SAVOURING THE FLAVOUR OF THE COMMENTARIAL OCEAN: CONSISTENCY AND 

DIVERSITY IN THE EARLY BUDDHIST COMMENTARIAL TRADITION 

 

 

 

 

 

 

 

A Dissertation 

Presented to the Faculty of the Graduate School 

Of Cornell University 

In Partial Fulfilment of the Requirements for the Degree of  

Doctor of Philosophy 

 

 

by 

Liyu Hua 

August 2023



 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

© 2023 Liyu Hua 

 

 

 

 

 

 

 

 

 

 



 

 

SAVOURING THE FLAVOUR OF THE COMMENTARIAL OCEAN: CONSISTENCY AND 

DIVERSITY IN THE EARLY BUDDHIST COMMENTARIAL TRADITION 

 

Liyu Hua, Ph. D. 

Cornell University 2023 

 

This dissertation focuses on the emergence of Buddhist commentaries in Pali and 

Gandhari as new exegetical genres in first-millennium South Asia, aiming at revealing the 

diversity of Buddhist commentaries and the multiple ways of interpreting Buddhist canonical 

texts. Comparing radically different commentaries and situating them in the history of religions, 

I argue that the Buddhist commentators maintained a tension between doctrinal consistency and 

methodological diversity in the multireligious world of the mid-first millennium. Specifically, I 

examine the commentarial styles in the Niddesa and compare them with the various methods of 

definition adopted in other commentaries, through a case study of the concept “suffering.” I find 

that, with the transition from oral to written commentaries, the scholastic tradition of 

Abhidharma/Abhidhamma contributed to the stylistic transition in commentaries and enabled the 

Buddhist commentators to maintain doctrinal consistency in their scholastic development. This is 

demonstrated through consideration of the Buddhist hermeneutical handbooks Nettipakaraṇa 

and Peṭakopadesa, and the Gandhari commentaries in the British Library Kharoṣṭhī fragments, 

which seek to limit the interpretative possibilities by formulating stringent structural analysis of 

the root texts. Additionally, the Guidelines (Pali naya) in the Nettipakaraṇa and Peṭakopadesa 

also shape the teaching event by targeting specific types of audiences. While tracing their literary 

root to nidāna (source, provenance) in the canonical texts, the narrative commentaries such as 

the Dhammapada-aṭṭhakathā and Jātaka-aṭṭhakathā, with their temporal structure and narrative 



 

art, function as effective didactic tools that convey karmic laws and soteriological goals to the 

audience. Buddhist semantic derivation and Pali language ideology suggest close interactions 

with the Vedic semantic derivation (Sanskrit. nirvacana/nirukta) and Pāṇinian grammatical and 

Mīmāṃsā tradition while offering serious challenges to their Brahmanical rivals. It is likely that 

the political and religious tensions in mid-first millennium Lanka motivated the Mahāvihārins to 

express a unique language ideology in their commentaries and to patronise more Pali literary 

production. The study of the individual texts and traditions refreshes our understanding of the 

socio-religious contexts that gave birth to variegated Indic Buddhist commentarial literature and 

demonstrates Buddhist commentaries as a fertile ground for strengthening doctrinal consistency 

and improving didactic efficacy in interreligious contexts. 



 

 

v 

BIOGRAPHICAL SKETCH 

Liyu Hua achieved his Bachelor of Arts Degree from Peking University and Master of 

Theological Studies from Harvard Divinity School. He defended his doctoral dissertation in the 

field of Asian Literature, Religion, and Culture in 2023.  



 vi 

ACKNOWLEDGMENTS 

This dissertation was an ambitious project if not a perfect one. I have devoted the past 

four years to this project and experienced a period of radical changes in my life. Each chapter 

was finished in a different place, from snowy Upper State New York to the Park Library near my 

home in beautiful Subei Plain. Thanks to the guidance, friendship, and love of many truly 

remarkable people, I can finally present this dissertation as an outcome. 

First, I am grateful for my special committee at Cornell University. Daniel Boucher, my 

committee chair, guided me in the study of early Buddhist history in South Asia, patiently gave 

me numerous comments on my dissertation and other papers, and always made me feel at home 

with the gourmet meals. His keen eyes made me a better writer, and his challenging questions 

trained me on how to tell intriguing stories in academic presentations. 

Anne Blackburn gave me constructive feedback and encouraged me to continue my 

exploration into the Pali textual world. Her encouragement boosted my morale when I was far 

away in China writing the dissertation alone with unpredictable situations under Covid. Her 

advice on the early Lankan history helped me polish the historical arguments in Chapter 4. 

Lawrence McCrea untiringly guided me to read vyākaraṇa, kāvya, Mīmāṃsaka, and 

Buddhist philosophy, and, most importantly, Alaṃkāraśāstra. His erudition of Sanskrit tradition 

opened my eye to a broader Indian literary world and its depth, which I would never have a 

chance to peek into had I never come to Cornell. 

I am deeply indebted to Maria Heim, whom I regret not having met earlier. She kindly 

read the drafts of this dissertation and sent me insightful comments on philological details and 

theoretical framework. Shenghai Li and Joseph Marino, III also read my Chapter 2 and gave me 

very useful feedback from different angles. Shenghai Li informed me of some interesting 

Mahāyāna materials related to commentarial study. Joseph Marino, III taught me Kharoṣṭhī 

scripts, shared a lot of Gandhari resource with me, and supported me both as a teacher and a 

friend. My cohort Bruno Shirley also shared his insights on Sri Lankan inscriptions with me and 

gave me useful comments on the Conclusion. 



 vii 

I am grateful to my teachers for my MTS programme at Harvard University, Anne 

Monius, Charles Hallisey, Michael Witzel, Janet Gyatso, Parimal Patil, and Nicholas Harkness. 

Their courses were mind-blowing and planted the seeds of this dissertation. I am indebted to 

Wang Hongsheng, Ye Shaoyong, and Guo Tong, whose support and instruction made it possible 

for me to continue my study abroad. Duan Qing and Wang Bangwei generously opened their 

Indology classes to me when I was in an academic limbo. Devendra Shukla from the Central 

Institute of Hindi (Delhi) first stimulated my interests in Indian philology and culture with his 

solid Brahmanical training. I am grateful to Norihisa Baba, who kindly accepted me as a visiting 

scholar at Tokyo University in 2019, although I could not make the journey because of Covid-

19. I also benefitted from the help and comments of many scholars in other institutions. They 

include Mark Allon, Collett Cox, and Richard Salomon, with whom I interacted in the Kharoṣṭhī 

Klub; Trent Walker and Maxwell Brandstadt, who came to my talk at IABSC 2022 and raised 

constructive questions. I also thank Justin McDaniel and Stefanie Jamison for helping me look 

for an unpublished paper. My editor Michael polished parts of the draft and improved it style. 

Several teachers at Cornell University supported my intellectual pursuit and helped me 

with learning and teaching skills. I am grateful to Jane-Marie Law, Suyoung Son, Nick 

Admussen, Stephanie Divo, and Kelly King-O’Brien. My cohorts and friends at Cornell made 

Ithaca a less gloomy place to live in. I am grateful to many friends in the Department of Asian 

Studies, especially, Tarinee Awasthi, Geethika Dharmasinghe, and Sirithorn Siriwan. 

Special thanks are reserved for my family in China who financially and emotionally 

supported me throughout the long process of my study abroad; for Ting Lee Ling and Diao 

Xiaolong for being my sister and brother and extending their hands to me whenever I needed; for 

Gu Li, Zhang Shujian, and Ni Rui, who have been my confidants and always taken my stress; for 

Yun, who inspired me with her bravery and visions and accompanied me on this thorny path for 

too long; and Yung-hui, who has been constantly patient and understanding and supported me in 

my busiest and weakest time.  

Certainly, all the shortcomings are mine alone.  



 viii 

TABLE OF CONTENTS 

ABBREVIATIONS ........................................................................................................... x 

INTRODUCTION............................................................................................................. 1 

The Idea of Commentary ................................................................................................ 3 

Immeasurability of the Buddha’s Words ...................................................................... 12 

Regulating Interpretation .............................................................................................. 15 

The Concern about Language ....................................................................................... 23 

Overview of the Dissertation ........................................................................................ 29 

CHAPTER 1: RECITATION AND DEFINITION: THE DIVERSITY OF EARLY 

BUDDHIST EXEGETICAL METHODS................................................................................. 33 

The Earliest Stage of Buddhist Commentary................................................................ 33 

The Multiple Methods of Definition: A Case Study of Suffering ................................ 51 

Conclusion .................................................................................................................... 86 

CHAPTER 2: THE GUIDE TO ENLIGHTENMENT: METHODS OF 

INTERPRETATIONS IN THE PEṬAKOPADESA AND NETTIPAKARAṆA ..................... 88 

Introduction ................................................................................................................... 88 

Preparation for Hearing................................................................................................. 91 

The Structural Analysis: Condensation and Conversion ............................................ 102 

Laying the Foundations for Guidelines: Types of Persons ......................................... 114 

Path to Enlightenment ................................................................................................. 126 

From Structural Analysis to Comprehension .............................................................. 132 



 ix 

Conclusion .................................................................................................................. 139 

CHAPTER 3: TRACING OF THE ORIGIN AND PLAYING WITH TIME: 

NARRATIVE COMMENTARIES ......................................................................................... 141 

How to Begin a Teaching: Nidānas and verses .......................................................... 142 

The Narrative Art of Illustrative Stories in the Jātaka-aṭṭhakathā and Dhammapada-

aṭṭhakathā ............................................................................................................................... 175 

CHAPTER 4: THE NATURAL LANGUAGE: KNOWLEDGE OF LANGUAGE IN 

THE BUDDHIST COMMENTARIES ................................................................................... 215 

Introduction ................................................................................................................. 215 

Alternative Semantic Derivations ............................................................................... 228 

Buddhist Languages and Ideology .............................................................................. 251 

CONCLUSION ............................................................................................................. 293 

BIBLIOGRAPHY ......................................................................................................... 303 

 



 

 

x 

ABBREVIATIONS 

A Aṅguttara Nikāya 

AKB Abhidharmakośabhāśyam 

AKK Abhidharmakośa-kārikā 

Ap Apadāna 

Ap-a Visuddhajanavilāsinī (Apadāna-aṭṭhakathā) 

Apte Revised and enlarged edition of Prin. V.S. Apte's The practical 

Sanskrit-English dictionary 

As Atthasālinī 

BHSD Buddhist Hybrid Sanskrit Dictionary 

Bv Buddhavaṃsa 

Bv-a Madhuratthavilāsinī (Buddhavaṃsa-aṭṭhakathā) 

Ch. Chinese 

Cp-a Paramatthadīpanī (Cariyāpiṭaka-aṭṭhakathā) 

CPD Critical Pali Dictionary (Treckner et al. 1924) 

CST4 Chaṭṭha Saṅgāyana Tipiṭaka Version 4 

D Dīgha Nikāya 

Dhp Dhammapada 

Dhp-a Dhammapada-aṭṭhakathā 

Dhs Dhammasaṅgāṇi 

Dīp Dīpavaṃsa 

DP A Dictionary of Pāli (Cone 2001–) 

EWA Etymologisches Wörterbuch des Altindoarischen (Mayrhofer 

1992-2001) 

G. Gandhari 

It Itivuttaka 

It-a Paramatthadīpanī (Itivuttaka-aṭṭhakathā) 

Ja The Jātaka together with its commentary 

Jat-a Jātaka-aṭṭhakathā, commentary only 

Khp Khuddakapāṭha 

Kv Kathāvatthu 

M Majjhimanikāya 

MBh Mahābhāṣya 

Mhv Mahāvaṃsa 

MIA Middle Indo-aryan 

Mp Manorathapūraṇī (Aṅguttaranikāya-aṭṭhakathā) 

MW Monier-Williams Sanskrit-English Dictionary 1899 

Nett Nettipakaraṇa 

Ngh Nighaṇṭu 

Nidd Niddesa 



 xi 

Nidd-a Saddhammapajjotikā (Niddesa-aṭṭhakathā) 

Nk Nirukta 

OIA Old Indo-aryans 

P. Pali 

Pā. Pāṇini’s Aṣṭādhyāyī 

Paṭis-a Saddhammappakāsinī (Paṭisambhidāmagga-aṭṭhakathā) 

Peṭ Peṭakopadesa 

Pj I Paramatthajotikā I (Khuddakapāṭha-aṭṭhakathā) 

Pj II Paramatthajotikā II (Suttanipāta-aṭṭhakathā) 

Pkt. Prakrit 

Ps Papañcasūdanī (Majjimanikāya-aṭṭhakathā) 

PTS Pali Text Society 

PTSD Pali-Text Society's Pali-English Dictionary (ed. Rhys Davids & 

Stede 1921–25) 

Pv-a Paramatthadīpanī (Petavatthu-aṭṭhakathā) 

Skt. Sanskrit 

SN Saṃyutta Nikāya 

Sn Suttanipāta 

SN-a Saṃyuttanikāya-aṭṭhakathā 

Sp Samantapāsādikā 

Spk Sāratthappakāsinī (Saṃyuttanikāya-aṭṭhakathā) 

Sv Sumaṅgalavilāsinī (Dīghanikāya-aṭṭhakathā) 

T. Taishō shinshū daizōkyō (Takakusu & Watanabe 1924) 

Th, Thī Theragāthā and Therīgāthā 

Th-a Paramatthadīpanī (Theragāthā-aṭṭhakathā) 

Thī-a Paramatthadīpanī (Therīgāthā-aṭṭhakathā) 

Ud Udāna 

Ud-a Udāna-aṭṭhakathā  

Vibh Vibhāṅga  

Vibh-a Sammohavinodanī (Vibhāṅga-aṭṭhakathā) 

Vin Vinayapiṭaka (ed. Oldenberg 1982–1997) 

Vism Visuddhimagga (ed. Davids, Caroline A. F. Rhys 1975) 

Vv-a Paramatthadīpanī (Vimānavatthu-aṭṭhakathā) 

 

For most of the Pali works, I follow the convention of abbreviations in Margaret Cone 

(2001)’s A Dictionary of Pāli. The words of foreign languages in the parentheses are Pali if not 

otherwise noted.



 

 

1 

INTRODUCTION 

Any scholar of Buddhist studies knows the vastness of the corpus of Buddhist canons, not to 

mention the commentaries. In the Vinaya Piṭaka, the corpus of the Buddha’s words 

(buddhavacana) is compared to a great ocean: 

Just as, bhikkhus, the great ocean has one flavour which is the salty flavour. In the same 

way, indeed, bhikkhus, this teaching (dhamma) and monastic law (vinaya) has one 

flavour which is the flavour of liberation. (Vin II 239)1 

Amazed by the great ocean of buddhavacana without limit, it is a great challenge for readers of 

the buddhavacana to find direction without getting lost in the endless ocean. As the quote points 

out, readers can avoid being overwhelmed by the voluminous canonical texts by understanding 

their internal consistency. In the same way as the ocean is of one single salty flavour, 

buddhavacana should reveal the single flavour of liberation. 

The commentarial projects were contemporaneous with South Asian Buddhist 

scholasticism and gave rise to much of the scholastic tradition. To maintain the tension between 

the limitless texts and their consistency, the commentators contributed to the consolidation of the 

ideological framework and enriched the Buddhist literary tradition with their adaptability to 

various intellectual and social contexts. By directing our attention to Buddhist commentaries, I 

would like to touch on the inner tension of Buddhist scholasticism that attempted both to reach a 

vast audience with its truth claim and methodological diversity and to preserve its own 

distinguishing features in a multireligious world. Limiting the temporal span of my research to 

 
1 “seyyathāpi bhikkhave mahāsamuddo ekaraso loṇaraso, evam eva kho bhikkhave ayaṃ 

dhammavinayo ekaraso vimuttiraso.” 



 2 

the period in and before the first millennium CE, I will show how the commentarial tradition 

both constitutes the early Buddhist scholastic practice and reflects Buddhist participation and 

adaptation in this multireligious world. As important witnesses to the transformation of 

Buddhism from a wandering community of renunciants to premodern corporates with sedentary 

scholars and magnificent architecture, commentaries record how the religious community in 

these changing milieux passed down the heritage from their predecessors and adapted to the new 

environments. In maintaining the consistency of Buddhist doctrines, the commentators 

functioned as tailors stitching Buddhist literature together, served as judges to decide the 

authenticity of the Buddha’s words, and laid exclusive claim to right spiritual path to 

enlightenment. Living in a competitive religious market and whimsical political climate, they 

had to devise attractive ways of teaching, increase their prestige, and struggle for patronage to 

ensure the Dispensation of Dhamma will last a long time. 

This study questions the validity and limitation of applying the concept “commentary” to 

various post-canonical Buddhist texts and broadens the understanding of what commentary is 

and how it functions. I will show the Buddhist creativity in stylistic innovations and hermeneutic 

reflection. With each chapter dedicated to one archetype of so called “commentary,” the 

dissertation aims to redirect our attention from the aṭṭhakathās on the Sutta Piṭaka to the 

Niddesa, the Dhammapada-aṭṭhakathā and Jātaka-aṭṭhakathā, and the Gandhari commentaries 

in relation to the Peṭakopadesa and Nettipakaraṇa. Although these commentaries appeared to be 

deviants from the stereotypical understanding of commentary, they demonstrate that textual 

interpretation can take various forms in accordance with their historical contexts, theoretical 

achievements, and practical functions. The oral-aural culture, the adoption of writing, the 

incorporation of stories, and the pursuit for doctrinal consistency drove the Buddhist 



 3 

commentators to reconfigure the interpretative styles and methods while maintaining a 

consistency with the canonical texts. 

Although the study draws on primary sources from the Pali textual tradition, it also 

attempts to excavate the often-neglected connections between the early Buddhist commentaries 

and the vibrant intellectual context of early first-millennium South Asia. As I will show in 

Chapter 4, Vedic semantic derivation (nirukta) and reflections on language from Vyākaraṇa and 

Mīmāṃsaka traditions have parallels in the Buddhist tradition in general and influenced the 

formation of the Pali language ideology. Buddhists in South Asia, like Buddhists elsewhere, are 

deeply involved in the intellectual interactions, even though some of the involvements are not 

readily discernible or are deflected in a unique Buddhist way. By transcending the disciplinary 

boundary of Buddhist studies, I hope to shed light on the intellectual exchanges among multiple 

traditions and explore how Buddhist commentaries deflected the non-Buddhist knowledge while 

remodelling this knowledge with their own distinctive concerns. 

The Idea of Commentary 

Scholars in Buddhist studies, especially we who study the earliest literary productions of the 

tradition, owe much to the abundant commentarial materials available to us. The Mainstream 

Buddhist schools, most prominently Theravāda and Sarvāstivāda, preserved much of the 

teachings in their own Tripiṭakas (or P. tipiṭaka, Three Baskets) which are close in structure and 

content.2 Structurally, the Tripiṭakas include the three collections of texts classified as Discourses 

 
2 I use the term “Mainstream Buddhist schools” to refer to Buddhists who were affiliated with the 

Buddhist schools, eighteen in total in traditional lists, although in fact there were more. These 

schools were often identified based on their different Vinaya transmissions and shared the 

spiritual goal of arahantship, which contrasts with the Mahāyāna, which often lived in the same 

monasteries with other Mainstream schools but aimed for becoming a buddha. 



 4 

(P. sutta), Disciplines (P. vinaya), and Higher Teachings (P. abhidhamma).3 Although the fifth-

century Buddhist commentator, Buddhaghosa, claimed that his commentaries were based on the 

old commentaries from the first Joint Recitation (P. saṅgīti) and later brought to Lanka by 

Mahinda,4 the old commentaries leave only some traces in the Pali aṭṭhakathās and exist no more 

in their original form. Some early commentaries composed before the close of the Pali canon 

may have been absorbed into the canon, such as the Niddesa which comments on part of the 

Sutta-nipāta and the Suttavibhaṅga section in the Vinaya Piṭaka as a commentary on the 

Pātimokkha. 

Even though the Buddhist schools on the South Asian continent composed a large 

number of texts around the beginning of the Common Era, our knowledge of the commentaries 

on the Tripiṭaka of this period is quite limited. The Gandhari manuscripts discovered in modern 

day Pakistan and Afghanistan give us access to the rare commentaries on canonical verses and 

sutta on the mainland. Richard Salomon reports commentaries on the Sūtra of Chanting Together 

(Saṅgīti Sūtra) and on canonical verses selected from the Kṣudraka Āgama collection in the 

British Library collection. Many canonical texts underwent the process of translation, which 

makes it difficult to distinguish the translator’s notes and their original Indic commentaries. The 

 
3 “Higher Teachings” is just an expedient translation of the polysemic term abhidhamma. The 

term is explained as “higher dhamma” in As 2,14. The Theravāda and the mainland tradition 

have different interpretations. “Though the interpretation of ‘abhi’ as ‘highest’ or ‘further’ is 

preferred by later Pali commentaries, the interpretation as ‘with regard to the teaching’ comes to 

represent the northern Indian Buddhist understanding of the term” (Willemen, Dessein, and Cox 

1997, 14). 
4 “atthappakāsanatthaṃ aṭṭhakathā ādito vasisatehi / pañcahi yā saṅgitā anusaṅgitā ca pacchā pi / 

sīhaḷadīpaṃ pana ābhatā 'tha vasinā Mahā-Mahindena / ṭhapitā Sīhaḷabhāsaṃ manoramaṃ 

bhāsaṃ” ([T]he commentary, I say, upon this Scripture was at the first Council rehearsed by five 

hundred holy elders, and in later times rehearsed again and yet again. And it was carried by the 

saintly Mahendra to the island of Ceylon, and for the sake of the dwellers in that isle translated 

into the Siṃhalese language) (Ps I 1). Translation from Vijasinha and Childers (1871, 290). 



 5 

translation process, accompanied by the oral exegesis, often conflates the root texts and the 

commentaries from the very beginning of Chinese Buddhism. Therefore, I will limit my 

discussion of the early commentaries to the texts in Indic languages.5 

Before delving into the specific cases of Buddhist commentaries, I would like to reflect 

on my theoretical approach and openly discuss my presuppositions. In this dissertation, I 

envision a study of Buddhist commentary that will “modulate” the category “commentary” as 

“characteristically employed by the historian of religions” and “redescribe” the category in “a 

new” and “unexpected light” (Smith 1982, 36). As a historian of religion, I do not regard the 

Buddhist truth or ideology as sui generis, primordial, and complete in the canonical period or 

even the Buddha’s lifetime and degenerated or devolved in a later age by the split, 

(mis)understanding, and adaptation of the commentators’ “degenerative character of ingenuity” 

(Smith 1982, 42). In contrast, Smith encourages students of religion to redirect our attention to 

the interpretive community’s agency within their self-imposed limits: 

It would mean that historians of religion must redirect their attention from their present 

equally romantic fixation on multivalent and condensed phenomena, such as symbols, 

which have more often served as eloquent testimony to the exegetical ingenuity of the 

researcher than of the community that has bound itself to them, and should rather become 

concerned with prosaic, expository discourse. It implies that historians of religion will not 

lose their freedom to study the all but limitless horizon of human religious experience and 

expression and objects of religious concern, but that they will take as a prime interpretive 

 
5 Stefano Zacchetti argued that Chen Hui (陳慧) and Kang Senghui (康僧會) edited the Da 

anban shouyi jing (大安般守意經) in the Taishō canon on the basis of an earlier Anban shouyi 

jing (安般守意經) by An Shigao and added glosses from their exegetical sources (Zacchetti 

2008, 460). 

 



 6 

and comparative task the understanding of the surrender of that freedom by the 

communities they study and the rediscovery of that freedom through the community’s 

exercise of ingenuity within their self-imposed limits. (Smith 1982, 44) 

Taking Smith’s advice, I redirect my attention from what certain concepts (such as suffering, 

nidāna, nirutti) mean in the Buddhist canon, if there were a consensus about their meaning 

anyway, to the history of the human ingenuity in explaining them, “represented by the rule-

governed exegetical enterprise of applying the canon to every dimension of human life” (Smith 

1982, 43). The multifaceted rules governing the Buddhist commentators include the canonical 

literature that not only provides the initiative for them to continue their theoretical creativity but 

also poses a challenge when silence and conflicts force them to “speak for the Buddha.” By 

focusing on the phenomena of theoretical continuity and creativity, we can treat Buddhist 

commentaries as the stage where interesting developments occur in the history of ideas and 

communities. It is in this tension between transmission and innovation that I find it beneficial to 

study commentary as a scholastic practice, to which I will come back later. 

Besides the canon that precedes the commentaries, the multifaceted “self-imposed limits” 

also include the socio-historical contexts of the ingenious commentators who exerted their own 

agency in the distinctive quality of their commentaries. Even though the commentaries all share 

the purpose of explaining the root text or the canon, they vary in the content on which they 

choose to comment and the methods they use. Scholars’ limited exposure to several texts from a 

specific tradition sometimes blinds them to the fact that different commentaries must adjust their 

expository styles and methods in accordance with the needs of different audiences, the material 

culture of their time, the social tension surrounding them, and so on. These variables in 

commentarial studies, especially the extra-content features of the commentaries, including the 



 7 

oral or written technology, the language used for composition, and the incorporation of 

narratives, all point to a complicated history of commentaries. The variation in styles and 

methods presented in the current study will “redescribe” or redefine commentary and broaden 

the limited understanding of commentary. 

The current research includes several case studies of various Buddhist commentaries and 

hermeneutical works. Despite the fact that they are arranged in a diachronic order, I do not intend 

to make a historical argument of the development of commentaries throughout the temporal span 

from the early oral period to the later days of the narrative commentaries. It would be fallacious 

to regard my research as advocating for a unidirectional, chronological evolution of 

commentaries. Except the Gandhari commentary, the Pali commentaries, and some hermeneutic 

manuals that I focus on for the current research are difficult to date. I will examine the details of 

their dating in their respective chapters. But the difficulty in dating does not hinder us from 

situating certain texts in their respective historical contexts. The internal evidence from Nidd 

shows that it was composed in an oral-aural culture or was at least a remnant of that culture. Its 

stylistic distinctiveness serves as perfect evidence and explanation for its incorporation into the 

Pali canon and its methodological particularity. It is by no coincidence that Nidd’s stylistic 

features never appear in the Gandhari commentaries, which date to the first century CE and 

without doubt belong to the earliest products of South Asian writing culture. The flourishing of 

Pali literature and the contemporaneous suppression of Mahāvihāra also imply a historical 

context that is conducive to the formation of a language ideology. Even if the narrative 

commentaries in Dhp-a and Jat-a and the hermeneutic handbooks cannot be definitively dated to 

a specific period, their unique commentarial practice definitely can refresh the understanding of 

the commentarial method and function and illuminate future studies on other commentaries. 



 8 

The reason I adopt the etic term “commentary” to denote the various texts that serve the 

purpose of explaining the canonical root texts is that they share a common purpose of attempting 

to help their readers understand the root texts better, even though their focuses and their methods 

can be different. However, the use of a single term for multiple indigenous categories risks 

levelling differences among stylistically diverse texts and obscuring the historical dynamics of 

the Buddhist commentarial tradition. Among the texts that are often classified as 

“commentaries,” the Niddesa represents an early phase of oral commentary that glosses the terms 

in the canon by providing synonyms and listing the categorical concepts. The Suttavibhaṅga in 

the Pali Vinaya that explains the Pātimokkha rules is close to the aṭṭhakathās in supplying both 

the original context (vatthu) of the rule and glosses of words (padabhājanīya). Although the 

Suttavibhaṅga resembles Buddhaghosa’s exposition of the canon, the commentators around the 

fifth century CE adopted the term aṭṭhakathā as a generic name for commentary. No emic term 

for “commentary” appears in the Gandhari manuscripts. 

The term aṭṭhakathā became the most used term for the commentaries on the canon in 

Buddhaghosa’s time, even though it seemed to refer to a general category of texts studying the 

canon or explaining its meaning. In the Pali canon, the term appears only in the title of Book 4 of 

the Dhammasaṅgāṇi, and the entire section of Book 4 could be a later addendum.6 Maria Heim 

summarises the general definition of aṭṭhakathā in Buddhaghosa’s commentary as the “doctrine 

of the ‘Teachers’ (ācariyavāda), and as something paired with or contrasted to the canonical text 

(pāḷi)” (2018, 65). As the major parts of the Pali canon (except the Khuddaka Nikāya) had been 

 
6 “[I]n an earlier passage [in the Atthasālinī] it was termed the fourth Vibhatti [(Part)] comprising 

the atthuddhāro” (Rhys Davids 1900, 360). The canonicity of this chapter is further complicated 

by the fact that it is reputed to be a work by Sariputta. Von Hinüber expresses the same doubt 

(2000, §135). 



 9 

closed before the fifth century CE, the commentators who were affiliated with Mahāvihāra were 

devoted to composing commentaries in Pali for the canonical texts and finally completed the 

project by ensuring that all of the canon had an aṭṭhakathā commentary. According to von 

Hinüber (2000, §307), except the undated Buddhavaṃsa-aṭṭhakathā and the very late Apadāna-

aṭṭhakathā, most aṭṭhakathās were composed during the time from the fourth to the late ninth 

century CE. If we leave out the miscellaneous collection of the Khuddaka Nikāya, all of the 

commentaries were composed in the first three centuries. In this way, the term aṭṭhakathā 

became exclusively reserved for the first generation of commentaries composed for the Pali 

canon. 

The internal division of another term, vaṇṇanā, which is often translated as 

“commentary,” reveals a self-reflective elaboration of two major commentarial methods. George 

Bond in his study of Pali commentaries distinguishes “word commentary” (padavaṇṇanā) from 

“meaning commentary” (aṭṭhavaṇṇā) in some of Buddhaghosa’s commentaries. According to 

Bond, the former glosses the words according to their conventional meaning, whereas the latter 

highlights the special meaning of the words in a Buddhological context (i.e., how they should be 

linked to Buddhist ideology and cultivation) (Bond 1982, 147–48). This distinction between 

word commentary and meaning commentary can apply retrospectively to the various 

commentaries I have discussed here. The Niddesa and “word commentary” focus on glossing 

individual words and are less about the extra-textual elements (origin of the text, intention 

behind the literal meaning, etc.) The Suttavibhaṅga and Buddhaghosa’s aṭṭakathās function more 

like a comprehensive explanation of the literal meaning, the contextual knowledge, and the 

doctrinal framework. For the current research, I will tentatively use “commentary” as a 

convenient term to include all these various styles of exegetical texts because they all share the 



 10 

function of explaining the root texts. 

In Buddhist studies, there is a tradition of prioritising the study of canonical texts over the 

study of commentaries and anthologies in vernaculars. The commentaries in Pali have also long 

been used as references to study the canonical literature because they are the earliest explanatory 

attempt to solve the obscurities in the canonical texts. However, compared with the immense size 

of the studies on the canonical texts, the studies that take commentaries as their object of 

research do not receive the same amount of attention. The imbalance in the attention paid to the 

commentaries is partly a result of scholars’ enthusiasm for reconstructing early Buddhism on the 

basis of the earliest texts; it is also because the early Buddhist scholars perceived the history of 

Buddhism as “a long process of decay” (Hallisey 1995, 38). Hallisey points out that the 

instrumentalist use of the commentaries and other Buddhist anthologies has blinded previous 

scholars to the value of these texts themselves: 

In this view, commentaries and translations were not the record of the growing 

understanding of a text, of the accumulation of evolving interpretation over the centuries; 

instead, they were signposts for those in the present to recover accurately the meaning 

that had already been promulgated in the past. They were instrumentally valuable, but 

were without interest in their own right. (Hallisey 1995, 43) 

In past decades, scholars gradually directed their interests to Buddhist commentaries. All 

scholarship on Buddhist commentaries is probably indebted to George D. Bond’s pioneering 

work, The Word of The Buddha: The Tipiṭaka and its Interpretation in Theravada Buddhism 

(1982). Bond’s work is the first systematic study of commentary in the anglophone world. He 

sorts out the interpretative methods in the Nettipakaraṇa and gives an insightful observation of 

Buddhaghosa’s aṭṭhakathā commentaries. Building on Bond’s achievement, I speculate on the 



 11 

gradual transition from denotative definition to analytical definition within the shift from oral to 

written commentarial culture. Moreover, I focus on a few Modes of Conveying (hāra) and 

examine their uniform roles in interpreting of the Buddha’s words. When the Modes of 

Conveying and Guidelines (naya) are combined as structural analysis of the root texts and 

adaptive teachings for various types of audience, they function as hermeneutic strategies to 

control the interpretations while being attentive to the individual predispositions. 

Richard F. Nance’s book, Speaking for Buddhas: Scriptural Commentary in Indian 

Buddhism (2015), focuses on the ideal mode of teaching and the epistemological possibility of 

successful explications in the Mahāyāna sūtras and late Mahāyāna treatises. His ground-breaking 

work explores the intersection between hermeneutics and Buddhist epistemology. I hope the 

current study, which focuses on the specific commentarial methods found in the early Pali and 

Gandhari texts, will provide a valuable foundation for understanding the Mahāyāna 

commentarial works and hermeneutics in the Indo-Tibetan tradition. 

Maria Heim’s recent work, Voice of the Buddha: Buddhaghosa on the Immeasurable 

Words (2018), clarifies, in brief and in detail, several important concepts in the Pali 

commentarial tradition, such as conventional teachings and teachings in the furthest sense, 

definitive and interpretable statements, and so on. Her case studies of Buddhaghosa’s 

commentaries on the Sutta, Abhidhamma, and Vinaya show the commentator’s dexterity in 

revealing the Buddha’s omniscience and his pedagogical enactment for specific occasions. 

Maintaining a historical awareness of the cultural contexts of the commentarial practices, such as 

the transition from oral to written culture, the arising of the Abhidharma/Abhidhamma, and the 

parallel science of language in the Brahmanical tradition, I strive to go beyond the Pali 

commentaries and situate the Buddhist commentaries in a broad South Asian intellectual world. 



 12 

In this broad context, the interplay of material culture and scholastic debates profoundly shaped 

the historical trajectory of the development of Buddhist commentary and enabled Buddhist 

commentators to maintain doctrinal consistency while being creative in both their methodology 

and philosophy. 

Attributing more than instrumental value to commentaries, I attempt in my current study 

to treat the various forms of commentaries as valuable textual products in themselves that 

emerged from a specific context in the history of religion. By studying the commentaries for 

their own sake, I will try to clarify the idea of “Buddhist commentary.” Even though it is difficult 

to determine definitively the chronological relationships among multiple commentaries, the 

variety of the commentarial styles reveals commentators’ creative visions of the commentaries in 

different contexts and the diversity of hermeneutic investigation. Never content with serving as 

obedient assistants or timid paraphrasers of the canonical texts, the commentators found 

commentaries to be a perfect way to integrate complementary materials, defend their 

authoritative positions, and put forward innovative ideas that were previously unseen in the 

canon. 

In my research, I also attempt to show that the dynamics and diversity in the 

commentarial styles are balanced by an internal force that seeks to control the interpretation of 

the root text in several ways. Commentaries routinised the audience’s engagement with the 

Buddha’s words, controlled the audience’s method of interpretation, and shaped the Buddhists’ 

perceptions of themselves and the world in which they lived. 

Immeasurability of the Buddha’s Words 

Commentaries presuppose the existence of a corpus of more authoritative texts that need to be 

explained and studied. The commentators, who live in a different times and places, must speak to 



 13 

their audiences in a legible and intelligible way. Maria Heim shows how Buddhaghosa unfolded 

the Buddha’s words (buddhavacana) in his commentaries on the Brahmajāla Sutta “as revealing 

the extent of [the Buddha’s] knowledge of particular beings” (2018, 49). The immeasurability of 

the Buddha’s words requires the commentators to attend to the original settings in which the 

Buddha preached. Heim convincingly shows how the Theravāda commentaries emphasise the 

occasion (samaya) or context (nidāna) of the Buddha’s instructions and often supply them in the 

commentaries if they are missing in the canon. In this way, the audience of the canonical text is 

better aware of the specificity of the Buddha’s words if they read the root text through the lens of 

the commentaries. 

The Buddhist commentators’ important task of unfolding the immeasurability of the 

Buddha’s words represents the role of commentary in the flourishing of Buddhist scholasticism. 

The commentarial practice and its product, the commentaries, are part of the Buddhist 

scholasticism on the basis of Jóse Ignacio Cabezón’s modification of Masson-Oursel’s definition 

of “scholasticism” (Cabezón 1994, 15). Buddhist scholasticism is an intellectual practice that 

bases its authority on the words of a sacred text, which includes the Buddha’s words and the 

other authoritative texts of the Buddhist tradition. Assuming that the authoritative texts are 

understandable, coherent, and construable, commentators aim at preserving and clarifying the 

religious truth, defending and distinguishing the tradition from others. The Buddhist 

commentators’ literary activity, as I will show, is centred on Cabezón’s list of four factors that 

“drive scholastic traditions toward rationalism and systemacity” (1994, 20). 

In light of Cabezón’s summary of the common features of scholasticism, we can relate 

the exposition of immeasurability to two characteristics of scholasticism—“proliferativity” and 

“completeness and compactness.” On the basis of the “scholastic tenet” that the “tradition is 



 14 

complete,” which means the canonical text is all-encompassing, “proliferativity” means 

“scholastics … opt for broader (even if inconsistent) canons and for minute and detailed forms of 

analysis that leaves no question unanswered, no philosophical avenue unexplored” (Cabezón 

1998, 5). In the same way, Buddhist commentators find infinity in the words of the Buddha, 

which are constantly subject to new interpretations. “[F]or Buddhaghosa, any particular 

expression in the Buddhist canon cannot be exhaustively described, and the interpretive task is to 

show the ‘immeasurability’ of the Buddha’s words and to shape the ideal reader’s response to it” 

(Heim 2018, 3). 

As questions emerge in a new context, commentators resort to the canonical texts to find 

their answers and can always be inspired by the Buddha’s words. Therefore, the productive 

source for raising new questions and answering them is “practice,” about which Heim continues, 

Second and relatedly, Buddhaghosa’s interpretive assumptions often treat the Buddha’s 

teachings not as declarative or discursive utterances so much as practice [(Heim’s 

emphasis)]. The Buddha’s words were uttered not so much to give final propositional 

accounts of the way things are, but to initiate a series of practices that are themselves the 

very work of insight and understanding that enact the tradition’s therapeutic and 

soteriological aims. (Heim 2018, 3) 

“Practice,” according to Heim, includes religious cultivation such as meditation or ritual 

performance, the real-world event where the interpretation actualises the referential function of 

language. But in scholasticism practice covers a wide range of activities that attempt to redress 

the Buddhist ideological framework through which a person interacts with the world. Therefore, 

when the Buddhist commentators enumerate all types of suffering, the definition of the term in 

the Niddesa subsumes all causes, symptoms of, and reactions to suffering in an encyclopaedic 



 15 

manner to provide the formulaic lists for teachers to make use of. In this way, the commentator is 

speaking for the Buddha—what the Buddha did not cover in his own teaching, the commentator 

completes for him. The commentator’s ability to speak testifies conversely to the completeness 

and compactness of the canonical corpus, the Buddha’s words. The completeness and 

compactness of the Buddha’s words require that “the tradition overlooks nothing and contains 

nothing that is unessential” (Cabezón 1998, 5), which the Pali commentators often humbly 

attribute to the Buddha’s omniscience, as Heim puts it, “[i]mplicit in Buddhaghosa’s theory of 

texts and evident in his actual practices is the idea that interpretive practice is itself an 

exploration of the case-by-case workings of the Buddha’s omniscience” (Heim 2018, 49).  

Regulating Interpretation 

To claim that the Buddhist commentaries reveal the immeasurability of the Buddha’s words does 

not contradict the Buddhist commentaries’ role in unifying the Buddhist ideology as opposite to 

non-Buddhist. Buddhism has diverse schools of thought and various practices. The border 

between Buddhism and non-Buddhism is vague, for we often encounter that situation when one 

self-identified Buddhist school accuses another self-identified Buddhist school as non-Buddhist. 

Furthermore, the schools or traditions that can be safely put under the name of Buddhism also 

differ in many ways in their ideas and practices, such as the Yogacāra and Madhyamaka 

Buddhism in late Indian Mahāyāna Buddhism, not to mention the Vajrayāna that embraced many 

erratic practices that substantially break monastic laws. 

We can safely conclude that, historically, Buddhism includes multiple subtraditions in 

terms of its philosophical abundance and diversity of practice. However, scholars are aware of 

the core teachings or doctrines that remain stable and are agreed on for a certain period or on a 

geographical scale. Certain teachings or doctrines, condensed from the canon, serve as criteria 



 16 

for drawing a clear line between Buddhist and non-Buddhist. For example, for the period when 

the canon was fixed, the four noble truths are well established in important texts like the 

Dhammacakkappavattana Sutta and are shared among various schools as a common language.7 

In this way, the consistency in the Buddhist doctrinal framework with a single flavour becomes 

the criterion for Buddhist community to determine whether a text is the Buddha’s words. 

Before I discuss the strategies used in the commentaries to regulate the interpretations of 

the canon in conformity with certain core teachings, I will first revisit the scholastic context in 

which the commentaries were composed. The emergence and development of the commentaries 

require a vibrant literary culture in which the canon or the root text is readily available as the 

precondition. It is in the aṭṭhakathā literature that we see the earliest reference to the organisation 

of the Buddhist canon, which is roughly the same as the Pali canon available to us today, 

including the five collections (nikāya) of the Sutta Piṭaka, Vinaya Piṭaka, and Abhidhamma 

Piṭaka.8 Although the collection of texts that are considered as the Buddha’s words 

(buddhavacana) grew throughout several centuries after the Buddha’s death, “[t]he canon was in 

all probability closed some time before the time of Buddhaghosa because stories and texts and 

verses from North India did not find their way into the canon” (Norman 1997a, 145–46).9 

Buddhaghosa’s commentaries in the fifth century CE, as Steven Collins surmises, is “the earliest 

date to which we can assign the Canon in the specific and final form” (Collins 2005, 76–77). 

The authority of the canonical root texts both provided the commentators with the 

 
7 André Bareau argues that the earliest texts, like Āriyapariyesana Sutta and the corresponding 

Chinese text, do not mention the four noble truths (de Jong 1993, 19). 
8 See As 18. The Therīgāthā is not included in the list, probably because it has not yet gained its 

authoritative status as a canonical text.  
9 K. R. Norman, on the basis of the internal evidence of the Pali tradition, surmised that the 

Vinaya Piṭaka dates between the second and third Joint Recitation (saṅgīti). The Abhidhamma 

Piṭaka including the Kathāvatthu was not closed until the third Joint Recitation. 



 17 

starting point of their thinking and pressed them to think intertextually. On the one hand, the 

literary proliferation around the start of the Common Era was good news for the Buddhist 

scholars because it might have preserved the teachings as well as making many texts available to 

the commentators, as we have seen in the Gandhari commentaries in the British Library 

collection. Various Buddhist schools each maintained textual transmissions and fixed their 

canons at certain periods, which contributed to the spreading of Buddhism to remote places. On 

the other hand, we can imagine that once texts were written down, collated into books, and 

carried around, the manuscripts often left the hands of their original composers and their 

authenticity needed to go through the test of their audience. As old languages became obscure 

and scribal strokes physically wore out, manuscripts and the teachings therein became 

increasingly difficult to understand. More importantly, the different literary styles and spiritual 

orientations in the newly composed texts, especially the Mahāyāna sūtras, led to the serious 

problem of textual authenticity and called for a new definition of the concept “buddhavacana” as 

early as the first century BCE.10 When the Buddhist canons, such as the Pali canon, were 

finalised, the gate to buddhavacana closed to the texts external to this tradition. The 

commentators needed not only to explicate the obscurities and remove the inconsistencies but 

also to establish a coherent teaching that unifies the discrete texts and serves as the criteria for 

deciding whether a text is the Buddha’s words. These needs compelled commentators to build an 

internal network of concepts scattered in various genres, including the highly contextualised 

suttas, the Vinaya texts, and the more abstract Abhidhamma texts. 

First, to construct a coherent framework of interpretation, the commentators defined 

 
10 The earliest evidence of Mahāyāna sūtra is a Prajñāpāramitā manuscript which Harry Falk and 

Seishi Karashima date to “the second half of the first century AD” (2012, 19). 



 18 

terms on the basis of the internal references within the canon, which may have started from an 

early period when the influence of oral culture was strong. Similar to the Vedic style of word 

commentary in the Brāhmaṇas and Upaniṣads and the Buddhist canonical texts, the early 

commentators provide strings of synonyms arranged according to phonetic rules or define the 

terms denotatively by giving examples. The latter lists examples of an abstract term to illustrate 

and define it. For example, the Niddesa uses the different names of diseases and so on to define 

the concept “suffering” (dukkha). However, as commentators had access to more texts and 

sought a more economical method of definition, they constructed internal references among 

canonical texts as part of their scholastic reflections, which the medieval Christians in the Latin 

world also did. 

Premedieval Christian thinkers had made some use of grammar and logic to interpret the 

Bible correctly. This practice was amplified in the Middle Ages. The expansion of 

scholarship demanded a more frequent recourse to grammar—in order to improve the 

quality of the liturgical idiom—and to logic, later to be called “dialectic”—in order to 

enhance precision in catechetics and theology. All began with juxtaposition of passages 

culled from ancient authors, both non-Christian and Christian. (Roy 1998, 23–24) 

The Christian case is unique because the Christian scholars aimed to apply the Graeco-

Roman heritage to the Bible and attune Christian theology to pre-Christian ideas. If the Buddhist 

commentaries available to us belong to the first five centuries of the Common Era, we must keep 

an eye on the contemporaneous intellectual achievements, especially those from outside 

Buddhism. During this period, the Abhidharma scholars gleaned more and more abstract 

concepts from the canonical text and built up a system of knowledge aimed at describing the 

phenomenal world and human psychology comprehensively and outlining the stages on the path 



 19 

to enlightenment. Although it is debatable to what degree the Abhidharma, especially Theravāda 

Abhidhamma, is an effort to make universal ontological claims, the Abhidharma/Abhidhamma 

approach seems to be an important intellectual revolution in and around Buddhism in this 

productive period of Indian history.11 Experts in Abhidharma dated the Abhidharma Piṭaka to “a 

period beginning after the first major schism and ending around the start of the common era” 

(Willemen, Dessein, and Cox 1997, 15). Bronkhorst asserts that Patañjali’s Mahābhāṣya, 

composed in the first century BCE, was influenced by the Abhidharma/Abhidhamma ways of 

thinking in a linear manner, which was not found in Pāṇini’s work (2016, 32–34). Although 

Bronkhorst’s claim that “there are two, and only two, independent traditions of rational debate 

and inquiry in the history of mankind” (1999, 11) (i.e., Greece and India) is dubious, he may be 

right in proposing that the vibrant scholastic culture in this peripheral region in north-western 

India around the beginning of the Common Era may be predicated on the vibrant cultural 

interaction within the region. Like the medieval Christian scholars, the Buddhists needed to 

maintain consistency in their textual traditions and defend against the outsiders’ questioning of 

the efficacy of their belief and the religious practice. In this sense, Bronkhorst hypothesises that 

“the Buddhists of North-West India adopted the method of rational debate and inquiry from the 

Greeks” (1999, 22–23). The commentaries written after the close of the Buddhist canon were 

products of this continuous multireligious or multi-traditional dialogue in which different 

scholastic participants drew from the knowledge available, challenged their rivals’ beliefs, and 

 
11 Maria Heim criticises “Ronkin’s ontological take of dhamma” in his book, Early Buddhist 

Metaphysics. She perceives Abhidhamma in the Pali canon as adopting “the method of 

‘comprehension by groups,’ that is assembling the modular data of experience in different sets 

and groupings to understand how phenomena are open to and shaped by other members in 

particular sets and how they are parts of systems of causes and conditions that operate in 

specifiable ways” (2018, 163). 



 20 

defended their own, like the traces of the Abhidharma mode of thinking in Patañjali’s 

commentary. Once a political ruler provides a religious group with a free and friendly 

environment, including financial support, political protection, and freedom from mundane 

labour, scholars in this group might produce more works in an openly discursive culture where 

other schools coexist.12 

The styles of commentaries vary from the Niddesa (Nidd) to the Gandhari commentaries 

and aṭṭhakathās because the Gandhari commentaries and aṭṭhakathās show the influence of the 

Abhidharma tradition. Not only is Nidd included in the Pali canon, but its style also presupposes 

mainly oral devices that facilitate its use in oral culture, focus on individual words and predate 

the analytical definition favoured in Buddhaghosa’s Vism and the northern Abhidharma 

tradition. The combination of the contextual and abstract teachings has its root in the canonical 

texts and was widely adopted in the commentaries of the scholastic tradition after an early phase 

of oral exegesis. Even if the commentators worked on the same root texts, the Gandhari 

commentaries and Pali aṭṭhakathās continued the phonetic rules and denotative definition and 

sought other strategies to illustrate concepts with the help of Buddhological frameworks. 

Although the post-Nidd commentaries preserved oral devices as convenient ways for 

glossing words, the later commentators integrated lexical exegesis into systematic knowledge 

composed of concepts arranged as numbered sets and lists. These numbered sets and lists that 

always appear in Abhidhamma/Abhidharma texts are called “teaching in the furthest sense” 

(paramatthadesanā) in the Pali aṭṭhakathās. Heim distinguishes the conventional teachings 

 
12 Jon Elster thinks consistency is key to rationality: “Consistency, in fact, is what rationality in 

the thin sense is all about: consistency within the belief system; consistency within the system of 

desires; and consistency between beliefs and desires on the one hand the action for which they 

are reasons on the other hand” (2016, 1). 



 21 

(sammutidesanā) from teachings in the furthest sense, which are usually Abhidhamma categories 

(2018, 85–86). According to Heim, “the notion of colloquial or transactional talk (vohāra) that 

predominates in the suttas suggests both the idiomatic speech of regular people and the language 

of trade, business, and legal transactions” (2018, 89). However, teachings in the furthest sense 

are associated with Abhidhamma knowledge, which “break[s] down composite entities into their 

parts as a ‘correct way of seeing’ (yathābhūtadassana), where such practice is embedded in a 

larger contemplative and analytical exercise of purifying view (diṭṭhivisuddhi)” (2018, 90). The 

practice of condensing the diverse root texts into the numbered sets and lists is common in the 

Gandhari commentaries in the British Library Kharoṣṭhī fragments. The two hermeneutic 

handbooks, Nettipakaraṇa and Peṭakopadesa, also make use of categorical terms, play with their 

related and contradictory concepts, analyse the structure of the canonical verses, and apply these 

concepts to various types of persons. 

The narrative commentaries in Jat-a and Dhp-a control the interpretation of the root texts 

by constructing the fictional world where a consistent moral law dominates and by extending the 

temporal framework from the time in the story to the time of the target audience. Besides 

building a coherent network of concepts within the canonical literature, the Pali commentators 

also exploited the concept of nidāna in the canon to construct a worldview that pervaded the 

Buddhist moral universe with their understanding of time. The nidāna section in the canon 

introduces the incident or cause of the Buddha’s teachings or the testimony of one of the great 

disciples (Ānanda in the case of sutta, Upāli in the case of vinaya) about the original teaching 

event. Buddhaghosa explores the literary function of the context or stories of origin and how it 

affects our reading of the root texts. Dhammapāla treats them as necessary background for 

understanding the authenticity of buddhavacana and the relationship between the Buddha and his 



 22 

remote audience. However, not confining their works to supply concise nidānas, commentators 

of Dhp and Jat innovatively expanded the nidānas and transformed the background introductions 

into the vivid narratives in the commentaries on verses. They incorporated the verses into the 

narratives karmically connected in a temporal sequence as the aphorisms that a character (usually 

the Buddha) proclaims. The Buddha, as the authoritative teacher, identifies the karmic 

connections between the characters in the story of the past and the story of the present. The 

structure of Dhp-a and Jat-a, like other multi-life accounts, Buddhavaṃsa and so on, subjects the 

characters in the narratives, some of which are probably of extra-Buddhist origins, to a moral 

universe based on Buddhist ethics. In this way, Buddhist commentators attempted to convey 

moral messages to their audiences on the basis of karmic law by presenting examples of ordinary 

people’s lives (with the good often achieving enlightenment) and constructing a fictional world 

where Buddhist commentators offer audacious if not always satisfactory answers to the problems 

humans encounter.13 The narrative complex connects the story of the past and the story of the 

present in the commentaries by karmic law to convey a specific moral message, but the moral 

message does not end in the time in the narrative. By recurrent references to the cyclical 

temporality in the narrative world, the commentators asserted the universal value of the karmic 

law to their own audience and merged the narrative world into the real world of their audience. 

In maintaining the structural consistency and the moral principles in these narratives, the 

commentators directed the interpretation of the canonical verses in a way both entertaining and 

didactic. 

 
13 Collins’s study of the Vessantararāja Jātaka reveals the “fear and pity” evoked in the audience 

and how that served its “transcendental goal” (1998, 554). 



 23 

The Concern about Language 

Similar to those in other scholastic traditions, commentators during this period maintained the 

consistency of the canon by showing “a concern with sacred language (scripture) and in exegesis 

and with language generally as medium of expression” (Cabezón 1998, 5). Although multiple 

Vinayas contain a famous story of the Buddha forbidding the two monks Yameḷu and Tekula 

from rendering the teachings into chandas, the northern schools and the Pali tradition clearly 

have different interpretations of the Vinaya account.14 Not only did the commentators show 

confidence “of the communicative ability of language” by applying the science of language to 

certain words (Cabezón 1998, 5), but the Pali commentators also developed a language ideology 

that elevated the Magadhan language to a status like that of Sanskrit. 

Just as Patañjali might have been influenced by the linear temporality in the Abhidharma, 

the Indian commentators widely glossed words by means of the semantic derivation (Skt. 

nirukta, P. nirutti) which the canon had adopted in their interpretation. Previous scholars, Ulrich 

Schneider (1954) and K. R. Norman (1991a), in their study of the Aggañña-sutta and Sābhiya-

sutta, have evidently shown that these canonical texts preserve a tradition of semantic derivation 

based on lost Middle Indo-aryan (MIA) forms. The MIA forms imply this tradition precedes the 

finalisation of the Pali canon. By comparing several specific cases of semantic derivation in 

Gandhari and Pali with their Brahmanical explanations in the Nirukta, I demonstrate that 

Buddhist semantic derivation consistently attunes the explanations of words to Buddhist 

ideology and philosophy. 

Even though Buddhists welcomed the semantic derivations as a form of knowledge 

 
14 Vin II 139. The classic study of the Vinaya story, see Brough (1980); for a recent revisiting of 

the question, see Levman (2009). 



 24 

transferable among religious traditions, the commentators limited the interpretation of these 

semantic derivations by adhering to those that best matched Buddhist ideology. The seemingly 

contradictory tendencies can be explained as a conscious scholastic choice in a multireligious 

context. On the one hand, the Buddhist scholars lived in a world where the Vedic tradition had 

relatively mature fields of knowledge that aimed at preserving the Vedic hymns and their ritual 

performance. They would not accept the ideas favourable to Brahmanical ideology to slip into 

their own textual tradition and cause potential interpretative difficulties. On the other hand, 

although they resisted the authority of the Vedas and their sacrificial rituals, the auxiliary 

sciences (vedāṅga), such as phonology, semantic derivation, and grammar, probably were still 

prestigious subjects that were effective in bolstering Brahmanical authority. The Buddhist 

scholars, some of whom were brahmin converts familiar with or well versed in these subjects, 

decided to adopt competitive tools that boasted a similar explanatory power. Therefore, instead 

of totally rejecting semantic derivation, they adapted semantic derivation and applied it to their 

own religious texts as a convenient way to serve their own religious ideology. 

At the same time, the capability of using the canonical language became an indispensable 

quality for people who had achieved the “four Discriminations” (paṭisambhidā) in the Pali 

textual tradition. The commentator of Vibh-a asserted that these four Discriminations are 

classifiable and “become manifest for five reasons” (Vibh-a 388)15: 

For which five reasons do [the four discriminations] become manifest? (1) Through 

attainment (adhigama), (2) through competency [in scriptures] (pariyatti), (3) through 

hearing (savana), (4) through being questioned, and (5) through previous work 

 
15 “imā pana catasso pi paṭisambhidā dvīsu ṭhānesu pabhedaṃ gacchanti, pañcahi kāraṇehi visadā 

hontīti veditabbāi.” 



 25 

(pubbayoga). (Vibh-a 388)16 

The four Discriminations (paṭisambhidā), which include the Discrimination of Language 

(nirutti-paṭisambhidā), the knowledge of language especially regarding the canonical language 

or Māgadha-bhāsā, constitute wisdom or understanding (paññā).17 It is unlikely that only one of 

the five reasons can lead a person to achieve the four Discriminations. More reasons make the 

person more likely to attain them. Besides the direct penetration—the attainment of arahantship 

(arahatta), Buddhists can achieve individual Discrimination by making efforts in learning the 

scriptures and engaging in the exposition of the root text. The commentator further explained, 

Here in (1) “attainment” is Arahantship (sainthood); for the Discriminations become 

manifest in one who has reached that. (2) “Competency [in scriptures]” is the Buddha’s 

word; for the Discriminations become manifest as one is learning that. (3) “Hearing” is 

the hearing the teaching (dhamma); for the Discriminations become manifest in one who 

listens to the teaching (dhamma) attentively. (4) “Being questioned” is the explanation of 

meaning (aṭṭhakathā); for the Discriminations become manifest in one who explains the 

meaning of the Pāḷi [(canon)] he has learnt. (5) “Previous work” is formerly practised 

meditation, having laid hold of the meditation subject in a past existence by the method 

of “carrying it forth and carrying it back”; for the Discriminations become manifest in 

one who was formerly a meditator. (Vibh-a 388–389)18 

 
16 “katamehi pañcehi kāraṇehi visadā hontīti? adhigamena, pariyattiyā, savaṇena, paripucchāya, 

pubbayogena.” 
17 For a brief summary of the Discriminations across the Buddhist traditions, see A. K. Warder’s 

introduction to Ñānamoḷi’s translation of the Paṭisambhidāmagga (Ñāṇamoli 1982, vii–xiii). 
18 “tattha adhigamo nāma arahattaṃ. taṃ hi pattassa paṭisambhidā visadā honti. pariyatti nāma 

 



 26 

When a practitioner achieved arahantship through practising morality and meditation, the person 

could attain the four Discriminations, or supposedly she may have practised in a previous life 

and the previous work comes to fruition in this life. Besides these two achievements that are hard 

to attain in this life or have been already determined according to the karma in the past life, the 

practitioner should be devoted to the study of the canonical texts so that the four Discriminations 

become manifest by the “competency [in scriptures],” or in Maria Heim’s translation, “the 

mastery of the Buddha’s words” (pariyatti) (Heim 2018, 50). As Heim shows, when 

Buddhaghosa uses “pariyatti,” it is directly connected to the knowledge of the Buddha’s words. 

Here the Buddha’s words are not a closed canon but the “enactment and further instantiation of 

the Buddha’s ken, suggesting a more ‘open’ and dynamic idea of canon than scripture as a 

container” (Heim 2018, 50). The person should actively hear the dhamma, which reveals the 

oral-aural transmission of the early textual tradition and situates the learning in a social context 

in which the hearer is expected to learn attentively under the guidance of a teacher. In the 

passage above, the commentator further glossed “being questioned” as “explanation of meaning” 

because the person who has Discrimination could explain the meaning of the texts as an 

instructor. Mastery of the Buddha’s words, learning them attentively, and revealing their 

meaning or intention by asking and answering questions are scholastic activities in which a 

practitioner can engage. Learning, questioning, and teaching constitute an organic whole for the 

commentator. By participating in these exercises, a practitioner attains the Discriminations, 

which constitute wisdom (paññā) in Vibh’s spiritual framework.  

The commentator discussed the scholastic activities after the explanation of the four 

 

buddhavacanaṃ. taṃ hi uggaṇhantassa paṭisambhidā visadā honti. savaṇaṃ nāma 

dhammasavaṇaṃ. sakkaccaṃ hi dhammaṃ suṇantassa paṭisambhidā visadā honti. paripucchā 

nāma atthakathā. uggahitapāḷiyā atthaṃ kathentassa hi paṭisambhidā visadā honti.” 



 27 

Discriminations, among which the Discrimination of Language (nirutti-paṭisambhidā) is 

explicitly said to be the knowledge of the Magadhan language. Here we can infer from the 

emphasis on the study of the canonical texts and the ensuing scholastic activities that the learning 

activities contribute to the mastery of the Magadhan language and may be carried out in this 

“perfect [language] usage” (P. brahmavoharo). In the Visuddhimagga, in addition to the 

discussion of the four Discriminations, Buddhaghosa provided the “others’” (apare) opinions: 

Others have said:  

A prior effort, and great knowledge,  

[Knowledge of] dialects, of scriptures, 

And questioning, and then achievement,  

and likewise waiting on a teacher, 

Success in friends—these are conditions. 

Productive of Discriminations.  

(Vism 442)19 

Here the “others” highlight the dialects or the knowledge of them as a contributor to the 

manifestation of Discriminations. The “others” may be some people who did not emphasise the 

importance of the Magadhan language but rather of “dialects” or “regional languages” 

(desabhāsā). Even though the “dialects” in this verse is in the plural form here, Buddhaghosa 

glossed it as “skill in the hundred-and-one tongues” (desabhāsā nāma ekasatavohārakusalatā) 

but added “particularly in that of the Magadhan language” (visesena pana Māgadhike kosallaṃ) 

 
19 “apare āhu: pubbayogo bāhussaccaṃ desabhāsā ca āgamo / paripucchā adhigamo 

garusannissayo tathā / mittasampatti cevā ti paṭisambhidapaccayā ti.” Ñāṇamoli’s translation 

from Buddhaghosa (2010, 438). 



 28 

(Vism 442). Buddhaghosa’s special concentration on the Magadhan language was a preference 

for language that had never occurred before in the history of Buddhism and had also been 

unknown elsewhere in the Buddhist world. 

The Pali commentators initiated a ground-breaking change in the theory of language. 

Buddhaghosa, patronised by the Mahāvihāra monastery in fifth-century Lanka, advocated using 

the language of the Pali canon and demonstrated its value in Buddhist intellectual practice. 

Historically speaking, the use of Pali language and the value judgment of it has neither an 

evident precursor in the canonical texts nor a known counterpart in other Buddhist textual 

traditions. Pali, or what Buddhists called the Magadhan language (P. māgadhabhāsā) is 

celebrated as the root language (mūlabhāsā) and the natural language (sabhāva-nirutti), which is 

“refined by perfect usage, the noble usage of the reality” (yathābhucca-brahmavohāra-

ariyavohārasaṅkhātā) and “does not change” (na parivattati) (Vibh-a 388). An ideology of the 

sacred language, which I would call “Pali language ideology,” emerged in the commentaries 

traditionally ascribed to Buddhaghosa when a whole new generation of Pali literature, including 

the early chronicles of Lanka and the aṭṭhakathās, were composed in this language and about this 

language. This ideology claims that the language is best suitable for conveying universal truth 

that originates spontaneously or persists continuously, in a way quite similar to Patañjali’s and 

the early Mīmāṃsaka theory of Sanskrit. Consequently, Pali in both theory and practice 

homogenised the literature of the Mahāvihāra tradition and became the standard language that 

anyone who wanted to participate in the Mahāvihārin scholastic tradition had to learn. 

The new Buddhist theory of language, largely elaborated and implemented in Pali 

commentaries, also played an important role in demarcating interreligious and intrareligious 

boundaries. Buddhist semantic derivation drew a line between Buddhist and Brahmanical 



 29 

semantics by maintaining a Buddhist perspective of semantic derivation and rejecting the 

Brahmanical authority. Pali language ideology is likely a Mahāvihārin attempt to assert its 

exclusive right over the canon and its interpretation. By claiming Pali is the language the Buddha 

spoke, the commentators sanctified their texts as the authoritative teachings of the Buddha, 

linguistically superior to the Buddhist texts in other languages. In this way, the discourse of Pali 

as the spontaneously arisen and perfect language set an explicit scholastic boundary between the 

Mahāvihārin lineage as an honest transmitter of the Buddha’s words and non-Mahāvihārin 

tradition as not so trustworthy.  

Overview of the Dissertation 

In this dissertation, I explore several subgenres of Buddhist commentaries. Because the dates of 

the most texts I study are uncertain, I do not want to make a historical claim on the level of the 

whole dissertation, even though I have made some historical arguments in limited scope. 

Scholars of Buddhist studies put the date of finalisation and composition of most of these texts 

from the first century BCE to the fifth century CE. Because the earliest Buddhist commentaries 

and scholastic works were composed in this period, I can safely claim my current research 

reveals the features of the “early” commentaries in Buddhism. 

The individual chapters and the various commentaries are connected by the root verses 

that find their parallels in canonical sutta collections, such as the Pali Khuddaka Nikāya, the 

Gandhari manuscripts of the British Library collection, and the corresponding translations in the 

Chinese canon. Because of their succinctness and lack of contextual background, these root 

verses serve as a central connecting point for various commentaries in different languages and 

styles. Therefore, I compare how these commentaries from various times and locations interpret 

the same root verses with various techniques and unique focuses. I also attempt to explain why 



 30 

certain strategies are used in a specific context and what functions they serve in their respective 

scholastic and textual communities. The differences between styles and focuses cannot be easily 

mapped onto school affiliation or language distinction. As I will show, even in the Pali tradition, 

the commentators have a wide range of approaches to expounding on the same root verses. 

In the first chapter, I will focus on the stylistic features of the Niddesa as the product of 

an oral culture, which refreshes our limited imagination of what a commentary is and how it is 

used in a writing culture. Taking the definition of suffering (P. dukkha, or Skt. duḥkha) as an 

example, I trace the various methods of definition in the commentaries and show how the 

discrete explanations in the canon cause interpretive difficulty for the Buddhists in the late 

canonical period. By the time of Buddhaghosa and other Abhidhammic/Abhidharmic tradition, 

commentators from both the northern and southern corners of South Asia strove for a new 

method of definition and reached a consensus on the meaning of the old terminology by 

integrating it into the doctrinal framework of Abhidhamma/Abhidharma. 

In the second chapter, I will describe the two Pali hermeneutic handbooks, the 

Nettipakaraṇa and the Peṭakopadesa, which guide their audience to interpret the Buddha’s 

teachings and unfold its meaning in a normative way. The handbooks do not expect the audience 

or the readers of the words of the Buddha to encounter the Buddha’s teaching freely from any 

background or with whatever techniques. Instead, they demand the audience to prepare 

themselves by developing “wise attention” (yoniso manasikāra) and to analyse the Buddha’s 

words in conformity with the doctrinal structures. Several instances of commentarial methods in 

Peṭ and Nett historically resemble the method called “categorial reduction” in the Gandhari 

commentaries in breaking the different parts of a canonical verse and mapping them onto basic 

doctrinal schemes such as the four noble truths. The commentators summarised and replaced the 



 31 

conventional language in the verses or expansive daily speech by “translating” them into 

Buddhist terminology by the “Modes of Conveying” (hāra). The other methods also include 

Modes of Conveying (hāra), like conversion (āvatta), that contrast antithetical concepts with the 

concepts in the root texts. In addition to the structural analysis, which unfolds the internal 

structure of the root texts, the hermeneutic handbooks devise interpretative strategies, Guidelines 

(naya), targeting various ideal types of persons, and actualising the soteriological function of the 

buddhavacana by guiding the learners to follow the spiritual progress. 

In the third chapter, I will explore the unconventional genre of “narrative commentary,” 

with Jat-a and Dhp-a as the representative works. I will begin by tracing the concept of nidāna in 

the canon and show how the focus of nidāna and the origin of a canonical teaching allow room 

for contextual narratives in the commentaries. After closely reading the two stories from Dhp-a 

and Jat-a, I will not only unfold the narrative art of the commentaries but also demonstrate that 

the move between two stories connected by a karmic link creates a unique narrative temporal 

framework with a strong Buddhist ethical concern. The narrative complex made of multiple 

stories that are connected by karmic links also situates its audience in this moral world where 

repetitive time and nonrepetitive time intertwine. By means of the narrative art and temporal 

structure, narrative commentary serves as an effective didactic tool that combines the everyday 

moral tensions, karmic laws, and soteriological transcendence. 

In the last chapter, I will delve into the theories and practices concerning language in the 

commentaries and demonstrate that the knowledge of language is used as an important tool in 

interreligious debate. I will look briefly at the commentarial device called nirukti in Pali or 

semantic derivation, which has long been used in the Brahmanical interpretive tradition in its 

study of the Vedas. Then I will conduct a case study of the semantic derivation of several terms 



 32 

in both Brahmanical and Buddhist literature. The study will show that Buddhist commentators 

treated the semantic derivation of these terms seriously and unanimously interpreted them within 

a strong Buddhist ideological framework. Their semantic derivation implies a theory of language 

that regards the signifying relationship between linguistic expression and its meaning (or object) 

as originary or constant and disputes the Brahmanical authoritative interpretation by positing 

alternative explanations. To substantiate my hypothesis, I will show the Pali commentators’ 

revolutionary new ideology that upheld Pali, or the Magadhan language, as the best language that 

is both “natural” (sabhāva) to its speaker and remains unchanging. I will substantiate that Pali 

language ideology had its predecessor in Brahmanical tradition and was motivated by sectarian 

competition in the socio-religious context of fifth century Lanka. 

  



 33 

CHAPTER 1: RECITATION AND DEFINITION: THE DIVERSITY OF EARLY BUDDHIST 

EXEGETICAL METHODS 

The Earliest Stage of Buddhist Commentary 

Scholars have used “commentary” to refer to a variety of texts in the Buddhist scholastic 

tradition that focus on the explication of a root text that is usually an authoritative teaching of the 

Buddha or another Buddhist intellectual. As I will show in this chapter, the stylistic features of 

the Niddesa (Nidd) are different from those of the Gandhari commentaries dedicated to similar 

root texts. Nidd are consisted of the commentarial styles that show deep connection with oral 

culture and predates the other commentaries that we will discuss later in this study. The 

commentarial styles represented by Nidd did not engage in doctrinal analysis, which was more 

prevalent in other scholastic and commentarial texts, such as Buddhaghosa’s aṭṭhakathās and the 

Abhidharmic treatises in the Sarvāstivāda tradition. By comparing how Nidd and later 

commentaries define an important concept, suffering (P. dukkha), I will show that the 

commentarial styles closely related to oral performance gradually decline and the later 

commentaries seek stylistic compactness and theoretical consistency by drawing new concepts 

from the Abhidharma/Abhidhamma tradition. 

The time when Nidd was composed is a difficult question, but the story of its finalisation 

implies that it could have been added through several centuries and gone through the transition 

from an oral culture to a writing culture. The relative date of Nidd as earlier than Dīp and the 

commentaries composed by Buddhaghosa means a terminus ante quem circa 350 CE (von 

Hinüber 2000, §183). If the traditional accounts in Dīp is reliable, Nidd could have been 

composed not long before the Second Joint Recitation (saṅgīti), which is probably circa 300 



 34 

BCE.20 K. R. Norman is inclined to believe in the traditional accounts and claims that “the 

beginning of the third century B. C. would seem to be quite suitable as the date of its 

composition” for Nidd (Norman 1983, 86). The dating of Nidd proposed by Sylvain Lévi, which 

relied on geographical information implied by the text, has not withstood later scholars’ 

scrutiny.21 Even though Himanshu Sarkar disputed Lévi’s dating of Nidd, neither did he offer a 

satisfactory solution to the question. He assumed that the Mahāniddesa (i.e., the first part of 

Nidd, which von Hinüber sees as a later arbitrary appellation) dates to Aśoka’s time based on the 

dubious Dīp accounts and the reference to some texts within the extant Suttanipāta (Sn) in the 

Bhābrū-Bairat edict (Sarkar 1981, 209).22 The fact that the Bhābrū-Bairat Edict contains titles of 

these texts does not prove the existence of Nidd. Although the Dīpavaṃsa (Dīp) and the 

Samantapāsādikā (Sp) refer to Nidd as an important text in the history of Buddhism, the dating 

of Nidd is difficult to determine.  

 
20 According to Dīp, the Second Joint Recitation occurred a hundred years after the Buddha’s 

death (ca. 400 BCE). On the dating of the Buddha’s death, see Berchert (1980). The aṭṭhakathā 

on Nidd, Saddhammapajotikā (from the ninth century CE) alleges that Nidd was composed by 

the Buddha’s disciple Sāriputta. The ninth-century account on its authorship cannot be 

considered reliable because it is not corroborated by other traditions (von Hinüber 2000, §287). 
21 Sylvain Lévi, in his examination of the place names in Nidd, gave a terminus ante quem of the 

second century CE (Nidd I 154–55). Lévi mapped the toponyms from Nidd to those in the 

Geography by Ptolemy and the lost work Bṛhatkathā. Himanshu Sarkar questioned Lévi’s 

reconstruction of many toponyms and pointed out Lévi’s methodological problems. First, Lévi 

presupposed that the place names referred to in Nidd follow an order based on geographical 

direction and located many places (from Gumba to Suvaṇṇakūṭa) in Southeast Asia. Second, 

Lévi’s ex nihilo argument that “the date of the list and consequently of the work [(Nidd)] that 

contains it is placed between the first century and the third century CE” because Nidd shows no 

knowledge of the continental countries such as Cambodia, Champa, and Borneo (1925, 51). 

However, the silence on these countries cannot exclude the possibility that the work was from a 

later period. Third, when looking for corresponding places in Sanskrit works, Lévi treated the 

Bṛhatkathā-Ślokasaṃgraha (ca. the ninth or tenth century CE) as a faithful descendent of the 

original Bṛhatkathā dated to the second century CE, which is hardly attestable. 
22 The division of Nidd into “great” (mahā) and “small” (cūla) is a twelfth-century one (von 

Hinüber 2000, §116). 



 35 

There is no question that Nidd can be placed in the earliest layer of extant Buddhist 

commentaries. In the first place, Nidd comments on texts which are probably the earliest 

Buddhist canonical works (the Aṭṭhakavagga, the Pārāyanavagga and the Khaggavisana Sutta of 

the Suttanipāta). Moreover, Nidd’s style is similar to the “old commentary” of the 

Suttavibhaṅga, which “defines the Pātimokkha rule word by word, giving for the most part, an 

explanation in the form of a list of synonyms” (Norman 1983, 19).23  

Even though Nidd could be an early commentary, the traditional accounts of its 

transmission suggest that it was not widely admitted as canonical. Nidd almost disappeared from 

the Pali tradition but was rescued in the first century BCE, when Pali canon was written down. 

Dīp records that as early as the second Joint Recitation Nidd was rejected by the Mahāsaṅghika 

school after the second Joint Recitation.24 If we believe the Dīp account is true, Nidd is passed 

down from the early oral tradition. Dīp and Buddhaghosa took the near extinction of Nidd in the 

first century BCE as the cause for which the Lankan bhikkhus wrote down the Pali canon in the 

first century BCE. Nidd seemed to have almost been lost during a great calamity (mahābhaye) 

according to the Samantapāsādikā (Sp 695–96). To connect Nidd to the writing down of the 

canon in Lanka, Norman hints at the great calamity (mahābhaya) as the one under the reign of 

King Vaṭṭagāmaṇī, when Dīp reported all the canonical texts were first written down after the 

Damiḷa invasions and “the decay of created beings” occurred (hāniṃ sattānaṃ) (Dīp 20.21).25 In 

 
23 On the Suttavibhaṅga, also see von Hinüber (2015, 419). 
24 “Rejecting the following texts, viz.: the Parivāra which is an abstract of the contents (of the 

Vinaya), the six sections of the Abhidhamma, the Paṭisambhidhā, the Niddesa, and some 

portions of the Jātaka, they [the Mahāsaṅghikas] composed new ones” (Oldenberg 1982, 141). 

(parivāraṃ atthuddhāraṃ abhidhammappakaraṇaṃ paṭisambhidañ ca niddesaṃ ekadesañ ca 

jātakaṃ ettakaṃ vissajjetvāna aññāni akariṃsu te) (Dīp 5.37; Sp 695–96). 
25 Oldenberg’s translation. It is unclear whether the great calamity is the same as the one occurred 

under the reign of Vaṭṭagāmaṇī. 



 36 

fact, Buddhaghosa reported that the reciters of the Majjhima Collection (Majjhima-bhāṇakas) 

put Nidd in the Khuddakagantha and included it in the Sutta Piṭaka while the Dīgha-bhāṇakas 

included it in the Abhidhamma Piṭaka (Sv 15,22–29). The uncertainty of Nidd’s position in the 

Buddhist canon even in the time of Buddhaghosa (ca. 4–5th centuries CE) implies that Nidd 

should belong to the layer of late canonical texts, that is later than the four major Nikāyas. 

Although the time span of seven hundred years is an extremely large one, it still situates 

Nidd among the earliest layer of commentaries in the Buddhist tradition. Its ambiguous status in 

the Pali canon suggests its complicated history of transmission, which is corroborated by the 

historical record of its near extinction. Nevertheless, Nidd without doubt precedes the period of 

massive production of commentarial literature during the lifetime of Buddhaghosa for about one 

hundred years. So in this study, it will be convenient for us to take Nidd as a representative of the 

oldest level of Buddhist commentaries and to explore the working of commentaries in the 

transition from oral to written culture.  

  



 37 

The Waxing Syllable Principle in the Niddesa 

Nidd shows several features of oral transmission, which were prevalent in the early canonical 

sutta literature. As a commentary which aims at explaining the terms from the root verses of Sn, 

Nidd adopted the expository styles from the canonical sutta literature, which include the 

synonymic progression and the categorial expansion.26 The synonymic progression comprises a 

string of words which are close in meaning and often derive from the same verbal stem. In many 

cases, the string of words are semantic equivalents of the terms commented upon, but each word 

emphasises the various aspects of the root terms with minor differences. Unlike the synonymic 

progression, the categorial expansion focusses on the referent of the term and its position in the 

whole system of Buddhist ideology. By situating the concept commented upon in relation to 

other Buddhist concepts, the categorial expansion represented an initial attempt to explore the 

meaning of a concept in relation to other concepts Although these two stylistic devices are 

commonly used in Nidd, I do not mean that they are mutually exclusive. In fact, the synonymic 

progression is a commonplace device that can form part of the categorial expansion, while 

categorial expansion often appears to consist of a hodgepodge of synonyms. 

The synonymic progression and categorial expansion in Nidd follow the rule of the 

Waxing Syllable Principle, which is an important sign of oral transmission widely seen in its 

canonical predecessors in Pali literature. The WSP in the canonical suttas was illustrated in Mark 

Allon’s detailed study of formulaic expressions in the Udumbarikasīhanāda Sutta of the Dīgha 

Nikāya. According to Allon, many formulaic expressions in the suttas follow the WSP: “as the 

sequence [of parallel word elements or units of meaning] progresses the syllable length of each 

 
26 I borrowed the term “synonymic progression” from William Stede’s introduction to Nidd II, 

see Stede (1918, xxv). 



 38 

subsequent element may be equal to or greater than what precedes it, but not less” (1997, 191). 

In his study of the orality of early Middle Indic Buddhist texts, Oskar von Hinüber refers to the 

WSP by borrowing a term from Helmer Smith’s Epilegomena of the Critical Pali Dictionary 

(CPD) “The rule of Waxing Components” (Trenckner 1924, 35). In his study published shortly 

before Allon’s, von Hinüber also dedicated a chapter to the WSP in the early commentary, 

including Nidd and the Suttavibhaṅga of the Vinaya Piṭaka. He noticed Nidd inherited the old 

formulas of synonymic progression from the suttas and “sondern auch neue gebildet, wobei alte 

gleichsam entfaltet warden” (new ones are also formed, with old ones developed) (1994, 26).  

The WSP is a trait of the oral culture that had a long history in the Brahmanical tradition 

before its adoption by the preachers of the early sutta literature. Its existence in the other early 

Indian texts preserved orally suggests that the WSP had been an important feature of the Indian 

oral tradition. Jan Gonda’s study showed that this WSP, what he called “tendency to ‘increasing 

magnitude,’” can be traced to the earliest Vedic hymns and the Brāhmaṇas, which had preceded 

the Buddhist texts and “attributed to an innate sense of rhythm and alternation” (J Gonda 1959, 

71).27 Caland also pointed out a similar principle of the components in a copulative (dvandva) 

compound, which had been theorised for the first time in Pā. 2, 2.34 (alpāctaram) and “was 

prevalent at the time when the Brāhmaṇas were a living literature” (1931, 59).28 Since the 

Brahmanical texts were dated to a period earlier than Nidd, the WSP prevalent in all the 

Buddhist and non-Buddhist literature could have influenced the author of Nidd. In the following 

example, Nidd follows WSP strictly as it tends to explain the root verses with long strings of 

words that share common verbal stems or express similar meanings. 

 
27 The early Jain canon also follows the WSP. See Allon (1997, 272). 
28 Pāṇini is dated “in or after the middle of the fourth century BCE” (Bronkhorst 2007, 177). The 

date of Pāṇini is around the earliest possible period of Nidd. 



 39 

In this way I know. In this way I understand. In this way I discern. In this way I 

penetrate. 

evaṃ jānāmi evaṃ ājānāmi evaṃ paṭijānāmi evaṃ paṭivijjhāmi. (Nidd II 14, 223)29 

Adhering to the WSP, the commentators who composed Nidd made use of the synonymic 

progression which fits the commentary’s need of exposition to intensify the repetition of sound 

and meaning. This line explains the phrase “maññāmi.” If we focus on the verbs and skip the 

unchanging evaṃ in the sentence, the pattern of the waxing syllables here is 3+4+5+5, which 

follows the rule. Moreover, since the first three verbs derive from the same stem √ñā, the 

syllables or phonemes in boldface create similar sounds which rhyme and intensify the focus of 

the exegesis in an oral-aural teaching context.30 The repetition of evaṃ also adds more sound 

similarities to the formula. The exegetical nature of Nidd allows the syllabic repetitions based on 

the words’ common roots or their similar conjugations. In comparison, it requires more prodigy 

to apply the WSP to a non-commentarial text. The Vedic formulas cited by Gonda contain fewer 

syllabic repetitions. For instance, “yathā vāto yathā vanaṃ yathā samudra ejati […]” (as the 

wind, as the forest, as the sea moves…) (Ṛg Veda 5.78.8ab; Gonda 1959, 70).31 The waxing 

components in the sentence are three different words: vāta (wind), vanaṃ (forest), and samudra 

(sea) and the rhythm is assisted by the repetition of yathā (as). In the Vedic example, an 

“increasing magnitude” of the words after yathā occurs in the semantic field while “an innate 

sense of rhythm and alternation” is also created by more syllables (Gonda 1959, 69–71). It is 

 
29 Since W. Stede did not give the full text of Nidd II in the PTS edition, the text I present here is 

often a reconstruction of the full text based on the PTS edition and the Chaṭṭha Saṅgāyana 

Tipiṭaka 4 (CST4) edition. 
30 The repetition of similar consonants (Skt. vyañjana) is named anuprāsa in the later Sanskrit 

poetic tradition. 
31 The version of the Ṛg Veda I use here is Nooten and Holland (1994). 



 40 

easier for the commentaries that repeat synonyms to apply the WSP and create sound similarities 

than the ordinary texts, which require more poetical prodigy. The Nidd commentor made the 

commentary easy to memorise and pleasant to hear by using the WSP. 

Evidence from other Buddhist canonical suttas and commentaries suggests that the 

Buddhist authors adhered to the WSP and deliberately expanded the synonymic progression from 

the early period of the canon to the later commentaries for other reasons. In some cases, new 

synonyms were added to the formulas probably to include words that are more familiar to a 

different audience so that the commentary like Nidd would have been more comprehensive and 

intelligible. For example, von Hinüber studied the formula explaining “to criticise” or “to deride” 

(khuṃseti, vambheti) and showed that Nidd borrowed synonyms from the Nikāyas and the 

Suttavibhaṅga while adding a new synonym (ghaṭṭeti) into the formula (1994, 28). In the case 

mentioned above, the commentator created more complicated words which are probably not 

familiar to the audience such as ājānāmi and paṭijānāmi. Although in many cases the lexical 

equivalents share the same verbal stem (jānāmi, ājānāmi, paṭijānāmi all derives from √ñā in the 

example), neither were the complex forms with prefixes commonly used outside the 

commentaries, nor did they convey any specific philosophical implication in a doctrinal context. 

The probable explanation is that such terms were artificially constructed to increase the length of 

the formula and to achieve the desired sound effect. Von Hinüber points out this artificially 

constructed synonymic progression had already occurred in canonical sutta texts, such as the 

case of “okiranti, ajjhokiranti, abhippakiranti” in the Mahāparinibbāna Sutta (D II 137) and 

words like “ajjhokirati, abhippakirati, pāhuneyya, gamma < Skt. grāmya or upassaṭṭha are 



 41 

limited to formulas” (1993, 106).32 As von Hinüber shows, the formula on okiranti is extended 

even longer by adding more synonyms in Sv (1994, 25). As a result of all the reasons mentioned 

above, the synonymic progression was still prevalent in the period when Nidd was composed.  

In Nidd, the commentator violated the WSP to create syllabic repetitions. The violation 

of the WSP, according to von Hinüber, is due to the “loss of stylistic feeling for those formulas” 

in some Buddhist Hybrid Sanskrit texts, such as the Mahāparinirvāṇa Sūtra and Divyāvadāna 

(von Hinüber 1993, 109). But the occasional violation of the WSP in Nidd often has the purpose 

of producing syllabic repetitions. For example, 

tvampi nipako paṇḍito paññavā buddhimā ñāṇī medhāvī (Nidd II 9, 184)33  

vimaṃsī ti paṇḍito paññavā buddhimā ñāṇī vibhāvī medhāvī (Nidd I 283, corresponding 

to Verse 22 in the British Library Kharoṣṭhī Fragments)34 

All the words above used to describe “you” (tvaṃ, i.e., the Buddha) mean “learned, wise.” 

Except paññavā and ñāṇī, the other terms do not share the common verbal stem √ñā. The pattern 

of syllables is 3+3+3+3+2+3 in a, 3+3+3+2+3+3 in b, neither of which follows the WSP strictly. 

This formula, as far as I know, does not occur in the Sutta or Vinaya Piṭakas. The reason why the 

WSP is violated in these cases would be easy to explain if we take the sound similarities into 

account—the terms can be divided into groups of two which have the endings in -o, -ā, -ī in a; In 

 
32 On the formula from the Mahāparinibbāna Sutta, “In der Sanskritfassung hat sich an der 

gleichen Stelle keine Entsprechung zu dieser Formel erhalten. Die beiden letzten Wörter 

scheinen für diese Formel geschaffen, da sie sich sonst nicht nachweisen lassen” (In the Sanskrit 

version, no equivalent to this formula has been preserved at the same place. The last two words 

seem to have been created for this formula, since they cannot be found otherwise) (von Hinüber 

1994, 25). 
33 “You are clever, learned, prudent, intelligent, having knowledge, and wise.” For a similar 

formula, see Nidd I 478; the formula corresponds to Verse 15 in the Gandhari manuscript of the 

British Library Kharoṣṭhī fragments studied by Stefan Baums. 
34 “Contemplative [means] learned, prudent, intelligent, having knowledge, powerful, wise.” 



 42 

b, the anomaly is a little different but understandable if we also look at the alliteration of pa-.35 In 

b, the first two words (paṇḍito paññavā) start with pa, then the second and the third (paññavā 

buddhimā) we put together because they both end in ā. The words in the last group (ñāṇī vibhāvī 

medhāvī) e