The most secret teaching of the Nāth tradition was revealed at a place called Candradvīpa — "Moon Island." Matsyendra, in the belly of a fish, overheard Śiva's teaching off the shore of Moon Island.1 So scholars went looking for Moon Island on the map of India. They never found it. The toponym "has never ceased to tantalize scholars who have attempted, without success, to localize" it.2
White's answer is that they couldn't find it because it isn't a geographic place. Moon Island — like the Fish Belly — is "a location that appears within the subtle body when a certain body of practice is carried to its conclusion."3 It's not a spot on the coast. It's an inner place, reached by doing the yoga, not by sailing anywhere. The search for Moon Island failed for the same reason a search for the "place" where you fall asleep would fail: it's a state, mapped as a geography.
This page is about geography internalized — the principle that the tradition's sacred place-names are often names of inner locations in the subtle body, not coordinates on land.
Candradvīpa, "Moon Island," and its variants Candragiri ("Moon Hill") and Candraparvata ("Moon Mountain") are the revelation-sites of the Western Transmission and the Yoginī Kaula.4 White's reading: "Moon Mountain and Moon Island — for what is an island if not a mountain with its feet in the water? — are, like the Fish Belly, locations that appear within the subtle body when a certain body of practice is carried to its conclusion."3 The "moon" is the key: in the yogic body, the moon and its nectar are located in the cranial vault, the destination of the rising practice. Moon Island is, on this reading, the inner moon-place — the lunar summit of the subtle body where the nectar pools — reached when the practice is completed.
The clincher comes from Benares again. During the fish-belly conjunction, when the rivers flood and reverse, the city of Benares — surrounded by water on all sides — becomes an island.5 And it's at exactly that moment (the fish-belly conjunction, the reversal, the opening of the central channel) that the secret is revealed. So Moon Island isn't somewhere you travel to; it's what the holy place becomes — an island — at the moment of inner reversal. The island appears when the practice culminates, just as Benares becomes an island when the flood comes.
White notes that some scholars did try to anchor it, and revealingly: Bagchi identified Moon Island with the Chinese "double-mountain" K'un-lun — which is itself a location within the Taoist subtle body.6 So even the comparative attempt to locate it lands on another tradition's inner location, which only confirms the point: Moon Island lives in the subtle body, whether Indian or Taoist.
Why would the tradition name an inner state "Moon Island" instead of just describing the state? Because the subtle body is imagined as a landscape, and landscapes have places.
Once you picture the inner body as a terrain — channels as rivers, the cranial vault as a moon-realm, energy-centers as stations — then the destinations of practice naturally become places: an island, a mountain, a forest, a city. Naming the lunar summit "Moon Island" isn't obscurantism; it's the consistent application of the body-as-landscape metaphor. The yogi journeys inward the way a pilgrim journeys outward, and the inner journey has a destination, so the destination gets a place-name.
This is why the same name attaches to real places and inner ones, and why the two keep getting confused. Real Benares becomes an island during the flood; the inner moon-realm is "Moon Island." A real coastal site might be called Candragiri; so is the inner lunar summit. The tradition doesn't always distinguish, because for it the real pilgrimage and the inner journey are homologous — walking the holy island and reaching the inner Moon Island are the same act at two scales. The scholar who insists on a single geographic referent is asking the wrong question; the tradition is using the place-name to name both the outer site and the inner state, deliberately.
And it explains the failure of the search. You can't find Moon Island on a map because it was never only on the map — it's the inner place that appears when the practice completes, the way Benares appears as an island when the flood comes. The "location" is real, but it's a location in the subtle body, accessible by yoga, not by boat.
This page generalizes the decoding method: sacred place-names in this tradition are often internalized geography, names of inner states.
It directly extends the Fish-Belly decoding — Moon Island is the next place-that-is-really-an-inner-location. It connects to the Western Transmission, whose revelation-sites these lunar toponyms are. And it pairs with the broader principle in sacred geography reproducing the alchemical body — the land mirroring the inner terrain.
The single sharpest clue is the moment Benares turns into an island.
In ordinary times, Benares is a city on a riverbank — not an island. But during the fish-belly conjunction, when the monsoon floods and the rivers (Ganges, Varaṇā, Matsyodarī, Asi) swell and reverse, the city becomes surrounded by water on all sides: an island.5 And this is precisely the moment of the great revelation, the moment the inner channel opens and the current reverses. So the holy city becomes Moon Island exactly when the inner reversal happens.
This tells you the islandness isn't a fixed geographic fact but an event — something that appears at the culminating moment and recedes after. And that's exactly what White says Moon Island is in the subtle body: a location that appears "when a certain body of practice is carried to its conclusion."3 The parallel is exact: just as Benares is only an island at the flood's peak, Moon Island is only "there" at the practice's culmination. The inner moon-realm isn't a permanent room you can visit anytime; it's a place that manifests at the summit of the rising practice, when the nectar pools in the cranial vault, and that you can't locate the rest of the time.
The case shows why the geographic search was doomed and the internalized reading is right. A place that only exists at a culminating moment can't be pinned on a map — but it can be reached, by carrying the practice to its end, just as Benares can be seen as an island by being there at the flood. Moon Island is real and locatable; it's just locatable in time-and-practice (at the culmination) rather than in space (on the coast).
You hit a sacred place-name that scholars can't pin down geographically. Run the internalized-geography test.
First, ask whether the name itself points inward. "Moon Island," "Moon Mountain" — the moon is a subtle-body location (the cranial vault, where the nectar pools). A place-name built from an element of the inner anatomy is a strong sign you're looking at an inner location wearing a geographic mask. Then ask when the place is "reached": if it's reached at the culmination of a practice (Matsyendra hears the secret at Moon Island; the revelation comes at the fish-belly conjunction), it's a state, not a coordinate — a place that appears when the practice completes.
Then stop trying to find it on the map and start reading it as a destination of the inner journey. The discipline is to recognize that the tradition maps inner states as places deliberately, because it pictures the subtle body as a landscape — so the "where" of a teaching's revelation may be a "when" and a "how far along" rather than a spot you could sail to. When the geographic search fails repeatedly, that failure is the data: it's telling you the place is internal.
The clean way to misread Moon Island is to insist it must be a real geographic place and keep hunting for the coordinates.
Generations of scholars did this, "without success."2 The failure isn't a lack of cleverness; it's a category error — treating an inner location as an outer one, demanding a single geographic referent for what the tradition uses as both an outer site and an inner state. The harder you search the map, the more the toponym slips away, because it was never only on the map. Insisting on localization mistakes the body-as-landscape metaphor for literal cartography.
The opposite failure is to declare the place purely internal and deny any geographic dimension at all — to say there's no real island anywhere, only an inner state. But the tradition genuinely does attach these names to real sites (real Benares really becomes an island; real coastal hills bear lunar names), and the outer pilgrimage really is performed.5 The honest reading holds the homology: Moon Island is both an inner location and (sometimes) a real place that mirrors it, because the tradition treats the outer pilgrimage and the inner journey as the same act at two scales. Insist it's only on the map and you'll search forever; insist it's only in the mind and you lose the real island Benares becomes. The place lives at the meeting of the two.
The reading rests on the Kaulajñānanirṇaya and Western Transmission sources (Candradvīpa/Candragiri as revelation-sites), the Benares-becomes-an-island datum (via Eck), White's "appears when the practice is completed" principle, and Bagchi's K'un-lun identification.156
The tension is that White's internalized reading competes with genuine geographic candidates. Scholars have proposed real locations — a Bengali coastal site, a Goan hill, Bagchi's K'un-lun.6 White doesn't claim there's no geographic dimension; he argues the primary meaning is the inner location, with real sites as homologues. But the balance between "real place that became symbolic" and "inner state that borrowed a place-name" isn't fully fixed — and for some of these lunar toponyms the geographic candidate may be more substantial than for others.
The open question is how systematically the tradition's place-names internalize. White establishes the principle for the Fish Belly and Moon Island; how far it extends — whether most of the tradition's revelation-sites and pilgrimage-names are primarily inner locations, or only a special class of them — remains a matter of case-by-case reading. The method (read the name for inner meaning, check whether it's "reached" at a practice's culmination) is general, but which places it applies to is not settled.
White reads with the geographers of Benares (Eck) and against the geographic-localizers (the scholars who hunted for Candradvīpa on the map), while engaging Bagchi's comparative identification with the Taoist K'un-lun. The convergence is suggestive: Bagchi, trying to locate Moon Island geographically, ends up identifying it with another tradition's inner location (the K'un-lun of the Taoist subtle body).6 Even the comparative attempt to pin it down lands inside a body, not on a map.
That accidental convergence supports White's thesis from an unexpected direction. The localizers' best cross-cultural candidate is itself an internal location — which suggests that "Moon Island" belongs to a class of lunar inner-body sites that recurs across traditions (Indian Candradvīpa, Taoist K'un-lun), all naming the inner summit reached at the culmination of practice. White's contribution is to read the Indian case as primarily internal; Bagchi's comparative work, meant to geo-locate it, instead reveals it has a Taoist twin that is also internal. The convergence is that two traditions independently mapped the inner lunar summit as a "place" — which is much harder to explain if these are merely garbled coordinates than if they're genuine, recurring inner locations that the body-as-landscape metaphor generates wherever it's used.
Plain version: a tradition that pictures the inner self as a landscape will name inner states as places — islands, mountains — so that what looks like geography is really a map of consciousness, and searching for it on land is a category error. That sits at the seam of how the mind is spatialized and how sacred geography works.
The sharpest version. Some places can't be found on any map because they were never on the map — they're inner states the tradition mapped as geography, reached by doing the practice, not by traveling. Moon Island is the inner lunar summit that appears when the yoga culminates, exactly as Benares becomes an island only at the flood's peak. Take this seriously and it reframes every sacred "place" of revelation: when scholars can't locate it and it's always "reached" at a practice's culmination, the failure to find it on land is the answer — the place is internal, and the search for coordinates is asking where you are when you fall asleep. The geography was always a map of consciousness.
Questions worth chasing: