Initially, preservation of the teachings with their word-
ing unaltered was not considered a necessary criterion
of authenticity, and this contributed greatly to foster-
ing linguistic diversity and spreading the teaching as
Buddhism left its homeland in the Ganges plain. The
texts that had hitherto been transmitted orally were
then transposed into other more or less supraregional
Middle Indian dialects to facilitate understanding and
wider dissemination. At present we know of only two,
Pali and Gandhar, Pali being a western dialect,
whereas Gandharwas widely used in the northwest-
ern part of the subcontinent and, with the growth of
the Kushan empire, in Bactria and Central Asia. Pali
became the canonical language of the Theravada
school, and Gandharthat of the DHARMAGUPTAKA.
Considerably later sources mention other Prakrits used
by various schools, namely Pais ́ac, Apabhrams ́a, and
Maddhyoddes ́ika. Apabhrams ́a is assigned to the
Sammatyas or to the Sthaviras, and Maddhyoddes ́ika
to either the Sthaviras or the MAHASAMGHIKA SCHOOL.
All schools must at first have transmitted their
canonical texts in Prakrit. Some of them, like the Ther-
avada, retained their Middle Indian language, while
others participated in the so-called Sanskrit renais-
sance and started to Sanskritize their received litera-
ture. Sanskritization was apparently a gradual process
permitting schools that were spread over a vast area to
undergo different regional developments. The litera-
ture of the (Mula-)Sarvastivada is preserved only in
Buddhist Sanskrit, but its older layers reveal many
traces of the underlying Prakrit. Surviving fragments
suggest that the Dharmaguptakas also took part in the
process of Sanskritization, at least in Central Asia. The
growing number of fragments found in Afghanistan
since 1994 supports the view that the Mahasamghikas,
and especially the Lokottaravada, used a specific mix-
ture of Prakrit and Sanskrit that may be termed Bud-
dhist Hybrid Sanskrit in the true sense, and which was
probably referred to as Maddhyoddes ́ika (intermedi-
ate recitation), a term not yet fully understood.
Retention or translation
When Buddhism began to spread beyond the Indian
subcontinent, missionaries and local followers were
confronted with the problem of how to communicate
the teaching, the rituals, and the literature in a totally
different linguistic environment. Basically, two possi-
bilities offered themselves: to preserve the Indian lan-
guage used so far, or to translate into the local
language. Preserving the Indian original offered prac-
titioners several advantages, among them a sense of
the sacredness of oral and written texts derived from
their use of the holy language supposedly spoken in
the homeland of the Buddha or even by the Buddha
himself; continuing access to other Indian sources;
and, very importantly, unambiguousness in termino-
logical matters. It also provided a useful common cur-
rency in a multiethnic and multilingual environment,
no small issue when a Buddhist missionary movement
came to be supported by the ruling powers for its uni-
fying potential. On the other hand, Indian languages
would have been incomprehensible to most followers
outside India and a deterrent to prospective converts,
especially in areas where non-Indo-European lan-
guages were spoken. This unintelligibility would have
facilitated their readiness to exchange the Indian orig-
inal for a more suitable vernacular, even if it neces-
sitated the gargantuan task of finding at least
approximate equivalents in the target language for dif-
ficult Indian Buddhist terminology. Discussions pre-
served in several Chinese and Tibetan treatises clearly
show that some translators were well aware of the
methodological, philological, and cultural problems
involved in the translation process; their reflections on
these problems resulted in attempts to establish guide-
lines for bridging the linguistic and cultural divide.In the course of history both these possibilities—
retention of the Indian original and translation into
the vernacular—were employed, sometimes side by
side. Several times the vernacular chosen for transla-
tion became itself a transregional “church” language
(i.e., the idiom used for canonical scriptures and litur-
gical purposes) when its specific form of Buddhism
crossed further linguistic borders, as in the case of Chi-
nese and Tibetan. Although no Buddhist tradition de-
veloped prescriptions for or against the use of a specific
language, in most cases one observes a slowly but
steadily increasing tendency to regard the language of
the written canonical texts as sacred, and this greatly
reduced the original openness to linguistic changes
characteristic of the early period of oral transmission
in India. Wherever the language of the canonical liter-
ature was not identical with the vernacular, sooner or
later the vernacular came to be used for the produc-
tion of a sometimes very rich noncanonical Buddhist
literature consisting of commentaries, story collec-
tions, manuals, poetry, devotional texts, and the like,
and sometimes this led to the development of a new
literary language in its own right. Examples are the use
of Newari, Tamil, and Old Javanese alongside Sanskrit,
and Thai, Japanese, and Mongolian alongside Pali,
Chinese, and Tibetan, respectively.LANGUAGES