सन्ध्यक्षर+उच्चारणम्॥
प्राचीना स्थितिः
पुरा - “एओ=अइ=अय्/ अउ=अw, ऐऔ → आइ=आय्/ आउ=आw”) इति।
- Paṇini-shixA has also asked to pronounce it ai, “अर्धमात्रा तु कण्ठस्य एकार+ओकारयोर्भवेत् …”
Any sonorant consonant can comprise the second part of a complex syllable nucleus; all can form diphthongs with any of the vowels *e, *o, *ē, *ō (such as *ey, *oy, *ēy, *ōy, *ew, *ow, *em, *en, etc.).
Proto-Indo-European phonology wiki article
परस्पर-विवेकः
आर्यभाषासु, यथा मम ग्रामभाषायां ba[y] (स एव), bā[y] (महिला) इत्यादयो बहवः शब्दाः सन्ति यत्र a[y]-ā[y] भेदेन a[w]-ā[w] भेदेन चार्थभेदोऽपि भवति। - उज्ज्वलः।
प्लुतशङ्का
आइ आउ इत्य् उच्चारणं प्लुतं भवेद्
इति सन्देहश् चेत् - “आय् आw” इति व्यञ्जनोत्तररूपेण ग्रहणेन परिहारः
सन्धि-निरूपणम्
- राम + इति → रामेति, चल + उत → चलोत
इत्यादि सन्धयः अय्-अव् उच्चारणेन सङ्गच्छन्तितराम्। - हरौ + इति → हराव् इति/ हरा इति, कुर्वै + इति = कुर्वाय् इति/ कुर्वा इति
इत्यादि सन्धयः आय्-आव् उच्चारणेन सङ्गच्छन्तितराम्।
पर-रूप-सन्धिः
वैदिकभाषायां रामो अत्र इत्येवं भवति सन्धिः। तत्र च ओकारो लघुर्भवति च्छन्दोऽनुसारेण। rāmaḥ atra > rāma[w] atra इति सरलतया प्रतिपाद्यते। छन्दसि लघुभावश्चापि युज्यते। ya[w]gaḥ इत्यत्र गुरुभावश्चापि स्पष्टो व्यञ्जनद्वये सति।
- उज्ज्वलः।
There are roughly about 90 अलोप अनुवाकंs in entire krishna yajur veda. Certain anuvakaas are termed as नालोप अनुवाकं where sandhi must happen. - KN Ramesh
सामगान-विकाराः
- अग्ने → अग्नाइ। एतादृशा विकाराः प्राचीनम् उच्चारणं सूचयन्ति।
ए-ओ-कारयोर् ay-aw-उच्चारणे स्तोभेषु लघ्वी समस्या ।
- मार्गीयवे यथा औउहोइ इत्यस्य आउहऊइ इति वचने उकारय् उचितः स्वरः स्यात्।
नव्योच्चारणम्
By the time of patanjali (classical saMskRta) the ए ओ were monophthongs, but ऐ औ were diphthongs:
किं प्रयोजनम्? Why ?
एचो ह्रस्वादेशशासनेष्वर्ध एकारोऽर्व ओकारो वा मा भूत् इति
So that half e’ and ‘half o’ may not be used when the rules relating to the shortening of long vowels operate.ननु च यस्यापि तपराणि तेनाप्येतद्वक्तव्यम्; इमावैचौ समाहारवर्णौ मात्रा- वर्णस्य मात्रेवर्णोवर्णयोः । तयोर्हस्वादेशशासनेषु कदाचिदवर्णः स्यात् कदाचिदिवर्णोवर्णो, मा कदाचित् अवर्णः भूत् इति
Oh! this has to be read even when one reads the diphthongs with at the end. The two letters which come under the pratyaharu are diphthongs having one matra for a and another matra for i or u. Hence when rules relating to shortening operate, sometimes a may be substituted and sometimes i or u. They have to avoid the substitution of a.
I don’t know about the RgvedaprAtishAkhya or the Rktantra - they would be prior to patanjali, but also when classical sanskrit had come to be spoken… patanjali (already classical) mentions that people from the राणायणीय stream pronounced a short version of ए. A short length vowel pretty much cannot be a diphthong. If this sect had been pronouncing a short version of ए since ancient times, that would argue against a diphthong ए.
In Pāṇini’s time the *āi > ai & *ai > e (also *āu > au & *au > o) chain shifts had not yet happened in sanskrit (which with its several dialects, was the commonly spoken bhāṣā of the day, even after being standardised by Pāṇini).
Also to the redactors of the Pali canon who (living not too distant from Pāṇini in time) were very likely not schooled in phonetics or morphology, the *āi and *ai would have sounded too similar to need separate akṣaras in the script so both were represented by the same letter 𑀏 (brāhmī symbol for ’e’). Thus while devaiḥ was pronounced *daivāiḥ by Pāṇini, it was pronounced in much the same way by the redactors of the pali canon, but written as devehi (where both e were expected to approximate the unified ai phoneme of the unschooled person.
अर्वाचीनेषु भ्रमः
- पेजावर-माध्व-मठस्य विश्वप्रसन्नतीर्थो ऽत्र।