Panini, by paying particular attention to sound, was the first to develop the science of phonetics, an inseparable part of philology. Although this was a long time ago, it was not that long ago. By then the Rgvedic and Yajurvedic periods had ended and the Atharvavedic period was in progress. His grammar was based on the Yajurveda, not on the Rgveda, because by then the Rgvedic style of pronunciation had disappeared. For this reason he encountered a good deal of difficulty.

Panini developed special rules for the use of ńa and śa based on phonetics. Earlier I pointed out that the Yajurvedic style of pronunciation was different from the Rgvedic style of pronunciation. Take, for example, the pronunciation of ńa and śa. In the case of Panjabi-speaking areas and nearby areas (in Paeshácii Prákrta-speaking areas) kśamá is pronounced khamá, not kśamá. This is the Yajurvedic style. Panini was from Gándhár (nowadays partially in Pakistan and partially in Afghanistan). The spoken language of this region during Panini’s time was Páshcáttya Prákrta. It was Panini’s mother tongue and the ancestor of the modern Pashto language.

Panini observed that in certain areas the pronunciation of ńa had metamorphosed naturally into ŕhan or nan. He accepted this natural pronunciation of ńa. For example, vipańi, báńijya, kalyáńa, and so on, are all pronounced in the Yajurvedic style.

Panini observed that if a word contained ra, r or śa then due to their influence the dental na would be pronounced like the cerebral ńa. Taking phonetics carefully into account, he legitimized this as the ńa rule, for example, dhárańá. Here he acted according to Yajurvedic phonetics, not Rgvedic, because by then the Rgveda was dúrast [remote].⁽¹⁾ In the Rgvedic style dhárańá becomes dháraŕhán. Many pandits pronounce dhárańá incorrectly.

Panini fixed certain rules in this regard by using psychologically sound means. If, in any compound word, ra, r, or śa are present in the first word then the na in the following word is often not affected. For example, harináma. Here there is a break between the two words, hari and náma. For this reason the ra of the first word hari does not affect the na of the second word náma. That is, the ná remains na. However, if the first word contains the prefix pra or pari then the na in the next word becomes ńa if it follows immediately afterwards. For example, parińáma, prańáma, prańaya, etc.

In some cases ra, r, or śa in the first word does not influence the following word. However there are some sounds in the following word whose pronunciation is easily affected even if separated by a gap.⁽²⁾ If the subsequent word contains a sound which extends the influence of the first word over the second then in such cases the ra of the first word affects the second word and the na changes into ńa. Some of these kinds of words have become exceptions – ayana (uttaráyańa), ahna (aparáhńa, púrváhńa, but madhyáhna), úhinii (akśaohińii), grámiińa, sarváungiińa, súrpańakhá.

In some cases the ra, r or śa of the first word does not affect the second word during pronunciation, for example, parinirváńa. If pra comes before the verbal root nash then it does not have any significant effect, thus prańásha can be written with either na or ńa.

If a vowel or light consonant falls just after or between ra, r or śa then it affects even a distantly situated na. In this case, the rule is that ra, r or śa affects the na even if it does not immediately follow it. For example, váśpáyańa. In the case of vowels, ka varga, ya, va, and ha the distant na becomes ńa even if they are situated far away and even if any of them occurs more than once.

Panini framed the rules of ńa and śa in such a way in his Sanskrit grammar that it is cent-per-cent acoustically-based. But it is worth pointing out that this style of pronunciation was already in existence, that is, people used to pronounce it that way. All that Panini did was to give this style of pronunciation a systematic form. He brought it within the scope of a formula.

The Bengali language differs from Sanskrit in many areas. Bengali has only the sha pronunciation, not that of śa or sa. However, although there is no separate pronunciation of śa, in special circumstances there is a slight difference in pronunciation. For example, śańd́a, rákśasa. In one case the śa is pronounced one way – śańd́a – and in the other it is pronounced differently – rákśasa (k + śa = kśa (kha)). Sa only survives as a written letter; it doesn’t have any separate pronunciation. Some people incorrectly pronounce words spelled with sha with a type of distorted sa sound. People like to make jokes about this. For example, syámavájárer Sasiivávu sasá khete khete sedin sásácchilen [Shyambazaar’s Mr. Shashii threatened while eating a cucumber that day].⁽³⁾ Another example of how Bengali and Sanskrit have deviated from each other in many instances in terms of pronunciation is that in Sanskrit śa can be substituted for sa when a verbal root beginning with sa follows the prefix ni. Thus in Sanskrit both niśáda and nisáda are correct (ni – sad verbal root suffixed by ghaiṋ). In Bengali only niśáda is used. If anyone spells niśáda with sa then some pandits may very well mark it wrong. The same goes for upaniśada (upa – ni – śad verbal root, kvip suffix = upaniśada in Bengali). In Sanskrit both upaniśada and upanisada are correct. Bengali has the letter sa but no separate pronunciation for it, so it is avoided as much as possible.

The suffix shatr has been used after the parasmaepadii verbal root to signify yá cale caleche [that which has been going]. At the time of the Rgveda there was a clear line of demarcation between parasmaepadii and átmanepadii. Whenever one’s own self was the main subject then átmanepadii was used and whenever another was the main subject then parasmaepadii was used. Wherever self [átma] or para [other] was not made clear then it was considered ubhayapadii. Átmanedpadii was used extensively in the Rgveda but its use diminished in the Yajurveda. Its usage disappeared for all intents and purposes in the Atharvaveda. Panini’s mother tongue was not Vedic Sanskrit, however he had to base his division of átmanepadii and parasmaepadii verbs on the Vedas, so he met with some difficulty in this regard.

Often it is difficult to determine which is átmanepadii and which is parasmaepadii. Take, for example, the verbal roots cal [to go] and ji [to triumph]. In the Rgveda they are átmanepadii, for example, jayate – satyameva jayate [truth alone triumphs]. However in modern Sanskrit the verbal root ji is parasmaepadii, thus jayati. Of course, after a prefix the verbal root ji is átmanepadii, for example, vi – jayati = vijayate. Due to these kinds of difficulties he accepted the use of ubhayapadii.

Vopadeva only accepted the use of parasmaepadii. Hence shatr will be added to cal, not shánac (to indicate present participle), because in Laokika Sanskrit the verbal root cal is parasmaepadii. However, after a lot of thought I finally recalled a Sanskrit shloka from the time of the Buddha, twenty-five hundred years ago, which uses the verbal root cal in átmanepadii (in the laokika sense). The example is:

Ihásane shuśyatu me shariiram
Tvagasthimáḿsaḿ pralayaiṋca yátu
Aprápyabodhiḿbahukalpadurlabháḿ
Naevásanát káyamatashcalisyate

Thus Panini was forced to accept ubhayapadii because cal verbal root + shatr = calat. For example, calat + citra = calaccitra. And cal + shánac = calamána. Both are correct. Bear in mind that shatr, shánac, átmanepadii, parasmaepadii, short, long and extended vowels, pronunciation – all of these change as the era changes. They have done so in the past and will continue to do so in the future.

In Sanskrit the suffix kvip is used in the sense of “natural inclination”. For example, gam + kvip = jagat [world] or “that whose nature is to move”. However, where it is haye caleche [in continuity] then the suffix is shatr. For example, gam + shatr = gacchat [going]. In the case of shánac, since it is átmanepadii then shánac is okay even if it does not preserve much rhythm. That is, shánac has a light rhythm. Either mán or án is added to shánac. Generally here, when the original verbal root ends in a consonant (for example, sev), then the third person dual ete is used in its lat́ (sevete).⁽⁴⁾ In this case, in place of this ete, mána is used with the aforementioned verbal root sev (sevamána). If the original verbal root ends in a vowel (for example, shii), then in lat́, third person, dual number, áte is used (shayáte). In this case, if shánac is added to the verbal root shii, then ána (shayána) is used in place of áte. In the case of svabháva [natural inclination] and in the sense of continuity, the two different forms of the same verb are not used in Sanskrit, but they are with the verbal adjectives. In English the verbal adjectives are formed by adding the suffixes “ing” and “ed”, etc. In the Indian language Bhojpuri, however, there are separate verbs for the verb forms in the sense of the natural inclination and for haye caleche. For example, Rámka bhái paŕelá and Rámka bhái paŕatá.

In this way Panini was able to do everyone a service by creating a harmonious adjustment amongst grammar, phonetics and rhythm, something which was not easy to do. Vopadeva added lightness, but in order to make it easy to add this lightness he ended up making the grammar complex. Panini’s composition and formation of grammar was quite strong and beautiful; it has a scientific framework. There is no latitude [jo] given it to meander here and there.⁽⁵⁾

Now let us see what happens with this phonetics in Arabic. In the majority of the European languages “h” at the beginning of a word is not pronounced. For example, “hotel”. In French it is pronounced otel. However in Arabic it is not silent. Sometimes it is pronounced and sometimes it is partially pronounced. In Arabic, if there is a vowel or an “h” after a word then the sixth case declension ul is inserted in place of a pronounced or unpronounced “h”. Where there is a clearly pronounced or silent consonant then the form of the sixth case declension is changed consistent with the sound of the consonant. For example, phajal + ul + rahaman = phajalur-rahamán. Here the consonant ra is partially pronounced thus we cannot say ul; it becomes phajalur. However, ájij + ul + hak = ájijuahak because the consonant ha in the word hak is fully pronounced. Thus the l is retained here making ájijulhak. Najar + ul + islám = najarulislám because the vowel l is clearly pronounced. However, ájim + ul + sán = ájimusyán because here the consonant sa is clearly pronounced, thus ul is not retained. In the same way, sháms + ul + náhár = shámsunnáhár.

Grammatical aphorisms are put together according to proper and natural pronunciation for the sake of phonetics. The style of pronunciation is created beforehand. Through careful listening the grammarians record this pronunciation at the syllabic level. For this reason all the languages of the world are bound together on a single thread; in the absence of phonetics no language can move a single step. Bengali has Turkish words as well, along with Arabic and Persian. Careful attention must be paid so that these words can be pronounced properly.

In all languages pronunciation changes in order to keep pace with vowel sounds. For example, in English there are twenty-one different pronunciations of the letter “a” but in French there is only one, the Bengali á. In French “Paris” is pronounced párii – the “s” is silent because the rule in French is that in most cases the final consonant is not pronounced at the end of a word; if the final consonant is “c”, “f”, “l”, or “r” then it is pronounced. To facilitate their pronunciation a vowel sound is added afterwards. Many people mistakenly spell the French word madame “madam”. If madame is spelled in French without the final “e” then its pronunciation becomes mádán. It has to be spelled with an “e” at the end in order to preserve the correct pronunciation. In French the masculine singular for “my” is mon and the feminine singular is ma. The plural is mes, thus the plural for madame is mesdames. Some people mistakenly read this as “may-s-day-m-s”.

“A” in English is a light vowel, thus if a root verb occurs after the prefix “a” then the first consonant is doubled. For example, “a + quire”. Here “quire” is the root verb and “a” is the prefix. In this case the “q” should be doubled, however since double “q” cannot be pronounced, a “c” is substituted instead; it becomes “acquire”. The same is the case with “acquisition”, but in the case of “require” and “requisition” the consonant is not doubled because “e” is not a light vowel.

This rule is followed in Sanskrit. It is doubled in accordance with the rules of phonetics. For example, tarań + apashyat = tarańńapashyat, that is, “he saw [dekhle] while swimming”. (Now, tell me, will it be dekhla or dekhle in the spoken language of Calcutta? It will be dekhle. In Rarhi Bengali it will be dekhlek in this case. Bear in mind that if the final k is dropped from the Rarhi Bengali it becomes the Calcutta Bengali. Of course, in this case the transitive and intransitive verb forms are different. Similarly, the spoken form of se balila [he spoke] will not be se balla but rather se balle.)

In French the masculine form of the definite article (the) is le, the feminine form is la and the plural form is les. However here the pronunciation of this le, la and les changes where there is elision. In such cases the spelling practice becomes paramount. For example, “the men” in French is not le am but lezam [les hommes].

The English word “knife” comes from the French word kanife. Here the English letter “k” is silent. However, since “knife” comes from the French word kanife, it is still spelt with a “k”.

“H” at the beginning of a word is silent in French, for example, “history” – the French pronunciation is istoire. “Hotel” in French is pronounced otel, however in English the pronunciation of the “h” is maintained, thus it is pronounced “hotel”. “Hospital” in French is pronounced opital, but in English it is “hospital”. However, in the case of “honour” the French style of pronunciation, onár, has been preserved in English. Even uneducated French people follow these natural pronunciation practices.

Not only is the pronunciation and spelling of prefixes influenced by the style of vowel pronunciation, suffixes are as well. The “pre” of “prefix” is a Latin prefix and “fix” is a verbal root. Thus “pre” + “fix” makes “prefix”. In Sanskrit there are twenty prefixes. “Suffix” is derived from placing “sub” before “fix”, that is, “sub” + “fix”. Here the “b” becomes an “f” because “b” is a medium consonant, thus “suffix” is spelled with two “f”s. The first “f” of “suffix” is a substitute for “b”; the second “f” is a hard sound.

If there is a suffix added to the root word then it is called taddhita, and in the case of a verbal root with an added suffix it is called krtprakarańa. In this krt or taddhita process of adding a suffix to a root word or verbal root, the original root word or verbal root occasions, in some cases, a change in consonant, and sometimes a change in vowel. For example, sarvajana + iin = sarvajaniin. If śńik is added to sarvajana it makes sarvajanik but instead it is pronounced sárvajanik.

Sometimes in the roots of Arabic words, if the verbal root is made into a noun then mu is added and a change occurs in the second consonant. For example, musháfir means “one who makes a journey [sháfar]”. In this case, according to the rule, mu is added before shafar and it becomes musháfir. Mu + shafar = musháfir. Similarly, mu + hájir = muhájer; mu + sáhib = musáheb. This change in the second is not a product of grammar but of the people. When Arabic arrived in India the Koráńasharif had already been written down. Arabic grammar came into being after the composition of the Koran. There are a great number of Arabic words in Bengali but they came much later.

These kinds of rules also exist in English. There is an English rule that when there are two vowels then the short vowel follows the long vowel. In Bengali also it is priiti → piiriti (not piriiti). Sanskrit also follows this rule. Panini said that whenever a consonant follows which requires the vowel to be lengthened in order to facilitate its pronunciation then the vowel becomes long. For example, cakśu + roga; the spelling will be cakśúroga. Similarly, ni + rava = niirava. The ra of cakśúroga is a hard ra. In the case of a light ra this does not take place.

When “i” and “e” are placed side by side in English then the “i” comes first because its pronunciation is longer and stronger than that of the “e”. The pronunciation of “i” is always i, but there are many pronunciations of “e” – i, e, á – thus “e” comes after “i”. However, there can be exceptions after a light consonant, for example, “ceive”. Here “c” is the original consonant, thus “i” comes after “e” in the verbal root “ceive”. This rule remains in force when the verbal root “ceive” follows different prefixes, for example, “receive”, “perceive”, etc. Similar is the case with “leisure”, “counterfeit”, and so on. Hindi also follows some of these same rules, for example, laŕkii, but laŕkián.

Footnotes

(1) The two Farsi words, dúr and ast are both Vedic in origin. The Farsi dúr comes from the Vedic dúraḿ; the Farsi ast comes from the Vedic asti.

(2) The people of Calcutta say gechi [“have gone-present perfect”]; people from Nadia say giyechi. In giyechi the ye collides with the cha. However, in gechi there is no ye to collide with it. From gechi comes gechlum [preterite tense] (Calcutta); from giyechi comes giyechilám (Nadia). I find gechlum to be more advantageous because it can be pronounced more rapidly, thus time and effort are saved.

(3) The play in sounds comes from replacing the initial shas with sa. –Trans.

(4) Sanskrit recognizes three variations in number: singular, dual and plural. Lat́ is present indicative or present continuous. – Trans.

(5) The word jo means “proper environment”. Farmers ask: Does the land have jo or not? In pure Bengali it is jo; in Rarhi Bengali it is batar – not bátar. Bátar means “tendency to itch” – in Sanskrit, kańd́úyana. In Sanskrit its meaning is “means”.