The Dhammapada: Read the Poem, Skip the Neuroscience

Verdict

Read it. The verdict is unchanged from our previous version. Almost every reason behind it has been replaced.

«The Dhammapada» is a religious anthology compiled and transmitted over centuries. It does not promise that eight weeks of anything will lower your cortisol. Holding it to a standard of falsifiable prediction is a category error, like scoring Ecclesiastes on labour-market forecasting. So this page does two jobs instead. It says what the text is and where it came from, so you can pick a translation rather than be sold one. Then it tests the one genuinely testable claim here, which nobody in the Pali canon ever made: that these verses anticipated cognitive science.

The text survives that intact. The marketing does not.

The claim on trial

The claim is not in the book. It is in the paragraph introducing the book: the Buddha opened his most famous work by stating that mind precedes and creates all things, which cognitive-behavioural therapy rediscovered 2,400 years later, which is why mindfulness science now validates him. Four assertions, each checkable. Was it an opening. Does the Pali say mind creates all things. Is the mechanism the one CBT proposes. Does the evidence carry the weight. In order: only in one tradition’s arrangement; no; not quite; and partly, far below the size implied.

One checked negative first. We searched PubMed on 11 August 2026 for «Dhammapada» and got a single record — PMID 12345273, a 1993 Asiaweek piece on the Buddhist view of family planning (PubMed E-utilities query, «Dhammapada»). «Dhammapada AND cognitive» returns zero. The proposition has never been posed as a hypothesis in the clinical literature, let alone tested. Untested is not refuted — but everyone repeating it repeats a literary impression, not a result.

What the text actually is

It is the second text in the Khuddaka Nikāya of the Sutta Piṭaka: 423 stanzas in 26 chapters, as catalogued chapter by chapter in Law’s «A History of Pali Literature», Kegan Paul, 1933. The verses are not unique to it — they recur elsewhere in the canon and across the parallel recensions, as the apparatus below shows. It is a curated selection, closer to a hymnal than an authored work. Nobody sat down and wrote it.

Several schools made their own selections from the same pool and ordered them differently. SuttaCentral’s parallels for Dhammapada 1–2 list a Gāndhārī version, a Patna Prakrit version, the Sanskrit Udānavarga, four Chinese recensions, a Tibetan one, and a prose parallel elsewhere in the Pali canon. Roebuck, Religions of South Asia, 2012, the Penguin Classics translator, maps those relationships and shows how much interpretation depends on which recension a scholar started from.

The physical record is younger than readers assume, because birch bark rots. Baums, Manuscript Cultures: Mapping the Field, 2014 counts more than 89 Gandhāran scrolls by over 51 scribes, dated by palaeography, linguistics and radiocarbon to the first and second centuries CE, some possibly first century BCE. Before those finds, exactly one such scroll was known: a Gāndhārī Dharmapada from near Khotan, edited by John Brough in 1962. The earliest physical witness to this family of texts is a Dharmapada in the wrong language and the wrong order.

One more date matters. The verses carry no context of their own; the stories that explain them come from a Pali commentary of the fifth century CE, which reached English alongside the verses only with Carter and Palihawadana’s edition, reviewed in Journal of the American Oriental Society, 1991. A charming backstory in your edition is a fifth-century monk’s gloss, not the verse.

Why the translations disagree

The Pali of the famous couplet reads «Manopubbaṅgamā dhammā, manoseṭṭhā manomayā». Two words carry the dispute: mano and dhammā. Max Müller, in the 1881 Sacred Books of the East volume, gave «All that we are is the result of what we have thought». Byrom in 1976 went further: «We are what we think. All that we are arises with our thoughts. With our thoughts we make the world.» The third line is the one on the mugs, and it is not in the Pali. His publisher describes the book as a «verse rendering», not a translation, and Bodhipaksa in Tricycle argues the world-making line imports an Advaita Vedanta idea the text does not hold — quoting Gil Fronsdal’s 2005 version by contrast: «All experience is preceded by mind, led by mind, made by mind.»

The sharpest correction comes from a translator working against the idealist reading. In his notes at SuttaCentral, Bhikkhu Sujato renders mano as «intention» and says why: it is the active aspect of mind, effectively meaning kamma, and «mind» favours an idealistic reading the rest of the verse does not justify. And the rest of the verse settles it. Speak or act with corrupt intent and suffering follows like the wheel behind the ox; with pure intent, happiness follows like a shadow. The subject is moral causation — intention, action, consequence. Not perception, and not the claim that thinking constructs reality.

Which editions do what. Technical standard: Norman’s «The Word of the Doctrine», Pali Text Society, 1997, reviewed in Bulletin of the School of Oriental and African Studies, 1998. Pali, verses and commentary together: Carter and Palihawadana. Modern and honest about the Pali: Fronsdal, Roebuck. Free and annotated: Sujato. Poetic and unmoored: Byrom. If a verse you love appears in only one of these, it is the translator’s.

What the meditation evidence shows

The honest summary: effects are real, modest, not specific to meditation, and produced by a literature with quality problems.

The anchor is Goyal et al., JAMA Internal Medicine, 2014, which screened 18,753 citations and kept only randomised trials with active controls — the reason its numbers are lower than the ones usually quoted. Across 47 trials and 3,515 participants: anxiety 0.38 (95% CI 0.12 to 0.64) at eight weeks, falling to 0.22 by three to six months; depression 0.30 (95% CI 0.00 to 0.59), an interval touching zero. For attention — the outcome any claim about training the mind depends on — they found low evidence of no effect or insufficient evidence of any effect, along with positive mood, sleep and eating habits. No evidence meditation beat any active treatment.

Goldberg et al., Clinical Psychology Review, 2018, across 142 samples and 12,005 participants, found d = 0.55 against no treatment, 0.23 against specific active controls, −0.004 against established evidence-based treatments. Two areas beat that average. Kuyken et al., JAMA Psychiatry, 2016, pooling patient-level data from nine trials and 1,258 patients, found mindfulness-based cognitive therapy cut depressive relapse over 60 weeks, hazard ratio 0.69 (95% CI 0.58 to 0.82). And Hoge et al., JAMA Psychiatry, 2023 randomised 276 adults with anxiety disorders to mindfulness-based stress reduction or escitalopram and found it non-inferior, difference −0.07 on clinician-rated severity.

Against that, the field’s own house critique. Van Dam et al., Perspectives on Psychological Science, 2018 — «Mind the Hype», fifteen authors including several prominent meditation researchers — documents that «mindfulness» has no agreed operational definition, that methodology is weak, and that the resulting misinformation leaves consumers harmed and misled. Not a hostile outsider piece, which is what makes it useful. The quality problem is measured: Goldberg et al., PLOS ONE, 2017 examined 171 randomised trials from 2000 to 2016 and found no improvement on any of six study-quality indicators.

Where the mechanism meets CBT

The bridge claim is that the verse’s mechanism — mental states precede and shape what follows — is the one cognitive therapy runs on. The verse is closer to a karmic claim, as above. Grant the mapping anyway, because the CBT evidence for cognitive mediation is itself contested.

Longmore and Worrell, Clinical Psychology Review, 2007 found little evidence that specifically cognitive interventions increase the therapy’s effectiveness, and little support for cognitive change being causal in the improvements CBT produces. Bell, Marcus and Goodlad, Journal of Consulting and Clinical Psychology, 2013 put numbers on it across 66 component studies from 1980 to 2010: dismantled treatments did not differ significantly from full ones, and adding a component bought d = 0.14 at completion and 0.28 at follow-up, only on the problems it targeted.

The other side gets the same standard. Lorenzo-Luaces, German and DeRubeis, Clinical Psychology Review, 2015 reviewed the same terrain and concluded that cognitive procedures do relieve depression and that cognitive change contributes to symptom change however it is produced. Defensible — and still weaker than the claim the verse is recruited for, because it does not show the cognitive route is necessary or privileged. Forty years on, the discipline that built its brand on thoughts driving change has not settled whether they do.

What the popular framing leaves out

Harm. Farias et al., Acta Psychiatrica Scandinavica, 2020 analysed 83 studies covering 6,703 practitioners. Sixty-five per cent reported at least one type of adverse event. Pooled prevalence was 8.3% (95% CI 5% to 12%), an average hiding a split: 3.7% in experimental studies against 33.2% in observational ones. Most common were anxiety (33% of studies reporting events), depression (27%) and cognitive anomalies (25%); suicidal behaviours appeared in 11%. The authors note these occur in people with no prior mental-health history. Britton et al., Clinical Psychological Science, 2021 measured it prospectively in 96 people completing an eight-week mindfulness-based cognitive therapy programme: 83% reported at least one meditation-related side effect, 37% a negative impact on functioning, 6% to 14% lasting bad effects. Their own conclusion is measured — rates comparable to other psychological treatments — which is the point. This is an intervention with a side-effect profile, not a benign default.

Measurement. The text’s central goods are equanimity and non-attachment, and both have been operationalised. Sahdra, Shaver and Brown, Journal of Personality Assessment, 2010 built the Nonattachment Scale from Buddhist scholarship; equanimity has the two-factor EQUA-S, validated by Juneau et al., PeerJ, 2020 in 265 adults. Instruments exist. Whether they measure the thing is another question: Grossman, Psychological Assessment, 2011 catalogues the problem — no external referent for validity, poor convergence between scales, self-reported competence that does not track behaviour, and response biases varying with practice — a ten-year meditator answers «I find it difficult to stay focused» differently from a beginner.

Who should actually read it

Read it as a primary source. Two hours, one of the most influential texts ever written, and it repays being read as literature and history. On that basis it earns its place among our reviewed shelf, most of which do not survive contact with their own citations. Read it for the ethical material, the bulk of the book and the part nobody quotes: the chapters on anger, speech and restraint are more specific and less flattering than the mind-makes-reality passages that get extracted.

Do not read it for a mechanism. Nothing in it tells you how attention works, and the habit of retrofitting neuroscience onto verse 1 says more about the retrofitters. And do not read Byrom if accuracy matters: read Fronsdal or Roebuck for a modern edition, Norman for the philology, Carter and Palihawadana for the commentary, Sujato for a free text annotated against the idealist reading.

For what happens when a modern author attaches a brand and a research claim to the same practice, see «Catching the Big Fish» and «Making Time».

For a far older tradition read the same way — strongest version first, then the mechanism — see the essay on Christianity as a religion of slaves and neurotics.

One thing to try

Read verses 1 and 2 in three translations side by side — Müller’s, Byrom’s, Sujato’s, all free online — before reading anything else in the book. Four minutes.

The point is not to pick a winner. The three are not describing the same claim: one says our thoughts constitute us, one says our thoughts construct the world, one says our intentions produce our actions and our actions our consequences. Watch that spread open in a text you thought you knew, and you will read every translated aphorism differently after it.

Get the book

Find «The Dhammapada» on Amazon — as an Amazon Associate, The Boring Work earns from qualifying purchases (disclosure).

The boring bottom line

«The Dhammapada» is a 423-verse anthology assembled from a shared pool of Buddhist verse, arranged one way by the Theravādins and other ways by everyone else, first attested physically in a Gāndhārī manuscript that puts the famous chapter two hundred verses in. It owes nobody a randomised controlled trial.

What it has been made to owe is another matter. The claim that it anticipated cognitive science depends on a translation choice Byrom made in 1976, and rests on a literature whose best active-controlled meta-analysis reports 0.38 for anxiety, 0.30 for depression, nothing for attention, and no advantage over any other active treatment.

Read the poem. Skip the neuroscience it never claimed.

When to see a professional

This page is general information about a book and a research literature. It is not medical advice and nothing here is a treatment recommendation.

Meditation is not risk-free. The evidence above shows adverse events including anxiety, depression, dissociation and cognitive disturbance at non-trivial rates, sometimes in people with no psychiatric history. If a practice is producing panic, intrusive memories, depersonalisation, sleep disruption that does not settle, or worsening low mood, stop and speak to a clinician — a GP, a clinical psychologist or a psychiatrist — rather than pushing through. Retreat teachers and app instructors are not qualified to manage that, and «sit with it» is not clinical guidance.

If you have a current mental-health condition, are in treatment, or have a history of trauma or psychosis, talk to the person treating you before starting an intensive programme. If you are having thoughts of harming yourself, contact emergency services or a crisis line immediately. Do not change prescribed medication on the strength of anything you read here.

Sources

• Law, «A History of Pali Literature», Kegan Paul, Trench, Trubner, 1933 — the Dhammapada as the second book of the Khuddaka Nikāya, 423 verses in 26 vaggas, listed chapter by chapter.
• SuttaCentral, Dhammapada 1–20, Pali text with Bhikkhu Sujato’s translation and notes — the Pali of verses 1–2, the parallels in Gāndhārī, Prakrit, Sanskrit, Chinese and Tibetan, and the case for reading «mano» as intention rather than mind.
• Roebuck, Religions of South Asia, 2012 — how the Pali Dhammapada, the other recensions and the Udānavarga relate, and what that means for translators.
• Baums, Manuscript Cultures: Mapping the Field, 2014 — more than 89 Gandhāran birch-bark scrolls dated to the first and second centuries CE; the Khotan Gāndhārī Dharmapada was the only one known before recent finds.
• Carter and Palihawadana’s Dhammapada, reviewed in Journal of the American Oriental Society, 1991 — the edition carrying the Pali text and the first English translation of the commentary’s explanation of the verses.
• Norman’s «The Word of the Doctrine», reviewed in Bulletin of the School of Oriental and African Studies, 1998 — the Pali Text Society translation of 1997, the scholarly standard.
• Müller, «The Dhammapada», Sacred Books of the East vol. X part I, Clarendon Press, 1881 — «All that we are is the result of what we have thought».
• Publisher’s page for Thomas Byrom’s Dhammapada — described by the publisher as a «verse rendering», not a translation.
• Bodhipaksa, «We are what we think», Tricycle — why Byrom’s «with our thoughts we make the world» imports an idea the Pali does not carry, with Fronsdal’s 2005 rendering for contrast.
• PubMed search for «Dhammapada», run 11 August 2026 — count 1: PMID 12345273, «Free to choose. The Buddhist view», Asiaweek, 1993; zero for «Dhammapada AND cognitive».
• Goyal et al., JAMA Internal Medicine, 2014 — 47 active-controlled trials, 3,515 participants: anxiety 0.38, depression 0.30 at eight weeks; no effect established for attention; no advantage over any active treatment.
• Goldberg et al., Clinical Psychology Review, 2018 — 142 samples, 12,005 participants: d = 0.55 versus no treatment, 0.23 versus specific active controls, −0.004 versus evidence-based treatments.
• Goldberg et al., PLOS ONE, 2017 — 171 randomised trials, 2000 to 2016: no evidence of improvement on any of six study-quality indicators.
• Van Dam et al., Perspectives on Psychological Science, 2018 — «Mind the Hype»: no agreed definition of mindfulness, weak methodology, consumers harmed and misled.
• Kuyken et al., JAMA Psychiatry, 2016 — individual patient data, 9 trials, 1,258 patients: relapse hazard ratio 0.69 (95% CI 0.58 to 0.82) over 60 weeks.
• Hoge et al., JAMA Psychiatry, 2023 — 276 adults randomised: mindfulness-based stress reduction non-inferior to escitalopram, difference −0.07; adverse events 15.4% versus 78.6%.
• Farias et al., Acta Psychiatrica Scandinavica, 2020 — 83 studies, 6,703 practitioners: 8.3% pooled adverse-event prevalence, 3.7% experimental versus 33.2% observational.
• Britton et al., Clinical Psychological Science, 2021 — structured interview, n = 96: 83% reported a side effect, 37% a negative impact on functioning, 6% to 14% lasting bad effects.
• Longmore and Worrell, Clinical Psychology Review, 2007 — little evidence that cognitive interventions add effectiveness, little support for cognitive change as causal.
• Bell, Marcus and Goodlad, Journal of Consulting and Clinical Psychology, 2013 — 66 component studies: dismantled treatments no worse; added components d = 0.14 at completion, 0.28 at follow-up.
• Lorenzo-Luaces, German and DeRubeis, Clinical Psychology Review, 2015 — the counter-case: cognitive procedures work and cognitive change contributes to symptom change.
• Sahdra, Shaver and Brown, Journal of Personality Assessment, 2010 — the Nonattachment Scale, built from Buddhist scholarship.
• Juneau et al., PeerJ, 2020 — the two-factor equanimity scale in 265 adults: even-minded state of mind and hedonic independence.
• Grossman, Psychological Assessment, 2011 — why self-report mindfulness inventories lack external referents, converge poorly, and drift with practice experience.