But there are so many possible scenarios for the origin and development of the word saṅghāṭi that relying on etymology is bound to fail. It could well be that the word originally referred to patchwork robes. But since the samaṇa tradition existed prior to the Buddha, I would guess that the idea of ascetics wearing robes given by householder also pre-dates Buddhism. The limit here is really our imagination: saṅghāṭi could refer to anything (well, you know). For this reason I think we need to judge the meaning from context.
One such context is the use of paṭapilotika to qualify saṅghāṭi. If saṅghāṭi really did refer to patch-work robes, then paṭapilotika would be redundant. I recognise that this argument is far from being final, since it all depends on the exact nuances of paṭapilotika and saṅghāṭi. But it is at least suggestive, especially if other evidence is pointing in the same direction.
An interesting use is found in MN40:
Seyyathāpi, bhikkhave, matajaṃ nāma āvudhajātaṃ ubhatodhāraṃ pītanisitaṃ. Tadassa saṅghāṭiyā sampārutaṃ sampaliveṭhitaṃ. Tathūpamāhaṃ, bhikkhave, imassa bhikkhuno pabbajjaṃ vadāmi.
Suppose the weapon called a mataja, well whetted on both edges, were enclosed and encased in a patchwork sheath. I say that such a bhikkhu’s going forth is comparable to that.
Here saṅghāṭi does not seem to refer to a robe at all, but apparently to a kind of cloth. This could mean that saṅghāṭi refers to patch-work cloth in general. Or it could just be that the word is used to facilitate the comparison with the monk who is wearing a saṅghāṭi.
By far the most common sutta idiom is saṅghāṭipattacīvaradhāraṇa, so whatever we do, it has to work there. How would you render this?
This compound occurs only once in the entire Vinaya Piṭaka, in what appears to be a late passage about Devadatta in the Sanghabhedakkhandhaka. Elsewhere in the vinaya the roughly equivalent expression is nivāsetvā pattacīvaraṃ ādāya, “having dressed (in the lower robe), he took bowl and robe …” Here cīvara must refer to an upper robe. Or it could refer to both upper robes, since the singular ending refers to the compound as a whole and not to its constituent parts.
An important fact about saṅghāṭipattacīvaradhāraṇa is that it is found only in a single context in the suttas, that is, in combination with the verbs abhikkamati, paṭikkamati, āloketi, viloketi, samiñjati, and pasāreti, and especially in the formula for full awareness. It is possible that in this context the meaning of cīvara is slightly different from the more common context referred to in the previous paragraph. It could well be that cīvara here refers to the lower robe and saṅghāṭi to one or both upper robe(s). It is also conceivable that the compound as we have it now is is a late interpolation, which may have started as a single occurrence and gradually become the norm for this particular context, for it is rather curious that the expression is only found in one context. On the other hand, saṅghāṭi is a relatively rare word in the suttas in any context, and so I prefer the former explanation.
The compound pattacīvara, on the other hand, is very common in the suttas, and most contexts suggest that cīvara here refers to the upper robe(s). A couple of examples: pattacīvaraṃ paṭiggahesuṃ and the very common nivāsetvā pattacīvaraṃ ādāya. However, there are some contexts where it seems cīvara refers to any of the three robes, such as: paripuṇṇaṃ pana te, bhikkhu, pattacīvaranti?
Another interesting passage is the following from the origin story to Sg.8:
Atha kho mettiyabhūmajakā bhikkhū pacchābhattaṃ piṇḍapātappaṭikkantā ārāmaṃ gantvā pattacīvaraṃ paṭisāmetvā bahārāmakoṭṭhake saṅghāṭipallatthikāya nisīdiṃsu …
After the meal they returned to the monastery, put their bowls and robes away, and squatted on their heels outside the monastery entrance, supported by their outer robe.
Again, cīvara can only refer to the upper robes, otherwise the monks would be naked. Moreover, this passage could be read as the saṅghāṭi not being included in “upper robe,” but the evidence is weak.
It seems to me that the precise meaning of many of these terms, especially cīvara, is strongly context dependent. As we have just seen, cīvara often refers to just the upper robes, but in some contexts, such as ticīvara, it refers to all three. And in the compound saṅghāṭipattacīvaradhāraṇa cīvara most likely refers to the lower robe (although it could refer to the upper robe, the uttarāsaṅga, or even both). I am not aware, however, of cīvara ever being used for lay people’s clothes. (But I could well be wrong about this.) So I think cīvara needs a broad translation, and I would suggest simply “robe.”
The lack of definite meaning for cīvara makes it difficult to pin down the exact meaning of saṅghāṭi in saṅghāṭipattacīvaradhāraṇa. But the context here certainly allows for cīvara to mean lower robe, and this seems to be the interpretation of the commentary, which says nivāsanapārupana (com. to DN2). If this is so, and it seems reasonable enough to me, then saṅghāṭi must refer to any upper robe in this context too. This would fit well with how saṅghāṭi is used elsewhere and would avoid the narrow interpretation of “outer robe,” which I feel quite certain must be a later development. This means that saṅghāṭipattacīvaradhāraṇa becomes “using his bowl and his upper and lower robes.”
So perhaps saṅghāṭi should be translated as “upper robe,” especially in the suttas. I am a bit uncertain whether saṅghāṭi ever included the lower robe (as I have suggested above), even at the very earliest stages, and so perhaps this can be disregarded.
In the suttas, by far the most common word for “robe” is cīvara. The words saṅghāṭi, uttarāsaṅga, and nivāsana are rarely used by comparison. A few examples:
Tapode gattāni parisiñcitvā paccuttaritvā ekacīvaro aṭṭhāsi gattāni pubbāpayamāno
Having bathed in the hot springs and come back out, he stood in one cīvara drying his limbs.
Here ekacīvara is used in the sense of “lower robe.”
upasaṅkamitvā ekaṃsaṃ cīvaraṃ katvā
After approaching, he put the cīvara over one shoulder
In the vinaya this is always uttarāsaṅga in place of cīvara. This suggests to me that the vinaya passages are later.
(1) Tasmātiha, cunda, yaṃ vo mayā cīvaraṃ anuññātaṃ, alaṃ vo taṃ – yāvadeva sītassa paṭighātāya, uṇhassa paṭighātāya, ḍaṃsamakasavātātapasarīsapa samphassānaṃ paṭighātāya, yāvadeva hirikopīnapaṭicchādanatthaṃ.
Therefore, Cunda, the kinds of cīvara I have allowed is sufficient to ward off heat and cold, to hinder contact with flies, mosquitoes, the wind, and creepy-crawlies, and to conceal the private parts.
(2) Idhāvuso, bhikkhu santuṭṭho hoti itarītarena cīvarena
In this case a monk is content with any kind of robe.
(3) Idha, bhikkhave, bhikkhu paṭisaṅkhā yoniso cīvaraṃ paṭisevati
In this case, a monk uses a cīvara after reflecting wisely.
(4) Yathārūpaṃ, bhante, cīvaraṃ sevato akusalā dhammā abhivaḍḍhanti, kusalā dhammā parihāyanti evarūpaṃ cīvaraṃ na sevitabbaṃ
Venerable Sir, if unwholesome qualities increase and wholesome qualities decline when one uses a certain kind of cīvara, one should not use that kind of cīvara.
(5) cīvara-piṇḍapāta-senāsana-gilānappaccaya-bhesajjaparikkhāra
The requisites of cīvara, almsfood, dwelling, and medicines for the sick
In all these cases cīvara again seems to refer to any of the three robes. I could go on at length, since this way of using cīvara in the suttas is very common. In fact this seems to be its basic meaning, that is, robe in general.
The ubiquitous use of cīvara to mean any kind of robe in almost any contexts does suggest to me that saṅghāṭi has a more specialised meaning. This, combined with the passage you quote from MN77, could mean that you are right about the saṅghāṭi being a patch-work robe, at least early on. But I still think it is more likely that it simply refers to an upper robe, regardless of how it has been made. This is enough to give it a specialised meaning and thus distinguish it from cīvara, which is more general.
That saṅghāṭi refers to upper robes has the added advantage, I think, that this fits better with how it is used in the vinaya. As in the suttas, by far the most common word for robe in the vinaya is cīvara. So again it seems saṅghāṭi must have a more specialised sense. In the case of the vinaya, however, this specialised sense is more likely to be “upper robe” than “patch-work robe.” In part this is so because of the common juxtaposition of saṅghāṭi with nivāsana and other instances where saṅghāṭi can only refer to upper robes. Moreover, the use of saṅghāṭi in some of the pātimokkha rules (bhikkhunī pārājika 8, bhikkhunī pācittiya 24, and bhikkhunī pācittiya 96) – rules in which the word must refer to robes of any quality – makes it unlikely that it is limited to those made of patches.
I am now leaning towards a two-fold distinction in the meaning of saṅghāṭi: “upper robe” in early usage, and “outer robe” in a few instances of later usage.