Sensual Pleasure

Identity

Segment range snp4.1:1.1–6.4. Sn 766–771 (6 verses). The opening sutta of the Aṭṭhakavagga and the shortest in the collection. Received title Kāmasutta — "Discourse on Sensuality." The title names the sutta's theme, not a character or occasion; 4.1 has no narrative frame in the Pāli and no canonical interlocutor. Standing where it does, it positions kāma as the AV's opening concern; Snp 4.7 Tissametteyya and 4.9 Māgaṇḍiya pick the theme up later in dialogue form.

Text and form

Śloka throughout (Norman 2001 p. 323). No metre partition, no composite-sutta seam, no speaker-tags. Single-speaker Buddha monologue with no framing envelope — no evaṁ me sutaṁ, no iti bhagavā, no dedication-close. The sutta's architecture is bilaterally symmetric:

• vv. 1–2 — the existential symptom (elation at getting kāma; the dart of losing it);
• v. 3 — the prescription stated abstractly (sidestep, mindfulness, transcendence);
• vv. 4–5 — the symptom re-stated concretely (an enumerative inventory of kāma-objects and the suffering that follows from greed for them);
• v. 6 — the prescription re-stated concretely, with the closing boat-image picked up from v. 5.

Verbal echoes hold the small structure together: v. 3's sato … parivajjeti returns in v. 6 as sato … parivajjaye; v. 5's broken boat inverts to v. 6's bailed-out boat; v. 2's jantu returns in v. 6.

Content

A mortal (macco) who desires a kāma and obtains it is elated; when kāmas fade from the same person, he is "hurt as if pierced by an arrow" (sallaviddhova ruppati). Whoever sidesteps kāmas as one sidesteps a snake's head with the foot, mindful, passes beyond "this stickiness in the world" (visattikaṁ loke).

Concretely: a man greedy for fields, land, gold, cattle and horses, slaves and servants, women, kin, many kāmas — the seemingly weak overpower him, dangers crush him, and suffering follows him "as water [follows] a broken boat." Therefore a creature, ever mindful, should sidestep kāmas; abandoning them, he might cross the flood, as one who has bailed out a boat reaches the far shore.

No dependent-origination, no four truths, no five aggregates, no developed doctrinal vocabulary. The sutta moves by image and by enumerative inventory.

Key passages

v. 2 (Sn 767) — the dart.

Tassa ce kāmayānassa, / chandajātassa jantuno; / Te kāmā parihāyanti, / sallaviddhova ruppati.

But for that creature, longing for the desired, in whom the urge has arisen — when those pleasures fade, he is hurt as if pierced by an arrow.

Kāmayānassa preserves a present-middle participle of the -āna type (Skt middle voice; cf. Whitney §1043f, Geiger §192) that Norman 2001 p. 324 flags as already going out of use by the Niddesa's period — the commentary (mnd1:7.2, mnd1:14.3) treats it as a synonym of kāmayamāna, the form just used in v. 1. The verbatim formula sallaviddhova ruppati is unique to this verse (close variants at SN 9.2:2.4 and Thag 17.1:19.2 use different syntactic cases).

v. 3 (Sn 768) — sidestepping.

Yo kāme parivajjeti, / sappasseva padā siro; / Somaṁ visattikaṁ loke, / sato samativattati.

Whoever sidesteps sensual pleasures as one sidesteps a snake's head with the foot — mindful, he passes beyond this stickiness in the world.

Three pādas (b, c, d) are preserved verbatim at Thag 6.14:5 (Sabbakāmin's verses), with cetā substituted for v. 3's kāme in pāda a. The four-pāda specificity argues for direct citation rather than shared stock. Sato here is bare, uncompounded mindfulness — not satipaṭṭhāna, not satindriya, not satibala. The Niddesa's gloss expands this single word into a taxonomic cascade (see Vocabulary and commentary).

v. 5 (Sn 770) — the broken boat.

Abalā naṁ balīyanti, / maddantenaṁ parissayā; / Tato naṁ dukkhamanveti, / nāvaṁ bhinnamivodakaṁ.

The [seemingly] weak overpower him; dangers crush him. Suffering then follows him as water [follows] a broken boat.

Pāda c — Tato naṁ dukkhamanveti — is verbatim with Dhp 1:5 (the first verse of the Dhammapada), where the same causal pāda is paired with the cart-following-ox image. Also at Thag 16.2:10.3 (Anuruddha). Three-text verbatim attestation marks the pāda as stock formula rather than a direct borrowing in either direction.

v. 6 (Sn 771) — the closing.

Tasmā jantu sadā sato, / Kāmāni parivajjaye; / Te pahāya tare oghaṁ, / Nāvaṁ sitvāva pāragū.

Therefore a creature, ever mindful, should sidestep sensual pleasures. Abandoning them, he might cross the flood, as one [who has] bailed out a boat reaches the far shore.

Ogha here is unitary — not yet split into the four-flood formula (kāmogha / bhavogha / diṭṭhogha / avijjogha) that the Niddesa (mnd1:77.7) imposes on the verse. The closing text-critical crux — nāvaṁ sitvā va (= sitvā iva, Norman's reading) versus nāvaṁ siñcitvā (PTS, Lee 2024) — is discussed below.

Choice-points

v. 2, kāmayāna vs. kāmayamāna. The Pāli preserves two distinct grammatical forms of √kam- in back-to-back verses: kāmayamānassa (present active participle, v. 1) and kāmayānassa (present middle participle of the -āna type, v. 2). The middle form was archaic by the time of the Niddesa, which glosses it as though equivalent. The Chinese Yizujing collapses both to 欲 ("to desire"). The distinction is preserved only in the Pāli — a quiet diagnostic of antiquity. The author of the verse could still hear the grammatical distinction; no surviving commentarial or recensional layer can.

v. 3, visattikā. Etymologically opaque. The Niddesa (mnd1:29.2–29.7) supplies nine kenaṭṭhena derivations; most translators settle for "attachment" (Norman, Bodhi, Sujato, Thanissaro, Lee) or "craving" (Fronsdal). The choice turns on whether to preserve the image-rooted etymology (cf. √sañj "to cling") or render it with the abstract term the Niddesa's expansion recommends.

v. 5, abalā naṁ balīyanti. Three readings are live: (a) "women [= the weak ones] overpower him" (Norman, following CPD and an Atharvaveda parallel at AV 5.19.8b, reading abalā as a Vedic polite-ironic designation for women); (b) "weak defilements overpower him" (Niddesa mnd1:47.2–3; Bodhi preserves it as a secondary reading); (c) "the [seemingly] weak overpower him" (Sujato, Lee, Bodhi primary) — preserving the abalā/balīyanti paradox without committing to a gendered referent. The Chinese Yizujing Y1.22 reads 倒羸爲強 ("inverted-weak becomes strong") — a direct lexical parallel to reading (c), no gender-specificity. As independent cross-recensional evidence against the Norman/CPD "women" reading, this effectively closes the ambiguity in favour of (c).

v. 5c, Tato naṁ dukkhamanveti. The stock pāda appears at Snp 4.1, Dhp 1, and Thag 16.2. Norman (p. 325–26), following CPD, argues for "enters" over "follows" for anveti in the boat context. Sujato keeps "follows" on the strength of Dhp 1's cart-following-ox parallel. Either is defensible; consistency across the three attestations argues for one choice throughout.

v. 6, nāvaṁ sitvā va / siñcitvā. Text-critical. The PTS edition reads siñcitvā (absolutive of √sic, "to bail out"; Lee 2024 follows). Norman reads sitvā va (= sitvā iva) on metrical and commentarial grounds — both the Niddesa and Pj II include yathā / iva in their gloss, suggesting they read iva. The reading affects cadence, not sense.

Vocabulary and commentary

Lexical profile. The sutta's vocabulary is archaic and sparse. It uses biological terms for the human subject — macco ("mortal"), jantu ("creature"), naro ("man") — not the psychologically or doctrinally loaded puggala, ariya, or bhikkhu. Its operative verbs are optatives (parivajjaye, tare) rather than imperatives; the hortatory mode is mild. Conspicuously absent from the verse, though omnipresent in the Niddesa's gloss: nibbāna, arahant, bhava/vibhava, diṭṭhi, upādāna, nīvaraṇa, bojjhaṅga, magga, khandha, āyatana, dhātu, paṭiccasamuppāda — the entire developed prose-Nikāya doctrinal apparatus. The verse uses none of it.

Mahāniddesa (Mnd 1). Covers all six verses in canonical Niddesa format: verse quoted, word-glossed, re-quoted with tenāha bhagavā. The commentary's interpretive moves cluster in two groups.

Where the Niddesa works with the verse. The snake-elaboration at mnd1:27.4–14 — twelve poetic epithets for sappa (serpent) drawn from compositional play on sarpati ("creeps"), bhujati ("bends"), urena gacchati ("goes on the breast") — and the framing of the snake-avoider as "one who wants to live, not die, wants pleasure, recoils from pain" (mnd1:27.15) — works in the verse's own image-register. Similarly the boat-elaboration at mnd1:71.2–7 (water entering a broken boat from four sides as suffering enters one whose pleasures have failed) elaborates without overwriting.

Where the Niddesa imports later doctrinal vocabulary. Seven cases are taxonomic impositions on bare verse-words:

• kāma → vatthukāma / kilesakāma split (the Niddesa's master move, applied at every occurrence of kāma).
• parivajjeti → vikkhambhana / samuccheda (Abhidhammic suppression-vs-eradication), further split onto the four magga-paths.
• sato → full satipaṭṭhāna / anussati / sambojjhaṅga / ekāyanamaggo cascade.
• visattikā → ~80-term taṇhā-megalist including the kāmataṇhā / bhavataṇhā / vibhavataṇhā tetrad (second-truth formulation).
• ogha → four-flood formula.
• pāragū → sevenfold typology (abhiññā / pariññā / pahāna / bhāvanā / sacchikiriyā / samāpatti / nibbāna-pāragū) plus a ~30-line arahant-portrait identifying pāraṁ with amataṁ nibbānaṁ.
• parissayā → pākaṭa / paṭicchanna split, importing the five-nīvaraṇa pentad.

The contrast is the structural diagnostic for any stratigraphic reading of the AV: the verse's own lexical economy is ~25 content-words; the Niddesa's expansion runs to several hundred lines of dense doctrinal prose. Chapter 5 treats this commentary-imports-doctrine pattern as a general feature of the Mahāniddesa; here only the sutta-specific case is registered.

Cross-recensional witnesses

Pāli: full; 6 verses.

Chinese Yizujing YZJ-1 桀貪王經 ("King Jietan Sūtra") at [T0198_p0174b09]–[T0198_p0175c25]: a 23-verse composite in which the 6-verse Kāmasutta core (Y1.18–Y1.23 = Sn 766–771) is embedded within a larger narrative-commentarial envelope. Structure per Lee 2024 pp. 143–152 and primary-text verification in notes/per-sutta/snp-4-1-kama-cross-recensional.md §2:

• Y1.1–Y1.4 — frame-story preamble with four verses paralleling AN 5.5.8 / AN 5.5.10 on the uselessness of grief (King Mahāvijaya, defeated in war).
• Y1.5 — the king speaks a verse equivalent to Sn 766 and offers a thousand gold coins to anyone who can explain it.
• Y1.6–Y1.13 — young man Yuduo (鬱多) delivers an eight-verse explanatory expansion of the king's opening verse.
• Y1.14–Y1.17 — king-Yuduo dialogue: the king confirms that "you spoke eight verses; [a reward of] a thousand [coins] per verse" (汝說八偈，偈上千錢, Y1.14–15).
• Y1.18–Y1.23 — the 6-verse Kāmasutta proper, Sn 766–771.

The 8-verse explanatory expansion is primary-text evidence, in a text translated 223–253 CE, that at least one branch of the AV-tradition understood the aṭṭhaka as a commentarial form in which a seed-verse is explicated by eight additional verses — not as a count of eight-verse suttas. This is a candidate etymology for the collection's name (§1.2 of Chapter 1) independent of the four titled aṭṭhaka-suttas reading, and worth its own treatment in Chapter 4 on composition.

Three further Chinese findings: (i) Y1.20's rendering of Pāli bare sato as a samādhi + jhāna doublet (定 + 行禪) shows the Niddesa's doctrinal expansion of sato already reaching the verse-text level in the Chinese transmission — an intermediate stratum between the Pāli's bare-sato and the Niddesa's full satipaṭṭhāna cascade. (ii) Y1.22's 倒羸爲強 closes the abalā reading-question for "seemingly weak" (see Choice-points). (iii) Y1.23 presents as prose in five of seven Chinese editions surveyed by Lee — structural evidence, alongside the Yogācārabhūmi's different closing verse (below), that Sn 771 was transmission-unstable.

Sanskrit: the Yogācārabhūmi (Asaṅga) preserves a Sanskrit version of the Kāmasutta whose closing verse has no Pāli parallel, replacing Sn 771 with a different verse (Enomoto 1989, cited Norman 2001 p. 323; Thanissaro 2016 note 1 to 4:1 independently flags). Primary source not in-repo; Enomoto 1989 flagged for acquisition. Independent-witness convergence with the Chinese's v. 6 instability is the load-bearing fact.

Gāndhārī: no text extant. SC's parallels.json advertises a parallel at sag#sag40.1–#sag40.6 (Senior 20 scroll, SAS6), but our in-repo TEI (gandhari/transcriptions/SAS6_CKM0252.TEI.xml) contains only a prose six-senses / four-truths sutta with zero verse-blocks; Lee 2024's Arthapada edition does not transcribe a Gāndhārī Kāmasutta either. The Senior scroll is evidently damaged or lacunary at the Arthapada opening — a scroll-end rot pattern consistent with Kharoṣṭhī manuscript preservation.

Coverage note. Snp 4.1 is effectively 3-recension (Pāli + Chinese + partial Sanskrit) with the Gāndhārī slot vacant. The richest cross-recensional feature is the Chinese composite-envelope around the 6-verse core.

Internal cross-references

Within the AV. No verbatim formulas of Snp 4.1 recur elsewhere in the AV. Thematic continuities are extensive: kāma as the AV's master concern is set up here and developed in 4.7 and 4.9; the salla image returns in 4.15 and 4.16 in different metrical contexts; ogha recurs at 4.2, 4.7, 4.15; visattikā recurs at 4.10.

Within the Khuddaka. The strongest cross-Khuddaka bond is v. 5c Tato naṁ dukkhamanveti = Dhp 1:5 verbatim (also at Thag 16.2:10.3 in Anuruddha's verses). This three-text attestation marks the pāda as stock formula — one of the cleanest early-verse formulaic units shared across the AV, Dhp, and Thag. Also: v. 3 preserves a three-pāda verbatim match with Thag 6.14:5 (Sabbakāmin), where only cetā substitutes for kāme in pāda a; the specificity argues for direct citation. The snake-head-with-foot image recurs at Thag 457.

Prose-nikāya uptake. No named canonical citation of Snp 4.1 appears in the prose Nikāyas (contrast Snp 4.9, cited by name at SN 22.3; Snp 4.16, named in the Bhabru edict). The sutta circulates thematically rather than by title. The Niddesa's choice of inset illustrative verses — especially Rāgo ca doso ca itonidānā at mnd1:68.1–4, known from DN 21 Sakkapañhasutta — shows that by the time of the Niddesa's composition, Snp 4.1 was being read in active connection with prose-Nikāya passages on the lobha / dosa / moha triad.

Reception and external attestation

Mahāniddesa: Mnd 1 covers all six verses. See Vocabulary and commentary.

Paramatthajotikā II: supplies a narrative frame (Bodhi 2017 pp. 1013–1014) — a Sāvatthī brahmin whose barley field is flooded; the Buddha teaches him the Kāmasutta, and he and his wife attain stream-entry. The frame is a 5th-century-CE construction, post-dating Buddhaghosa, and is structurally different from the Yizujing's (roughly 200 years older) King Mahāvijaya frame. The Pāli commentarial frame and the Chinese Indic-source frame appear to be independent later constructions on the same frameless verse-core.

Aśoka Bhabru edict: not identified with any of the seven dhamma-paliyāyāni.

Peṭakopadesa: no verse of Snp 4.1 is cited in Pe chapter 1's AV-extraction (Chapter 8); the Pe's AV-citation density concentrates on 4.2, 4.6, 4.7.

Reading

Snp 4.1 opens the AV on kāma in the terms that organise the collection's attention: concrete, image-driven, image-bodied. No character, no occasion, no doctrinal apparatus. The bilaterally symmetric 6-verse structure (symptom / prescription / worked symptom / restated prescription) is tight, and the three images it introduces — arrow, snake-head, boat-and-flood — recur across the AV in later metrical registers. Nothing about the verse requires a larger frame to make sense.

The Yizujing preserves such a frame anyway. Its 23-verse King Jietan composite surrounds the 6-verse core with an AN-derived grief-verse preamble, a king's opening verse, an eight-verse explanatory expansion by Yuduo, and a king–Yuduo dialogue — a full narrative-commentarial envelope that the Pāli has shed. Whether this envelope descends from a pre-bifurcation tradition or is a parallel later expansion in the Indic-to-Chinese transmission is a question this entry does not resolve. What is certain is that the Pāli-specific minimalism of Snp 4.1's received form — pure didactic verse, no frame — is a preservation-choice, not a compositional default.

The Yizujing's own internal arithmetic is more striking. The king rewards Yuduo at "a thousand coins per verse" for eight verses, and the narrator names this counting explicitly. Read this way, "Aṭṭhakavagga" may name the commentarial form itself: a seed-verse explicated by an eight-verse commentary. The four titled aṭṭhaka-suttas of the Pāli collection (4.2–4.5) would then be the surviving cases where both verse-core and explicating eight survived together; Snp 4.1's Chinese witness preserves such a pairing where the Pāli has kept only the core. The hypothesis is not settled by the Chinese evidence alone — it is one reading of a composite text whose compositional history is uncertain — but it is a better grounded candidate etymology than any available from the Pāli alone. Chapter 4 treats the question.

Two further recensional signals complicate Sn 771 specifically. The Chinese v. 6 presents as prose in five of seven editions surveyed by Lee; the Yogācārabhūmi preserves a different closing verse with no Pāli parallel. Independent witnesses pointing to the same instability at the closing verse is evidence that Sn 771 settled into its received Pāli form later than Sn 766–770. The Pāli's closure is a local stabilisation, not a universal one. Whether this matters for the sutta's meaning is a separate question; as a transmission-historical datum, it is the one piece of the AV's opening that is genuinely unstable.