The 'body' is one of the main objects of attachments, and as an attachment one of the main obstacles in the development of the mind in the Buddhist framework. What follows is less a reflection on the various body-meditations but more fundamentally the implications the choice of words in the suttas has on correctly understanding the context.
kāya in the dictionary
The entry is very long (why can't it just have one meaning??), but the meanings of kāya are basically
group, heap, collection, aggregate, body
This range resonates with khandha, the other major doctrinal term which emphasizes the anattā-doctrine. Where the worldling is tempted to identify with the body-unity and to find in that body-image the cherished 'me', the term kāya brutally goes against this temptation and calls this beloved body simply a heap, a collection. The simile that most clearly exemplifies this - and might well be the root of kāya-as-body - is the 'bag of grain':
Just as though there were a bag with an opening at both ends full of many sorts of grain... so too, a bhikkhu reviews this same body as full of many kinds of impurity...
(MN 10, MN 119, DN 22, Thag 19.1 verse 1134)
kāya in pre-buddhist compositions
This is a simple matter because kāya just doesn't appear in pre-buddhist literature and is a real novelty of the nikāyas! Due to the influence of the Buddhist texts kāya-as-body is later well-established - both in the meaning of 'body' and 'collection' - in the mahabhārata and manusmrti.
Pāṇinī seems to root kāya in ci- with the meanings: to arrange in order, heap up, pile up
How did the Bṛhadāraṇyaka Upaniṣad (BU) and the Artharva Veda call 'the body' then?
Surprisingly mostly ātmān, and if in a pure material sense then śarīra (pali: sarīra), and sometimes aṅga (Pali: aṅga, hardly to be found in the suttas as 'body').
śarīra / sarīra
The Pali term appears a number of times in suttas, and I'm proposing the simplified theory: An almost complete replacement of such a basic term as 'body' didn't come over night, and the Buddha could not have started out calling the 'body' something else, simply because it would have confused his audience. So I see the general use of kāya-as-body as a late developmental stage in the Buddhist texts. Hence suttas contaning sarīra-as-body might represent an old layer of suttas.
I could identify about 20 different original contexts for sarīra-as-body (sometimes with the connotation of corpse or 'empty shell') but would like to single out three contexts that fit into the above theory.
Body vs. soul
Most often we find sarīra in the pericope that summarizes all possible pre-buddhist philosophical positions that a bhikkhu should give up:
‘The world is eternal’ or ‘The world is not eternal’; ‘The world is finite’ or ‘The world is infinite’; ‘The soul (jīva)and the body (sarīra) are the same’ or ‘The soul (jīva) is one thing, the body (sarīra) another’; ‘The Tathāgata exists after death,’ or ‘The Tathāgata does not exist after death,’ or ‘The Tathāgata both exists and does not exist after death,’ or ‘The Tathāgata neither exists nor does not exist after death’.
(AN 4.38, AN 10.20, AN 10.93, AN 10.95, AN 10.96, SN 12.35, SN 12.36, SN 24.13, SN 24.14, SN 33.1, SN 33.55, SN 41.3, SN 44.7, SN 44.8, SN 56.8, SN 56.41, MN 25, MN 63, MN 72, DN 6)
Three things speak for the antiquity of this passage: The use of sarīra for body, and the use of the vedic jīva for soul (instead of attā). And that this passage represents one of the fundamental differences between any pre-Buddhist conception and the Budhha's novel insights.
The second most often used passage with sarīra is when the bodhisatta sat with relentless determination for his last pre-bodhi meditation.
Willingly, let only my skin, sinews, and bones remain, and let the flesh and blood dry up in my body (sarīra), but I will not relax my energy so long as I have not attained what can be attained by manly strength, energy, and exertion.
(AN 2.5, AN 8.13, SN 12.22, SN 21.3, MN 70)
Here as well we cannot understand sarīra as corpse but rather as 'body' in a more general sense. What is striking about this passage is that it's mostly used as a quote, as if it was an intertextual insertion, quoted from another source. This, next to the use of sarīra, indicates to me the older layer of text.
The break-up of the body
In the Dhammapada we come across the two verses 138 and 140 that speak for a transitional period when sarīra and kāya were still in equal use
Dhp 138: sarīrassa va bhedanaṃ
Dhp 140: kāyassa bhedā duppañño
Both lines mean a break-up of the body, and the poet sees no problem in using both terms equally, thereby quoting two different sources for the same expression. A later, more doctrinally oriented author would have chosen one term over the other in both verses, emphasizing either the aspect of the corpse (sarīra), or the collection-aspect of the body (kāya).
kāya vs. ātman
Reading Olivelle's contemporary translation of the BU I was surprised how often he translated ātman as 'body'. My preconception was that ātman generally means 'essence' or 'soul'. And in some cases I would have still chosen that over 'body'. But in most cases it's undeniable that 'body' is the straight forward translation for ātman, and only sometimes means 'essence'/'soul'.
What does it mean then that in the suttas we don't find ātman any more in this meaning? We find it of course as anattā, the negated soul-principle. To me it's very significant that when the BU speaks about the body it uses ātman which neighbors 'soul', whereas the suttas mostly favor kāya which neighbors 'no-soul'. This cannot be a coincidence but must have been originally a doctrinal stand (before kāya became just the general term for 'body').
Compounds with kāya meaning 'collection' / 'group'
In compounds kāya mostly occurs as the first part, but in a few cases also at the end, with the specific meaning 'a group of'. I give more extended references here to show that kāya-as-group is by no means a rare occurance in the suttas.
nikāya - collection, group, assemblage. This is of course the best known example of kāya meaning collection, since every Buddhist pali text collection is called a nikāya. The word doesn't appear alone in the suttas though, only at the end of a compound --> dighanikāya, etc.
asurakāya - an assembly of Asuras (AN 3.37, DN 18, DN 19, DN 21, DN 24, DN 33)
balakāya - (is my favourite in this list) a collection of strength --> an army (AN 3.14, AN 5.133, AN 5.136, AN 6.54, SN 2.23, SN 3.18, MN 92, DN 26)
brahmakāyika - one-with-the-Brahma-group --> one who becomes part of the entourage of Brahma (AN 3.70, AN 4.123, AN 4.125, AN 6.10, AN 6.25, AN 7.44, AN 7.52, AN 7.56, AN 7.69, AN 8.35, AN 9.24, AN 11.11, SN 56.11, MN 31, MN 42, DN 11, DN 15, DN 33, DN 34)
devanikāya - an assembly of devas (AN 2.36, AN 4.191, AN 5.206, AN 7.40, AN 8.29, AN 8.64, AN 9.72, AN 10.14, SN 35.241, MN 16, MN 127, DN 14, DN 20, DN 33, DN 34)
mahājanakāya - a group of people (AN 4.244, AN 8.8, SN 3.10, SN 7.14, SN 35.244, SN 42.6, SN 45.160, SN 47.20, SN 52.8, SN 56.41, MN 77, MN 86, DN 14, DN 16)
hatthikāya - a group of elephants / assakāya - a group of cavalry / rathakāya - a group of chariots / pattikāya - a group of foot soldiers (AN 5.139, SN 3.5, SN 3.15, MN 82)
Implications for the sutta transmission
- It was the Buddha's innovation to refer to the body as kāya. Not just that in the pre-buddhist era kāya was not used (instead mostly ātmān, śarīra or aṅga). But kāya has implications that are connected with the anattā doctrine, since kāya emphasizes not the unity of the body but on the contrary its fragmentation, and hence its unworthiness as an object of identification.
- With this in mind the use of kāya would have been first limited to contexts where the body's assembled nature was emphasized. This would have been made clear to the unaccustomed listener by the context, or a simile like 'the bag of grain'.
- Only later, when kāya became the general de facto Buddhist term for the body (replacing śarīra) it would have been applied to doctrinal and general contexts where the anattā connotation was secondary. It would have been that time when the suttas were fixed and edited.
In a way I used this research to show that the sometimes grim discussions about what 'the real words of the Buddha were' are misguided. To those accustomed to text critical studies it might sound trivial, but I tried to show that we can never know based on the suttas which words the Buddha really used. What we have are the texts portraying the message of the Buddha.
In these texts we can sometimes witness the evolution of terms and a solidification of ideas, or a mythical superelevation of gods and mystical powers etc. These are elements that are worth seperating from older layers.
More specifically I wanted to contribute to the discussion of how fundamentally different the Buddhist view of the body became compared to the mystics of the late vedic period. Whereas in many contemporary languages the original meaning of 'body' 'Körper' etc. are lost, the Indic choice of words was purposeful and resonated the worldview of its saints.