Historical Notes:
The Text & the Translation

There are many versions of the Dhammapada now extant: several recensions of the Pali Dhammapada from Burma, Cambodia, Laos, Sri Lanka, and Thailand; two incomplete manuscripts of a Gandhari Dharmapada found in central Asia; and a manuscript of a Buddhist Hybrid-Sanskrit Dharmapada found in a library in Tibet, called the Patna Dharmapada because photographs of this manuscript are now kept in Patna, India. There is also a Chinese translation of the Dharmapada made in the third century C.E. from a Prakrit original, now no longer extant, similar to–but not identical with–the Pali Dhammapada. Parts of a Dharmapada text are included in the Mahavastu, a text belonging to the Lokottaravadin Mahasanghika school. In addition, there are Sanskrit, Tibetan, and Chinese versions of a text called the Udanavarga, which is known in at least four recensions, all of them containing many verses in common with the Dhammapada/Dharmapada (Dhp) texts. To further complicate matters, there are Jain anthologies that contain verses clearly related to some of those found in these Buddhist anthologies as well.

Despite the many similarities among these texts, they contain enough discrepancies to have fueled a small scholarly industry. The different recensions of the Pali Dhp contain so many variant readings that there isn’t yet–even after more than a century of Western scholarship on the topic–a single edition covering them all. The discrepancies among the Pali and non-Pali versions are even greater. They arrange verses in different orders, each contains verses not found in the others, and among the verses in different versions that are related, the similarity in terms of imagery or message is sometimes fairly tenuous.

Fortunately for anyone looking to the Dhp for spiritual guidance, the differences among the various recensions–though many in number–range in importance from fairly minor to minor in the extreme. Allowing for a few obvious scribal errors, none of them fall outside the pale of what has long been accepted as standard early Buddhist doctrine as derived from the Pali discourses. For example, does the milk in verse 71 come out or does it curdle? Is the bond in verse 346 subtle, slack, or elastic? Is the brahman in verse 393 happy or is he pure? For all practical purposes, these questions hardly matter. They become important only when one is forced to take sides in choosing which version to translate, and even then the nature of the choice is like that of a conductor deciding which of the many versions of a Handel oratorio to perform.

Unfortunately for the translator, though, the scholarly discussions that have grown around these issues have tended to blow them all out of proportion, to the point where they call into question the authenticity of the Dhp as a whole. Because the scholars who have devoted themselves to this topic have come up with such contradictory advice for the potential translator–including the suggestion that it’s a waste of time to translate some of the verses at all–we need to sort through the discussions to see what, if any, reliable guidance they give.

Those who have worked on the issues raised by the variant versions of Dhp have, by and large, directed the discussion to figuring out which version is the oldest and most authentic, and which versions are later and more corrupt. Lacking any outside landmarks against which the versions can be sighted, scholars have attempted to reconstruct what must have been the earliest version by triangulating among the texts themselves. This textual trigonometry tends to rely on assumptions from among the following three types:

1) Assumptions concerning what is inherently an earlier or later form of a verse. These assumptions are the least reliable of the three, for they involve no truly objective criteria. If, for instance, two versions of a verse differ in that one is more internally consistent than the other, the consistent version will seem more genuine to one scholar, whereas another scholar will attribute the consistency to later efforts to “clean up” the verse. Similarly, if one version contains a rendition of a verse different from all other renditions of the same verse, one scholar will see that as a sign of deviance; another, as a sign of the authenticity that may have predated a later standardization among the texts. Thus the conclusions drawn by different scholars based on these assumptions tell us more about the scholars’ presuppositions than they do about the texts themselves.

2) Assumptions concerning the meter of the verses in question. One of the great advances in recent Pali scholarship has been the rediscovery of the metrical rules underlying early Pali poetry. As the Buddha himself is quoted as saying, “Meter is the structural framework of verses.” (SN 1:60) Knowledge of metrical rules thus helps the editor or translator spot which readings of a verse deviate from the structure of a standard meter, and which ones follow it. Theoretically, the obvious choice would be to adopt the latter and reject the former. In practice, however, the issue is not so clear-cut. Early Pali poetry dates from a time of great metrical experimentation, and so there is always the possibility that a particular poem was composed in an experimental meter that never achieved widespread recognition. There is also the possibility that–as the poetry was spontaneous and oral–a fair amount of metrical license was allowed. This means that the more “correct” forms of a verse may have been the products of a later attempt to fit the poetry into standard molds. Thus the conclusions based on the assumption of standard meters are not as totally reliable as they might seem.

3) Assumptions concerning the language in which the original Dhp was first composed. These assumptions require an extensive knowledge of Middle Indic dialects. A scholar will assume a particular dialect to have been the original language of the text, and will further make assumptions about the types of translation mistakes that might have been common when translating from that dialect into the languages of the texts we now have. The textual trigonometry based on these assumptions often involves such complicated methods of sighting and computation that it can produce an “original” version of the text that is just that: very original, coinciding with none of the versions extant. In other words, where the current variants of a verse might be a, b, and c, the added assumption about the Dhp’s original language and the ineptitude of ancient translators and copyists leads to the conclusion that the verse must have been d. However, for all the impressive erudition that this method involves, not even the most learned scholar can offer any proof as to what the Dhp’s original language was. In fact, as we will consider below, it is possible that the Buddha–assuming that he was the author of the verses–composed poetry in more than one language, and more than one version of a particular verse. So, as with the first set of assumptions, the methods of triangulation based on an assumed original language of the Dhp tell us more about the individual scholar’s position than they do about the position of the text.

Thus, although the scholarship devoted to the different recensions of the Dhp has provided a useful service in unearthing so many variant readings of the text, none of the assumptions used in trying to sort through those readings for “the original” Dhp have led to any definite conclusions. Their positive success has been limited mainly to offering food for academic speculation and educated guesses.

On the negative side, though, they have succeeded in accomplishing something totally useless: a wholesale sense of distrust for the early Buddhist texts, and the poetic texts in particular. If the texts contain so many varying reports, the feeling goes, and if their translators and transmitters were so incompetent, how can any of them be trusted? This distrust comes from accepting, unconsciously, the assumptions concerning authorship and authenticity within which our modern, predominately literate culture operates: that only one version of a verse could have been composed by its original author, and that all other versions must be later corruptions. In terms of the Dhp, this comes down to assuming that there was only one original version of the text, and that it was composed in a single language.

However, these assumptions are totally inappropriate for analyzing the oral culture in which the Buddha taught and in which the verses of the Dhp were first anthologized. If we look carefully at the nature of that culture–and in particular at clear statements from the early Buddhist texts concerning the events and principles that shaped those texts–we will see that it is perfectly natural that there should be a variety of reports about the Buddha’s teachings, all of which might be essentially correct. In terms of the Dhp, we can view the multiple versions of the text as a sign, not of faulty transmission, but of an allegiance to their oral origins.

Oral prose and poetry are very different from their written counterparts. This fact is obvious even in our own culture. However, we have to make an active effort of the imagination to comprehend the expectations placed on oral transmission between speakers and listeners in a culture where there is no written word to fall back on. In such a setting, the verbal heritage is maintained totally through repetition and memorization. A speaker with something new to say has to repeat it often to different audiences–who, if they feel inspired by the message, are expected to memorize at least its essential parts. Because communication is face-to-face, a speaker is particularly prized for an ability to tailor his/her message to the moment of communication, in terms of the audience’s background from the past, its state of mind at present, and its hoped-for benefits in the future.

This puts a double imperative on both the speaker and the listener. The speaker must choose his/her words with an eye both to how they will affect the audience in the present and to how they will be memorized for future reference. The listener must be attentive, both to appreciate the immediate impact of the words and to memorize them for future use. Although originality in teaching is appreciated, it is only one of a constellation of virtues expected of a teacher. Other expected virtues include a knowledge of common culture and an ability to play with that knowledge for the desired effect in terms of immediate impact or memorability. The Pali Dhp (verse 45) itself makes this point in comparing the act of teaching, not to creating something totally new out of nothing, but to selecting among available flowers to create a pleasing arrangement just right for the occasion.

Of course, there are situations in an oral culture where either immediate impact or memorability is emphasized at the expense of the other. In a classroom, listening for impact is sacrificed to the needs of listening for memorization, whereas in a theater, the emphasis is reversed. All indications show, however, that the Buddha as a teacher was especially sensitive to both aspects of oral communication, and that he trained his listeners to be sensitive to both as well. On the one hand, the repetitious style of many of his recorded teachings seems to have been aimed at hammering them into the listener’s memory; also, at the end of many of his discourses, he would summarize the main points of the discussion in an easy-to-memorize verse.

On the other hand, there are many reports of instances in which his listeners gained immediate Awakening while listening to his words. And, there is a delightful section in one of his discourses (the Samaññaphala Suttanta, DN 2) satirizing the teachers of other religious sects for their inability to break away from the formulaic mode of their teachings to give a direct answer to specific questions (“It’s as if, when asked about a mango, one were to answer with a breadfruit,” one of the interlocutors comments, “or, when asked about a breadfruit, to answer with a mango.”) The Buddha, in contrast, was famous for his ability to speak directly to his listeners’ needs.

This sensitivity to both present impact and future use is in line with two well-known Buddhist teachings: first, the basic Buddhist principle of causality, that an act has repercussions both in the present and on into the future; second, the Buddha’s realization, early on in his teaching career, that some of his listeners would attain Awakening immediately on hearing his words, whereas others would be able to awaken only after taking his words, contemplating them, and putting them into prolonged practice.

A survey of the Buddha’s prose discourses recorded in the Pali Canon gives an idea of how the Buddha met the double demands placed on him as a teacher. In some cases, to respond to a particular situation, he would formulate an entirely original teaching. In others, he would simply repeat a formulaic answer that he kept in store for general use: either teachings original with him, or more traditional teachings–sometimes lightly tailored, sometimes not–that fit in with his message. In still others, he would take formulaic bits and pieces, and combine them in a new way for the needs at hand. A survey of his poetry reveals the same range of material: original works; set pieces–original or borrowed, occasionally altered in line with the occasion; and recyclings of old fragments in new juxtapositions.

Thus, although the Buddha insisted that all his teachings had the same taste–that of release–he taught different variations on the theme of that taste to different people on different occasions, in line with his perception of their short- and long-term needs. In reciting a verse to a particular audience, he might change a word, a line, or an image, to fit in with their backgrounds and individual needs.

Adding to this potential for variety was the fact that the people of northern India in his time spoke a number of different dialects, each with its own traditions of poetry and prose. The Pali Cullavagga (V.33.1) records the Buddha as insisting that his listeners memorize his teachings, not in a standardized lingua franca, but in their own dialects. There is no way of knowing whether he himself was multi-lingual enough to teach all of his students in their own dialects, or expected them to make the translations themselves. Still, it seems likely that, as a well-educated aristocrat of the time, he would have been fluent in at least two or three of the most prevalent dialects. Some of the discourses–such as DN 21–depict the Buddha as an articulate connoisseur of poetry and song, so we can expect that he would also have been sensitive to the special problems involved in the effective translation of poetry–alive, for instance, to the fact that skilled translation requires more than simply substituting equivalent words. The Mahavagga (V.13.9) reports that the Buddha listened, with appreciation, as a monk from the southern country of Avanti recited some of his teachings–apparently in the Avanti dialect–in his presence. Although scholars have often raised questions about which language the Buddha spoke, it might be more appropriate to remain open to the possibility that he spoke–and could compose poetry in–several. This possibility makes the question of “the” original language or “the” original text of the Dhp somewhat irrelevant.

The texts suggest that even during the Buddha’s lifetime his students made efforts to collect and memorize a standardized body of his teachings under a rubric of nine categories: dialogues, narratives of mixed prose and verse, explanations, verses, spontaneous exclamations, quotations, birth stories, amazing events, question and answer sessions. However, the act of collecting and memorizing was pursued by only a sub-group among his monks, while other monks, nuns, and lay people doubtlessly had their own individual memorized stores of teachings they had heard directly from the Buddha or indirectly through the reports of their friends and acquaintances.

The Buddha had the foresight to ensure that this less standardized fund of memories not be discounted by later generations; at the same time, he established norms so that mistaken reports, deviating from the principles of his teachings, would not be allowed to creep into the accepted body of doctrine. To discourage fabricated reports of his words, he warned that anyone who put words in his mouth was slandering him (AN 2:23). This, however, could in no way prevent mistaken reports based on honest misunderstandings. So, shortly before his death, he summarized the basic principles of his teachings: the 37 Wings to Awakening (bodhi-pakkhiya dhamma–see the note to verse 301) in the general framework of the development of virtue, concentration, and discernment, leading to release. Then he announced the general norms by which reports of his teachings were to be judged. The Mahaparinibbana Suttanta (DN 16) quotes him as saying:

Thus, a report of the Buddha’s teachings was to be judged, not on the authority of the reporter or his sources, but on the principle of consistency: did it fit in with what was already known of the doctrine? This principle was designed to ensure that nothing at odds with the original would be accepted into the standard canon, but it did open the possibility that teachings in line with the Buddha’s, yet not actually spoken by him, might find their way in. The early redactors of the canon seem to have been alert to this possibility, but not overly worried by it. As the Buddha himself pointed out many times, he did not design or create the Dhamma. He simply found it in nature. Anyone who developed the pitch of mental strengths and abilities needed for Awakening could discover the same principles as well. Thus the Dhamma was by no means exclusively his.

This attitude was carried over into the passages of the Vinaya that cite four categories of Dhamma statements: spoken by the Buddha, spoken by his disciples, spoken by seers (non-Buddhist sages), spoken by heavenly beings. As long as a statement was in accordance with the basic principles, the question of who first stated it did not matter. In an oral culture, where a saying might be associated with a person because he authored it, approved it, repeated it often, or inspired it by his/her words or actions, the question of authorship was not the overriding concern it has since become in literate cultures. The recent discovery of evidence that a number of teachings associated with the Buddha may have pre- or post-dated his time would not have fazed the early Buddhists at all, as long as those teachings were in accordance with the original principles.

Shortly after the Buddha’s passing away, the Cullavagga (XI) reports, his disciples met to agree on a standardized canon of his teachings, abandoning the earlier nine-fold classification and organizing the material into something approaching the canon we have today. There is clear evidence that some of the passages in the extant canon do not date to the first convocation, as they report incidents that took place afterwards. The question naturally arises as to whether there are any other later additions not so obvious. This question is particularly relevant with regard to texts like the Dhp, whose organization differs considerably from redaction to redaction, and leads naturally to the further question of whether a later addition to the canon can be considered authentic. The Cullavagga (XI.1.11) recounts an incident that sheds light on this issue:

In other words, Ven. Purana maintained–and undoubtedly taught to his followers–a record of the Buddha’s teachings that lay outside the standardized version, but was nevertheless authentic. As we have already noted, there were monks, nuns, and lay people like him even while the Buddha was alive, and there were probably others like him who continued maintaining personal memories of the Buddha’s teachings even after the latter’s death.

This story shows the official early Buddhist attitude toward such differing traditions: each accepted the trustworthiness of the others. As time passed, some of the early communities may have made an effort to include these “external” records in the standardized canon, resulting in various collections of prose and verse passages. The range of these collections would have been determined by the material that was available in, or could be effectively translated into, each individual dialect. Their organization would have depended on the taste and skill of the individual collectors. Thus, for instance, we find verses in the Pali Dhp that do not exist in other Dhps, as well as verses in the Patna and Gandhari Dhps that the Pali tradition assigns to the Jataka or Sutta Nipata. We also find verses in one redaction composed of lines scattered among several verses in another. In any event, the fact that a text was a later addition to the standardized canon does not necessarily mean that it was a later invention. Given the ad hoc way in which the Buddha sometimes taught, and the scattered nature of the communities who memorized his teachings, the later additions to the canons may simply represent earlier traditions that escaped standardization until relatively late.

When Buddhists began committing their canons to writing, approximately at the beginning of the common era, they brought a great change to the dynamic of how their traditions were maintained. The advantages of written over oral transmission are obvious: the texts are saved from the vagaries of human long-term memory and do not die out if those who have memorized them die before teaching others to memorize them as well. The disadvantages of written transmission, however, are less obvious but no less real. Not only is there the possibility of scribal error, but–because transmission is not face-to-face–there can also be the suspicion of scribal error. If a reading seems strange to a student, he has no way of checking with the scribe, perhaps several generations distant, to see if the reading was indeed a mistake. When confronted with such problems, he may “correct” the reading to fit in with his ideas of what must be right, even in cases where the reading was correct, and its perceived strangeness was simply a result of changes in the spoken dialect or of his own limited knowledge and imagination. The fact that manuscripts of other versions of the text were also available for comparison in such instances could have led scribes to homogenize the texts, removing unusual variants even when the variants themselves may have gone back to the earliest days of the tradition.

These considerations of how the Dhp may have been handed down to the present–and especially the possibility that (1) variant recensions might all be authentic, and that (2) agreement among the recensions might be the result of later homogenization–have determined the way in which I have approached this translation of the Pali Dhp. Unlike some other recent translators, I am treating the Pali Dhp as a text with its own integrity–just as each of the alternative traditions has its own integrity–and have not tried to homogenize the various traditions. Where the different Pali recensions are unanimous in their readings, even in cases where the reading seems strange (e.g., 71, 209, 259, 346), I have stuck with the Pali without trying to “rectify” it in light of less unusual readings given in the other traditions. Only in cases where the different Pali redactions are at variance with one another, and the variants seem equally plausible, have I checked the non-Pali texts to see which variant they support. The translation here is drawn from three editions of the text: the Pali Text Society (PTS) edition edited by O. von Hinüber and K.R. Norman (1995); the Oxford edition edited by John Ross Carter and Mahinda Palihawadana, together with its extensive notes (1987); and the Royal Thai edition of the Pali Canon (1982). The pts edition gives the most extensive list of variant readings among the Pali recensions, but even it is not complete. The Royal Thai edition, for example, contains 49 preferred and 8 variant readings not given in the PTS version at all. Passages where I have differed from the PTS reading are cited in the End Notes.

Drawing selectively on various recensions in this way, I cannot guarantee that the resulting reading of the Dhp corresponds exactly to the Buddha’s words, or to any one text that once existed in ancient India. However, as I mentioned at the beginning of this note, all the recensions agree in their basic principles, so the question is immaterial. The true test of the reading–and the resulting translation–is if the reader feels engaged enough by the verses to put their principles into practice and finds that they do indeed lead to the release that the Buddha taught. In the final analysis, nothing else really counts.