SN 1.36 presents a number of difficulties. Unlike the vast majority of texts in the Sagathavagga it has no reply by the Buddha. Furthermore, the identity of the speaker of the verses is unclear.
There are four verses. The first is attributed to one deva. Then another deva addresses the Buddha with “these verses”. Despite the plural here, the text supplies a close -ti at the end of the second verse. The Mahasangiti edition punctuates accordingly, enclosing the second verse in one set of quotes, and the third and fourth in another; thus three devas are speaking (even though the text specifies two).
Ven Bodhi follows this arrangement without comment. He inserts an attribution of the third and fourth verses to a third deva. But this doesn’t explain why the plural “these verses” is used of the deva who only spoke one verse, nor why the Buddha is absent.
Comparing with the two Chinese parallels at SA 1286 and SA2 284, the text has a decidedly different structure. There, one deity gives an opening verse; the Buddha responds with a series of 5 or 6 verses, following which the deva concludes with a verse in praise of the Buddha. The verse on faith, which opens the Pali text, is attributed not to a deva, but is the second or third verse in the series spoken by the Buddha.
This gives a much more satisfying structure to the text, and it seems to me likely that the Pali text has suffered serious loss. I’m not sure that it’s justified to attempt to recreate the original, but at least we should not impose additional details to try to justify the current form of the text.