Marpa the Translator: Life, Journeys and Lineage

By Tim Yu, Founder & Editor · Published 27 October 2026

Checked against the primary texts · How we check · Report a correction

Marpa the Translator (Marpa Chökyi Lodrö, mar pa chos kyi blo gros) was an eleventh-century Tibetan layman from Lhodrak in southern Tibet who travelled to Nepal and India, received tantric teachings from Indian and Newar masters, among them, by every account, Nāropa, and became the first Tibetan master of the Kagyu school and the teacher of Milarepa. He is usually dated 1012–1097, the dates of the Blue Annals (1476), but his dates, like most of his life, are uncertain.

“No contemporary historical document attests to his existence,” writes Cécile Ducher, whose 2017 study Building a Tradition compares every life of Marpa she could find. What we have are sacred biographies (namthar) by his spiritual heirs, and songs attributed to him, which Ducher thinks circulated among his disciples in the decades after his death. This page reports each episode as its source tells it.

Marpa at a glance

• Names: Marpa Chökyi Lodrö; also Lhodrak Marpa and Marpa Lotsāwa, “the translator”.
• Dates: usually given as 1012–1097 (the Blue Annals, 1476); other Tibetan histories give c. 1000–1081, 1000–1085 or 1024–1107, and most later Kagyu historians preferred the earlier dates.
• Home: Drowolung in Lhodrak.
• Teachers (as his songs and lives name them): Nāropa, Maitrīpa, Jñānagarbha, Śāntibhadra and Newar masters in Nepal.
• Best-known disciple: Milarepa.
• School: the Kagyu, or Marpa Kagyu (Himalayan Art Resources).

How we know about Marpa

Ducher counts thirty-two biographies of Marpa. None, she notes, “was composed during his lifetime or by a direct disciple”, and all were written by Tibetans who had never been to India. The first was written two generations after him.

The oldest layer is the songs attributed to him, the biographers’ main building material. The earliest prose lives date from the twelfth century: by Ngamdzong Tönpa Changchub Gyalpo, a disciple of Milarepa; by Ngok Dode, son of Marpa’s disciple Ngok; and, a few decades later, by Martön Tsultrim Jungne.

The version most readers know came much later. Tsangnyön Heruka (1452–1507), best known for his Life of Milarepa, dictated a Life of Marpa on returning from Nepal in 1505 and had it printed the same year. It became the standard account, translated in full by the Nālandā Translation Committee in 1982 as The Life of Marpa the Translator. Ducher finds it “remarkably faithful to its sources” — but also reshaped, so that everything turns on Nāropa. A second long life, by Pawo Tsuglag Trengwa (Pawo II), abbot of Drowolung, was compiled in 1563 from thirteen earlier biographies, its colophon says.

We may not know, Ducher writes, “when he was born, where, exactly, he went in India or whom he met there”, but what survives comparison is more solid than any single version.

When did Marpa live?

The familiar pair of dates, 1012–1097, is found in no text earlier than the Blue Annals of Gö Lotsāwa (1476), Ducher shows; the birth year 1012 derives from Gö’s comparing the lives of other translators of the period. Those dates were widely adopted, in the West above all. The Treasury of Lives heads its entry “b.1012? – d.1097” and Lotsawa House gives c. 1012–1097; Himalayan Art Resources gives 1012–1096.

Other Tibetan historians disagreed. Opening Eyes, a Phagmodru Kagyu history of 1418, gives 1024–1107. Pawo II listed rival birth years in older sources ranging from 982 to 1024, and tentatively chose 1000, with a life of eighty-six years. Ducher follows the early tradition, as most later Kagyu historians did, and writes his dates “1000?–1081?”, because they “cannot be ascertained on conclusive grounds”. Marpa himself probably did not know his birth year, she adds: he was born before the Kālacakra calendar was introduced and its first sixty-year cycle began in 1027. One of the rare anchors is Nāropa’s death, around 1040.

Early life, as the biographies tell it

Ngamdzong Tönpa’s twelfth-century life places his birth at Chukhyer in lower Lhodrak. He was an unruly boy who beat the neighbours’ children, and his parents, afraid he would kill someone, sent him to study with the translator Drokmi Lotsāwa at Mugulung. He stayed a year. Drokmi gave no teaching without a fee, but Marpa learned the translators’ language from visiting Indians and decided to go to India himself.

Not every early life has this episode. Most versions of Martön’s life leave out the training under Drokmi, and Ducher finds the chronology hard to square with Drokmi’s dates. The family’s standing varies too. Ngok Dode says it was rich enough to own “two or three valleys”, while Ducher’s summary of the evidence is that Marpa “did not belong to a famous family” and raised his own funds.

In the song attributed to him that opens Ducher’s book, Marpa says he “met with the Dharma” in his thirteenth year, learned first the alphabet and then translation, and stayed three years in the Nepal Valley before India.

The journeys to Nepal and India

How many journeys? Ducher shows that Marpa’s songs “clearly and repeatedly mention three”, and the historian Butön repeats that count in 1322. Martön’s life and the Blue Annals give only two. Tsangnyön, whose main source had two, added a second journey, “a rather bland one”, to match the songs. Between journeys the biographies have him back in Tibet, gathering gold to offer his teachers.

He travelled with Nyö Lotsāwa of Kharag. Tsangnyön popularised the picture of Nyö as a jealous rival who plotted to destroy Marpa’s texts; some sources ignore that episode, and some rehabilitate him.

His teachers

The song names “the lineage rivers of the four directions”. In the song, Marpa received the Guhyasamāja in the west from Jñānagarbha; in the south, the Mahāmāya from Śāntibhadra; in the east, Mahāmudrā from Maitrīpa; and in the north, at Phullahari, the Hevajra, the six doctrines and the Aural Tantra from Nāropa. Before that, the song says, Newar masters in Nepal gave him the Catuṣpīṭha.

In Ngamdzong’s family of texts, Marpa meets Jñānagarbha first and Nāropa later; in the group of Ngok Dode, Martön and Tsangnyön, he goes to Nāropa first, introduced by Paiṇḍapā. Ngamdzong names his Nepal teacher as the Newar Chitherpa (spyi ther pa); Ngok Dode names Phamtingpa and Paiṇḍapā. Opening Eyes treats Śāntibhadra and Kukuripa as one person; the eighteenth-century Kathok Rigdzin counts two. BDRC and the Treasury of Lives add Prajñārakṣita and Niguma.

Did Marpa really meet Nāropa?

No biography has Marpa fail to meet Nāropa, and the meeting anchors the Kagyu’s claim to its Indian source. A song in Ngok Dode’s life puts it directly:

Even the place is disputed: most biographies say Magadha, in central India; some say Vikramaśīla, others Nālandā; and many colophons and some biographies put Nāropa’s monastery, Puṣpahari, in Kashmir. Puṣpahari is the Sanskrit form of Phullahari, as Pawo II explains, and Ducher suspects that placing it in Kashmir is a later transposition of an Indian sacred site to the Himalayan world.

The meeting itself was questioned around 1200. In a letter, the Sakya master Drakpa Gyaltsen quoted Naktso Lotsāwa, Atiśa’s disciple and translator, who reported a story passed down from a man who claimed to be Marpa’s disciple (his name appears in none of Marpa’s biographies): that Marpa said he “never arrived there” and had received Nāropa’s instructions from one of Nāropa’s disciples. Kagyu historians replied: Opening Eyes in 1418, and in 1506 the Fourth Shamarpa, who called the letter a Sakya zealot’s addition. The historian Ronald Davidson calls the association “a monument of Tibetan fictionalization”, arguing that Marpa probably received much of his instruction from the Phamtingpa brothers, Nāropa’s disciples, at Pharping, and that Nāropa’s death around 1040–42 makes the twelve or more years with him that the tradition describes hard to justify. Some Kagyu writers, he adds, even had Marpa meet Nāropa after Nāropa’s death, on the third journey, a device he traces to Tsangnyön’s Life of Marpa.

No documents survive to settle it, Ducher notes, and rival lineages had reasons to stress or shrink Nāropa’s role. She is “quite convinced” they met, while admitting that her only proof is the traditional narratives, and suggests the meetings “may only have consisted in a few months and a few transmissions”. Most lives say that on the last journey Nāropa had “entered the practice” and stopped teaching, so Marpa found him only after a quest full of visions, which Ducher reads as “likely more symbolic than ‘historical’”. Tsangnyön has Milarepa urge that journey; Ducher finds this hard to fit the chronology and “not widespread in the biographies”. Nāropa’s own life is told on our Nāropa page.

What Marpa brought back

Butön lists the transmissions Marpa mainly received: Guhyasamāja, Cakrasaṃvara, Hevajra, Mahāmāya, Catuṣpīṭha and “many key instructions”. The lineage traces to Nāropa the practices later called the six doctrines, or Six Yogas, of Nāropa. Ducher shows the set had not been fixed when Marpa received it: the song in which the biographies have Nāropa sum them up for Marpa (in Tsangnyön’s version, trans. 1982, pp. 94–97) lists seven — inner heat (tummo), illusory body, dream, luminosity, the intermediate state, transference and entering another’s body. The Karmapa’s office says Marpa finally gave Milarepa the empowerment of Vajravārāhī and instructions on the Six Dharmas of Nāropa, “primarily on tuṃmo (inner heat)”. This page only names these Vajrayana practices, which the lineages pass on from teacher to student.

The last of the seven fills his biographies with scenes of Marpa entering the bodies of dead animals; the Blue Annals says only that he did it four times. Tsangnyön says these teachings “were not destined to spread in Tibet”; Ducher thinks the practice probably continued informally but never widely.

How much did he translate?

His title can mislead. In the survey of Tibet’s translators by Chomden Raltri, a thirteenth-century abbot of Narthang, neither Marpa nor Nyö, Ducher reports, “seems to have produced a significant number of translations”; it names Guhyasamāja traditions, Hevajra and key instructions. In Tibet, she writes, he was known for the wealth of his instructions and began a tradition centred on practice rather than study. We have not verified a list of canonical translations under his name, so we give none.

Pawo II credits him with writings, as BDRC records: a commentary on part of the Hevajra Tantra, a Six Dharmas: Stream of Nectar composed for his son, the Fifteen Scrolls and many short works. The fifteen scrolls were said to have been hidden in his house, Sekhar, and revealed generations later as treasure (terma) by Guru Chöwang (1212–1270). One short text from his collected works (Tibetan edition, 2014) is on Lotsawa House:

Householder and “hidden yogi”

Marpa was not a monk. Ngamdzong Tönpa says that back from India he married and had children, and that outwardly he seemed to strive “for power and wealth” while being “actually a hidden yogi”.

The household grows in the telling. His wife Dagmema is first named in Gyaltangpa’s thirteenth-century life; her name echoes Nairātmyā, Hevajra’s consort. A later life by Khachö Wangpo introduces nine consorts. Gyaltangpa gives seven or eight sons; a text of the Ngok tradition gives nine.

The best-known son is Darma Dode, whom Himalayan Art Resources calls “probably his most beloved student”. Gyaltangpa is the first to tell of Nāropa’s prediction of his death and says he died after a fall from a horse. The Lhorong Chöjung (1446) adds that his consciousness entered a pigeon that flew to western India, near Oḍḍiyāna, where it entered the body of a newly dead young brahmin, who became known as the siddha Tipupa; a life of the Karmapa Dusum Khyenpa instead identifies Darma Dode with Yönten Zangpo, a contemporary of the Karmapa (Ducher, p. 130). A hostile version stands outside Marpa’s biographies: the hagiography of the rival translator Rwa Lotsāwa, a text that may have been compiled as late as the 17th century, claims that Rwa “liberated”, that is killed, Darma Dode.

Early lives carry a prophecy that Marpa’s “familial lineage will disappear like a celestial flower, but his Dharma lineage will reach as far as a river flows”. Ngamdzong Tönpa already ties that prophecy to a scene in which Nāropa manifests the maṇḍala of Hevajra and Marpa bows to the maṇḍala rather than to his guru. Tsangnyön links Darma Dode’s death and the end of the family line to that error; the eighteenth-century Rosary of Crystal Gems by Belo Tsewang Kunkhyab holds that the link cannot be established (Ducher, p. 206).

Disciples and the “four pillars”

Marpa’s four main disciples are remembered as the four pillars: Tsurtön Wangi Dorje in the east, Ngok Chöku Dorje in the south, Metön Tsönpo in the west and Milarepa in the north. Among the lives and histories of Marpa, the grouping first appears in the Red Annals of 1346. The song of the four pillars itself, Ducher shows, was not originally part of Marpa’s biographical tradition: it is first found in one of the earliest lives of Milarepa, possibly the Twelve Great Disciples, attributed to Ngamdzong Tönpa and eleven other disciples of Milarepa. In Tsangnyön’s telling the four pillars first appear in a prophetic dream of Milarepa’s, which Marpa interprets; the Rain of Wisdom, compiled some forty years later by the Eighth Karmapa, Mikyö Dorje, makes both the dream and its interpretation Marpa’s.

The transmission Ngok received is still passed on, Ducher notes, as the Mar-Ngok lineage within the Kagyu Ngakdzö collection, and Metön established an institute where the tantras he inherited from Marpa were studied. In Tsangnyön’s telling, Marpa set Milarepa to build and tear down stone towers before teaching him; the office of the 17th Karmapa Ogyen Trinley Dorje says he “endured immeasurable hardships under Marpa”. That story is told on our Milarepa page, and its link to preliminary practice under ngöndro.

The Kagyu after Marpa

Ducher stresses that Marpa founded no monastery and that the early Kagyu “did not develop within any structured locus of power”. Through Milarepa’s disciple Gampopa (1079–1153), whom the same office presents as the founder of the Dakpo Kagyu, Marpa’s line grew into what it calls “the four elder and eight younger lineages”. Ducher counts Marpa the first Tibetan master of every Kagyu branch but one, the Shangpa Kagyu, which came through Khyungpo Naljor from Indian teachers, chiefly the ḍākinīs Niguma and Sukhasiddhi. The schools are compared on the four schools of Tibetan Buddhism; the translators’ age is told in Buddhism in Tibet, the tradition as a whole in Tibetan Buddhism, and Marpa’s peers among the most influential Buddhist teachers.