Professional Documents
Culture Documents
net/publication/263962646
Philosophy of language
CITATIONS READS
0 2,844
1 author:
Johannes Bronkhorst
University of Lausanne
263 PUBLICATIONS 1,046 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Johannes Bronkhorst on 16 July 2014.
The role of language in Indian philosophy is great, recited at special, mainly ritual, occasions and
and indeed, much of this philosophy remains were believed to be vital for the efficacy of those
unintelligible without an awareness of the role rites. The belief in the efficacy of mantras had a
that language plays in it. It can reasonably be determining effect on the Brahmans’ attitude
maintained that an important part of Indian phi- toward language.
losophy is philosophy of language, to be under- The language of the Brahmanical mantras was
stood in the double sense of “philosophy inspired Sanskrit. Modern linguists may be tempted to
by language” and “reasoned inquiry into lan- point out that the language of many of the vedic
guage.” An exclusive concentration on the latter of mantras is an older form or even a predecessor of
these two, however, would not do full justice to the classical Sanskrit, which they may give a different
role of language in Indian philosophy. What is name – for example, “Vedic” or “Old Indo-Aryan”;
more, such a presentation would risk leaving out for the Brahmanical tradition, the Vedic language
essential elements and remaining, to at least some and classical Sanskrit were one and the same lan-
extent, unintelligible. This article will therefore guage, even though the Sanskrit of the Veda
deal with both these aspects. admittedly often used words and grammatical
The point of departure is, and has to be, that the forms that were not commonly used in classical
main actors in Indian philosophy held a number Sanskrit.
of views about language, which they considered This conviction that Vedic and classical San-
either self-evident or too deeply anchored in their skrit were one and the same language was impor-
worldviews to allow for critical discussion. These tant for the Brahmans. It allowed them to think of
views were therefore not normally objects of, but this language as unchangeable and even as eternal
rather starting points for, inquiry. and without beginning. Indeed, in postvedic
These main actors were for a long time pri- times, the vedic corpus itself came to be looked
marily Brahmans and Buddhists. Brahmans and upon as eternal and beginningless. This raised
Buddhists did not share the same views about questions that will be discussed below.
language, and indeed, their positions were often At this point I will consider two consequences
radically opposed to each other. In spite of these of the belief in the efficacy of mantras. This belief
divergences, a remarkable convergence took place was particularly important for Brahmans, for their
in the course of time, which led to a situation in livelihood depended on it. Brahmans owed
which Brahmans and Buddhists agreed on a num- their special position in society primarily to their
ber of far-from-obvious points. In order to under- knowledge of the Veda: they could recite portions
stand this convergence, it will be necessary to of it and use its mantras at appropriate occasions.
discuss the different starting points and early But to be efficacious, these mantras had to be
developments of Brahmanical and Buddhist recited in a phonologically correct manner. One
thought (see also → Hinduism and Buddhism). of the consequences of the belief in the efficacy of
mantras was therefore that major attempts were
made to describe the phonology of those mantras,
Brahmanical Presuppositions and and secondarily of Sanskrit in general. A rich lit-
the Birth of Sanskrit Linguistics erature on the phonology of Sanskrit arose, much
of which has survived until today (Deshpande,
Scholars are well informed about religious con- 1997, 31ff.; Scharfe, 1977, chs. 5, 7).
ceptions that were prevalent in Brahmanism at an There was a second consequence of the belief in
early age, because this tradition has preserved for the efficacy of mantras, less immediately visible in
us a corpus of texts, at least part of which goes the form of a surviving literature, but equally
back to a time before any script was used in the influential. The efficacy of Sanskrit mantras
Indian subcontinent. This corpus of texts, collec- showed, at least to the Brahmans, that Sanskrit,
tively known as the → Veda, consists in part of sac- their eternal language, was very close to objective
rificial formulas, or → mantras. Mantras had to be reality. Sanskrit mantras could have an effect on
Philosophy of Language 673
the objective world because a close link connected which itself became the starting point for an
Sanskrit with that objective world. Sanskrit distin- extensive grammatical literature that consisted
guished itself in this regard from the dialects spo- partly in commentaries on Pāṇini’s text, the
ken by less eminent people. These dialects from Aṣtạ̄ dhyāyī, and partly in independent treatises
our point of view were languages in their own that were yet profoundly influenced by Pāṇini’s
right, and for the Brahmans no more than corrup- grammar (Bronkhorst, 1981; see also → language
tions of the only real language, Sanskrit. No such and linguistics).
thing as historical linguistics could therefore exist We will see below that the beliefs about the
from the Brahmanical point of view. The one real Sanskrit language just described came to exert
language, Sanskrit, did not change and had not an influence on Brahmanical philosophy. These
changed from beginningless time. All other lan- beliefs themselves did not yet amount to a
guages were not real languages but corruptions philosophical position, but they shaped future
due to people too lazy or ignorant to learn the one thinking.
real language. Some scholars have made a somewhat different
The conviction that Sanskrit is close to reality is claim. They have drawn attention to the central
central to Brahmanical thought. It finds expres- significance of language study in Indian culture,
sion in various ways. Most notable among these is and on the sophisticated way in which the gram-
the belief that the form of a word tells us some- marian Pāṇini in particular succeeded in dealing
thing about the object it designates. This belief is with complex linguistic issues. They have con-
behind the numerous semantic etymologies found cluded from this that Pāṇini’s grammar played a
in Vedic literature, primarily in the Vedic texts role in India comparable to Euclid’s geometry in
called Brāhmaṇas (→ Vedas and Brāhmaṇas). These Europe in becoming a methodical guide for more
etymologies have nothing whatsoever to do with recent philosophers. In support of this claim, they
the history of words, as in the case of historical can point out that the study of Pāṇini, or one of his
etymologies studied nowadays by historical lin- successors, was fundamental in the classical for-
guists. Semantic etymologies are something dif- mation of scholars, so much so that all of them
ferent altogether. They start from the assumption were acquainted with the grammatical method
that the similarity between words reveals a com- developed in these works. However, it is very hard,
mon feature shared by the things designated. perhaps impossible, to give more substance to this
Semantic etymologies assume, at least in vedic lit- claim. No one has yet been able to show in what
erature, the presence of hidden links between way philosophical thinking followed a pattern
objects, links that can only be brought to light derived from Pāṇini’s grammar. It is, however,
through the analysis of the words that designate beyond doubt that grammatical rules and argu-
them. ments are frequently cited in certain philosophi-
Interestingly, the belief in the significance of cal developments, some of which will be considered
semantic etymologies gave rise to an early attempt below. The influence of the brahmanical presup-
at systematization known as nirukta, “etymology.” position as to the close connection between lan-
Its classical treatise, also called Nirukta, was com- guage (i.e. Sanskrit) and reality will become clear
posed by a certain Yāska, who must have lived as the discussion proceeds.
after Pāṇini but before Patañjali (see below), and
therefore in or around the 3rd century BCE. This
discipline, considered a vedic auxiliary science Buddhist Systematic Philosophy
(vedāṅga), attempted to formulate the rules that and the Role of Language
govern semantic etymologizing. One of its princi-
pal rules, unsurprisingly, is that in establishing Buddhist thinkers had no special attachment to
links between different words, considerations of the Sanskrit language, nor to any other language
meaning have to play a central role. for that matter. They preserved what they believed
It is against this same background that we have were the words of the Buddha in different lan-
to understand the interest in Sanskrit morphology guages, often a language close to the one spoken in
and linguistics in general that characterized the the region where they had settled. It is true that in
late vedic period. This interest culminated initially the long run, some Buddhists came to think that
in the famous grammar of Pāṇini (4th cent. BCE), the language in which they preserved the words
674 Philosophy of Language
of the Buddha was the language in which the objects in the external world: for example, there is
Buddha had preached. Some went further and no such thing as a chariot; there is only this vast
claimed, no doubt under Brahmanical influence, sum of momentary constituent elements (Bronk-
that the language of their canon was the original horst, 2009, 55ff.).
language from which all other languages derived; This remarkable vision of the world has but lit-
this happened in the case of Pali, a language that tle in common with the original teaching of the
the Buddhists concerned call Magadhi, namely, Buddha, and the Buddhists of northwestern India
the language of Magadha, the region where the had some difficulty anchoring it in the inherited
Buddha had preached. Brahmanical influence Buddha words. However, as a philosophy, this
took a different shape in northern India, where the vision came to flourish and exert a determining
Buddhists, having preserved the Buddha word in influence on most of the subsequent philosophi-
local languages for half a millennium, changed cal developments in Indian Buddhism. My ques-
over to Sanskrit and even came to believe that the tion is: What has this vision to do with the relation
ancient dialect in which some of their texts had between language and reality?
been preserved was a form of vedic Sanskrit This relation enters into the picture in a manner
(Bronkhorst, 1993). that can be usefully illustrated with the example of
But whatever the language they used for a chariot. We see and believe in the existence of a
religious purposes, the Buddhists were not, at chariot, when in reality there is no such thing.
least not initially, inclined to believe that any of How are we misled into entertaining such an
these languages – or any other language for that incorrect idea? As a result of language. We believe
matter – had a particularly close connection with there is a chariot, because there is the word “char-
reality. However, developments now to be sketched iot.” The same applies to all the numerous other
led to a situation where a number of Buddhists things we believe populate the world. In reality,
came to believe that there was a close connection they do no such thing, but we are tricked by the
between language and our experience of the language we speak.
world. Two observations have to be made at this
This connection with language was initially not point. The relation postulated by these Buddhists
obvious in the philosophical developments that between language and experience does not con-
the Buddhist school, called Sarvāstivāda, in north- cern reality as it really is. In deepest reality there
western India underwent during the final centu- are only vast numbers of momentary dharmas, as
ries preceding the Common Era. For reasons we have seen. The experience, whose parts corre-
that may be connected with the Hellenistic sur- spond to the words of language, is not a reliable
roundings in which they found themselves, these representation of reality. In other words, from this
Buddhists made an attempt to systematize the Buddhist perspective, language is not closely
teachings they had inherited, ultimately from the related to reality. Quite on the contrary, language
Buddha, they thought. These inherited teachings tricks us into believing in a world that is ultimately
contained lists of items that had been extracted not real. The parts of language that do the tricking
from the sermons of the Buddha. These items are its words. Chariots, persons, and so many
frequently referred to mental states, some others other things are ultimately unreal, but are believed
to physical elements. The name that came to to be real because of the words that supposedly
be used for these items was → dharma, and the designate them.
Buddhists of northwestern India developed the The systematizations of the Sarvāstivāda school
idea that these dharmas are the ultimate constitu- of Buddhism appear to be the first manifestations
ent elements of human beings, and by extension of of systematic philosophy on the Indian subconti-
everything else as well. For various reasons, these nent. They were not primarily concerned with
Buddhists also concluded that the dharmas must language, yet language played an important role in
be momentary. On top of this, they stated that them, as we have seen. The words of language are
composite objects have no separate existence. In responsible for the fact that we believe, incorrectly,
other words, a person is nothing but a collection that we live in a world in which there are chariots,
and sequence of a large number of small and persons, and other such things. However, this
momentary entities; beside these numerous enti- vision of the Sarvāstivādins had to face an obvious
ties, there is nothing that one might call a person. difficulty. Words, like everything else, are them-
The same applies by extension to macrocosmic selves accumulations – most specifically, succes-
Philosophy of Language 675
sions – of dharmas. As such, they have no existence In this situation Brahmanism developed two
of their own, like all other accumulations of ontologies of its own. The one that is most inter-
dharmas. This is problematic in that our incorrect esting in the present context is called → Vaiśeṣika.
beliefs in objects, such as chariots and persons, It is most interesting in that it takes over the
depend on words. If there are no words, what then Sarvāstivāda idea that the world of our experience
is responsible for our mistaken conceptions? has a close connection with the words of language,
It seems that the early Sarvāstivādins experi- but adjusted to its own Brahmanical presupposi-
enced this as a problem. This we may conclude tions. In Vaiśeṣika the world of our experience
from the fact that they included among their lists coincides with the real world, and the only lan-
of existing dharmas a number of items that cor- guage that is taken into consideration is Sanskrit.
respond to words and sentences. In other words, For Vaiśeṣika, then, language is not a source of
they postulated, beside the succession of auditory confusion that makes us believe in the external
and other elements that make up the word “chariot,” reality of a world that does not in that form exist,
the existence of a momentary dharma that some- as it is for Sarvāstivāda. Quite on the contrary, for
how represents the whole word “chariot.” This Vaiśeṣika language (i.e. the Sanskrit language) is a
went of course against their general principle of source of information about the real world. If we
not recognizing the existence of wholes, but wish to develop an ontological scheme, namely, a
strictly speaking they managed to avoid this scheme of what there is, the Sanskrit language is
inconsistency. The “wholes” postulated in the case the means par excellence to find it.
of words (and sentences) were, strictly speaking, This fundamental conviction of the Vaiśeṣikas
not wholes, but entities (dharmas) in their own finds expression in their philosophy. Vaiśeṣika
right, which somehow accompanied the succes- claims that all things (artha) in the world fall into
sion of auditory elements that gave expression to three categories: substances, actions, and quali-
that meaning. In this complicated way, the ties. These three categories correspond to the three
Sarvāstivādins had again words, and they could main types of words already distinguished by the
again claim that our illusion concerning the world grammarian Patañjali (2nd cent. BCE): nouns,
was due to these words. These linguistic dharmas verbs, and adjectives. In order to make this onto-
of the Sarvāstivādins might be looked upon as an logical scheme coherent, various Vaiśeṣikas added
idiosyncrasy of the philosophy of these Buddhists a number of further categories, arriving at a vary-
without deeper significance. Such a conclusion ing total of six, seven, ten, or even more. However,
would not be justified, if for no other reason than even in their own understanding of their system,
that more recent brahmanical thinkers happily these further categories were add-ons.
drew inspiration from these Sarvāstivāda inven- The three main categories had subdivisions.
tions in order to deal with problems they had to The fundamental rule followed to find those sub-
face (Bronkhorst, 1987, ch. 3/7). divisions was that they had to correspond to
words. The list of substances illustrates this par-
ticularly well: it corresponds by and large to the
Reality as Language Incorporated nouns of the Sanskrit language. The list of quali-
ties raises more difficult issues, for they are not all
The Sarvāstivāda philosophy constituted a chal- referred to by adjectives or other words; some of
lenge that Brahmanical thinkers could not leave them, in particular, are designated by nouns.
unanswered. Buddhism possessed in this philoso- Questions about their existence are resolved by
phy a coherent and well-thought-out system of considering language use. The quality “number”
thought, at a time when Brahmanism had nothing must exist, because expressions such as “one pot,”
of the sort. It appears that Brahmanism had to “two pots,” and so on are common usage. The exis-
come up with something comparable, so as to be tence of the quality “dimension” (pramāṇa) is
able to defend itself in the public debates that shown by the use of the word “measure” (māna).
sometimes took place at the royal courts. The out- There are further examples of this kind in the
come of such debates could be more than aca- classical surviving standard work of Vaiśeṣika,
demic and might affect the court’s willingness to the Padārthadharmasaṃ graha of Praśasta (or
support one group rather than another (Bronk- Praśastapāda, c. 6th cent. CE). Vaiśeṣika is in this
horst, 2007). way not so much a philosophy of language, as it is
676 Philosophy of Language
a philosophy inspired by language, or rather: a phi- expressed in language, and a listener. From the
losophy inspired by certain beliefs about the San- point of view of the listener, the received message
skrit language (Bronkhorst, 1992). will be reliable if two conditions are fulfilled:
Vaiśeṣika ontology became preponderant in (1) the speaker is reliable, and (2) the message
Brahmanism, especially in those schools of is correctly interpreted by the listener. In the
philosophy that developed ideas on language case of the Veda, there is no speaker who formu-
(→ Mīmāṃ sā, → Nyāya, → Bhartṛhari). The other lated the message initially. The chance that the
main early Brahmanical ontology, → Sāṃ khya, Veda is used to mislead us is therefore excluded,
was not inspired by conceptions about language in and we must conclude that the message of the
the way Vaiśeṣika was. This does not mean that it Veda is reliable, if only we know how to interpret
denied the close connection between language the text correctly. This is the self-assigned task
and reality. This will become clear in the discus- of Mīmāṃ sā.
sion below of the “linguistic crisis” that hit all How does one interpret a text correctly? Once
Indian schools of philosophy – Brahmanical, Bud- again, it is a matter of eliminating sources of
dhist, and even Jain – in the early centuries of the error. The aim is to get as close as possible to
Common Era. the text, letting it speak for itself. Mīmāṃ sā devel-
oped a whole list of principles of interpretation,
which are in the end nothing but concrete mani-
The Word as Source of Knowledge festations of this general aim and are justified
in this manner.
Before discussing the “linguistic crisis,” we have to The belief that the Veda is without beginning
pay attention to another important development has some consequences that have to be taken into
in Brahmanism that shaped the way it came to consideration. Being without beginning, it is not
think about language: the interpretation of the posterior to any event that has ever taken place in
Veda, known as Mīmāṃ sā, whose main classical the history of the world. Yet many vedic passages
text is the commentary of Śabara (c. 5th cent. CE). give the impression of relating such events, usu-
Mīmāṃ sā did not initially present itself as a school ally in the form of what we might call myths. None
of philosophy, but as a school of Vedic hermeneu- of these stories can be taken literally, because the
tics. Recall that Brahmanism had preserved an Veda had already existed for an eternity at the time
enormous corpus of literature, the Veda, at least in when these events supposedly took place. Stories
part, in the form of memorized texts. The vedic and other references to historical events must
corpus contains mantras, as we have seen. It therefore be interpreted differently. They must be
also contains texts globally referred to by the read in context and have their place in the Veda in
Mīmāṃ sakas as Brāhmaṇas: texts in prose related order to illustrate or lend support to vedic con-
to the solemn vedic sacrifices. The Mīmāṃ sakas tents that can be taken literally.
were primarily interested in these Brāhmaṇas, in What are the vedic contents that should be
which they expected to find matter useful or taken literally? The process of elimination shows
essential for the performance of those rites: injunc- what they are. All statements of fact in the Veda
tions, information as to how to proceed and must be interpreted metaphorically. This is partly
resolve apparent contradictions, and so on. They due to the fact that the Veda cannot refer to his-
based their endeavors on some global assump- torical events, for the reason indicated above.
tions about the Veda. However, other, nonhistorical, statements of fact
Most important among these assumptions was contained in the Veda have another disadvantage.
the claim that, like its language, the Veda itself, They might turn out to be in contradiction with
too, had no beginning: the Veda had always been our sense experience. This would be serious, for it
there. This is of course only possible if the Veda would imply competition between two different
had had no author. Being without author, the Veda means of knowledge: the vedic word, and percep-
is pure speech. This, from the Mīmāṃ sā point of tion. This difficulty does not exist in the case of
view, has important consequences. To understand injunctions. Injunctions tell us what we must do.
these, consider first ordinary verbal communica- No sense experience can ever be in conflict with
tion between human beings. In ordinary verbal an injunction, because perception informs us
communication, there will be a speaker, a message about states of affairs, not about obligations. With
Philosophy of Language 677
regard to obligations, we have only one ultimate correctly interpreted by the listener) retains its
source of information, namely, the Veda. pertinence even in the case of the Veda. Unlike the
Mīmāṃ sā arrives in this way at an interpreta- Mīmāṃ sakas, the Naiyāyikas had to argue that the
tion of the Veda in which injunctions are central. Veda had a trustworthy author, and indeed that
They are what the Veda is all about. Everything the reliability of the Veda depended on the trust-
else is to be interpreted in connection with those worthiness of that author. In an important sense,
injunctions. In practice, the task is complex, but they had to admit that the reliability of the Veda
the theoretical basis is relatively simple and depends on the same factors that make an ordi-
straightforward. In the end it amounts to this – nary statement reliable, that is, the trustworthi-
that the word par excellence, the Veda itself, tells ness of the speaker. They solved this problem by
us what to do (Bronkhorst, 1997). attributing the Veda to the most reliable person
The second half of the 1st millennium CE saw there is: God himself. This did not change the fact
the rise into prominence of a school that claimed that their position – to the extent that the word
to be no more than an improved version of classi- (namely, linguistic utterances) is a reliable source
cal Mīmāṃ sā, but which opened up a path for of knowledge – was bound hands and feet by the
completely new developments. This school – which supplementary requirement that the speaker, who
referred to itself by various names, among them uttered these words, is himself reliable (NyāS. and
Śārīrakamīmāṃ sā and Brahmamīmāṃ sā, later also NyāBh. 2.1.50–67; trans. Jha, 1939).
Uttaramīmāṃ sā, and one of whose most impor- We saw in the preceding section that the
tant early authors was → Śaṅkara (c. 700 CE) – came Vaiśeṣika ontology, too, is ultimately based on the
to be known by the name → Vedānta, or Vedāntism. word as source of knowledge. The situation here is
It owes this name to the fact that the parts of the nonetheless different from the one described in
Veda known as vedānta (end of the Veda), better connection with Mīmāṃ sā and Nyāya. There (in
known as → Upaniṣads, played an important role Mīmāṃ sā and Nyāya) I talked about the reliability
in its reflections. The importance of these portions of certain linguistic utterances: primarily those
of the Veda is, however, in a certain sense due to contained in the Veda, further those pronounced
coincidence. The Brahmamīmāṃ sā applied the by reliable speakers. In the case of Vaiśeṣika, how-
same principles as classical Mīmāṃ sā, but slightly ever, I considered the relationship between the
improved them with the result that certain upani- structure of reality and the structure of the
shadic statements gained center stage. Recall that Sanskrit language; the Veda or other reliable state-
classical Mīmāṃ sā had maintained that only in ments did not play a role here. Interestingly,
the case of injunctions can we be sure that there some Buddhist thinkers, most notably Dignāga
will be no conflict with other means of knowledge. (6th cent. CE), came to adopt an ontological posi-
The Vedāntists added vedic statements that teach tion similar to that of Vaiśeṣika, ultimately for
knowledge about → brahman; these vedic state- almost the same reason. Dignāga could not, of
ments happen to occur in the Upaniṣads. They course, claim that language would tell us some-
argued that knowledge about brahman can only thing about the world as it really is, for he, like
be obtained from these vedic statements, not by most other Buddhists, thought that language is
any other means of knowledge. Since no conflict rather the source of our mistaken ideas about real-
between different means of knowledge is therefore ity. However, language, being the reason why we
possible, these statements have to be taken liter- experience the world the way we do, is capable of
ally (Bronkhorst, 2007). providing us with information about the world
The Nyāya school of thought was willing to of our experience. In order to find out more
agree with Mīmāṃ sā in its claim that the word – about this, Dignāga undertook an analysis of
and especially the linguistic expressions that are the relationship between words and things, which
found in the Veda – is a reliable source of knowl- I will discuss below. Here it must suffice to observe
edge. Unfortunately, and for reasons that will be that he arrived at an ontological scheme similar
discussed below, it adopted the position that the to that of Vaiśeṣika, be it that in his case this
relation between words and things was conven- ontological scheme was limited to the aspect of
tional. For them, therefore, the first of the two reality (ultimately unreality) ruled by language
conditions specified above (a received message is (Katsura, 1979).
reliable if the speaker is reliable and the message is
678 Philosophy of Language