A Philological Approach To Buddhism

The Bukkyō Dendō Kyōkai Lectures 1994, K. R. Norman, School Of Oriental And African Studies University Of London (1997)

Published in 1997, the lectures were originally intended for an audience familiar with Buddhism but not with philology.
The contents cover the relationship between Buddhism and philology, its origins, oral tradition, regional dialects, writing, Sanskritisation, Aśoka, canonicity, and the commentarial tradition.

I - Buddhism and Philology

At a 1993 Buddhist Studies conference, many scholars felt marginalized, while the author, a philologist, felt his discipline was central to the subject, even if unpopular.
Modern scholarship often prioritizes new, “trendy” topics over the fundamental work of re-editing and re-translating core texts, which is crucial as our linguistic and lexicographical tools have improved.
To seriously investigate a religion like Buddhism, one must examine its foundational texts, which requires a philological approach to understand how language and meaning have changed over time and to identify later interpolations.
Early Western knowledge of Pāli came from envoys and Christian missionaries, but academic study flourished with the 19th-century rise of comparative philology and the acquisition of manuscripts from Asia.
The philological approach asks not just “what does a text mean?” but “how or why does it mean it?”, focusing on grammar, syntax, and etymology rather than relying on intuition.
Translating by intuition or guesswork can lead to significant errors, as demonstrated by T.W. Rhys Davids’ mistranslation of a passage in the Dīgha-nikāya.
Philology clarifies the grammar of fundamental concepts like the “four noble truths,” showing the correct translation is “The noble truth that ‘this is suffering’,” not “this suffering is a noble truth.”
A philological analysis of key terms provides more accurate understanding; for example, amata (an epithet for nibbāna) means “without death” because it is “without birth,” not “immortal” in the sense of eternal life.
Comparing parallel versions of texts in different languages (Pāli, Sanskrit, Gāndhārī) helps identify and correct textual corruptions and misunderstood grammatical forms.
Many commonly accepted translations of titles and terms are likely incorrect. Philology suggests “the truth of the noble one (the Buddha)” is a better translation for ariya-sacca (noble truth), and Mahāparinibbānasutta means “the large text about the decease,” not “the book of the great decease.”
The prefix pari- in parinibbāna does not mean “final,” and the word kusala is better translated as “good” rather than “skilful” in many doctrinal contexts.
Scholars often rely on outdated and flawed resources like the Pali-English Dictionary (PED), treating its definitions as infallible and perpetuating errors.

II - Buddhism and Its Origins

A philological approach, which examines the meaning and origin of words, can provide insights into the political, economic, social, and religious environment in which Buddhism emerged.
Early Pāli texts suggest a context of increasing urbanisation, trade, and a rising merchant class (vaiśyas) who, along with the ruling class (ksatriyas), were receptive to a religion that challenged the social superiority of the brahmanical caste.
Philological evidence suggests the Buddha’s family were likely minor tribal chieftains of non-Indo-Aryan origin who had recently been assimilated into the caste system.
The Buddha accepted, modified, or rejected existing religious beliefs of his time, including an early idea of a single afterlife, the later doctrine of samsāra (endless rebirth), and the existence of Vedic gods (whom he demoted to super-men subject to rebirth).
A core part of the Buddha’s teaching was the rejection of the Upanisadic concept of a universal Self (ātman or brahman). His doctrine of anattā (“not self”) was a direct refutation of this, arguing that since the world is impermanent and characterized by suffering, it cannot be part of a permanent, blissful Self.
The Buddha’s meditative practices and terminology were developed from, and shared with, other contemporary non-brahmanical ascetic (śramana) movements.
The Buddha’s awakening (bodhi) was an awakening to the attainment of nibbāna. The concept of nibbāna and many meditative states (jhānas) likely pre-dated Buddhism and were incorporated into its system.
After his awakening, the Buddha analysed his experience through the 12-fold chain of dependent causation, concluding that his new knowledge (vijjā) broke the chain that begins with ignorance (avijjā) and leads to suffering, thus ending his future rebirths.
Buddhism offered two paths to release: an immediate, meditative path for monastics and a gradual, merit-based path for lay followers, whose generosity (dāna) was essential for supporting the monastic community.
Buddhism shares significant terminology, concepts (like previous prophets), and anti-brahmanical literature with Jainism, suggesting both emerged from a common śramana background rather than one being an offshoot of the other.
Unlike the Jains, who believed in a permanent individual soul, the Buddha denied both a universal and an individual self. This made nibbāna difficult to describe, leading to its definition through negatives as a state free from the suffering, birth, and death of samsāra.
Philological and archaeological evidence suggests a later date for the Buddha’s life (c. 480-400 BCE) than traditionally accepted, aligning better with the social conditions described in the texts.
While the Buddha’s teaching was unique, it held much in common with contemporary movements, with its most original feature being the combination of a belief in transmigration with the denial of a self to transmigrate.

III - Buddhism and Oral Tradition

The Buddha’s teachings were initially spread by word of mouth, and early Buddhist scriptures were transmitted orally.
Evidence for the oral tradition includes the absence of rules for writing materials in the Vinaya, a vocabulary centred on “hearing” and “speaking”, and the Theravādin tradition that the canon was first written down in the first century B.C.E.
Early recitations may have been flexible, but the need for consistency led to codification, beginning with the Pātimokkha (monastic rules).
After the Buddha’s death, a joint recitation (sangiti) was held, and the bhānakas (“speakers”) system was established to preserve and transmit the texts.
Different groups of bhānakas were responsible for different sections (nikāyas) of the canon, working to homogenize the language and structure.
The traditional account of the bhānakas system being established immediately after the first recitation is likely an anachronism; it was probably a later development to authenticate the Theravādin canon.
Various mnemonic devices were used to aid memorization and recitation:
- Stock phrases and repetition: Standardized openings and the verbatim repetition of sentences and paragraphs.
- Lists and indexes (mātikās, uddānas): Tables of contents used to structure recitations.
- Numerical arrangement: Organizing teachings by number, as seen in the Anguttara-nikāya.
- Waxing Syllables Principle: Ordering strings of words based on their increasing number of syllables.
The oral tradition was vulnerable to errors and could be broken if the chain of teachers and students was interrupted.
An instance in Sri Lanka where a text was almost lost due to war and famine highlighted the risks of relying solely on oral transmission, likely prompting the decision to write the canon down.

IV - Buddhism and Regional Dialects

The common belief that the Buddha preached in Prakrits and refused translation into Sanskrit is based on a difficult-to-interpret passage and may be an example of an academic idea becoming “fact” through repetition.
Middle Indo-Aryan (MIA) dialects, or Prakrits, were in use during the Buddha’s time, and it is likely he preached in various local vernaculars, including Old Māgadhī.
The language of the Theravādin canon, Pāli, is not Māgadhī, despite what commentators claimed. Pāli is predominantly a “western” dialect, which indicates the texts were translated at least once.
Pāli contains many “anomalous” forms (e.g., eastern nominative endings in -e, voicing of consonants, l for r) which are remnants from the eastern dialects through which the teachings were originally transmitted.
These anomalous forms can be traced using Aśokan inscriptions and grammarians’ works, suggesting the texts passed through the Magadha region and that a diversity of sub-dialects existed even in that area.
The existence of “hyperforms” (incorrect back-formations like Yamataggi for Jamadagni) shows that translators were aware of dialectical differences but sometimes made errors.
Anomalous forms were retained for several reasons: they were important technical terms (bhikkhu, nibbāna), they were kept by oversight, or the meaning was ambiguous.
Pāli is a composite, literary language, not a pure historical vernacular. It also contains later “Sanskritisms” (re-introduced Sanskrit forms), which were likely added in Sri Lanka.

V - Buddhism and Writing

The Pāli canon was first written down in Sri Lanka during the reign of King Vattagāmini Abhaya (first century B.C.E.) due to a perceived “loss of living beings,” which may refer to a breakdown of the oral recitation (bhānaka) system, famine, invasion, and political pressures.
Later sources mention a joint recitation (sangīti) before the writing, possibly a later invention to grant the written canon traditional authority.
The process of writing down the canon provided an opportunity to homogenize the language, although this was incomplete, as variations between different texts (nikāyas) remain.
The date of writing’s introduction to India is debated, but Emperor Aśoka used Kharoṣṭhī and Brāhmī scripts around 260 B.C.E. The author supports the view that Brāhmī was adapted from a Semitic script by traders before Aśoka’s time, used initially on perishable materials for commerce and administration.
Some textual errors, such as the metathesis of syllables, suggest that manuscripts were being copied even before the canon was officially written down.
Writing introduced new types of errors, including dittography (repetition), haplography (omission), and those caused by deficiencies in early scripts.
Early Brāhmī script did not write double consonants or consistently mark long vowels and nasalization (anusvāra), leading to ambiguity between words like upecca (approach) and uppacca (leap up), and confusion between active (muñcati) and passive (muccati) verb forms.
Similar ambiguities in other scripts, like Kharoṣṭhī or later Burmese script, also led to scribal errors, which can sometimes help trace a manuscript’s transmission history.
Writing led to the use of abbreviations (peyyāla) for repeated passages and forced scribes to choose a single spelling for puns that relied on oral ambiguity, sometimes resulting in a loss of the original meaning.
Once written, the Theravādin canon became relatively fixed, with few later additions or linguistic changes. However, the oral tradition continued for some time, allowing variant readings to persist.
The shift to writing broke the control of the bhānakas, who had previously been able to reject new teachings inconsistent with the established oral canon.
This change made it easier for new schools, including Mahāyāna, to produce their own written texts and claim them as the Buddha’s word, as they no longer needed to be validated by a long-standing oral tradition.
The decision to write down the Theravādin canon may have been a response to the need to establish its authenticity against the emerging written scriptures of these other Buddhist schools.

VI - Buddhism and Sanskritisation

Sanskritisation is defined in two ways: the broader adoption of Sanskrit by Buddhism, which initially opposed it, and the specific process of replacing Middle Indo-Aryan dialects in texts with Sanskrit.
This process occurred in stages, from adding Sanskritisms to Prakrit, to translating texts into Buddhist Hybrid Sanskrit, to composing new works directly in Sanskrit.
The trend likely began in the 2nd century B.C.E. as a competitive response to new Buddhist sects that were composing their teachings in Sanskrit, the language of culture.
Sanskritisation involved restoring Sanskrit phonology, morphology, vocabulary, sandhi (word juncture rules), and correcting metre in verse.
In the Pāli canon, this process occurred in two distinct phases: an early phase around the time the canon was written down, and a much later phase (after the 12th century C.E.) influenced by grammarians aiming to make Pāli more like Sanskrit.
The process created significant problems, particularly with homonyms—words in Middle Indo-Aryan that had multiple meanings derived from different Sanskrit originals. Translators had to choose one meaning, which could lead to errors or the loss of intentional ambiguity.
Examples of translation issues include confusion between dīpa (lamp) and dvīpa (island), and pīti (drinking) and prīti (joy).
Puns and etymologies that worked in Middle Indo-Aryan were often lost in Sanskrit. For instance, the pun linking samana (ascetic) to the root sam- (to be at rest) is lost when the word is Sanskritised to śramana (from śram- “to make an effort”).
Incorrect back-formations occurred, such as the Middle Indo-Aryan sutta (“well-spoken”) being Sanskritised to sūtra (“thread”), and bodhisatta becoming bodhisattva.
Sanskritisation had a profound effect on Buddhism. While it made texts accessible to a wider, educated audience across different schools, it also distanced the religion from the common people.
This shift towards an academic, elite language, in contrast to Jainism’s continued use of vernaculars, may have contributed to Buddhism’s eventual decline in India after its monasteries were destroyed.

VII - Buddhism and Aśoka

Common portrayals of Aśoka as the “first Buddhist Emperor” who made Buddhism a state religion are largely based on Buddhist chronicles like the Mahāvamsa, not his own inscriptions.
According to the Mahāvamsa, Aśoka killed his brothers, converted to Buddhism, exclusively supported the Buddhist sangha, built 84,000 vihāras, and personally settled a schism before patronizing the Third Buddhist Council.
Aśoka’s own edicts present a different picture: his conversion was prompted by remorse over the Kalinga war, an event the chronicles ignore. He refers to his living brothers and sisters, contradicting the fratricide story.
While inscriptions confirm Aśoka was a Buddhist layman (upāsaka), his personal ‘dhamma’ that he promoted was a universal moral and ethical code, not Buddhist doctrine. It focused on obedience, non-killing (ahimsā), and truthfulness, and lacked core Buddhist concepts like nibbāna or the Four Noble Truths.
Aśoka’s goal for his people was happiness in this world and heaven in the next, not the Buddhist goal of escaping the cycle of rebirth.
Contrary to the chronicles, the edicts show Aśoka was impartial, honouring and supporting all religious sects, including brāhmanas, Jains, and Ājīvikas, to whom he donated caves.
His policy of dhamma-vijaya (“dhamma victory”) was not a spiritual conquest to spread Buddhism, but an effort to persuade neighbouring kingdoms to adopt his moral principles of peace and non-violence as an alternative to military conquest.
The Schism Edict suggests he took action to unify the sangha, but the chronicles likely embellished this event, attributing actions to him that were probably carried out by his ministers (mahāmātras).
The Buddhist chronicles likely exaggerated Aśoka’s role to legitimize their religion in the face of rivalry from Brahmanism and Jainism, and to establish the Theravāda school as the orthodox tradition.
The expansion of Buddhism during his reign was more likely an indirect result of the peace, prosperity, and trade routes he established, rather than a direct, state-sponsored missionary effort.

VIII - Buddhism and Canonicity

The term “canon” is a Western concept that must be used carefully, as it can be anachronistic for early Buddhism. A canon can be an open collection of scriptures (allowing additions) or a closed one.
The Theravādin canon is now closed, whereas Mahāyāna traditions in Tibet and China had more open canons, allowing for later additions.
Early classifications, such as the nine aṅgas (types of texts), do not represent a formal canon. Similarly, terms like Tipiṭaka (“three baskets”) and pāli (“text”) are not synonymous with a closed canon, but rather describe divisions or the texts themselves as distinct from commentaries.
The most relevant early concept is Buddhavacana (“the words of the Buddha”), which was interpreted broadly to include not only the Buddha’s direct teachings but also the approved words of his disciples and later works like the Kathāvatthu, which was justified as having been foreseen by the Buddha.
The Theravādin canon evolved from an open collection of Buddhavacana to a closed one, with different sections closing at different times.
- The Vinaya-piṭaka likely closed after the second council but before the third (c. 250 BCE), as it mentions the former but not the latter.
- The Abhidhamma-piṭaka remained open until at least the third council, as it includes the Kathāvatthu, composed at that time.
Scholars debate the final date of the canon’s closure. Some argue for the time of Buddhaghosa (5th century CE), who listed its contents. Others suggest it was effectively closed when written down in the 1st century BCE, supported by the fact that later authoritative texts like the Milindapañha were not included.
Aśoka’s Bairāt inscription (3rd century BCE) names seven specific texts, indicating that individual suttas were authoritative, but it does not prove the existence of a fully formed, closed Tipiṭaka at that time.
Even after the list of texts was fixed, some texts with a flexible structure (e.g., Theragāthā, Kathāvatthu) may have had additions made until commentaries were written, which finalized their content.
In later periods, “quasi-canonical” or “apocryphal” texts were composed in Southeast Asia, mimicking the style of canonical suttas. These were likely created as teaching materials rather than forgeries intended to deceive.

IX - Buddhism and the Commentarial Tradition

Commentaries likely began with the Buddha explaining words with synonyms to aid understanding across different regions with varying dialects.
The earliest extensive commentary is the pada-bhājanīya (“analysis of words”) on the Pātimokkha, which is embedded within the Vinaya-piṭaka itself.
The Niddesa, a commentary on parts of the Sutta-nipāta, is the earliest separate commentarial text and has canonical status. It is characterized by repetitive, verbatim explanations, typical of an oral tradition.
Commentaries serve two functions: explaining the meaning of individual words and explaining the meaning of the phrases or sentences in which they appear.
According to tradition, Mahinda brought Buddhism and its commentaries to Sri Lanka, translating them into the local Sinhalese language for the islanders.
Evidence from dialectal variations preserved in the Pāli commentaries suggests that the commentarial material originated in North India and predates its introduction to Sri Lanka.
The Tipiṭaka and its commentaries (aṭṭhakathā) were transmitted orally until they were written down in Sri Lanka during the first century BCE.
In the fifth century CE, Buddhaghosa translated the existing Sinhalese commentaries (such as the Mahā-aṭṭhakathā and Andhaka-aṭṭhakathā) into Pāli, as these texts were reportedly no longer available in India.
Additions were made to the Sinhalese commentaries in Sri Lanka until at least the early fourth century CE.
Parallels between Pāli commentaries and the canonical texts of other Buddhist schools (e.g., Gāndhārī, Chinese, Tibetan versions) suggest that some commentarial material is very ancient, possibly dating back to a time before the sects separated.
Commentators used various etymological methods, including folk etymologies (nirukti) for religious purposes, which sometimes reveal information about the dialects in which they were composed.
Interpretations of canonical terms evolved over time; for example, the phrase tasā vā thāvarā vā (“moving or unmoving creatures”) was later interpreted by commentators to mean “those with craving and arahats”.
The original Sinhalese commentaries were eventually lost and superseded by Buddhaghosa’s Pāli works, which themselves later required sub-commentaries (ṭīkā).
The commentarial tradition is valuable for explaining difficult passages, preserving superior textual readings, and showing the development of Buddhist thought, but it can also be obscure, incorrect, or even introduce errors into the canonical texts.

X - Philology and Buddhism

A philological approach to Buddhism provides a better understanding of the texts, the transmission of the tradition, cultural influences, and the development of Buddhism itself.
Previous lectures in the series covered the background of the Buddha’s teaching, its oral nature, regional dialects, the impact of writing, Sanskritisation, the role of Aśoka, the concept of canonicity, and the commentarial tradition.
Future philological work could help date early Buddhist texts and define the development of the religious environment in India from the 5th century B.C.E. to the 5th century C.E.
There is a fear that the number of philologists with the necessary skills in Indian, Tibetan, and Chinese is decreasing due to university funding models that disadvantage subjects with low student numbers.
Much modern work in Buddhist studies is done by those without adequate philological training, leading to scholarship based on translations, with errors in the use of Pāli/Sanskrit terms and diacritics.
Many general books on Buddhism, including recent ones, are outdated or misleading because they rely on a Pāli-centric view and do not incorporate findings from Sanskrit, Tibetan, and Chinese sources.
Effective research requires collaboration between specialists in different traditions and a rigorous understanding of the original languages, enabling scholars to critique translations and dictionaries rather than just relying on them.
The author expresses pessimism about the future of Buddhist studies, fearing that it will increasingly be taught by and to people with no knowledge of the original languages, leading to oversimplification and a loss of scholarly depth.