoka for the specific purpose of writing his inscrip- late (i.e., Mauryan) date for Brahmi and Kharosthi and
tions" (p. 246). His objections to what may be referred the postulation of the former as an "invention" under the
to as the "invention theory" of the origin of Brahmi stimulus of one or the other of the Mauryan emperors is
mainly concern the irregularities and inconsistencies of by no means entirely new. For instance, as noted by Falk
the graphic system: for instance, inconsistencies in the (p. 163), Max Muller in 1892 (before Biihler!) opined
formation of the graphs for aspirate consonants, some that Brahmi was probably "das Werk einer Kommission
of which are clearly based on the corresponding non- von Gelehrten, die, wahrscheinlich im Auftrage des
aspirates (e.g., C ta and 0 tha) while others (e.g., l Konigs [Asoka], aus fremden Quellen ein Alphabet ent-
ta and ( tha [not cited by Norman]) are not so deriv- warfen, ... die Laute der gesprochenen Sprache aus-
able. Such patterns lead Norman to conclude that Brahmi zudriicken." The old invention theory, which had largely
"evolved [my emphasis] in a haphazard way, with some fallen out of favor after Bihler, was revived in 1979 by
of its aksaras being borrowed from some other source" S. R. Goyal, in his essay, "Brahmi: An Invention of the
(p. 245). Early Mauryan Period,"9 who argued "that the Brahmi
But von Hiniiber, in Der Beginn der Schrift ..., in- script was invented in the first half of the third century
terprets the patterns of textual variation in Pali which B.C., and that the Indians of the Vedic and early Buddhist
underlie Norman's theory quite differently, noting that periods were illiterate" (p. 4), and that "in all probability
geminate consonants were still not regularly noted in Brahmi was invented in the age of Asoka and the idea
Indian inscriptions of the first century B.C. when the Pali ... of writing came from the West" (p. 17). Though not
texts were presumably first written down (p. 64), and entirely original, the data and arguments invoked by
that long a was often left unindicated in early Brahmi Goyal-the persistent failure of efforts to find and iden-
inscriptions from Sri Lanka (p. 66). Von Hiniiber's ar- tify actual specimens of pre-Agokan writing, the testi-
guments are persuasive if we can assume that the ortho- mony of Greek authors (especially Megasthenes) to the
graphic standards of early inscriptions also prevailed in absence of writing in India in the early Mauryan period,
contemporary (pre-Christian) religious or literary texts the evident influence of Indian phonetic and grammati-
in manuscript form. However, although we do not have cal theory on the structure of the early scripts, and the
any manuscripts this old, it is not impossible that stricter primitive and uniform appearance of Asokan Brahmi-
orthographic standards, including the notation of gemi- prefigure the positions developed at greater length in the
nates, might have been applied to them, in contrast to newer works. Goyal's essay seems to have served as a
the standards of inscriptions which at this period were stimulus to the recent re-thinking of and revival of inter-
often still treated quite casually in terms of orthography est in these questions, and his essay should be (re-)read
and layout. Nonetheless, it must be conceded that Nor- in conjunction with those being reviewed here.
man's arguments rest on a largely hypothetical basis and Turning now to these new publications, in chronologi-
that underlying orthographic inconsistencies reflected in cal order: Gerard Fussman, in "Les premiers systemes
much later manuscripts of the Pali canon are hardly co- d'ecriture en Inde," briefly10 argues (pp. 513-14) that the
gent grounds for the reconstruction of a proto-Brahmi of Kharosthi script originated in the northwest before the
Mauryan period and provided both the model and in-
the pre-Mauryan era. Norman's position is essentially a
reaffirmation of the more moderate version of the oldspiration for the development of Brahmi. This latter
school of thought, which places the origin of Brahmi in
or around the fifth century B.C. But his arguments for
9 In The Origin of Brahmi Script, ed. S. P. Gupta and K. S.
Ramachandran (History and Historians of India Series, vol. 2;
Delhi: D. K. Publications), 1-52. (Reviewed in JAOS 102
8 Here Norman (n. 24) cites Fussman, "Les premiers
[1982]: 553-55.)
syst6mes . . . ," 513. His subsequent allusion to the theory of 10 Fussman's ideas on these subjects will presumably be de-
the invention of Brahmi under Asoka is made without reference veloped at greater length in his article, "lcritures indiennes,"
to the works of von Hiniiber and Falk, which were evidently not in D. Arnaud's Histoire de l'ecriture, cited in Falk's bibliogra-
yet in print when Norman wrote his article, though he was phy (p. 31) as "im Druck," and not available at the time of this
probably aware of their ideas on the subject. writing.

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms
274 Journal of the American Oriental Society 115.2 (1995)

creation was undertaken, probably during the time of Aramaic (p. 58); he also suggests that the word lipi here
Candragupta Maurya, in order to serve administrative may not refer to writing at all, but rather to painting or
needs in Magadha, and was carried out by "pandits qui the like (p. 57).
furent charges de creer une ecriture pour une region de However, the main focus, comprising six of fifteen
l'Inde qui ne la connaissait pas" (p. 514). Fussman's short chapters, of von Hiniiber's discussion is the refer-
grounds for this chronological reconstruction are princi- ences, real or apparent, to writing in the Pali canon, es-
pally the testimony of the Greek historians; for while pecially in the Vinaya-pitaka. Here he argues at some
there are references to writing, presumably Kharosthi, in length (including several highly technical excursuses)
the northwest at the time of Alexander the Great, Me- that terms such as the verb likh and its derivatives
gasthenes, who lived in Pataliputra late in the fourth cen- either do not actually refer to writing, or, if they do,
tury B.C., declared the Indians to have no writing at all. are attributable to later strata of the Vinaya literature
Fussman admits the "faiblesse intrinseque" of these (pp. 36-40), which in any case reflects, on the whole,
sources, but declares that they "se combine pourtant de a later stage of development than the Sutta-pitaka
faqon a former un faisceau de pr6somptions acceptables" (pp. 46-54).
(p. 513). Brahmi, he concludes, is "heritiere de l'Iran Thus von Hiniiber finds no cogent evidence whatso-
pour l'idee, tributaire des premiers modeles arameens et ever for any kind of writing in the Indian heartland in
arameo-indiens pour sa technique, purement indienne en the early Buddhist era or at any other time in the pre-
ce qui concere sa lisibilite et son adequation a la Mauryan period. The creation of Brahmi thus evidently
langue" (p. 514). took place during Mauryan times, with the erstwhile
In his view of Brahmi as an artificial conglomerate of Achaemenid model of Old Persian Cuneiform as the in-
Iranian, Semitic, and Indian elements, Fussman is in spiration for the creation of a new "imperial" script, and
general agreement with von Hiniiber and Falk, although with Kharosthi as a systemic model: "Die Zentralver-
his chronology differs, particularly with respect to the waltung des Maurya-Reiches konnte also in der Pho-
time of its development, which they are inclined to at- netik bewanderte Brahmanen mit dem Einwurf einer
tribute to Asoka rather than to Candragupta. In his neuen, fur eine monumentale Epigraphik besser als die
monograph, Der Beginn der Schrift undfruhe Schriftlich- Kharosthi-Schrift geeignten und zugleich dem Bediirf-
keit in Indien, von Hiniiber carefully reexamines the nis nach einer leichteren Deutbarkeit des Geschriebenen
principal sources which have been invoked by earlier entgegenkommenden Schrift beauftragt haben" (p. 59).
scholars in their efforts to prove the existence of writing This undertaking is attributed by von Hiniiber, some-
in pre-Mauryan India, and finds that none of them stand what cautiously (p. 60), to Asoka himself.
the test: "Fremde Beobachtungen sprechen also in Many of the themes introduced and discussed, if some-
Ubereinstimmung mit den Zeugnissen aus Epigraphik what briefly, by von Hiniiber are taken up again in greater
und Numismatik eindeutig dafiir, daB es in Indien vor detail by Harry Falk in his much more voluminous study,
Asoka keine Schrift gegeben hat, wenn man von den in- Schrift im alten Indien. The connections are no doubt to
dischen Provinzen des Achamenidenreiches absieht" be explained by the authors' common participation in
(p. 22; cf. also p. 72). For example, his analysis Sonderforschungsbereich
of Me- 321, "Ubergange und Span-
gasthenes' alleged references to Indian writingnungsfelder
(ch. 4) zwischen Miindlichkeit und Schriftlichkeit"
leads him to the conclusion (p. 20) that the only at
authen-
the University of Freiburg. Von Hiniiber himself (p. 5)
tic one is his statement that they have none (oue6 yaphis readers to the then-forthcoming report of Falk
refers
ypdalictLaa ei&vat auoo6g). This he interprets asfor a more comprehensive study, and Falk (p. 12) ac-
a blan-
knowledges
ket denial, rejecting the interpretation of J. D. M. Der- the "sehr fruchtbar" exchanges with von
rett and others that it refers only to written documents Hiniiber. This is not to say, however, that the two publi-
in connection with legal procedures, which are the im-are repetitive or imitative of each other. For al-
cations
mediate context of Megasthenes' discussion. As for thethey do reflect generally similar points of view
though
reference by Nearchos, quoted by Strabo, to the and come to substantially the same conclusions, they
Indians'
practice of writing letters (crtoXzo&K;) on cloth, he con-
cover different ground, present different modes of argu-
siders it "sehr wohl moglich, daB Nearch iranische, d.h.
mentation on some issues, and disagree significantly, if
wohl aramaische Briefe meint, wenn er von 'indischen' not fundamentally, on several important questions.
spricht, weil er sie eben in Indien gesehen hat" (p. 21). As the subtitle (Ein Forschungsbericht mit Anmerkun-
Similarly, he suggests that the lipi mentioned by P.anini gen) indicates, Falk's study is arranged in the form of a
(3.2.21), if it really refers to a 'script' at all, must be comprehensive review in chronological order of all the
either "eine sehr frihe Form der Kharosthi-Schrift," or scholarly literature on issues relevant to the origin and

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms
 SALOMON: Origin of the Early Indian Scripts 275

early development of writing in India. The summaries of all details. Thus he agrees with von Hiniiber that "Me-
previous writings are supplemented, wherever appropri- gasthenes sagt klar und eindeutig, daB in Magadha [sic]
ate, by the author's own evaluative comments ranging zu seiner Zeit Schrift ganz allgemein nicht in Gebrauch
from brief notes to discussions ranging over several war" (p. 293), and that the writing observed by Nearchos
pages. 1 The basis of the work is therefore a comprehen- in northwestern India was probably Aramaic (p. 290).
sive bibliography (pp. 15-66) which is followed by six- Likewise, he takes Panini's lipi to refer to Aramaic, con-
teen analytical sections, among which the longest and cluding that "fur eine eigene, einheimische Schrift zu
most important cover such topics as "Theorien zum Ur- seiner Zeit gibt es aber keinerlei Anhaltspunkte" (pp.
sprung der Brahmi" (? 6.3), "Archaeologische Argu- 258-59). But he does not share von Hinuber's doubts,
mente" (? 8), "Literarische Zeugnisse fur Schrift" (? 9), cited above, as to whether lipi in Panini refers to writing
and "Berichte von Auslandern zur Existenz von Schrift" at all. Similarly, in his discussion of alleged early
(? 11). references to writing in the Pali Vinaya, he comes to
The "Forschungsbericht" format of Falk's book is onthe same basic conclusion, namely negative, as von
the whole successful, though it inevitably has both ad-Hiniiber, but his specific interpretations often differ. The
vantages and disadvantages. Among the latter, Falk's in-term likhitako coro, for example, was explained by von
tention to present comprehensive coverage of all theHiniiber (p. 38) as referring to a thief identified by a pic-
ture, rather than a written document such as an arrest
literature ("versuchte ich alle Publikationen zu erfassen,
die den Umstanden der Einfiihring der Schrift in Indien warrant; Falk, however, quite plausibly suggests (pp.
gewidmet sind," p. 12) leads him to include, not only in276-77) that the phrase alludes to a branded or other-
the bibliography but also in the analytic text, numerouswise physically "marked" thief. And while von Hiniiber
specimens of the sort of pseudo-scholarship that is all (pp. 39-40) does accept the expression lekham chindati
too well represented in this field. While Falk generallyas one of the authentic references to writing in the Vi-
manages to dispose of such amateurish efforts conciselynaya but dismisses it as relatively late, Falk disagrees,
and effectively (see, e.g., pp. 144-47 and 157-60), cautiously suggesting (p. 279) that it may refer to an as-
considerable space might have been saved simply bycetic practice of cutting off pieces of one's own flesh.
skipping over the chaff and concentrating on serious Regardless of the differences in the details of their in-
publications; we hardly need another discussion of Cun-terpretation of particular passages, Falk is in general
ningham's theory of the hieroglyphic origin of Brahmiagreement with von Hiniiber as well as with Goyal and
(p. 143) or Shamasastry's tantric theories (p. 144). other recent writers that most if not all of the terms and
Nonetheless, since Falk's main purpose was to set thepassages in early Indian literature and in the writings of
record straight, the comprehensive approach is justifi-foreign observers which were cited by Biihler et al. in
able. As he notes (p. 11), the literature on the subject of
support of an early date of origin for the Indian scripts
the origins of writing in India (as on many other sub-in fact prove no such thing. Their arguments are indeed
jects) is full of misattributions, misinterpretations, andpersuasive, but not completely decisive. For instance,
simple ignorance of previous works; see, for example, there is no concrete evidence that the writing referred to
pp. 126-27, showing how Diringer's incorrect accounts by observers such as Nearchos and Panini in northwest-
of opinions on the Indian scripts have found their wayern India in and around the fourth century was Aramaic,
into the specialist literature. Unlike most previous writ-rather than Indian, i.e., Kharosthi; here von Hiniiber
ers on this topic, Falk has obviously taken the trouble to (quoted above) has wisely left the door open by allowing
locate, read and understand all of the literature, rather for the possibility of a "very early form" of Kharosthi,
than relying on what others have said about it, and thus while Falk, perhaps a little rashly, excludes this possi-
has succeeded in clarifying authoritatively, for once and
bility. For Kharosthi, according to Falk, must have been
for all, who said what and when. Just for this, we arecreated at one stroke at some time later, not before 325
much indebted to him. B.C. (p. 104). The argument for this date is based on the
In his review of evidence for the antiquity of writing theory that the new script could only have originated
in historical India, Falk follows, as noted above, the when the professional monopoly of the Aramaic scribe-
positions of von Hinuber in broad outline, though not in bureaucrats of the Achaemenid empire (cf. pp. 78-81)
had broken down in the wake of the Greek conquest.
While the introduction of social and economic consid-
1 The model for the structure of the book is, as the author erations into the discussion of the origin of Indian
notes (p. 11), George Cardona's Pdnini: A Survey of Research scripts is a welcome addition, this particular argument is
(The Hague: Mouton & Co., 1976). speculative at best and hardly constitutes firm grounds

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms
276 Journal of the American Oriental Society 115.2 (1995)

for a late date for Kharosthi. The stronger argument for vehemently and not unconvincingly against the exis-
this position is that we have no specimen of the script tence of Brahmi before Agoka. His comprehensive re-
before the time of Asoka, nor any direct evidence of in- view (ch. 8, pp. 177-239) of the archaeological, i.e.,
termediate stages in its development; but of course this epigraphic and numismatic, evidence confirms the re-
does not mean that such earlier forms did not exist, only cent trend of opinion, developed by such authorities as
that, if they did exist, they have not survived, presum- D. C. Sircar and A. H. Dani, according to which none of
ably because they were not employed for monumental the several early documents such as the Piprahwa reli-
purposes before Asoka. quary inscription, the Sohgaura bronze plaque, and the
Likewise, Falk's conclusion (p. 103) that Kharosthi Mahasthin stone inscription, which had in the past been
was a conscious creation loosely modelled on Aramaic, presented as pre-Asokan in date, can in fact be proven
rather than the product of a gradual evolution and differ- to be so. Another strong argument in favor of an Asokan
entiation from it, is intriguing and at least partly origi- origin for Brahmi is the formal one adduced by Goyal
nal, but by no means immune to objections. The most and others and endorsed by Falk (pp. 164-65), which
important of these, which Falk himself anticipates, is proposes that the simple and symmetrical geometric
that in several cases Kharosthi characters have different forms which predominate in early Brahmi (e.g., Q tha
phonetic values from the Aramaic letters that they most and + ka) are indicative of a recent origin and an ar-
closely resemble in shape (e.g., Aramaic 2 pe /bitrary creation. Falk thus concludes (pp. 337-39), like
Kharosthi 7 a). This problem he attributes to a hypo- Miiller, Goyal, and von Hiniiber before him, that
thetical situation in which the inventor of Kharosthi was
Brfhmi was a conscious creation of the Mauryan period,
familiar with, but did not actually know, Aramaic script:
probably designed during the reign of Agoka for the ex-
"jemand die Kharosthi entwickelt hat, dem man zwar press purpose of the monumental presentation of his
einmal die Funktionsweise und die Lautwerte der edicts. This new script was designed, according to Falk,
aramaischen Zeichen erklart hatte, der sich die Er- primarily on the systemic model of Kharosthi, but with
klarungen aber nur teilweise richtig merkte und deshalb significant input, especially as far as the overall monu-
spater einige Zeichen neu bewertete und andere neu ent- mental ductus (cf. p. 82 and 111) and the left-to-right di-
warf. Nur ein Entwickler ohne profunde Kenntnis der rection was concerned, from Greek. Since both of these
aramaischen Schrift wiirde so groBziigig mit dem Vor- latter scripts are ultimately derived from Phoenician
bild umgehen" (p. 103). I must confess this I find this (via its derivative Aramaic, in the case of Kharosthi),
explanation unconvincing, all the more so in view of Brahmi in Falk's view is ultimately an Indian adaptation
Falk's emphasis elsewhere on the important role and of Semitic scripts.12
wide usage of Aramaic, now so well attested by the However, Falk does not attempt to work out in full
Asokan Aramaic inscriptions, in the eastern regions of detail the derivations of the individual characters of
the Achaemenid empire. Even if we accept that Aramaic Brahmi from their presumptive prototypes, as Buhler
was something of a guild monopoly, its basic syllabary (following the lead of Albrecht Weber'3) and others have
is so simple and straightforward that it is hard to imag- in the past attempted to do, albeit with limited success.
ine that someone clever enough to invent Kharosthi He does stress the anomalous form of Brahmi () tha,
could have so badly misunderstood it. Hence I am still which not only resembles the Greek theta (0) in form
inclined to accept Buhler's principle of deriving the in-
dividual characters of Kharosthi from the phonetically
corresponding Aramaic consonants. For the results of 12 The composite theory of the development, or invention, of
the application of this principle, if inevitably not en- Brahmi is not, as Falk notes (p. 338), entirely unprecedented.
tirely satisfactory, are on the whole successful and per- The French Semiticist J. Halevy, in particular, tried to establish
suasive, far more so than in the case of Brahmi. Buhler's a composite derivation from Aramaic, Greek and Kharosthi.
explanations of the alterations of individual Aramaic Falk feels it was largely because he published in French that
prototypes into the Kharosthi aksaras on the grounds of Halevy's contributions have not received the credit they de-
reasonably consistent principles of inversion, reversal, served (p. 127); but it seems to me that it was the serious flaws
and differentiation still seem distinctly preferable to in- in argumentation, reflecting profound misunderstandings of the
voking the deus ex machina of an ignorant genius in- Indian cultural background, as well as his intemperate tone of
venting Kharosthi out of a vague acquaintance with argumentation (cf. p. 132), rather than the language of his pub-
Aramaic script. lications, that cost Hal6vy much of his credibility.
Even more important and thought-provoking are 13 In "Ueber den semitischen Ursprung des indischen Alpha-
Falk's theories on the origin of Brahmi. He argues bets," ZDMG 10 (1856): 389-406.

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms
 SALOMON: Origin of the Early Indian Scripts 277

and phonetic value but also is the only Brahmi letter, taw, but not from Kharosthi '2 ta; several other
except for initial i ( . ) that consists of more than one such examples could be cited. What this boils down to
unconnected stroke: "Das heiBt, man opferte in der is the old problem that each of the proposed prototypes
Brahmi das sonst iiberall erkennbare Prinzip von der gra- for Brahmi, viz., Kharosthi, Aramaic, Phoenician, and
phischen Einheit jedes Zeichens im Rahmen einer Uber- Greek, can provide models for some of its characters,
nahme" (p. 111). I do not find this to be a particularly but no one of them can explain all of them; to do so, one
cogent argument, since it is hard to know exactly what must revert to rather far-fetched combined derivations
the nature and significance of the alleged "Prinzip von of the sort proposed by Halevy (see n. 12). Falk does not
der graphischen Einheit" is, especially when not one but address these problems head-on, and perhaps would be
two Brahmi characters violate it. By the same logic, the inclined to dismiss them, as have some others, on the
Greek script should have a similar principle, also with grounds that the characters of Brahmi were essentially
two exceptions (? theta and E xi), but it is hard to see arbitrary creations with a general input from Greek and
any special significance in this. This is not to deny the Kharosthi, but not systematically patterned on either of
striking similarity between theta and Brahmi tha, or even them. This too is not impossible, but still the resem-
the possibility that the former influenced the formation blance of many of the Brahmi characters to phonetically
of the latter, especially since this is one of the Brahmi cognate ones in one or the other scripts is troubling. It
characters for which it is difficult to find a suitable pro- may not ever be possible fully to establish the deriva-
totype in late Aramaic (although similar shapes are tions of each Brahmi character, and this was clearly not
available in earlier Semitic scripts, e.g., Phoenician @ Falk's intention, but I cannot help feeling that in this
tet). But I doubt whether this single example deserves regard he has overestimated the role of Greek at the
the special significance attributed to it by Falk, who per- expense of Aramaic.
haps overemphasizes the influence of Greek on Brahmi. All of the above discussion assumes that Brahmi is in
Much the same can be said about his analysis of the fact a derived script, created from or loosely modeled on
vocalization system of Brahmi. That the basic system one or more Semitic or Semitic-derived scripts. While it
of indication of post-consonantal vowels by diacritic is true that the historical and geographical circum-
marking was originally developed in and adapted from stances point strongly in this direction, it must be
Kharosthi seems well established. But Falk's suggestion remembered that this point of view is not at all widely
(pp. 111, 339) that the introduction into Brahmi of dis- accepted in South Asia, and should not be taken for
tinct diacritics for short and long vowels was inspired granted; with his assumption of a Semitic derivation and
by the model of Greek script is doubtful, since the nota- especially his strong emphasis on the role of Greek, Falk
tion of vowel quantity in Greek operates on entirely may leave himself open to charges of a Eurocentric
different principles. Whereas Greek uses distinct alpha- viewpoint. Such questions of indigenous development
betic characters, mostly derived from Semitic conso- versus borrowing from outside will also arise in connec-
nants, to represent, incompletely and inconsistently, short tion with Falk's analysis of the numerical notation sys-
and long vowel pairs, Brahmi has a complete and regular tem of Brahmi. Because the use of distinct signs in
set of matched short-long pairs of post-consonantal dia- Brahmi for each of the digits (1 to 9) and the decades (10
critic signs. Thus at best one might suggest that Greek to 90) is "eine radikale Abkehr vom semitischen Sys-
provided an example or inspiration for the development tem" (p. 175), he looks elsewhere for its prototype and
of a system of notation of vowel quantity. But I hardly finds a similar system in early Chinese numerals, which
see the necessity for even this much, since, given their he thinks could have been brought to India by Chinese
well-established tradition of phonetic analysis, the Indi- merchants travelling to Gandhara in ancient times. But
ans could certainly have thought of this on their own. So I find it hard to accept archaeological evidence of Chi-
here again, the weight of Greek influence seems to be nese wares in the Swat Valley in the early second mil-
overemphasized. lennium B.C. (ibid.) as any sort of evidence for a possible
A further problem in deriving Brahmi as a composite borrowing of a system of numerical notation. It is sur-
of Greek and Kharosthi involves the several Brahmi prising that Falk does not take into serious consideration
characters that are more readily explained by reference
the striking similarities, discussed by Biihler and others,
to the presumptive Aramaic prototype of Kharosthi thannot only in system but also (unlike Chinese) in the actual
to the Kharosthi (or Greek) characters themselves. form of several of the numerical signs, between Brahmi
Among these are Brahmi Lr ha, which can reasonablyand hieratic and demotic Egyptian. Though I am not
convinced that the Brahmi characters are in fact
be derived (by inversion) from an Aramaic 7\ he, but
hardly from Kharosthi , ha, and l ta from Aramaic borrowed from Egyptian, this seems a far more plausible

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms
278 Journal of the American Oriental Society 115.2 (1995)

possibility than borrowing from China. Thus while it is ence to the empirical evi
not strictly correct, as Falk states, that "Die einzige Al- think that Brahmi did no
ternative zu einem chinesischen EinfluB auf die Brahmi- B.C., and that it was creat
Zahlzeichen ... ist die Annahme einer Neuschopfung in adaptation of one or mor
Indien mit zufalliger Parallelitat" (p. 175), the possibil- with Kharosthi playing a
ity of an indigenous origin should be seriously consid- itself almost certainly di
ered. Since numerical signs, unlike phonetic signs, are Falk et al., and probably
not wholly arbitrary but rather tend to develop by fourth century, and (co
cursive simplification from collocations of countingbefore then.
strokes, coincidental similarities in their forms are not One final and important problem remains. According
nearly as unlikely as it might seem at first glance. In theto the position espoused in these books-which, given
case of the Egyptian systems, we have sufficient mate-the authority of their authors and the quality of their
rials to see how the separate hieratic decade signs orig-scholarship, is likely to be hereby established as the pre-
inally developed from the cursive writing of additivevailing point of view, at least in the West-the heart-
groups of the single hieroglyphic sign for 10. A similarland of India was preliterate until the third century B.C.
system could well have developed separately in India,But can we imagine such a state of affairs, given what
independent of the influence of any outside system andwe know (admittedly not too much) of the state of so-
even apart from the development of linguistic writingciety and culture in India, especially in the northeast,
(which would explain the persistent problems in estab-before this time? If we can put any trust at all in the tra-
lishing phonetic or systemic linkages between earlyditional lore of the puranas and the testimony of the Pali
Brahmi and the numerical system associated with it, cf.canon, Magadha was the site of great and prosperous
Falk 169-73). Certainly we do not need to look all theempires, notably that of the Nandas, decades if not cen-
way to China. turies before the foundation of the Mauryan dynasty
around 320 B.C. Can we believe that these dynasties with
* * *
their legendary riches, and the remarkable intellectual
and cultural life of India in the time of the Buddha and
Such disagreements over details notwithstand
Mahavira, existed in a totally illiterate sphere? It is cer-
studies discussed in this review are works of h
tainly true that intellectual activity in India has always
arly quality which will be of lasting impact an
strongly favored oral over written means of expression,
Indeed, it is no exaggeration to say that these
and both von Hiniiber and Falk have effectively put to
especially Falk, have raised the level of disc
rest the already discredited skepticism about the possi-
the old problems of early literacy and the orig
bility of oral composition and preservation of the Veda,
scripts in ancient India to an entirely new lev
Panini's grammar, etc. (see, e.g., Falk 321-27). But the
them have succeeded in casting off the old p
fact that Panini did not use writing in composing the As-
and lingering effects of Buhler's outdated arg
tadhydyi does not necessarily mean that he was illiterate
and Falk in particular has brought a wealth of
(cf. Falk 259); it may only mean that writing was not
formation and erudition to the field. In particul
considered an appropriate vehicle for intellectual en-
for the first time looked seriously from a mo
deavors of this kind. Even given the very different cul-
point at the Iranian, Semitic, and Hellenist
tural role of writing in India as compared to many other
ground to developments in Mauryan India, w
ancient civilizations, it is hard to conceive that practical
objectively and perceptively reevaluating the e
affairs, such as the keeping of records and accounts in
pus of Indological and classical data. Thus wh
a wealthy empire like that of the Nandas, could have
agree on several secondary points of interpret
been kept in order without any form of writing at all, or
method, and while I would maintain that much
at least without some alternative system of memory aids
mains to be done, particularly with regard to
like the Inca quipu. Thus one is tempted to think along
reexamination the development of the in of
the lines of William Bright (cited by Falk, p. 290) of
characters of Brahmi,
I find myself more convi
some type of writing that was "perhaps used for com-
ever by Falk's arguments, bolstered by thos
mercial purposes, but not for religious or legal texts."14
Hinuber, Goyal, and Fussman, for a late or
Brahmi in the Mauryan, and probably the
period. In the light of new evidence such as th
14 Similar
inscriptions of Asoka and arguments
the are also presented in K. R. Norman's
reinterpretation o
review of von
faulty claims for evidence ofHiniiber's Der Beginnwriting,
early der Schrift ... in JRAS,it m
ser. 3, vol. 3 (1993):
be admitted that, as long as 277-81
one (esp.agrees
p. 279). to give

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms
 SALOMON: Origin of the Early Indian Scripts 279

Admittedly, we have not a shred of concrete evi- purely hypothetical) connection with the Indus script.
dence for this, and it is perhaps better to stick with what Nevertheless, it would be unwise to rule out surprises
we have and assume that business affairs, like cultural in the future, and we should leave the door open, as
ones, were conducted in pre-Mauryan Magadha simply does Falk (p. 340), to discoveries that could revive
on the basis of the highly-developed memory skills so theories of an early development of Brahmi. But we
well attested in ancient and modem India, perhaps with must also agree, if reluctantly, with his final sentence:
the assistance of a system of numerical notation such as "Zur Zeit erscheint dieser Fall jedoch kaum zu er-
that hypothesized above. This-it would be hard to warten" (ibid.).
deny, in light of the evidence that Falk, von Hinuber
et al. have laid out before us-is the most likely sce- * * *

nario on the grounds of the unfortunately meager evi-

dence that is left to us. Still, we should not fall into the One final note. It is unfo
trap of thinking that the last word has been spoken. Ad- man (in French) and von
mittedly, it hardly seems likely, after all the years of will not be available to a
waiting, searching, and the dashing of false hopes, that audience. These works, e
some major archaeological discovery will reveal a Indien, merit publication
whole new picture of the origins of writing in the In- to bring them to that wi
dian heartland, or reveal a sustainable (rather than influence future discussi

This content downloaded from

130.192.98.170 on Tue, 24 Nov 2020 08:42:19 UTC
All use subject to https://about.jstor.org/terms