उच्चारण-परिवर्तनम्

परिवर्तनक्रमः

“The orthodox Indian tradition has greatly respected the relationship of a teacher and his disciple and has continued to believe that nothing ever changed in this oral transmission. It is believed that the oral text has been handed down from one generation to the next without any changes. However, we must view a preliterate oral tradition with a fresh linguistic and analytical approach. We must distinguish a “pseudo-connection” from a “real connection” between two successive synchronic states of a preliterate oral text.”
Within the veda-s, one finds great inconsistency in applicaiton of phonetic rules (“प्र णो " इत्यपि दृश्यते, “प्र नो” इत्यपि।) - so that “the preredaction oral traditions were extremely irregular, imperfect, flexible and, therefore, ironically, more “natural” to ancient oral literature, and that a major amount of the phonetic “information” which we find in the existing Vedic recensions can hardly be considered to represent the original compositions.”

प्रादेशिक-परिवर्तनानि

वर्ण-व्यत्ययो ऽन्यत्र द्रष्टव्यः।

Kashmiri īḍe -> ēḍe, purohitam -> purohētam
yajñasya : jña -> Maharasthra dnya, N. India gny,

लल्लावाक्यानाम् उदाहरणम्

“Grierson (1929) has given a fascinating account of how the Kashmiri text of Lalla-vAkyAni, originally composed in the latter half of the fourteenth century, has been preserved orally to the present day without ever having been written down. Grierson (1929:74) points out that “save for a few forms that have remained unchanged…her verses are in what is practically modern Kashmiri.” However, the text of the Mahanayaprakasha was composed in the fifteenth century and was written down at the time it was composed, and hence its language is preserved without any further change. "
“Grierson describes the ‘unconscious’ change taking place in a precodification oral tradition: “Each hymn [of Lalla-vAkyAni] was handed down from teacher to pupil through five centuries, care being taken to preserve the text unchanged. But during all this time the language was insensibly changing, and, as there was no written record of the originals in the form in which they were first uttered, the language of the hymns insensibly changed at the same time. The reciters, it is hardly necessary to point out, were unaware of the change of language that was going on. In each generation that was very slight, and was not noticeable, but the total of the changes at the end of five centuries was very great indeed….It was so gradual that no one was ever aware that any change was taking place at all” (1929:75). Grierson clearly perceived that the same must have happened in the case of the Vedic texts before they were codified by the redactors.”

Loanword approximation

Western European pronunciation of T, D is NOT retroflex (tongue curl).
- “The British had to use hundreds of local Indian words in English. However, they did not pick up the retroflex sounds, but rather approximated them to other English sounds. Thus, the Marathi place name पुणॆ became “Poona” and खड्कि came to be called “Kirkee.*””
- “When native speakers of Indian languages heard English with their “Indian ears,” English alveolars were naturally felt to be closer to Indian retroflexes and were approximated to Indian retroflexes. For example, English ‘table’ becomes tebal in Marathi.”

वेदेषु परिवर्तनम्

“The three volumes of Vedic Variants noted by Franklin Edgerton and Murray Emeneau amply attest to the variants found among the various Veda Śākhās, and some of these are probably attributable to regional variation in the pronunciation of Sanskrit.”
वर्णव्यत्ययः पृथग् उक्तः।
स्वरोच्चारणे विकाराः स्वरभाग उक्ताः।
बहुत्र सूक्ष्म-भेदानां केषाञ् चन प्रातिशाख्यादौ न वर्तते प्रमाणम्।
अजितस्याभिप्रायो ऽत्र। (तत्र जात्यस्वरितादेर् उत्पत्तौ व्याकरणविशेषास् सन्ति।)
न्यूङ्खवचने शस्त्राणि बहुदीर्घ+ओकारेणोच्यन्ते।

ऋग्वेदे पादशो वचनम्

अनुदात्तं सर्वमपादादौ ८.१.१८ इति पाणिनिः।
- तेन विषमपादादाव् अपि न स्वरनिघातः, यथा- “जी॒वात॑वे प्रत॒रं सा॑धया॒ धियोऽग्ने॑ स॒ख्ये मा रि॑षामा व॒यं तव॑ ॥” इत्यत्र। तथा, “the enclitic forms such as ते and मे, as they come under the Pāṇinian restriction of अपादादौ cannot occur at the beginning of any pādas” इत्यपि वर्तते।
It is interesting to note that lots of Rk-s preserved in other vaidika texts, such as naxatra sUkta from TB are indeed split pAda-wise. This can still be heard in current recitations. Ditto in case of sAma reciters (at least the one tradition I’ve heard) chanting Rks such as “आ꣡क्र꣢न्दय कु꣣रु꣡ घोष꣢꣯म् म꣣हा꣢न्तम्… “. They were probably received at a similar time.
Timing. This pAda-wise recitation may have been before the production of current Rk saMhitA as well as kramapATha: “However, beginning with the Kramapāṭha, the half-verses [ardharca] are treated as units, and the end of the half-verse [ardharca] is marked with a repetition of the last word like X iti X.”

वेदेष्वेवोल्लेखः

स-ण-षकारपाठ-विषये ब्राह्मणे चर्चा दृश्यते- मूर्धाभागे दृश्यताम्।

पुराणोल्लेखः

लिङ्गपुराणे - एको वेदश्चतुष्पादस् त्रेतास्विह विधीयते । संक्षयाद् आयुषश् चैव व्यस्यते द्वापरेषु सः ॥ १,३९.५७ ॥ ऋषिपुत्रैः पुनर्भेदा भिद्यन्ते दृष्टिविभ्रमैः । मन्त्रब्राह्मण-विन्यासैः स्वर-वर्ण-विपर्ययैः ॥ १,३९.५८ ॥ संहिता ऋग्यजुःसाम्नां संहन्यन्ते मनीषिभिः । सामान्या वैकृताश्चैव द्रष्टृभिस्तैः पृथक्पृथक् ॥ १,३९.५९ ॥
वायुपुराणे - संरोधाद् आयुषश् चैव दृश्यते द्वापरेषु च । वेदव्यासैश् चतुर्धा तु व्यस्यते द्वापरादिषु ॥ ५८.११ ॥ ऋषिपुत्रैः पुनर्वेदा भिद्यन्ते दृष्टिविभ्रमैः। मन्त्रब्राह्मण-विन्यासैः स्वर-वर्ण-विपर्ययैः ॥ ५८.१२ ॥ संहिता ऋग्यजुःसाम्नां संहन्यन्ते श्रुतर्षिभिः। सामान्याद्वैकृताच्चैव दृष्टिभिन्नैः व्कचित्व्कचित् ॥ ५८.१३ ॥ ब्राह्मणं कल्पसूत्राणि मन्त्रप्रवचनानि च। अन्ये तु प्रहितास्तीर्थैः केचित्तान् प्रत्यवस्थिताः ॥ ५८.१४ ॥
मत्स्य-पुराणे - संक्षेपादायुषश्चैव व्यस्यते द्वापरेष्विह ॥ १४४.१० ॥ वेदश्चैकश्चतुर्धा तु व्यस्यते द्वापरादिषु । ऋषिपुत्रैः पुनर्वेदा भिद्यन्ते दृष्टिविभ्रमैः ॥ १४४.११ ॥ ते तु ब्राह्मणविन्यासैः स्वरक्रमविपर्ययैः । संहृता ऋग्यजुःसाम्नां संहितास्तैर्महर्षिभिः ॥ १४४.१२ ॥ सामान्याद् वैकृताच् चैव दृष्टिभिन्नैः क्वचित्क्वचित् । ब्राह्मणं कल्पसूत्राणि भाष्य-विद्यास् तथैव च ॥ १४४.१३ ॥ अन्ये तु प्रस्थितास्तान्वै केचित्तान् प्रत्यवस्थिताः । द्वापरेषु प्रवर्तन्ते भिन्नार्थैस्तैः स्वदर्शनैः ॥ १४४.१४ ॥ एकमाध्वर्यवं पूर्वमासीद्द्वैधं तु तत्पुनः । सामान्यविपरीतार्थैः कृतं शास्त्राकुलं त्विदम् ॥ १४४.१५ ॥ आध्वर्यवं च प्रस्थानैर्बहुधा व्याकुलीकृतम् । तथैवाथर्वणां साम्नां विकल्पैः स्वस्य संक्षयैः ॥ १४४.१६ ॥ व्याकुलो द्वापरेष्वर्थः क्रियते भिन्नदर्शनैः । द्वापरे संनिवृत्ते ते वेदा नश्यन्ति वै कलौ ॥ १४४.१७ ॥

Role of shixA texts

From vyAkaraNa and prAtishAkhya to later shixA texts, one observes an evolution. At various stages of drift in pronunciation, these texts aimed to “justify” the current pronunciation and freeze it (ie prevent further drift). They were partially successful.

Example:

किमर्थम् अनुवारस्य दीर्घपूर्वस्य पद-मध्ये वर्तमानस्य वज्झभिः+++(??)+++ श्लोकैर् लक्षणं क्रियते । - ननु पाठाद् एव सिद्धं यथान्येषां वर्णानां । सत्यम् । किं तु दुराम्नान-निवृत्त्यर्थं अनयोः शिक्षापटलयोर् बहूनां वर्णानां लक्षणं क्रियते । यथा समापाद्यान्य् उत्तरे षट्-पकार - शनः शेपो निष्षपौ शासिनिष्षालविक्रमान् । एयरित्येकार मकारम् आहुर् इत्य् एवमादौनि ।

कथं पुनर् दुराम्नान-प्रसंगः? सन्ति हि - तीर्थोषिताः अलसाः अगुणदोषज्ञाः अन्योन्याध्यापकाः । सर्ववर्णान्यत्व-जनयितारस् - तद्-दोष-निवृत्यर्थं आचार्येण शिष्यहितार्थम् इदम् आरब्धम् । अनुवारस्य स्थाने तावत् ङकारं जनयन्ति - तस्मात् ङकारात् परं ककारम् अन्तःपातं जनयन्ति - हवींषि, सर्पींषि, भासांसि, अवांसीत्य् एवम्। तन्निवृत्त्यर्थम् अनुस्वार-लक्षणं क्रियते ।

इत्य् उव्वट ऋक्प्रातिशाख्यभाष्ये।

To this day, you find (particularly N Indian) pAThaka-s mispronouncing anusvAra as indicated.

Hasta-svara-s, keraLa-style mudra-s and vikRti-pATha-s serve similar conservative purpose.
“Once personalities like Sakalya and Mandukeya had fixed the texts of their respective recensions, the orally preserved texts were as if quickly frozen with all the changes that had taken place so far, and then texts like PratishAkhyas were composed to describe in detail the features of these “frozen” texts. Staal (1967:17) rightly points out that the PratishAkhyas were “not interested in the Vedic language as such, but in the utterances handed down ” by the oral tradition. However, phenomena like retroflexion had set in (but not completely pervaded) before the process of “text freezing” had begun. The same phenomenon is seen in the application of other sandhi rules in the preserved recensions.”

तकारस्य लकारः

केरळब्राह्मणानां पदान्ते , व्यञ्जनपरत्वे च तकारस्थाने लकारोच्चारणं परम्परया प्रमादेन संजातम् ।

ह्म →म्ह

Many traditionalists transpose ह्म् ह्न् ह्ण् into म्ह् म्न् ण्ह् -

They cite -

तैत्तिरीय-प्रातिशाख्ये -

हकारान् +++(शाखान्तरे लोकप्रसिद्धप्रयोगे वा)+++ नणम-परान् नासिक्यम्॥२१.१४॥ +++(पूर्वं निपात्यते)+++
After h, when followed by n, ṇ, or m, is inserted nāsikya.

If this is not dismissed as a record of later practice, it is to be taken as being a compulsion specific to a particular vedic branch only.
Besides, it itself is proof of the ह्म् ह्न् ह्ण् pronunciation commonly accepted by shiShTas outside taittirIya shAkhA at that time.

यतो लिखित-पाठः कथम् पठनीय इति वक्तुं न प्रवृत्तम् प्रातिशाख्यम् इति तद्विदः।
तेन सम्मतोच्चारणान्तरापेक्षयैवात्र परिवर्तनं विहितं स्यात्।

Some sAmpradAyikas quote a pANinIya-shixA verse in isolation to justify their traditional (ह्म →म्ह) pronunciation -

हकारं पञ्चमैर्युक्तम्
अन्तस्थाभिश्च संयुतम् ।
उरस्यं तं विजानीयात्
कण्ठयम् आहुर् असंयुतम् ॥

but this makes no such transposition recommendation.
On the contrary, a parallel verse in shaunaka shixA indicates that hm hn etc.. were pronounced exactly in that order:

हकारं पञ्चमैर्युक्तं
अन्तःस्थाभिश्च संयुतम् । उरस्यं तं विजानीयात्
परं चेत् पृथग् उच्चरेत् ॥

रङ्गनाथशर्मा यथा -

संयोगश्च हकारस्य एभ्यः पूर्वत्वेन, न तु परत्वेन ।
तद् उक्तं शौनकशिक्षायाम् …
यथा - “रथान् हयाः वहन्ति । "
अत्र नकारात् परो हकारः इति तस्य कण्ठ्यत्वम् एव ।
प्रकृते उरस्यं हकारस्य ।
इमानि उदाहरणानि - पूर्वाह्नः, मध्याह्नः, ब्रह्मा, ब्राह्मणः इति पञ्चमवर्णात् पूर्वस्य ।
…
तत्र वायोः अभिघातस्य शिथिलत्वात्
अत्यन्त-शैथिल्येन हकारस्य उच्चारणम् उरसि भवस्य इति ब्रूमः ।
…
एवं च ब्रह्म, ब्राह्मणः, मध्याह्नः इति हकारं कर्णकठोरं पठन्तः पाठयन्तः च परास्ताः वेदितव्याः ।

So, correct tradition, as per shixA texts alone, is that in ह्म , h is lighter/ less emphasized.