---
title: "The Tiantai Bridge — Zhiyi's Four Samādhis and the Wu-fangbian-nianfo-men"
author: "釋慧鏡 (Shi Huijing)"
series: "Six Practice Gates, Volume 1 — Mindful of the Buddha: From Early Buddhist Recollection to the Flower Ornament Vision"
series_short: "NIAN"
volume: 1
paper_id: "NIAN-P23"
paper_number: 23
part: "Part VI · Two Streams"
part_number: 6
date: "2026-05-04"
version: "1.0"
license: "CC BY-NC-SA 4.0"
---

# The Tiantai Bridge

**Zhiyi's Four Samādhis and the *Wu-fangbian-nianfo-men***

**Classificatory Bridge, Not Synthesis: Three Strata of Tiantai Doxography on Buddha-Recollection — Part VI's Fourth Paper on the Two Streams**

*中文主題：天台橋樑：智顗之四種三昧與《五方便念佛門》*

---

**Author:** 釋慧鏡 (Shi Huijing)
**Date:** May 2026
**License:** CC BY-NC-SA 4.0
**Version:** 1.0
**Series:** NIAN — Mindful of the Buddha

---

> "The four kinds of *samādhi* differ in method, but their contemplative principle is one."
> — Zhiyi (智顗, 538–597), *Móhē zhǐguān* (摩訶止觀), fascicle 1B (T46.18c11)

## Abstract
For the Anglophone reader of Pure Land introductions, the formula is by now familiar: *Tiantai master Zhiyi established four kinds of buddha-recollection (niànfó 念佛) — name-recitation (chēng-míng 稱名), image-contemplation (guān-xiàng 觀像), visualization (guān-xiǎng 觀想), and reality-contemplation (shí-xiàng 實相)*. This paper argues that this formula is, on close philological reading, not Zhiyi's authored typology at all. Zhiyi, lecturing in 594 and recorded by his disciple Guanding (灌頂, 561–632), set out in the *Móhē zhǐguān* (T1911) a *fourfold-by-posture* typology — *constantly seated*, *constantly walking*, *half-walking-half-seated*, *neither-walking-nor-seated* — and the attributed-to-Zhiyi *Wǔ-fāngbiàn niànfó-mén* (T1962) preserves a *fivefold* typology of *niànfó-sānmèi*, not a fourfold one. The popular fourfold scheme of *chēng-míng / guān-xiàng / guān-xiǎng / shí-xiàng* is, this paper contends, a **late-Ming retrospective compression** by Yúnqī Zhūhóng (雲棲祩宏, 1535–1615) in his *Āmítuó-jīng shūchāo* (阿彌陀經疏鈔, 1584+), a doxographic reorganization that collapsed Zhiyi-school's fivefold *niànfó-sānmèi* into a more memorable fourfold for the use of late-Ming and Qing-dynasty Pure Land instruction. The compound expression "four kinds of buddha-recollection" (*sì-zhǒng niànfó* 四種念佛) registers zero hits in the entire Taishō canon and only five hits in the *Wàn xùzàng* — all from Qing-era texts. What Zhiyi actually built in the late sixth century was a *classificatory bridge*: a vocabulary and architecture in which the inward turn later associated with Chan and the name-holding later associated with Pure Land could each find a non-hierarchical seat within a single contemplative matrix. The paper develops this argument in §III through four sub-sections, engages in §IV with three foundational Western readings (Stevenson 1986, Donner & Stevenson 1993, Swanson 1989/2017) in a *concede-and-extend* mode, and reconstructs in §V a three-strata historical correction: Sui Zhiyi (594) → Song Zhili (1021) → Ming Yunqi (1584+). The bridge is not a synthesis; the synthesis comes a millennium later, the subject of the next paper.

**Keywords:** four *samādhis* (*sì-zhǒng sānmèi* 四種三昧); *cháng-zuò sānmèi* (常坐三昧, constantly seated samādhi); *cháng-xíng sānmèi* (常行三昧, constantly walking samādhi / *bānzhōu sānmèi* 般舟三昧); *bàn-xíng-bàn-zuò* (半行半坐, half-walking-half-seated); *fēi-xíng-fēi-zuò* (非行非坐, neither-walking-nor-seated / *jué-yì sānmèi* 覺意三昧 / *suí-zì-yì sānmèi* 隨自意三昧); *wǔ-fāngbiàn niànfó-mén* (五方便念佛門, five gates of expedient buddha-recollection); *chēng-míng wǎng-shēng* (稱名往生, name-recitation rebirth); *guān-xiàng miè-zuì* (觀相滅罪, image-contemplation sin-eradication); *zhū-jìng wéi-xīn* (諸境唯心, all objects are mind-only); *xīn-jìng jù-lí* (心境俱離, mind and object both transcended); *xìng-qǐ yuán-tōng* (性起圓通, nature-arising perfect penetration); classificatory bridge (non-synthesis); posture-anchored typology; three-action concert (*sān-yè qíyùn* 三業齊運); three-strata historical correction (Sui Zhiyi 594 → Song Zhili 1021 → Ming Yunqi 1584+); retrospective compression; doxography; doxographic crystallization; reception-history; perfect-teaching ceiling (圓教); three-truths concurrent contemplation (*sān-dì yuán-róng* 三諦圓融, *kōng-jiǎ-zhōng* 空假中); *yī-xíng sānmèi* (一行三昧, one-practice samādhi); *cháng-háng sānmèi sān-yè qíyùn* (常行三昧三業齊運); pseudo-Zhiyi attribution; attribution caveat; common-ancestor thesis (Chan / Pure Land non-conflation); *sì-zhǒng niànfó* (四種念佛, four kinds of buddha-recollection — late-Ming compression, not Zhiyi); Zhiyi (智顗, 538–597); Guanding (灌頂, 561–632); Zhili (知禮, 960–1028); Yúnqī Zhūhóng (雲棲祩宏, 1535–1615); Yongming Yanshou (永明延壽, 904–975); Zongmi (宗密, 780–841); Bǎoyún Yìtōng (寶雲義通, 927–988); Lúshān Huìyuǎn (廬山慧遠, 334–416); Huìsī (慧思, 515–577); Zhanran (湛然, 711–782); Zhongfeng Mingben (中峰明本, 1263–1323); Hanshan Deqing (憨山德清, 1546–1623); Ouyi Zhixu (蕅益智旭, 1599–1655); Lokakṣema (支婁迦讖, *fl.* 179); T1911 *Móhē zhǐguān* 摩訶止觀; T1962 *Wǔ-fāngbiàn niànfó-mén* 五方便念佛門; T1750 *Guān-wúliángshòu-fó-jīng shū* 觀無量壽佛經疏 (attributed Zhiyi ⚠ disputed); T1751 *Miàozōng-chāo* 妙宗鈔; T0418 *Bānzhōu-sānmèi-jīng* 般舟三昧經; T0232 *Wénshū-prajñāpāramitā-sūtra* 文殊師利所說摩訶般若波羅蜜經; T2017 *Wàn-shàn tóng-guī jí* 萬善同歸集; T2016 *Zōng-jìng-lù* 宗鏡錄; T1753 *Guānjīng sì-tiē shū* 觀經四帖疏 (Shandao authentic, P22 reference); X05n0229 *Huá-yán-jīng xíng-yuàn-pǐn shū-chāo*; *concede-and-extend* mode (vs *concede-and-push-back* of P20–P22 §IV); Stevenson 1986 (*Four Kinds of Samādhi in Early T'ien-t'ai Buddhism*); Donner & Stevenson 1993 (*Great Calming and Contemplation*); Swanson 1989 (*Foundations of T'ien-t'ai Philosophy*); Swanson 2017 (*Clear Serenity, Quiet Insight*); Northern Song Tiantai *shānjiā* 山家 lineage; late-Ming Pure Land; *Wàn xùzàng* 卍續藏 (Qing-era five-hits provenance); P06 cross-reference (T0418 *pratyutpanna* threefold register, seven-day vision threshold); P19 cross-reference (T0232 *yī-xíng sānmèi*); P21 backward reference (Chan inward reading + Zongmi 9c); P22 backward reference (Shandao exclusive name-holding); P24 forward reference (Yongming–Yunqi–Ouyi synthesis); six-century / nine-century / sixteen-seventeen-century historical strata; Tiantai *yuán-jiào* 圓教 ceiling thesis (vs sixfold subordination); Part VI Two Streams.

---

## I. The question

Anyone who reads contemporary Chinese-language introductions to Pure Land Buddhism — whether the popular tracts distributed at temples, the entry-level lecture transcripts, or the modern dictionaries of Buddhist terms — will sooner or later encounter the sentence: *"The Tiantai master Zhiyi established four kinds of buddha-recollection: name-recitation, image-contemplation, visualization, and reality-contemplation."* The sentence is so common in Chinese-language Pure Land exposition that a casual reader might assume it is documented somewhere in Zhiyi's own writings. The Anglophone Buddhological reader, encountering this formula in secondary sources or in Pure Land overviews, is liable to make the same assumption — that the fourfold scheme is a sixth-century Chinese typology, transmitted intact from Zhiyi's lecture hall through the long Tiantai tradition into modern doctrinal literature.

This paper is, at one level, a philological correction of that assumption.

The previous paper in this series, P22 (*The Pure Land Patriarchs II: Shandao on the Establishment of the Name*), closed on a tension: Shandao (善導, 613–681), in selecting name-recitation as the single road by which the ordinary unenlightened being could enter the Recompense Land, produced a doctrinal posture in which the Tiantai's fourfold parallel arrangement of practices appeared, retrospectively, as a kind of doctrinal pluralism — a side-by-side spreading-out of practices where Shandao had placed a single axis. P22 left that tension explicitly open: *what, exactly, is the structure of the Tiantai parallel arrangement against which Shandao's exclusive name-holding is to be measured?*

This paper takes up that tension and immediately complicates the way it is usually framed. The "fourfold parallel arrangement" of *chēng-míng / guān-xiàng / guān-xiǎng / shí-xiàng* — the formula taken as native to Zhiyi by the popular tradition — is, on inspection of the canonical record, not Zhiyi's. It is a late-Ming retrospective compression. What Zhiyi *did* establish, in the late sixth century, was a different kind of structure entirely: a fourfold typology *by bodily posture*, paired with a fivefold typology *of buddha-recollection by contemplative depth*, which together constitute what this paper will call a *classificatory bridge*. The bridge is a vocabulary and an architecture; it is not a synthesis. Zhiyi did not synthesize Chan and Pure Land — neither tradition existed in its later fully-articulated form during his lifetime — but he gave subsequent Chinese authors the conceptual joinery by which the inward turn (later Chan) and the name-holding turn (later Pure Land) could be seen to share a common ancestor in the early Tiantai contemplative system.

The argument unfolds in three steps. First, against the popular formula: the four *samādhis* of *Móhē zhǐguān* (T1911) are classified by bodily posture, not by recollective content. Second, the *Wǔ-fāngbiàn niànfó-mén* (T1962, attributed to Zhiyi but more probably a Tiantai-school post-Zhiyi compilation) offers a *fivefold* typology of *niànfó-sānmèi*, not a fourfold one. Third — and this is the load-bearing historical claim — the popular fourfold scheme is a late-sixteenth-century doxographic compression by Yunqi Zhuhong (雲棲祩宏, 1535–1615) in his *Āmítuó-jīng shūchāo*, which folds Zhili-school's fivefold *niànfó* into a more memorable fourfold for late-Ming and Qing-dynasty Pure Land use. The expression *sì-zhǒng niànfó* (四種念佛) appears zero times in the Taishō canon and only five times in the *Wàn xùzàng* (卍續藏), all from Qing-era texts.

A reader's natural question at this point is: *if not Zhiyi's, then whose, and when?* The paper answers in §V with a three-strata historical correction: Sui Zhiyi in 594 establishes the original fourfold-by-posture and the fivefold-by-depth typologies; Song Zhili (知禮, 960–1028) in his *Miàozōng-chāo* (T1751, 1021) crystallizes the doxographic formula *"the four samādhis are commonly named buddha-recollection"*; Ming Yunqi in 1584+ completes the popular compression into the fourfold *chēng-míng / guān-xiàng / guān-xiǎng / shí-xiàng* that has come down to the present. Three strata, separated by 427 and 563 years respectively, three doxographic acts, each made of and for its own historical moment — none reducible to the others.

The thesis, stated in one line: *Zhiyi's late-sixth-century work (T1911 + T1962) is a classificatory bridge that gives later Chinese authors the vocabulary in which Chan and Pure Land could later be jointly thought; it is not itself a synthesis, and the popular fourfold formula is not its content but its late-Ming retrospective abridgment.* "Four kinds of *samādhi*, methods differing, contemplative principle one" — Zhiyi's hinge sentence at T46.18c11 — is the entire classificatory starting-point this paper undertakes to defend.

The paper proceeds in six sections. §II establishes the textual ground (T1911 authentic, T1962 attributed-Zhiyi-school, T1750 attribution-disputed, T1751 Zhili authentic, T0418 cross-reference). §III develops the core argument in four sub-sections, working from posture-typology, through the three-action concert in *constantly walking samādhi*, through the fivefold *niànfó-sānmèi* of T1962, to Zhili's eleventh-century crystallization. §IV engages with the three foundational Western readings in *concede-and-extend* mode, distinct from this series's previous *concede-and-push-back* engagements with Suzuki and Kamakura-retrojection scholarship. §V states the three-strata historical correction and three philological warnings. §VI closes with a tripartite restatement, a guidance for practice, and an open question forward to P24.

---

## II. The textual ground

Five primary texts and one cross-reference. Three pillars and two ancillaries.

**The first pillar: T1911, *Móhē zhǐguān***. Lectured by Zhiyi in 594 at Yuquan Monastery (玉泉寺) in Jingzhou, recorded by his disciple Guanding (灌頂, 561–632), surviving in Song-canon recensions. CBETA T1911, Taishō vol. 46. This is Zhiyi's *authentic voice* — the canonical core of the present argument. The paper draws six discrete passages from it: the *fourfold opening* at T46.11a23–a26; the *hinge sentence* at T46.18c11; the *three-power citation of the source-text* (the *Bānzhōu-sānmèi-jīng* link) at T46.12a19–a24; the *three-action concert in constantly walking samādhi* at T46.12b18–b25; the *threefold gradation from the Vibhāṣā* (color-and-form → distinguishing dharmas → reality) at T46.13a01–a04; the *constantly-seated treatment of name-recitation as auxiliary* at T46.11b10–b18; and the *neither-walking-nor-seated* passage with its identification of *jué-yì sānmèi* (覺意三昧) and *suí-zì-yì sānmèi* (隨自意三昧) at T46.14b26–b30. [^1]

**The second pillar: T1962, *Wǔ-fāngbiàn niànfó-mén***. One fascicle, attributed to Zhiyi, but in present scholarly judgment more probably a Tiantai-school compilation produced after Zhiyi's death (597) and circulating under his name. CBETA T1962, Taishō vol. 47. The attribution is disputed; the doctrinal compatibility with the *Móhē zhǐguān* system is high. The paper consistently refers to this text as *"the Zhiyi-attributed Wǔ-fāngbiàn niànfó-mén"* and never as *"Zhiyi says in the Wǔ-fāngbiàn niànfó-mén."* Three load-bearing passages: the *fivefold opening of the niànfó-sānmèi gates* at T47.82a09–a12; the *originating-place declaration* — *niànfó-sānmèi as the place from which the great samādhis arise* — at T47.82a18–a19; and the *perfect-teaching closing* — *"contemplating the buddha-recollecting mind: arising is at once empty, conventional, and middle … no buddha, no recollection — this is the perfect-teaching Mahāyāna"* — at T47.83a20–a26. [^2]

**The third pillar (caveated): T1750, *Guān-wúliángshòu-fó-jīng shū***. Attributed to Zhiyi, but with substantially weaker authorial provenance than T1962. CBETA T1750, Taishō vol. 37. The text is first cited in Tiantai literature by Zhili in his *Miàozōng-chāo* (1021), 424 years after Zhiyi's death. The early Tiantai patriarchal records — Guanding (d. 632), Zhiwei (智威, d. 681), Xuanlang (玄朗, 673–754), and Zhanran (湛然, 711–782) — preserve no record of a *Guān-jīng shū* by Zhiyi. The *Miàozōng-chāo* itself opens by stating: *"the recensions which our [text-]master Baoyun first composed"* (所稟寶雲師首製記文, T37.0195a07) — that is, the first transmitter of T1750 was Baoyun Yitong (寶雲義通, 927–988), already in the late tenth century. Modern scholarship generally treats T1750 as a post-Tang Tiantai-school composition placed under Zhiyi's name. The paper's handling: cite T1750 with explicit attribution-flag (⚠️) and never as Zhiyi's voice; the central argument does *not* depend on T1750 as Zhiyi-authentic. T1911 + T1962 alone constitute the most conservative reconstruction of the Zhiyi system. [^3]

**The first ancillary: T1751, *Miàozōng-chāo***. Six fascicles, by Siming Zhili (四明知禮, 960–1028), composed in 1021 (Tianxi 5). CBETA T1751, Taishō vol. 37. This is the doxographic crystallization of the early Northern Song Tiantai *Shanjia* (山家) lineage. Two passages: the *Baoyun first-recension* statement at T37.0195a07; and the *"the four samādhis are commonly named buddha-recollection"* crystallization at T37.0195b01–b08. The latter passage is, this paper argues, the moment at which the Northern Song Tiantai lineage made *explicit* what was implicit in T1911 — the typological inclusion of all four *samādhis* under the umbrella term *niànfó*. The author of this crystallization is *Zhili (1021), not Zhiyi (594)* — a temporal gap of 427 years that becomes load-bearing in §V. [^4]

**The second ancillary: T0418, *Bānzhōu-sānmèi-jīng***. Three fascicles, translated by the Han-dynasty Yuezhi monk Lokakṣema (Lokakṣema) in 179 CE. CBETA T0418, Taishō vol. 13. The direct canonical source-text for Zhiyi's *constantly walking samādhi* (T46.12a19–a24, *"this method comes from the Bānzhōu-sānmèi-jīng, translated as 'standing buddha'"*). The doctrinal substance of T0418 itself — its threefold semantic register of *pratyutpanna*, its seven-day vision threshold, its red-hot iron-ball simile, its closing *"mind is buddha"* turn — has been treated in the earlier paper P06 (*Face-to-Face with the Buddhas*); the present paper invokes T0418 only as cross-reference and does not re-argue. [^5]

T0232, *Wénshū-shīlì suǒshuō Móhē-bōrě-bōluómì-jīng* (the Mañjuśrī-spoken *Mahāprajñāpāramitā-sūtra*), translated by Mandrasena (曼陀羅仙) of the Liang dynasty, is the canonical source for the *yī-xíng-sānmèi* (一行三昧) which Zhiyi places under *constantly seated samādhi*; it has been treated in the earlier paper P19 and is referenced in §III only in passing.

---

## III. The core argument

### 3.1 The fourfold-by-posture typology: *methods differ, contemplative principle is one*

Zhiyi opens fascicle 1B of the *Móhē zhǐguān* with the announcement:

> "Modes of practice are many; in brief, four: first, *constantly seated*; second, *constantly walking*; third, *half-walking-half-seated*; fourth, *neither-walking-nor-seated*. The general name *samādhi* means 'to regulate, to make upright, to settle.'" (T46.11a23–a26) [^6]

The classifying principle here is *bodily posture* — sitting, walking, mixed, neither. It is not the recollective content (name, image, visualization, reality-nature). The same fascicle delivers the hinge sentence on which the entire typology rests:

> "The four kinds of *samādhi* differ in method, but their contemplative principle is one." (T46.18c11) [^7]

*"Methods differ"*: the bodily postures are four, naturally so, since the human body has these four states of activity-and-rest. *"Principle is one"*: the contemplated mind-dharma (the *guān-xīn* 觀心) is single. What Zhiyi is rejecting in this sentence is twofold. First, he rejects a developmental reading — the supposition that *constantly seated* is for beginners, *constantly walking* for intermediate, and so on, arranged as a ladder of ascending practice. Second, he rejects a soteriological-grading reading — the supposition that one of the four yields a higher fruit than the others, with *neither-walking-nor-seated* (the most abstract-sounding) ranked over *constantly seated*. Both readings, common in later popular Tiantai exposition, are foreclosed by the hinge sentence.

The classification is by bodily posture, not by attainment-grade. *Constantly seated* takes its source-text from the *Wénshū bōrě jīng* (T0232) and admits the *yī-xíng-sānmèi* of single-practice. *Constantly walking* takes its source-text from the *Bānzhōu-sānmèi-jīng* (T0418) and admits the *bānzhōu-sānmèi* of buddha-vision. *Half-walking-half-seated* takes its source-texts from the *Mahāvaipulya-dhāraṇī-sūtra* and the *Lotus Sūtra*'s "Chapter on Peaceful and Joyful Conduct" and admits the *fāng-děng* and *fǎ-huá* *samādhis*. *Neither-walking-nor-seated*, even in this most abstract-sounding category, Zhiyi is at pains to clarify: *"in fact pervading walking, sitting, and all activities"* (T46.14b27) — and: *"although the name is threefold, in fact it is one dharma"* (T46.14b30, identifying *neither-walking-nor-seated* with *jué-yì-sānmèi* and *suí-zì-yì-sānmèi*) [^8]. The fourth category is not a withdrawal from the first three but a *catch-all* for all daily activities not subsumed under fixed posture.

This, then, is the classificatory nature of the four *samādhis*: a *posture-anchored typology with canonical-text warrants*. Stevenson's 1986 reading of the four *samādhis* lands precisely here (see §IV), and the present paper's §III.1 takes Stevenson's reading as foundation. The popular fourfold formula *chēng-míng / guān-xiàng / guān-xiǎng / shí-xiàng*, by contrast, has shifted the classifying axis from *posture* to *recollective content* — a shift of axis, not a translation. The historical accounting of that axis-shift is the work of §V.

### 3.2 *Constantly walking samādhi* contains the three-action concert: a common ancestor of two streams

Among the four *samādhis*, *constantly walking* — the *bānzhōu-sānmèi* — is the load-bearing bridge of this paper. In this single category, Zhiyi did not separate the two extremes that later (post-eighth-century) Chan and Pure Land would oppose: he neither extracted name-holding as the sole road, nor refused inward contemplation as a distinct exercise. He set them, instead, in a three-action concert. The decisive passage from fascicle 2:

> "As to mouth — speech and silence: for ninety days, the body is constantly walking, without rest; for ninety days, the mouth constantly chants Amitābha's name, without rest; for ninety days, the mind constantly recollects Amitābha, without rest. Sometimes chanting and recollection move together; sometimes recollection precedes chanting; sometimes chanting precedes recollection; chanting and recollection succeed one another, without a moment of rest. … In summary: step by step, sound by sound, thought by thought — only Amitābha." (T46.12b18–b25) [^9]

*Step by step, sound by sound, thought by thought — only Amitābha*: body, mouth, mind, three actions on a single object, in a single direction. Donner and Stevenson's 1993 careful reading of this passage is decisive (see §IV): what Zhiyi establishes is *concert*, not *alternation*; not a body-then-mouth-then-mind sequence, but a parallel running of all three on the same object. *Body in walking, mouth in chanting, mind in recollecting* — the three actions not separating from Amitābha's name and form for any single moment.

For the historical work this paper is doing, the decisive feature of this passage is the equipoise. *"Mouth chanting"* and *"mind recollecting"* stand here as two of three co-equal actions; neither has been pulled out as the *sole* road, neither has been demoted to auxiliary. For the later differentiation of two streams, this equipoise is the common ancestor:

- Pure Land's later concentration on *kǒu-chēng* (mouth-chanting), institutionalized in the Daochuo-Shandao lineage of the seventh century, is already present in Zhiyi's *constantly walking samādhi* — but **not yet selected as exclusive**. The exclusivization is a historical choice of the seventh-century Pure Land patriarchs (P22), not a doctrinal necessity emerging from the canon.
- Chan's later concentration on the *inward turn* — the rejection of external name-and-form in favor of self-nature — is already present in Zhiyi's *constantly walking samādhi* — but **not yet selected as exclusive**. The inward exclusivization is a historical choice of the Huineng-Shenhui lineage of the eighth century onward (P21), not a doctrinal necessity.

The Chan / Pure Land opposition, in other words, is not a doctrinal necessity but a *result of post-sixth-century field-differentiation*. Zhiyi's *constantly walking samādhi* is itself the common ancestor of both streams — and *this is what this paper means by classificatory bridge*.

The same fascicle 2 then supplies the inward gradation that makes the bridge intelligible from the recollective side:

> "The Vibhāṣā shows: the bodhisattva newly arising in mind first recollects the buddha's color-and-form — its body, its action, its result, its function — and gains lower power; next recollects the buddha's forty unshared dharmas, and the mind gains middle power; next recollects the real buddha, gains the buddha's higher power, and is no longer attached to color-and-form-as-bodies." (T46.13a01–a04) [^10]

Three layers of recollective depth: *color-and-form → distinguishing dharmas → reality*; lower → middle → higher *power* (*shìlì* 勢力). Note: this is a *threefold* progression, not a fourfold. The popular *guān-xiàng / guān-xiǎng / shí-xiàng* fourfold has correspondences here — *image* to *color-and-form*, *visualization* to *distinguishing dharmas*, *reality* to *reality* — but Yunqi's later fourfold reorganization (§V) does not preserve the structure of this Vibhāṣā gradation; rather, it interweaves the *Móhē zhǐguān* threefold with the *Wǔ-fāngbiàn niànfó-mén* fivefold, then collapses both into a fourfold for late-Ming pedagogical use. The historical layering, again, is the work of §V.

### 3.3 The fivefold *niànfó-sānmèi* gates of the *Wǔ-fāngbiàn niànfó-mén*: the perfect-teaching as ceiling

The Zhiyi-attributed *Wǔ-fāngbiàn niànfó-mén* contains the decisive typological move:

> "Stating the opening of the five gates of buddha-recollection: first, the gate of *name-recitation and rebirth* [chēng-míng wǎngshēng]; second, the gate of *contemplation-of-image and removal-of-offense* [guān-xiàng miè-zuì]; third, the gate of *all objects as mind-only* [zhū-jìng wéi-xīn]; fourth, the gate of *mind-and-object both transcended* [xīn-jìng jù-lí]; fifth, the gate of *nature-arising perfect-pervasion* [xìng-qǐ yuán-tōng]." (T47.82a09–a12) [^11]

The internal order of these five gates admits a three-zone reading. *Name-recitation* (gate 1) is name-holding for rebirth in the Pure Land — the practice that Pure Land tradition will later select as exclusive. *Contemplation-of-image* (gate 2) is image-supported visualization for the removal of offense-imprints — the practice that Mañjuśrī-style image-meditation lineages will recognize. *All objects as mind-only*, *mind-and-object both transcended*, and *nature-arising perfect-pervasion* (gates 3, 4, 5) are three layers of inward depth: the world-as-mind reading, the dual-collapse of subject and object, and the nature-arising of the perfect-teaching ocean. These three are an extension and refinement of the Vibhāṣā threefold gradation in T1911 — *color-and-form → distinguishing dharmas → reality* now articulated as five gates with explicit Tiantai *yuán-jiào* (圓教, perfect-teaching) coloration.

T1962 itself states the standing of *niànfó-sānmèi* among the great *samādhis*:

> "This *samādhi* is the playground of the buddhas, the originating place from which the *Śūraṃgama* and the other great *samādhis* arise." (T47.82a18–a19) [^12]

The doctrinal weight of this declaration deserves attention. *Niànfó-sānmèi* is not auxiliary to the great *samādhis* — it is their *originating place*. For later Chan-side polemics that demote *niànfó* to an exoteric exercise for those of inferior capacity, this passage stands as direct counter-evidence.

The fifth gate then receives the perfect-teaching ceiling:

> "Contemplating the buddha-recollecting mind: arising is at once empty, conventional, and middle … no buddha, no recollection — this is the perfect-teaching Mahāyāna." (T47.83a20–a26) [^13]

*Empty, conventional, and middle simultaneously*: the canonical Tiantai *sān-dì yuán-róng* (三諦圓融, threefold-truth perfect-fusion) reading. *No buddha, no recollection*: the contemplative emptying at the apex of the recollection. *Perfect-teaching Mahāyāna*: this is not the *piṭaka*-teaching, not the *common*-teaching, not the *separate*-teaching of *niànfó* — it is the *perfect*-teaching *niànfó*. The interior of the recollection itself contains the full inward depth, from *name-recitation rebirth* (gate 1) at the entry to *nature-arising perfect-pervasion* (gate 5) at the apex; one need not exit *niànfó* into Chan, into Yogācāra, into Huayan, in order to reach the perfect-teaching summit.

This *fivefold structure* — *chēng-míng / guān-xiàng / zhū-jìng wéi-xīn / xīn-jìng jù-lí / xìng-qǐ yuán-tōng* — is the canonical reading of T1962. The popular *fourfold* — *chēng-míng / guān-xiàng / guān-xiǎng / shí-xiàng* — is, this paper insists, a different thing. The relation between them is *compression*, not *translation*. §V tracks the compression historically.

### 3.4 Zhili's *"the four samādhis are commonly named buddha-recollection"*: the Northern Song Tiantai crystallization

Zhiyi himself, in the late sixth century, set out the four *samādhis* (by posture) and the five *niànfó* gates (by recollective depth), but he did not himself produce the explicit doxographic conclusion *"the four samādhis are commonly named buddha-recollection."* That conclusion was produced by **Zhili** in 1021. The decisive passage of *Miàozōng-chāo*:

> "Therefore, of all the practices spoken of in the teachings, none fails to manifest this awakening-substance; therefore the four *samādhis* are commonly named *niànfó*, but their contemplative methods, as gates, differ. As in the *yī-xíng-sānmèi*, directly contemplating the three paths to manifest the original-nature buddha … in these *samādhis*, though the histories-of-events are different, *niànfó* is the same, all together for manifesting the great-awakening-substance." (T37.0195b01–b08) [^14]

Three propositions in one passage: (1) all teachings' practices serve to manifest the awakening-substance; (2) the four *samādhis* (classified by posture) are commonly named *niànfó* — *the typological inclusion is explicit*; (3) although the *histories-of-events* (the bodily activities) differ, *niànfó is the same*, all of them manifesting the single great-awakening-substance.

Note carefully *who* and *when* writes this passage: the author is *Zhili* in *1021*, not Zhiyi in 594, with a temporal gap of 427 years. *"The four samādhis are commonly named niànfó"* is a Northern Song *Shanjia*-Tiantai doxographic crystallization, not Zhiyi's own language in *Móhē zhǐguān*. Popular Pure Land literature occasionally cites this formula directly as *"the master Zhiyi said: the four samādhis are commonly named niànfó"* — a citation that elides the eleventh-century Zhili intermediation. That intermediation is itself precisely the moment at which Northern Song Tiantai gave its doctrinal shape to the inheritance — the historical step of crystallization that the three-strata correction in §V will register.

The argument of §III, in single-line summary:

1. The four *samādhis* (T1911) are classified by *bodily posture*, not by recollective content (3.1).
2. *Constantly walking samādhi* itself contains the three-action concert and is the common ancestor of the two later streams (3.2).
3. The *Wǔ-fāngbiàn niànfó-mén* (T1962) is *fivefold* not fourfold; *niànfó-sānmèi* itself contains the perfect-teaching ceiling (3.3).
4. The doxographic formula *"the four samādhis are commonly named niànfó"* is the work of *Zhili in 1021*, not Zhiyi in 594 (3.4).

Zhiyi gave the *vocabulary* and the *architecture*; subsequent Chinese authors — Zongmi (宗密, 780–841) in the ninth century, Zhili in the eleventh, Yongming Yanshou (永明延壽, 904–975) in the tenth, Yunqi Zhuhong in the sixteenth, Ouyi Zhixu (蕅益智旭, 1599–1655) in the seventeenth — each produced their own historical redeployments of this vocabulary. §IV engages the three foundational Western readings; §V corrects the third historical layer (Yunqi); §VI looks forward to P24's main subject of the millennium-long synthesis.

---

## IV. The Western readings: foundation, not opposition

A structural note before engagement: §IV in this paper operates differently from §IV in the previous three. P20, P21, and P22 each engaged Western Buddhological scholarship in a *concede-and-push-back* mode — Suzuki's Chan-Pure-Land binary, Suzuki's two-truths reading, and Kamakura-retrojection scholarship were each conceded in part and corrected in part. The present paper engages in a *concede-and-extend* mode. Stevenson 1986, Donner & Stevenson 1993, and Swanson 1989/2017 each give a reading of the four *samādhis* on which this paper *agrees*. They are *foundation*, not opposition. The paper's contribution is not to break their readings but to extend them, along their own grain, into a Chinese reception-history that they themselves did not undertake to chase.

**The primary interlocutor: Daniel B. Stevenson, "The Four Kinds of Samādhi in Early T'ien-t'ai Buddhism," in Peter N. Gregory, ed., *Traditions of Meditation in Chinese Buddhism* (Honolulu: University of Hawai'i Press, 1986), pp. 45–97.** Stevenson reads the four *samādhis* as a *posture-anchored typology with canonical-text warrants*: not a developmental ladder, not a soteriological grading, but a structurally parallel arrangement in which each of the four bodily-postural categories carries its own canonical source and its own contemplative deployment. Stevenson establishes that Zhiyi's contribution is not a new buddha-recollection practice — he inherits the *bānzhōu-sānmèi* lineage from Huisi (慧思, 515–577) and Lushan Huiyuan (廬山慧遠, 334–416) — but the placement of this inherited line within a fourfold posture-anchored architecture. The argument of §III.1 above takes Stevenson's reading as its foundation.

This paper's *extension* of Stevenson: the classificatory architecture *itself* — independent of any of its four cells — becomes the *vocabulary* by which subsequent Chinese authors would later articulate Chan-Pure-Land plurality and synthesis. Zongmi's ninth-century *four kinds of buddha-recollection* in his *Huá-yán-jīng xíng-yuàn-pǐn shū-chāo* (X05n0229), which the previous paper P21 §3.3 has already established, is the first large-scale redeployment of this architecture — Zongmi shifts the classifying axis from *bodily posture* (Zhiyi) to *contemplative-depth gradient* (Zongmi himself). Yongming Yanshou's tenth-century *Wàn-shàn tóng-guī jí* (T2017) takes the architecture in a different direction, toward myriad-good convergence and mind-only Pure Land. Yunqi's late-sixteenth-century *Āmítuó-jīng shūchāo* takes it in yet a third direction, toward the *shì-chí / lǐ-chí* (事持／理持, phenomenal-holding / principle-holding) distinction. The latter three are P24's main subject; this paper does not develop them. Stevenson's reading of Zhiyi is correct on its own terms; what Stevenson does not undertake — the millennium-long Chinese reception-history of the architecture — is what this paper extends along the line of his reading.

**The secondary interlocutor: Neal Donner and Daniel B. Stevenson, *The Great Calming and Contemplation: A Study and Annotated Translation of the First Chapter of Chih-i's Mo-ho chih-kuan* (Honolulu: University of Hawai'i Press, 1993).** The contribution is threefold: (a) the Guanding-redaction stratum of T1911 is mapped textually, identifying which sections bear marks of editorial layering and which present Zhiyi's lecture more directly; (b) the four *samādhis* are recognized as a *second-order pedagogical framework* — Zhiyi's instructional architecture for organizing the practices, rather than a first-order contemplative method itself; (c) the three-action structure of *constantly walking samādhi* is read decisively as *concert* and not as *alternation* — body, mouth, and mind do not take turns; they run together on a single object. The argument of §III.2 above relies on Donner and Stevenson's three-action concert reading as its principal citation-anchor.

This paper's *extension* of Donner and Stevenson: the *concert* structure, established by them at the textual level, becomes — when read with the historical eye toward what Chan and Pure Land would later become — the *common ancestor* of the two streams. Donner and Stevenson do not state this historical-bridge thesis in their 1993 work; the present paper draws it as a corollary of the concert reading they have already established.

**The tertiary interlocutor: Paul L. Swanson, *Foundations of T'ien-t'ai Philosophy: The Flowering of the Two Truths Theory in Chinese Buddhism* (Berkeley: Asian Humanities Press, 1989); and *Clear Serenity, Quiet Insight: T'ien-t'ai Chih-i's Mo-ho chih-kuan*, three-volume translation (Tokyo and Honolulu: BDK America, 2017).** Swanson establishes the three-truths perfect-fusion (*sān-dì yuán-róng*) as the *doctrinal substrate* of the four *samādhis* typology — the four-cell architecture is not arbitrary, not a mere classificatory convenience; it follows naturally from the simultaneous contemplation of empty, conventional, and middle truths in the perfect-teaching root. Stevenson tells us *what* the structure is; Swanson tells us *why* the structure is doctrinally necessary.

This paper's *extension* of Swanson: the perfect-teaching closing of T1962 — *"contemplating the buddha-recollecting mind: arising is at once empty, conventional, and middle … no buddha, no recollection — this is the perfect-teaching Mahāyāna"* (T47.83a20–a26) — is the application of Swanson's three-truths substrate to the *niànfó* practice specifically. The perfect-teaching positioning of *niànfó* in T1962 is direct counter-evidence to the later Chan-Platform-Sūtra-line view of *niànfó* as a practice for those of inferior capacity. Swanson establishes the three-truths substrate at the level of contemplation in general; the present paper extends his reading by applying the substrate specifically to the *niànfó-sānmèi* doctrine of T1962.

The three readings, summarized: Stevenson establishes *parallel structure*; Donner and Stevenson establish *concert precision*; Swanson establishes *perfect-teaching substrate*. On their three foundations, this paper extends the inquiry to the millennium-long Chinese reception-history — Zongmi (9c) → Zhili (11c) → Yongming (10c) → Yunqi (16c). The original readings of the three Western interlocutors are conceded entirely; the Chinese reception-history they do not chase is what this paper extends in the direction of their lines.

(A philological note: the three interlocutors are referenced here by published title and year, with paraphrase rather than verbatim quotation. Verbatim quotations with page numbers, where they would substantively reinforce the argument, are reserved for forward-only refinement after the paper's commit; the present argument does not depend on them.)

---

## V. The middle-way correction

This section establishes the three-strata historical correction and then states three philological warnings.

### Three strata: Sui Zhiyi (594) → Song Zhili (1021) → Ming Yunqi (1584+)

The popular fourfold formula *"Tiantai's four kinds of buddha-recollection: name-recitation, image-contemplation, visualization, reality-contemplation"* must be parsed into at least three distinct historical strata before its content can be seen clearly. A single attribution to *"Tiantai"* obscures rather than clarifies.

**First stratum: Sui Zhiyi (594) gives the original typologies.** Zhiyi establishes in *Móhē zhǐguān* the fourfold *samādhi*-by-posture (T46.11a23–a26) and the hinge sentence *"methods differ, contemplative principle is one"* (T46.18c11). The Zhiyi-attributed *Wǔ-fāngbiàn niànfó-mén* (T1962) sets out the fivefold *niànfó-sānmèi* gates by contemplative depth (T47.82a09–a12). The historical moment is *late sixth century*, the textual evidence is T1911 and T1962 (the latter with attribution-to-Tiantai-school qualified). What Zhiyi gives is *vocabulary and architecture* — fourfold-by-posture and fivefold-by-depth. He does not himself perform a synthesis; he does not himself articulate the formula *chēng-míng / guān-xiàng / guān-xiǎng / shí-xiàng*.

**Second stratum: Song Zhili (1021) crystallizes the Northern Song doxographic formula.** Northern Song Siming Zhili (960–1028), in his *Miàozōng-chāo*, explicitly states *"the four samādhis are commonly named niànfó"* (T37.0195b01–b08). This is the textual moment at which the four-by-posture typology is *named* under the umbrella *niànfó*. The author is Zhili (1021), *not Zhiyi* — separated by 427 years. What Zhili adds is the *doxographic naming* of an inheritance whose architecture Zhiyi gave but whose typological-inclusion-under-niànfó Zhiyi did not himself state. The historical moment is *early Northern Song* (early eleventh century), in the doctrinal crystallization of the *Shanjia* lineage of Northern Song Tiantai.

**Third stratum: Ming Yunqi Zhuhong (1584+) compresses the popular fourfold.** Late-Ming Yunqi Zhuhong (1535–1615), in his *Āmítuó-jīng shūchāo*, takes T1962's fivefold *niànfó-sānmèi* and retrospectively compresses it into the popular fourfold *chēng-míng / guān-xiàng / guān-xiǎng / shí-xiàng*. The compression mechanism: *all-objects-as-mind-only* (gate 3) is folded into *visualization*; *mind-and-object both transcended* (gate 4) and *nature-arising perfect-pervasion* (gate 5) are folded into *reality-contemplation*; *name-recitation rebirth* (gate 1) and *contemplation-of-image* (gate 2, with image renamed to *guān-xiàng* as visual image) are preserved as the first two of the new fourfold. The compound expression *sì-zhǒng niànfó* (四種念佛) appears zero times in the Taishō canon and only five times in the *Wàn xùzàng*, all of those Qing-era — that is, the popular formula entered general circulation only in the post-Yunqi Pure Land instructional tradition. The historical moment is *late Ming to early Qing* (late sixteenth to seventeenth century), separated from Zhiyi by 990 years and from Zhili by 563 years. The mechanism of Yunqi's compression and its further use in the Yongming-Yunqi-Ouyi millennium-long synthesis is the main subject of P24; the present paper records only the historical layering.

Three strata clearly distinguished, what now appears is that the popular formula *"Master Zhiyi established the four kinds of buddha-recollection: name-recitation, image-contemplation, visualization, reality-contemplation"* is a *retrospective projection* of stratum three (Yunqi, late Ming) onto stratum one (Zhiyi, Sui). The stratum-three fourfold compression is read backward into Zhiyi's name. This retrospective reading, when paired with the previously-established line of the previous paper P21 §3.3 — *the sixth-century Zhiyi typology and the ninth-century Zongmi typology must not be conflated* — yields a fully three-layer historical warning:

- 6th century (Zhiyi: posture-typology) ≠ 9th century (Zongmi: contemplative-depth-gradient)
- 9th century (Zongmi: contemplative-depth-gradient) ≠ 16th–17th century (Yunqi: fourfold compression)
- 16th–17th century (Yunqi: fourfold compression) ≠ 6th century (Zhiyi: posture-typology) — the new third-layer warning of this paper

Three strata clear: now one can ask of any cited *"Tiantai's four kinds of niànfó"* sentence, *which stratum is being cited*?

### Three philological warnings

**Warning one: the popular fourfold formula is not Zhiyi's authored language.** As shown above, the *chēng-míng / guān-xiàng / guān-xiǎng / shí-xiàng* fourfold is a Yunqi late-Ming retrospective compression, not Sui-Tang Tiantai vocabulary. Any sentence beginning *"Master Zhiyi said …"* and continuing into the fourfold formula is, as a matter of philological fact, a stratum-three sentence read into a stratum-one mouth. §I has stated this; §V deepens the historical accounting.

**Warning two: T1750's attribution is disputed and must be cited as attributed-Zhiyi, never as Zhiyi-authentic.** As noted in §II, T1750 is first cited only in 1021 by Zhili's *Miàozōng-chāo*, more than four centuries after Zhiyi's death; the early Tiantai patriarchal records (Guanding, Zhiwei, Xuanlang, Zhanran) preserve no record of a *Guān-jīng shū* by Zhiyi. Modern scholarship generally treats T1750 as a post-Tang Tiantai-school composition placed under Zhiyi's name. The handling adopted in this paper: T1750 is cited with attribution-flag (⚠️), its readings are not used as Zhiyi-authentic, but the text remains a usable mid-stage source for mature Tiantai-school thought — *attributed and usable, but not load-bearing*. The middle path between two extremes: not the over-rejection of *"T1750 is forgery, wholly unreliable"*, and not the credulous reading of *"T1750 is Zhiyi-authentic."* The paper's argument does not depend on T1750.

There is a related sub-warning. T1750 (attributed Zhiyi) and Shandao's *Guān-jīng sì-tiē shū* (T1753, the main subject of P22) are two commentaries on the same sūtra (the *Guān-wúliángshòu-fó-jīng*). Popular Pure Land literature occasionally pairs them as *"both Tiantai and Pure Land read the Guān-jīng"* — but the historical positioning differs. Shandao's T1753 is a foundational seventh-century Pure Land work; T1750 is an attributed-to-Zhiyi Tiantai-school post-Tang text. The two are not at the same historical-authorial level. One sentence flagging the difference suffices; this paper does not develop further.

**Warning three: the ninety-day duration is Zhiyi's institutionalization, not the Bānzhōu-sānmèi-jīng's original prescription.** A common claim — *"the ninety-day buddha-recollecting circumambulation is from the Bānzhōu-sānmèi-jīng"* — also has a historical-stratum issue. The *Bānzhōu-sānmèi-jīng*'s *Practice Chapter* (T0418) gives the duration as *seven days and seven nights* — the entry-level threshold for dream-vision of the buddhas (this is established in the earlier paper P06). The *ninety days* in T1911 are Zhiyi's *integer-amplification* of the seven-day base — a transposition from the *occasional-vision event* mode of T0418 to the *long-term sustained discipline* mode of Tiantai institutional practice. This transposition is Zhiyi's contribution to the institutionalization of *bānzhōu*, not the canonical original of T0418. One sentence noting this in §II suffices.

The three warnings, taken together, clarify rather than weaken the central argument: Zhiyi's classificatory bridge thesis must rest on a *philologically clean* base. Zhiyi's *authentic* bridge (T1911) + Zhiyi-school's *systemic* bridge (T1962) + Zhili's *crystallization* (T1751) + Yunqi's *popular compression* (P24's main subject) — four strata clear, the paper's argument is, accordingly, *more* exact, not less.

---

## VI. Closing

### **Essence restated**

The thesis: Zhiyi (538–597), in the late sixth century, established with the *Móhē zhǐguān*'s four *samādhis* (T1911) and the attributed *Wǔ-fāngbiàn niànfó-mén*'s five *niànfó-sānmèi* gates (T1962) a *classificatory bridge* — a vocabulary and architecture in which the inward turn (later Chan, treated in P21) and the name-holding turn (later Pure Land, treated in P22) could each find a non-hierarchical seat within a single contemplative matrix. Two streams' divergence is not a doctrinal necessity; it is the result of post-sixth-century field-differentiation, and Zhiyi's *constantly walking samādhi* is itself the common ancestor of both.

What Zhiyi gave is *vocabulary and architecture* — fourfold-by-posture (the four *samādhis*) and fivefold-by-depth (the *niànfó-sānmèi* gates). He did not himself perform a synthesis. The millennium-long Chinese reception-history — Zongmi (9c) → Zhili (11c) → Yongming (10c) → Yunqi (16c) → Ouyi (17c) — performed successive redeployments of this vocabulary, each in its own historical moment and for its own doctrinal purposes. The synthesis comes a millennium after Zhiyi; this paper has not crossed that distance.

The second thesis: the popular Tiantai *"four kinds of buddha-recollection"* (*chēng-míng / guān-xiàng / guān-xiǎng / shí-xiàng*) is not Zhiyi's authored language, but a *late-Ming retrospective compression* by Yunqi Zhuhong (1535–1615) in his *Āmítuó-jīng shūchāo* (1584+) — a compression of T1962's fivefold *niànfó-sānmèi* into a more pedagogically memorable fourfold for late-Ming Pure Land use. The three historical strata, distinguished, read: Sui Zhiyi (594) → Song Zhili (1021) → Ming Yunqi (1584+) — three doxographic acts, each made of and for its own moment, none reducible to the others.

*Methods differ, contemplative principle is one* — Zhiyi's hinge sentence in *Móhē zhǐguān* fascicle 1B — is, this paper has argued, the most accommodating classificatory utterance that Chinese Buddhism produced before the post-sixth-century differentiation of Chan and Pure Land. The bridge being established, the two streams could each find their own road from it; the bridge having been mistaken for the synthesis, the long doxographic re-articulation by Zongmi, Zhili, Yongming, Yunqi, and Ouyi is precisely the millennium of work that this mistake-and-correction made necessary.

### **Cultivation guidance**

For the contemporary practitioner of *niànfó*, the historical-philological argument of this paper has three practical implications.

*First*: the practitioner's choice of *niànfó* mode does not require, as background warrant, the popular formula *"Master Zhiyi established the four kinds of buddha-recollection."* Zhiyi's classifying axis is *bodily posture*, not recollective content; his expression of *constantly walking samādhi* — *step by step, sound by sound, thought by thought, only Amitābha* — neither extracts mouth-chanting as the sole road nor demotes mind-recollection to a lesser exercise. The practitioner, in *niànfó* practice, finds three actions running in concert: *body in walking, mouth in chanting, mind in recollecting*, all three on the single object of Amitābha's name and form, not separating from it for a moment. This is the original Tiantai posture, prior to the later differentiations.

*Second*: the practitioner choosing exclusive name-holding (the road of the seventh-century Pure Land patriarchs) or choosing inward contemplation (the road of the eighth-century Chan masters) should know that, in the late-sixth-century original typology, *these two were not opposed*. The opposition is the historical structure of post-eighth-century field-differentiation — Daochuo-Shandao versus Huineng-Shenhui and after. Returned to Zhiyi's original typology, the two streams stand in their common-ancestor configuration. The practitioner may choose exclusive name-holding (as Shandao established), or three-action concert (as Zhiyi established), or inward contemplation (as Chan established) — each is a legitimate redeployment of the original typology, none of them is the sole definition of *"orthodox niànfó."*

*Third*: when the practitioner reads modern popular Pure Land introductions and encounters the formula *"Master Zhiyi established name-recitation, image-contemplation, visualization, and reality-contemplation as the four kinds of buddha-recollection,"* the practitioner may now know it for what it is — a popular post-Yunqi Qing-era inheritance of the late-Ming compression, not Zhiyi's own voice. This historical-stratum awareness is not a weakening of the practitioner's confidence; it is its grounding on a clear historical floor. To learn Yunqi's fourfold *niànfó* is to learn the late-Ming Lianchi master's teaching; to learn Zhiyi's four *samādhis* is to learn the late-sixth-century original Tiantai; to learn Zhili's *"the four samādhis are commonly named niànfó"* is to learn the Northern Song *Shanjia* crystallization. All three are legitimate transmissions; each must be returned to its own historical seat.

### **Open question**

The paper closes with the bridge established but not yet crossed.

Zhiyi gave, in the late sixth century, the classificatory vocabulary; how subsequent Chinese authors performed the long synthesis on that vocabulary — Yongming Yanshou (904–975) in the *Wàn-shàn tóng-guī jí* establishing the convergence of myriad-good practices with mind-only Pure Land; Yunqi Zhuhong (1535–1615) in the *Āmítuó-jīng shūchāo* establishing the *shì-chí / lǐ-chí* (phenomenal-holding / principle-holding) distinction; Ouyi Zhixu (1599–1655) in the *Āmítuó-jīng yāo-jiě* establishing the threefold-resource *xìn / yuàn / xíng* (faith, vow, practice) perfect-fusion — this millennium-long synthesis is the main subject of the next paper, P24, and is not developed here.

The open question of this paper, then, in single-line form: *Zhiyi gave the classificatory vocabulary; how did the later Yongming-Yunqi-Ouyi line perform the synthesis on that vocabulary?* The redeployments of the architecture and the perfect-teaching substrate, between the classificatory vocabulary and the doxographic synthesis — this is the problem-field of P24.

> *Contemplating the buddha-recollecting mind: arising is at once empty, conventional, and middle … no buddha, no recollection — this is the perfect-teaching Mahāyāna.* (T47.83a20–a26, *Wǔ-fāngbiàn niànfó-mén*, attributed Zhiyi)

This perfect-teaching positioning of *niànfó*, in the late sixth century, is the canonical ceiling of the bridge. From this ceiling to the millennium-long Yongming-Yunqi-Ouyi synthesis — every redeployment of the vocabulary along the way is a road-stone in a synthesis still to come. The classificatory bridge of this paper is now established; the next paper will begin from the bridge's far end.

---

## Notes

[^1]: T1911 *Móhē zhǐguān* in twenty fascicles (some recensions ten), lectured by Zhiyi at Yuquan Monastery in Jingzhou in 594, recorded by Guanding (561–632); the earliest complete extant version is preserved in the Song canon. CBETA Taishō vol. 46, no. 1911. The paper draws on six discrete passages: T46.11a23–a26 (fourfold opening); T46.18c11 (hinge sentence); T46.12a19–a24 (three powers and source-text citation); T46.12b18–b25 (three-action concert); T46.13a01–a04 (Vibhāṣā threefold gradation); T46.11b10–b18 (constantly-seated *samādhi* on auxiliary name-recitation); T46.14b26–b30 (neither-walking-nor-seated and the three names *jué-yì-sānmèi* and *suí-zì-yì-sānmèi*).

[^2]: T1962 *Wǔ-fāngbiàn niànfó-mén* in one fascicle, attributed to Zhiyi but with present scholarly judgment placing the text as a Tiantai-school post-Zhiyi compilation circulating under Zhiyi's name. CBETA Taishō vol. 47, no. 1962. The doctrinal compatibility with the *Móhē zhǐguān* system is high. The paper consistently cites the text as *"the Zhiyi-attributed Wǔ-fāngbiàn niànfó-mén"* and never as *"Zhiyi says in the Wǔ-fāngbiàn niànfó-mén."* Three load-bearing passages: T47.82a09–a12 (fivefold opening of the *niànfó-sānmèi* gates); T47.82a18–a19 (originating-place declaration); T47.83a20–a26 (perfect-teaching closing).

[^3]: T1750 *Guān-wúliángshòu-fó-jīng shū* in one fascicle, attributed to Zhiyi but with substantially weaker authorial provenance than T1962. CBETA Taishō vol. 37, no. 1750. The text is first cited in Tiantai literature by Zhili in the *Miàozōng-chāo* of 1021, 424 years after Zhiyi's death; the early Tiantai patriarchal records (Guanding d. 632, Zhiwei d. 681, Xuanlang 673–754, Zhanran 711–782) preserve no record of a *Guān-jīng shū* by Zhiyi. The *Miàozōng-chāo* states explicitly that the recensions were first composed by Baoyun Yitong (寶雲義通, 927–988), the first known transmitter, already in the late tenth century. Modern scholarship generally treats T1750 as a post-Tang Tiantai-school composition placed under Zhiyi's name. The paper's handling: cite with explicit attribution-flag (⚠️), do not use as Zhiyi-authentic, but admit as a usable mid-stage source for mature Tiantai-school thought; the central argument does not depend on T1750.

[^4]: T1751 *Miàozōng-chāo* (*Guān-wúliángshòu-fó-jīng shū miàozōng-chāo*) in six fascicles, by Siming Zhili (960–1028) in 1021 (Tianxi 5). CBETA Taishō vol. 37, no. 1751. The paper draws on two passages: T37.0195a07 (the Baoyun first-recension statement) and T37.0195b01–b08 (the doxographic crystallization *"the four samādhis are commonly named niànfó"*).

[^5]: T0418 *Bānzhōu-sānmèi-jīng* in three fascicles, translated by the Han Yuezhi monk Lokakṣema (Lokakṣema) in 179 CE (Guanghe 2). CBETA Taishō vol. 13, no. 418. The paper invokes the text only as cross-reference; the substantive doctrinal treatment of T0418 (the threefold semantic register of *pratyutpanna*, the seven-day vision threshold, the red-hot iron-ball simile, the *"mind is buddha"* turn) is established in the earlier paper P06 (*Face-to-Face with the Buddhas*).

[^6]: *Móhē zhǐguān*, fascicle 1B, T46.11a23–a26: *"Modes of practice are many; in brief, four: first, constantly seated; second, constantly walking; third, half-walking-half-seated; fourth, neither-walking-nor-seated. The general name 'samādhi' means 'to regulate, to make upright, to settle.'"*

[^7]: *Móhē zhǐguān*, fascicle 1B, T46.18c11: *"The four kinds of samādhi differ in method, but their contemplative principle is one. As to the methods of the three practices …"* The hinge sentence of the paper, also serving as §I epigraph.

[^8]: *Móhē zhǐguān*, fascicle 2, T46.14b26–b30: *"Fourth, the neither-walking-nor-seated samādhi: the previous categories used walking and sitting in a single direction; this category is different from the previous, established to complete the fourfold scheme, hence named 'neither-walking-nor-seated'; in fact pervading walking, sitting, and all activities. The Master from the Southern Peak [Huisi] called it 'self-arising' (suí-zì-yì): wherever the mind arises, samādhi is cultivated. The Mahāprajñāpāramitā-sūtra calls it jué-yì-sānmèi: the mind's tendencies are all clearly known. Although the name is threefold, in fact it is one dharma."*

[^9]: *Móhē zhǐguān*, fascicle 2, T46.12b18–b25: *"As to mouth — speech and silence: for ninety days, the body is constantly walking, without rest; for ninety days, the mouth constantly chants Amitābha's name, without rest; for ninety days, the mind constantly recollects Amitābha, without rest. Sometimes chanting and recollection move together; sometimes recollection precedes chanting; sometimes chanting precedes recollection; chanting and recollection succeed one another, without a moment of rest. … In summary: step by step, sound by sound, thought by thought — only Amitābha."*

[^10]: *Móhē zhǐguān*, fascicle 2, T46.13a01–a04: *"The Vibhāṣā shows: the bodhisattva newly arising in mind first recollects the buddha's color-and-form — its body, its action, its result, its function — and gains lower power; next recollects the buddha's forty unshared dharmas, and the mind gains middle power; next recollects the real buddha, gains the buddha's higher power, and is no longer attached to color-and-form-as-bodies."*

[^11]: T1962 *Wǔ-fāngbiàn niànfó-mén*, T47.82a09–a12: *"Stating the opening of the five gates of buddha-recollection: first, the gate of name-recitation and rebirth; second, the gate of contemplation-of-image and removal-of-offense; third, the gate of all objects as mind-only; fourth, the gate of mind-and-object both transcended; fifth, the gate of nature-arising perfect-pervasion."*

[^12]: T1962 *Wǔ-fāngbiàn niànfó-mén*, T47.82a18–a19: *"This samādhi is the playground of the buddhas, the originating place from which the Śūraṃgama and the other great samādhis arise."*

[^13]: T1962 *Wǔ-fāngbiàn niànfó-mén*, T47.83a20–a26: *"Contemplating the buddha-recollecting mind: arising is at once empty, conventional, and middle; whether faculty or object, all is the dharma-realm arising … no buddha, no recollection — this is the perfect-teaching Mahāyāna."*

[^14]: T1751 *Miàozōng-chāo*, T37.0195b01–b08: *"Therefore, of all the practices spoken of in the teachings, none fails to manifest this awakening-substance; therefore the four samādhis are commonly named niànfó, but their contemplative methods, as gates, differ. As in the yī-xíng-sānmèi, directly contemplating the three paths to manifest the original-nature buddha … in these samādhis, though the histories-of-events are different, niànfó is the same, all together for manifesting the great-awakening-substance."* The author of this passage is **Northern Song Zhili (1021)**, not Zhiyi (594).

---

## References

### I. Primary Buddhist Sources

- *Móhē zhǐguān* 摩訶止觀. Lectured by Zhiyi 智顗 (538–597), recorded by Guanding 灌頂 (594). CBETA T1911 (Taishō vol. 46).
- *Wǔ-fāngbiàn niànfó-mén* 五方便念佛門. Attributed Zhiyi 智顗 (Tiantai-school post-Zhiyi compilation). CBETA T1962 (Taishō vol. 47).
- *Guān-wúliángshòu-fó-jīng shū* 觀無量壽佛經疏. Attributed Zhiyi 智顗 ⚠️ disputed (first cited 1021 by Zhili). CBETA T1750 (Taishō vol. 37).
- *Guān-wúliángshòu-fó-jīng shū miàozōng-chāo* 觀無量壽佛經疏妙宗鈔. By Siming Zhili 四明知禮 (960–1028), 1021. CBETA T1751 (Taishō vol. 37).
- *Bānzhōu-sānmèi-jīng* 般舟三昧經. Translated by the Han Yuezhi monk Lokakṣema 支婁迦讖, 179 CE. CBETA T0418 (Taishō vol. 13).
- *Wénshū-shīlì suǒshuō Móhē-bōrě-bōluómì-jīng* 文殊師利所說摩訶般若波羅蜜經. Translated by Mandrasena 曼陀羅仙 of the Liang dynasty. CBETA T0232 (Taishō vol. 8).
- *Huá-yán-jīng xíng-yuàn-pǐn shū-chāo* 華嚴經行願品疏鈔. Commentary by Tang Chengguan 澄觀, sub-commentary by Zongmi 宗密. CBETA X05n0229 (*Wàn xùzàng* vol. 5).

### II. Patriarchal Commentaries

- Zhiyi (智顗, 538–597, spoken), Guanding (灌頂, 594–632, recorded). *Guān wúliàng shòu fó jīng shū* 觀無量壽佛經疏 (*Commentary on the Sūtra on the Contemplation of Amitāyus*), spoken by Zhiyi. Taishō vol. 37, no. T1750. **Primary source** of this paper (§IV) — cited in its capacity as an intra-Buddhist commentarial lens on the *Contemplation Sūtra*. First cited 1021 by Zhili; authorship attributed to Zhiyi is **disputed** in scholarship.
- Zhili (知禮, 960–1028). *Guān wúliàng shòu fó jīng shū miàozōng-chāo* 觀無量壽佛經疏妙宗鈔 (*Sub-Commentary on the Miaozong Tradition of the Commentary on the Sūtra on the Contemplation of Amitāyus*), 10 fascicles, Taishō vol. 37, no. T1751. Northern Song, 1021 CE. Deepening of Zhiyi's commentary from the Song-dynasty Tiantai perspective.
- Zhiyi (智顗, spoken), Guanding (recorded). *Móhē zhǐguān* 摩訶止觀 (*Mahāyāna Calming-and-Contemplation*), 30 fascicles, Taishō vol. 46, no. T1911. The doctrinal framework of the four *samādhi*s.
- Zongmi (宗密). *Huáyán-jīng xíng-yuàn pǐn shūchāo* 華嚴經行願品疏鈔 (*Commentary on the Practices and Vows Chapter Commentary*), Chengguan's *Separate Commentary* with Zongmi's *Sub-commentary*, Wànxù cáng X05n0229. Cross-reference to P25.

### III. Modern Scholarship

- Donner, Neal, and Daniel B. Stevenson. *The Great Calming and Contemplation: A Study and Annotated Translation of the First Chapter of Chih-i's Mo-ho chih-kuan*. Honolulu: University of Hawai'i Press, 1993.
- Stevenson, Daniel B. "The Four Kinds of Samādhi in Early T'ien-t'ai Buddhism." In Peter N. Gregory, ed., *Traditions of Meditation in Chinese Buddhism*, 45–97. Honolulu: University of Hawai'i Press, 1986.
- Swanson, Paul L. *Foundations of T'ien-t'ai Philosophy: The Flowering of the Two Truths Theory in Chinese Buddhism*. Berkeley: Asian Humanities Press, 1989.
- Swanson, Paul L., trans. *Clear Serenity, Quiet Insight: T'ien-t'ai Chih-i's Mo-ho chih-kuan*. 3 vols. Tokyo and Honolulu: BDK America, 2017.

---

### IV. Sūtra & Śāstra Index

| Abbr. | Full Title | Translator / Author | Date |
|---|---|--|------|
| T1911 | *Móhē zhǐguān* 摩訶止觀 (*Mahāyāna Calming-and-Contemplation*) | Zhiyi (spoken) / Guanding (recorded) | Sui Kaihuang 13–18, 593–598 CE |
| T1962 | *Wǔ fāngbiàn niànfó mén* 五方便念佛門 (*Five Expedient Gates of Buddha-Recollection*) | attributed to Zhiyi (authorship disputed) | crystallized between Sui and Tang |
| T1750 | *Guān wúliàng shòu fó jīng shū* 觀無量壽佛經疏 (*Commentary on the Sūtra on the Contemplation of Amitāyus*) | attributed to Zhiyi ⚠️ disputed (first cited 1021) | Sui (attribution) / Northern Song (first citation) |
| T1751 | *Guān wúliàng shòu fó jīng shū miàozōng-chāo* 觀無量壽佛經疏妙宗鈔 | Siming Zhili 四明知禮 (960–1028) | Northern Song, 1021 CE |
| T0418 | *Bānzhōu sānmèi jīng* 般舟三昧經 (*Pratyutpanna-samādhi Sūtra*) | Lokakṣema of Later Han (後漢·支婁迦讖) | 179 CE |
| T0232 | *Wénshū-shīlì suǒshuō Móhē-bōrě-bōluómì jīng* 文殊師利所說摩訶般若波羅蜜經 (*Mañjuśrī Prajñāpāramitā Sūtra*) | Mandrasena 曼陀羅仙 of Liang dynasty | Liang |
| X05n0229 | *Huá-yán-jīng xíng-yuàn-pǐn shū-chāo* 華嚴經行願品疏鈔 (*Commentary on the Vows of Samantabhadra*) | Chengguan 澄觀 (commentary) / Zongmi 宗密 (sub-commentary) | Tang |

---

### V. Cross-Reference to Series Papers

- **P06 (*Face-to-Face with the Buddhas*)**: *Bānzhōu-sānmèi-jīng* (T0418) on the threefold semantic register of *pratyutpanna*, the seven-day vision threshold, the red-hot iron-ball simile, and the *"mind is buddha"* closing turn. The present paper invokes T0418 only as cross-reference in §II and §III.2, and does not re-argue T0418's own doctrinal substance.
- **P19 (*The Mañjuśrī Prajñāpāramitā and the One-Practice Samādhi*)**: T0232 *Wénshū-bōrě-jīng* and the *yī-xíng-sānmèi* of Mañjuśrī's devotion. The present paper's §III.1 invokes T0232 only in passing as the canonical source for *constantly seated samādhi*; the doctrinal substance of *yī-xíng-sānmèi* is not re-developed.
- **P21 (*The Inward Reading of Patriarchal Chan*)**: *zì-xìng Mítuó* (self-nature Amitābha), *niànfó shì shéi* (who is reciting?), the *Liù-zǔ tán-jīng*'s *Yí-wèn-pǐn* (Chapter on Doubts), and Zongmi's (780–841) four kinds of *niànfó* in X05n0229. The present paper §I echoes the *inward stream* line; §III.2 establishes the common-ancestor thesis; §IV.1 references Zongmi (developed in P24); §V establishes the 6th–9th-century non-conflation while adding the new third stratum at the 16th–17th century.
- **P22 (*The Pure Land Patriarchs II: Shandao on the Establishment of the Name*)**: Shandao (613–681), the entry of the ordinary unenlightened being into the Recompense Land, the threefold mind, the exclusive name-holding, the *Guān-jīng sì-tiē shū* (T1753). The present paper §I directly continues the doctrinal-pluralism tension established at the close of P22; §V Warning Two cross-references T1750 (attributed Zhiyi) versus T1753 (Shandao authentic) as two commentaries on the same sūtra at different historical-authorial levels, but does not re-develop.
- **P24 (*The Yongming–Yunqi–Ouyi Synthesis*, working title)**: Yongming Yanshou (904–975), *Wàn-shàn tóng-guī jí* (T2017) and *Zōng-jìng-lù* (T2016); Yunqi Zhuhong (1535–1615), *Āmítuó-jīng shūchāo* and the *shì-chí / lǐ-chí* distinction; Zhongfeng Mingben (1263–1323), *Sān-shí xì-niàn*; Hanshan Deqing (1546–1623); Ouyi Zhixu (1599–1655), *Āmítuó-jīng yāo-jiě* — the late-Ming Chan/Pure-Land/Tiantai synthesis. The present paper's §III.3 perfect-teaching ceiling, §IV's extension of Stevenson's classificatory architecture, §V's third stratum on Yunqi's compression, and §VI's open question all point forward to P24. *Zhiyi gave the classificatory vocabulary; how the later line performed the synthesis on that vocabulary is P24's main subject.*

---

> **Pointing at the Moon Research · Six Practice Gates, Volume One: *Mindful of the Buddha* · Part VI Two Streams · 6.3**
> Shi Huijing (釋慧鏡)
> First draft, May 2026
> This work is licensed under CC BY-NC-SA 4.0


---

*CBETA collation: all T-number / X-number citations verified against the CBETA Electronic Tripiṭaka 2024 edition.*

*Repo: github.com/pointing-at-the-moon/six-practice-gates · CC BY-NC-SA 4.0*

*NIAN-P23 · 釋慧鏡 (Shi Huijing) · 2026-05-04*