What is a pāsāda?

Tags: #<Tag:0x00007fc45559c060>


A pāsāda is a certain kind of building, but its exact nature has never been properly investigated. This may seem like an unpromising topic for a short essay, but it is remarkable how little things can sometimes have quite significant ramifications.

Pāsāda is normally translated as “palace”, sometimes as “mansion”. Apart from the phonetic similarity between pāsāda and “palace”, which does not seems to have any linguistic significance, this is based, it seems, mostly on commentarial explanations. Even a fairly cursory look at the canonical texts, however, shows us that there is not much basis for these renderings.

Let’s take a brief look at a couple of commentarial descriptions. At AN 3.39 the Buddha-to-be is said to have had three pāsādas before he left the household life, one for each of the Indian seasons. According to the commentary these were respectively nine, seven, and five storeys tall. A famous pāsāda was the one given by Visākha to the Sangha at Sāvatthī. According to the commentary this was seven storeys tall. Tall stories, all of them!

These commentarial descriptions are hard to take seriously. There are no obvious mentions in the four Nikāyas of multi-storey buildings, except for occasional references to a single upper storey or a loft, such as in certain kuṭis. Moreover, the archaeological evidence from the period suggests that the towns were modest in size, even the big ones, like Sāvatthī. Academic estimates for the size of the largest cities is typically a few thousand inhabitants, or perhaps a few tens of thousands. It seems unlikely that these would have had enough material resources and wealth to build very large or tall structures. Even more important is the fact that all buildings were built of perishable materials – this is why there are no material remains from the period, except embankments and city walls and that sort of thing. Once again, it seems unlikely they would have built very large buildings, especially very tall ones, out of wood. Everything in the archaeological record from the Ganges plain for the relevant period points to small-scale buildings.

So if the commentarial notions are fanciful, what were these pāsādas? A suitable point of departure is to note that pāsādas were allowed by the Buddha for monastics. Now one of the main principles of the vinaya is that monastics should not live in luxury. On this basis alone we can conclude that pāsādas were not generally luxurious. This immediately eliminates “palace” and “mansion” as suitable translations.

Another thing that stands out about pāsādas is that they are always ascended to and descended from. This means that they must have been relatively tall buildings. An interesting point is that the suttas do not seem to mention any other way of accessing these buildings – apart from ascent/descent – which must mean that they were not directly accessible from ground level. They must have been built on some sort of high foundation. A passage from SN 3.21 is particularly instructive:

Suppose, great king, a man would climb from the ground on to a palanquin, or from a palanquin on to horseback, or from horseback to an elephant mount, or from an elephant mount to a pāsāda

According to a number of passages pāsādas had external staircases (sopānakaḷevara, MN85) or ladders (nisseṇi, DN9 and DN13), including the pāsāda given to the Sangha by Visākha (MN107), and they seem always to have been entered with the help of these. Similes in DN9 and DN13 speak of making the ladder before building a hypothetical pāsāda. These similes seem to imply that all pāsādas were accessed by climbing. In addition, the use of the word “ladder” shows that pāsādas often were quite humble buildings.

The need for climbing to get access also seems to be why – in one of the case stories to pārājika two – Ven. Pilindavaccha was able to make use of a pāsāda to hide some children from kidnappers. If pāsādas were high off the ground and the access ladder was removed, then the children were presumably safe. We also find the fairly common compound uparipāsāda, “up the pāsāda”, which fits this description. No other names for types of buildings are compounded with upari in this way.

The final clue to the meaning of pāsāda is that they seem to have had a space underneath. In the passage from AN 3.39 mentioned above, where the Buddha speaks of his life before going forth, it is said that he stayed in the rainy-season pāsāda for the full four months of the rains, without heṭṭhāpāsādaṃ orohati. Orohati means “descends” and heṭṭhā means “below”. Heṭṭhā is used in other compounds such as heṭṭhāmañcaṃ which unambiguously means “below the bed (mañca)”. It follows that heṭṭhāpāsādaṃ orohati can only really mean “descends to below the pāsāda,” which must mean that pāsādas had open spaces under them. My conclusion, then, is that pāsāda is likely to refer to a stilt house.

Now a stilt house is a fairly humble building compared to a palace. But it seems likely to me that in an ancient Indian setting they would have had considerable advantages, and thus were probably regarded as high-end housing. In a climate where heavy rain and flooding were common, stilt houses would be particularly sought after. It is also possible, as the story with Ven. Pilindavaccha shows, that these houses were considered relatively safe from burglars and the like.

We still need to consider a couple of passages in the suttas were pāsāda clearly does refer to a “palace”, such as Sakka’s palace in MN37 and Mahāsudassana’s palaces in DN17. In both of these cases the pāsādas were very grand structures, and even the word “palace” may not do full justice to their magnificence. At the same time, neither of these contexts refers to historical India, and they are perhaps best regarded as mythological. A possible explanation for the use of pāsāda in these cases is that they didn’t have any other suitable word. Pāsādas were perhaps the best houses of that time, and thus the term came to be used to describe any kind of nice house, even grand visions of the imagination.

Does any of this really matter? It gives us a more realistic picture of the life of the Buddha-to-be, a picture that is more consistent with known facts, such as archaeological findings. It is so hard to escape the grip of post-canonical exaggerations and the tendency to unreasonably glorify the Buddha. Such glorification tends to remove us from the historical Buddha and to make him into a character of fiction. Once the Buddha becomes fictitious and unrealistically portrayed, it becomes much more difficult to relate to him as a human being, and his example to us becomes much diminished. To be able to draw the maximum benefit from the Buddha’s teachings, it is important, I think, that we get as realistic a picture of him as possible. “Stilt house”, rather than “palace”, is a small but not insignificant contribution to this end.

The Translation of (pli-tv-bu-vb-np2/pli/ms#25)

Wow, that’s great, and very persuasive. I recall being a bit unsure about the whole pasada idea, but I never solved it. Incidentally, as so often, the PTS dict comes pretty close to the sense that you’re getting at, and the sanskrit dicts are not so far off either. Probably the meaning evolved into “palace” in later texts. I have a few questions.

I wonder if the divine pāsādas are inferred because of the whole “floating in the sky” thing? In which case, what’s the relation to vimāna?

Another term that’s translated as palace in the Suttas is rājantepura. Of course, it’s most obviously used in the sense of “harem”, but we also have passages like AN 3.60 and SN 42.10. Here you have a group of, apparently, court officials or dignitaries of various sorts sitting around the rājantepura and chatting. It’s also the focus of gossip in MN 87#5 and MN 90#6.

But I’m not really sure what kind of building we’re talking about, and where they are chatting. I tried to check up the known architecture of old Indian palaces, but came up blank. Anyway, the oldest is probably Ashoka’s and that’s going to be on a completely different scale.

To complicate things, in SN 3.19 Pasenadi is said to have brought a deceased estate into the rājantepura.

In DN 2, we have Ajatasattu sitting chatting upari­pāsāda­va­ra, which we might render “upstairs in the stilt hall”, using “hall” to stand for pāsāda­va­ra(?) Was this pāsāda a building inside the rājantepura?

Perhaps it was something like the old monastery design, with a square of building surrounding a central courtyard, and the pāsāda as the central building in the courtyard. Or maybe not a square of buildings, just a walled compound for the royal buildings.


The ‘floating in the sky’ bit reminds me of this passage in an3.63:

“But, brahmin, there are three kinds of high and luxurious beds that at present I gain at will, without trouble or difficulty. What three? The celestial high and luxurious bed, the divine high and luxurious bed, and the noble high and luxurious bed. These are the three kinds of high and luxurious beds that at present I gain at will, without trouble or difficulty.”

  • celestial = jhāna
  • divine = brahmā vihāras
  • noble = nibbāna



Dear Bhantes,

I think this is very plausible. Just to share something from the culture I grew up in. We call a structure that can fit an extended family a “balai”. In old Malay, it means a “hall”. Basically it was a sturdy built longhouse (rectangular in shape and was an open hall with minimal separation and privacy), mildly ornate, on top of supporting hardwood pillars to avoid monsoon floods, alleviate tropical heat (with the winds passing underneath to make it cooler) and provide safety from wild animals and rogues at night. It was also used as a gathering place for important occasions, especially the ones that dealt with community issues and passing of laws (usually the raja’s house or any important person within the community).

(Eventually as generations passed, any sizeable family house were now called balai. The only exemption is the simple and humble “kalapao” (hut), similar to the size of a kuti, or even smaller, and only used in the forests or remote areas now as temporary lodging. In the past, my ancestors had the habit of going to the forest for retreats for days on end and kalapaos where used, which boggled the European colonizers because they couldn’t stand the jungle heat or just didn’t understand the purpose of such retreats LOL :stuck_out_tongue_winking_eye: )

Perhaps, there is a similarity here with the pasada and a large multifunctional hall?

with reverence, respect, and gratitude,



Stilt houses is common type of houses used in ancient time in our country (and most countries in South East Asia) and it is still used as family residential in villages which is still keeping the traditional way of life until now. The ancient palaces is using this architectural too and many this type of ancient palaces are preserved as historical sites and tourism object here.

This is a picture of traditional residential from a stilt house:

And this is a picture of ancient palace from a stilt house:

Above is a multi-storied stilt house usually built for royal palace, but not all ancient royal palaces have many stories like this.


Very good point; the “long houses” could be grand structures most worthy of a king.

Edit: I see Seniya beat me to it!


But one thing I don’t get it, bhante: if ancient India really has stilt houses as their housing structure, why we can’t find ancient stilt houses still being used as family residential or at least ancient royal palaces built from stilt houses as historical remains (as ones in the South East Asia)? Or maybe I just overlooked it…


just a hypothesis

at some point in the ancient history occurred transition to using stone and bricks for construction, especially of temples and palaces

shortage and/or expensiveness of timber, especially for poor population, since stilt houses seem to be mostly wooden


Do we have any wooden structures from that era anywhere else? Wood doesn’t last nearly as long as stone.


I’m really not aware of any wooden structures so old. The building in Haeinsa temple for the Tripitaka Koreana (on which the modern CBETA edition is based) is 750 years old. Which is a lot less, and a much cooler climate.

There’s probably wooden remains of the age of the Buddha, but not much.


I mean that why nowadays we can’t find traditional India people who still use stilt houses as their housing?


You can. Go to Bangladesh, Assam or West Bengal. They are choke full of stilt houses:

I think earlier when people lived in the direct vicinity of large rivers and didn’t have that much engineering know-how, the stilt houses were much more wide-spread in the East India. Maybe, there are still many of the in the rural areas in the Ganges valley, not sure.


Another idiom I came across that reinforces this interpretation. At SN 51.14 several mendicants are said to be staying heṭṭhāmigāramātupāsāde. Ven Bodhi translates this as “on the ground floor of the mansion”. But it makes much more sense for this to mean “beneath the stilt house”.


I’ve been having a look at (rāj)antepura, since it occurs in the Vinaya as well. Having looked at the evidence, I think “compound for royal building” comes close.

Some of the best evidence for this word comes from AN10.45 (which parallels the origin story to bhikkhu pācittiya 83). According to this sutta the rājantepura included the harem, secret deliberations, a father longing for his son, someone being promoted/demoted, the army being dispatched, the trampling of elephants, horses, and chariots. This is more than a mere royal residence. It seems to suggest the seat of government and perhaps a little town in its own right.

This picture is supported by evidence from other passages. Criminals are sentenced in the antepura (Mv.1.43); the king’s home is only one building inside the antepura (Mv p.272); a carriage reaches the gate of the antepura (MN24); the antepura seems to be a place within a town/city (Cv p.184); at MN24 we find: atha kho, āvuso, rājā pasenadi kosalo sāvatthiyā nikkhamitvā antepuradvārā paṭhamaṃ rathavinītaṃ abhiruheyya (BB: “leaving Sāvatthī through the inner palace door”), which seems to mean that the whole of Sāvatthī was the antepura; the King’s assembly being seated in the antepura (AN3.61).

Then there is the compound uparipāsādavaragata, which you also mention. The element vara, which means “best” or “excellent”, has never been satisfactorily explained, as far as I am aware. If the antepura is a whole compound, then presumably it included a number of pāsādas, one of which may have been the best one, perhaps the king’s favourite.

Moreover, the fact that the antepura had a a gate would seem to mean that it was a clearly defined area, perhaps surrounded by walls or at least a fence. According to bhikkhu pācittiya 83, monastics would not normally go for alms to the antepura, yet they would go for alms in Sāvatthī. From this it would again appear that the antepura was a compound within a city, or some such thing, which fits the literal meaning of antepura, “within the city”. My impression is that rājantepura and antepura are used synonymously.

In the end, it s hard to draw any absolute conclusion, but I feel the evidence points to the antepura being a district/enclave/precinct within a city or town. So far I have settled for “the royal precinct”.


two prominent meanings of the Sanskrit antaḥpura are a king’s palace and women’s quarters

and since antaḥ means inside this must be a sort of an inner city


It sounds a lot like the medieval castles. There’d be a village or town spread out around it, and an inner walled area, which was normally used for government business, but in times of war the citizens would temporarily be housed there for protection. I think this is the original meaning of pura, which if I’m not mistaken is cognate with “fort”, “burg” and “borough”, as well as “peri(metre)” and Pali pari-. Originally it would have meant a village with a defensive fence or wall; or perhaps for the Indo-European horse masters, a caravan with the wagons drawn around as a defensive perimeter.

So we can envisage a secure walled compound of some size, within which a variety of activities would be carried out. Given the different layouts, geographies and histories of the various cities, though, it may well have differed significantly from place to place. Perhaps as cities became larger and more secure, the sense of “fort” faded away, and the term became used for the “hidden city”, i.e. the harem.

That seems like a very likely sense for uparipāsādavaragata.

Is this the harem? Or a more general usage?

I agree, although I think “enclave” or “compound” captures the sheltered aspect better. “Precinct” or “district”, at least in modern English, refer to a general administrative region (more like a sīmā).


Yes, something like this. According to OED “fortress” would perhaps be the best word. But I am not sure if it was as fortified as a fortress or even a castle. The word used at AN7.67 for what is really a fortress is nagara. Here we have all the usual characteristics of a fortress, such as a moat and a rampart. I don’t see any indication that the antepura was fortified to this extent.

I did actually consider enclave , but according to my OED it often refers to large pieces of land, even a whole country. However, the context will probably make it clear what is meant. Either “precinct” or “enclave” would be fine by me, but “compound” has too much military association, I feel.


Yes, a nagara in that sense would seem to be more specifically a fortress; since it’s often used of a border town, probably the whole thing was enclosed; I use citadel.

You’re right with enclave, I misunderstood how it was used. But “compound” doesn’t really have a specifically military connotation. Google has:

an open area enclosed by a fence, for example around a factory or large house or within a prison.

Which seems pretty accurate. Incidentally, I didn’t know this, but “compound” in this sense is derived from the Malay kampong. Huh.


The OED also mentioned the the connection to kampong. Perhaps “compound” will work …


Another phrase for ‘stilt house’ is ‘pile dwelling’.

This one has a wall:

This could be the mansion/chateau version: