Reading a Five Hundred Arhats set: how 500 rakan are distributed across a numbered program

A gohyaku rakan set distributes 500 enlightened disciples across 100 scrolls of five figures each. The reading is positional: identify the scroll’s place in the numbered hundred, read the five rakan as a working group performing a miracle, find the attendant acolytes, then read the donor inscription that paid for that one scroll. The Daitoku-ji Song set (1178) and the Kano Kazunobu Zōjō-ji set (1854–63) are the two great programs; a single surviving scroll is always a fragment of a count.

What a gohyaku rakan set is, and why it is never one painting

The thing on the wall is one scroll. The work is a hundred of them. That gap is the first thing to hold onto, because almost every Five Hundred Arhats (五百羅漢, gohyaku rakan) scroll a reader meets in a museum is hung alone, captioned alone, and read alone, when it was made to be the eighty-second of a hundred, or the third, or the fifty-ninth. It is a numbered unit in a program that no gallery can hang in full and that most temples could only ever unroll a few of at a time.

The arithmetic is fixed and worth saying once. A canonical set is one hundred scrolls, five rakan to a scroll, five hundred figures. The hundred is the structural number; the five is the compositional one. A reader who knows the convention counts the principal figures in any scroll first, expects five, and treats the scroll’s number, when it survives (inked on the silk or recorded on the box), as the single most informative fact about it.

Rakan is the Japanese reading of the Chinese luohan (羅漢), itself the contraction of aluohan, the transliteration of the Sanskrit arhat: one who has done what was to be done, the enlightened disciple who remains in the world. The textual armature that turns a vague “many disciples” into a counted set is narrow and identifiable. Richard K. Kent, writing the Arhat Images entry for the Encyclopedia of Buddhism, traces organized arhat veneration to the mid-seventh-century Chinese translation of the Record of the Abiding Law attributed to Nandimitra, which fixed sixteen “great” arhats charged with remaining until the next Buddha; the set then “expanded from sixteen to eighteen and then to five hundred in number.”¹ The sixteen are a named, individuated canon, the subject of a separate reading on bodhi. The five hundred are the opposite: a crowd whose power is precisely that it is uncountable in practice and exactly countable in principle. The set form exists to make the crowd legible without making it individual.

The first move: find the scroll’s number, not the scroll’s saint

With the sixteen arhats the reading question is which one: Piṇḍola, Nakula, the named bearer of a named attribute. With the five hundred it is almost never that. Outside the few scrolls that carry an identifying cartouche, the individual rakan in a 500-set are not meant to be named, and trying to name them is the genre’s commonest misreading. The reading question is where in the hundred does this scroll sit.

That number is load-bearing because the hundred is sequenced, not random. The Daitoku-ji Song set, the best-documented program, runs its hundred scrolls as a loose narrative arc (daily monastic life, encounters with donors and rulers, miracles and supernatural displays, the rakan crossing into other realms) so that a scroll’s position predicts its subject the way a chapter number predicts a chapter. The number is recorded in three places, in descending order of reliability: an inscription on the painted surface, a label on the storage box and its wrapping, and the institutional catalogue that reconstructs the order after the set has been split. When a museum scroll is captioned “no. 22 of 100,” that is not a curatorial nicety; it is the load-bearing datum, and a reading that does not lead with it has read the object as a picture when it is a page.

The five as a working group, not five portraits

Inside the frame the convention is just as fixed. The five rakan are not five separate icons sharing a sheet. They are a working group, posed in a landscape, jointly doing something, and the something is usually a demonstration of the powers the arhat is canonically held to possess.

The vocabulary of the demonstration is stable across five centuries. Rakan summon dragons from water and clouds; cross rivers on a reed or on the back of a beast; emit light or transformation-bodies from the crown of the head; manifest a bodhisattva above them as a vision; receive offerings from kings, monks, and barbarian envoys; or, at the quiet end of the range, simply wash robes, read, doze, and sweep, the supernatural held in reserve. Kent distinguishes the two visual traditions the convention carries: the Guanxiu line (after the monk-painter Guanxiu, 832–912), in which the rakan are “distinctly exotic, even grotesque” (bushy-browed, slack-jawed, big-nosed foreign sages), and the sinicized Ningbo line, dominant by the twelfth century, in which they become decorous Chinese elders set in “elaborate landscape settings” with “narrative implications.”¹ A reader who can place a scroll on that scale before reading anything else has already done most of the iconographic work: the grotesque type is making a point about the arhat’s strangeness to the world; the decorous type is making a point about the arhat’s mastery within it.

The Cleveland handscroll read here is the decorous type taken to its limit. Wu Bin’s Five Hundred Arhats (Cleveland 1971.16) is not a hanging-scroll set at all but a single continuous handscroll more than sixty-six feet long, and it makes the genre’s structure unusually visible by laying the whole count out in one register instead of splitting it across a hundred sheets. The Cleveland catalogue records the population precisely: 447 luohan, 72 attendants, and the bodhisattva of compassion at the very end.² Stand at the case where the scroll is opened to its middle and the reading method declares itself: the figures do not arrive as a uniform file but in legible knots. Here a seated rakan under a wisteria-hung tree with a boy crouched at his feet; there a procession of elders with attendants shouldering bundles across a stony ground; throughout, the brush holding to a pale, even, almost colorless line that never lets any one face become a portrait. The uniformity is deliberate, not lazy. The line is the argument. This is baimiao, the fine-outline manner the eleventh-century lay Buddhist Li Gonglin made canonical (Wu Bin’s box title names the scroll a copy “after Li Gonglin”), and its refusal of bravura is doctrinal: the figures are not to be admired individually, they are to be counted devotionally.²¹

Find the acolytes: the attendant is part of the iconography

The small figures matter. In a Five Hundred Arhats scroll the rakan are almost never alone; they are attended by acolytes (novice monks, lay boys, servants carrying sutra boxes, fans, alms bowls, animals), and the attendant is not staffage. It is a scale device and a status marker, and counting it is part of the reading.

Two things follow from looking at the attendants first. The proportion fixes the type: in the Ningbo-line sets the attendants are markedly smaller than the rakan, a deliberate hierarchic scaling that tells the reader the painting is a devotional ranking and not a genre scene of monks at work. And the attendant’s task names the rakan’s act: the boy holding the unrolled scroll is there because the rakan is expounding; the servant restraining a beast is there because the rakan has summoned or subdued it; the figure with the alms bowl is there because the scene is an offering, which, in a set whose scrolls were themselves offerings, is the genre quietly depicting its own commissioning. Wu Bin’s 72 attendants against 447 rakan hold roughly the canonical proportion; the Cleveland record’s separate count of the two populations is itself a sign that the distinction is structural, not incidental.²

Read the inscription: the merit-economy of a hundred scrolls

The most consequential text on a Five Hundred Arhats scroll is usually not iconographic at all. It is the donor inscription, and it is the key to why these sets exist in the form they do.

A hundred-scroll set is enormous, expensive, and slow. The economic solution was to break the commission into a hundred parts and sell them, scroll by scroll, to individual donors, each of whom paid for one painting and received the merit of having endowed it. The surviving Daitoku-ji set carries this on its silk. Wen C. Fong’s foundational study of the program, his Princeton dissertation published by the Freer as The Lohans and a Bridge to Heaven (1958), established that the set was a Ningbo workshop product, commissioned through the monk Yishao for the Hui’an cloister near Ningbo and executed by the professional painters Lin Tinggui and Zhou Jichang, dated by inscription to 1178.³⁴ One of those gold dedicatory inscriptions, on the scroll now in the Freer as Luohan Laundering, reads in translation: the wife of a minor court official “contributed the funds to paint this picture and presented it to the Hui’anyuan as a permanent offering, that this meritorious deed may safeguard the wife’s welfare,” dated the fifth year of the Chunxi reign, with the monk-in-charge Yishao named and the painter Lin Tinggui signing.⁵⁴

That single inscription contains the whole logic of the set. The donor is private and named; the gift is one scroll, not a hundred; the merit is specific and personal (“safeguard the wife’s welfare”); the recipient is an institution that will hold the scroll “as a permanent offering,” meaning the hundred scrolls were never expected to be one owner’s possession but a hundred separate acts of merit pooled into one program by the temple that received them. The University of Alberta study The Daitokuji Five Hundred Arhats Paintings and Their Beholders reads the set on exactly this footing: a ritual object whose hundred parts were beheld in rotation, the program reconstituted in the act of display rather than ever shown whole.⁶ A reader who finds a donor’s name and a merit-dedication on a Five Hundred Arhats scroll has not found a footnote. They have found the reason the object is a scroll-of-five and not a wall.

The two great programs, and how a fragment points back to its count

Two sets anchor the genre at its two ends, and any single scroll is read against them.

The Daitoku-ji Song set is the continental archetype. One hundred hanging scrolls, five luohan each, painted at Ningbo and dated 1178, carried to Japan in the thirteenth century and held for centuries by the Kyoto Rinzai monastery Daitoku-ji, which is the source of the Japanese term and the Japanese reading habit. The set is now itself a fragment: eighty-two scrolls remain at Daitoku-ji, ten were sold to the Museum of Fine Arts, Boston, in the 1890s, and two reached the Freer. The dispersal is the genre’s condition made literal: a program designed as a hundred separable acts of merit has been separated, and each scroll in Boston or Washington is read by the institution back into a count it is no longer physically part of.³⁴

The Kano Kazunobu Zōjō-ji set is the Japanese culmination. Between 1854 and his death in 1863, the Edo painter Kanō Kazunobu (1816–63) produced one hundred hanging scrolls of the Five Hundred Arhats for Zōjō-ji, the Tokugawa-sponsored Jōdo-shū Pure Land monastery in Edo. The structure is the inherited one, a hundred scrolls with the disciples in groups, but the content is wrenched toward catastrophe: the rakan are shown traversing the cosmos to rescue the suffering from earthquake, flood, and fire, alongside the humbler register of bathing, shaving, and burying the dead. The set survived the anti-Buddhist disestablishment of the 1870s and the Tokyo air raids, was effectively unseen for most of the twentieth century, and was reconstituted for public viewing only in the twenty-first, in the exhibition the Smithsonian toured as Masters of Mercy.⁷ Its history is the Daitoku-ji condition in fast-forward: a hundred-part program that is read whole only in the rare moment a museum can put the count back together.

The Cleveland Wu Bin scroll sits between them as the third term. It is not a hundred-scroll set; it is the count compressed into one continuous handscroll, painted around 1600 by a lay Buddhist for whom, as the Cleveland record and the Late Ming scholarship both stress, the act of painting five hundred figures was itself a penitential and meditative practice akin to copying a sutra.² That is the same merit-logic as the Ningbo donor’s, relocated from the patron to the painter: the Daitoku-ji donor bought one scroll for merit; Wu Bin painted all five hundred figures for it. Read against the hanging-scroll sets, the Cleveland handscroll is the genre’s structure laid bare, the hundred made visible as one unbroken count, ending, as the Ningbo and Edo programs do, with the bodhisattva of compassion who gathers the disciples in.

Where this reading commits, and where it is open

bodhi commits to the positional method: a Five Hundred Arhats scroll is read first as a numbered unit of a hundred, then as a five-figure working group, then for its attendants and its donor text, and only last — if a cartouche allows it — for any individual rakan’s identity. The commitment is against the museum-default reading that treats a single scroll as a self-contained painting, which is the reading the genre’s commission economy and display practice both contradict.

Two uncertainties are named rather than smoothed. First, the identification systems for which rakan is which within the five hundred are not one system: the China-and-Japan tradition derives its names from a Southern Song stele, the Luohan zunhaobei, while Korea used a separate Goryeo-period system, so a “no. 153” in one tradition is not the “no. 153” of another, and a reading that asserts a specific rakan’s identity from position alone is overreaching unless the scroll itself carries the cartouche. Second, the page-level scholarship behind the Daitoku-ji program here is cited at work level: Fong 1958 and the Alberta study are followed for the Ningbo workshop attribution, the 1178 dating, and the donor-inscription economy, but the specific pages are not pinned in this draft and an operator pass against the physical volumes is owed before any of those claims is given a page. The donor-inscription text is quoted from the InscrChArt transcription of the Freer scroll, not collated against Fong’s printed translation; the two are consistent in substance, and the divergence, if any, is at the level of wording, not of meaning.

Zōjō-ji set, 1854–63, 100 scrolls |

Related

• The Sixteen Arhats — reading the named canonical set
• Reading a nehan-zu — the death of the Buddha
• Reading a suijaku mandara
• Rakan / luohan (entity)
• Jōdo-shū (entity)

Footnotes

1. Richard K. Kent, “Arhat Images,” in Encyclopedia of Buddhism, ed. Robert E. Buswell (Macmillan Reference, 2004), via Encyclopedia.com. Claim borrowed: the Nandimitra Record of the Abiding Law as the textual basis for the sixteen, and the expansion “from sixteen to eighteen and then to five hundred”; the Guanxiu grotesque type vs the sinicized Ningbo type with elaborate landscape settings. Reference-tier; cited at entry level. ↩ ↩² ↩³

2. The Cleveland Museum of Art, accession 1971.16, The Five Hundred Arhats (五百羅漢圖), Wu Bin (Chinese, active c. 1591–1626), handscroll, ink and color on paper, CC0. Claim borrowed: the population (447 luohan, 72 attendants, Guanyin at the end), the baimiao / “after Li Gonglin” lineage, and Wu Bin’s lay-Buddhist devotional practice. clevelandart.org/art/1971.16, accessed 2026-05-16. ↩ ↩² ↩³ ↩⁴

3. Wen C. Fong, The Lohans and a Bridge to Heaven (Freer Gallery of Art, Smithsonian Institution, 1958), and Fong’s subsequent Five Hundred Lohans at the Daitokuji. Claim borrowed: the Daitoku-ji set as a 1178 Ningbo workshop product (Lin Tinggui and Zhou Jichang), commissioned through the monk Yishao for the Hui’an cloister, with gold dedicatory inscriptions. Pages NOT pinned — cited at work level pending an operator pass against the printed volume. ↩ ↩²

4. National Museum of Asian Art (Freer Gallery of Art), accession F1902.224, Luohan Laundering (羅漢洗濯圖), Lin Tinggui, 1178. Claim borrowed: the set of one hundred hanging scrolls of the five hundred luohan, the 1175 commission / 1178 completion, the transport to Japan and Daitoku-ji, and the 1890s dispersal (82 Daitoku-ji / 10 MFA Boston / 2 Freer). asia-archive.si.edu/object/F1902.224, accessed 2026-05-16. ↩ ↩² ↩³

5. InscrChArt (Miami University Libraries), item 5019, transcription and English translation of the gold donor inscription on Luohan Laundering. Claim borrowed: the verbatim merit-dedication formula (donor’s wife “contributed the funds to paint this picture and presented it to the Hui’anyuan as a permanent offering, that this meritorious deed may safeguard the wife’s welfare”), dated Chunxi 5 (1178), Yishao monk-in-charge, Lin Tinggui painter. inscrchart.lib.miamioh.edu/s/InscrChArt/item/5019, accessed 2026-05-16. Quoted from the InscrChArt transcription, not collated against Fong’s printed translation; substance consistent. ↩

6. The Daitokuji Five Hundred Arhats Paintings and Their Beholders, University of Alberta thesis (ERA / Scholaris repository). Claim borrowed: the set as a ritual object whose hundred parts were beheld in rotation, the program reconstituted in display rather than shown whole. ualberta.scholaris.ca/handle/123456789/60723, accessed 2026-05-16. Cited at work level; abstract consulted. ↩

7. National Museum of Asian Art (Smithsonian), Masters of Mercy: Buddha’s Amazing Disciples. Claim borrowed: Kanō Kazunobu (1816–63), one hundred hanging scrolls of the Five Hundred Arhats for Zōjō-ji (Jōdo-shū, Edo), 1854–63, the disaster-rescue scenes (earthquake, flood, fire), and the twentieth-century invisibility / twenty-first-century reconstitution. asia.si.edu, accessed 2026-05-16. ↩