The History of the Bengali Language/Lecture 8
SOME FACTORS RELATING TO PHONOLOGY AND ACCENT
LECTURE VIII
Section 2
Bengali Metrical System
We should do well to proceed now to ascertain the character of the Bengali accent of olden days, by examining the metrical system, preserved in the poetical works of old Bengal. Adverting to the fact, that the early Vaiṣṇava poets of Bengal treated each and every letter as a syllable, and made the final consonantal sounds non-hasanta, by imitating the old fashionable poet Vidyāpati of Mithilā, many people have formed two wrong notions; one is that Bengali was derived from Maithili, and the other is that our mode of pronunciation and of counting syllables, was of the type presented by those poets. Mithilā became no doubt, at one time, a portion of old Gauḍ which extended to the foot of Nepal if not into Nepal itself, but the Maithili speech of the days of the Vaiṣṇava poets, had nothing to do with our Bengali language. As to elements, common to Bengali and Maithili, we have to look to the older Māgadhi speech of which, notice will be taken later on. From the earliest known time, our Bengali poets (excepting those who followed the Maithili fashion) have uniformly composed their poems, not by counting letters but by counting syllables. Looking to the fact, that the fourteen letters of the পয়ার verse for example, are the same as fourteen mātrās of fourteen syllables, the পয়ার may be seemingly regarded as composed of fourteen letters; but that it is a syllable (which may consist of more than one letter) and not a letter which is the unit in a Bengali word, should never be lost sight of. How is it that consistently with or rather in harmony with the phenomenon, that our metrical system is grounded on a syllabic and accentual basis, a fixed number of letters come in a chanda, will be explained presently, after considering some facts leading to the point.
To explain the character of Bengali syllables, let me set forth some words with their syllable divisions by marking the syllables off, by the sign hyphen: মা-নুষ, আ-সা-মের, বন্-ধু (বন্ধু), অ-হং-কার, etc., will show that there may be less number of syllables than the number of letters in a word. Mr. J. D. Anderson has rightly remarked with reference to our phrase accents as well as in respect to the syllables in a word, that this special aspect of phrase accent in a word, "is sufficiently dominant to be the basis of accentual verse in Bengali." Mr. Anderson has very successfully demonstrated what I once feebly pointed out in a Bengali essay, that the assertion of our Paṇdits, that the পয়ার metre is not composed of syllables but of fourteen letters, is wrong. The two lines quoted by Mr. Anderson from Kṛttivāsa in J. R. A. S., 1913, 861, may be cited to show, that the lines of fourteen letters consist of twelve and thirteen syllables respectively, and that the verse moves on with syllables and not with letters. The lines, as accented by Mr. Anderson stand as:
র্এক-দিন র্শু-ক্র গে-ল ত-র্প-স্যা ক-রি-তে
হে-ন কা-র্লে র্দণ্-ড রা-জা গে-র্লেন র্প-ড়ি-তে।
It will be noticed that it is the accent on লেন of গেলেন in the second line, which has given the easy motion required by the rhythm, and by virtue of accents, lines of seemingly unequal syllables have agreed in the metre. I adduce now another example, to illustrate our metrical system. Take first a line of verse which accommodates fifteen letters which coincide with fifteen syllables:
তুমি অতি শিশু ছেলে কোথা যাবে একাকী?
Let us then notice, that maintaining this very metre, lines of fifteen syllables may be introduced in a verse, though counting by letters, the lines may contain twenty or twenty-two letters:
আ-শ্বিন মা-সের ভো-রের বে-লায়
- বা-গান ত-খন ফুল-প-রা
- বা-গান ত-খন ফুল-প-রা
স-তেজ শ্যা-মল ত-রুর ত-লায়
- গ-ঙ্গা ছি-ল কূল-ভ-রা।
To indicate cæsura, I have divided the lines above keeping eight syllables in the first foot. The first line of the verse contains twenty-two and the second twenty letters. Refer to whichever poet of the past time you please, excepting the Vaiṣṇava poets, who imitated non-Bengali forms, you will find that our verses are all based on accentual basis. I quote here only some lines from Dāsarathi Ray and Iśvar Gupta, who preceded our immortal poet Madhusudan.
খেয়ে কিনা মোচার ঝালঝাল লেগে বানরের পাল
আপ্নার গাল আপ্নি চড়া চড়ি।
If read according to the natural accents of the words, it will be seen that the syllables being taken as units, there are eight distinct sounds of eight letters in each of the first two feet, and sounds of ten letters occur in the third, fulfilling the requirements of number of letters for such a ত্রিপদী metre. Next after this verse of Dāsarathi, a verse of Iśvar Gupta in our indigenous রেক্তাছন্দ may be noticed.
"নানারকি নানা কেলে
নানারকি নানা কেলেরাজ্য পেলে
তাতেই এত জারি।"
Before proceeding to show, how from our indigenous songs which unmistakably disclose our accent system of old times, not only the literary Bengali verses but the Sanskrit verses as well originated, let me notice here the wrong opinions of some eminent Bengali writers, regarding the character of our versification. It has been wrongly held by some, that in the old poetry of Bengal, হসন্ত words practically do not exist. It has been wrongly asserted, that in our old poetry, natural accent was not cared for and the "unnaturalness of recitation was made up for by chanting the verses to a tune." This is only true of those writers, who imitate the old Vaiṣṇava poets in the matter of their versification and diction, that an adventitious artificial jingle has been introduced.
In addition to what I have quoted from the old poets, I quote below another verse from Kṛttivāsa to substantiate my statement and to show the incorrectness of the opinion just referred to.
অতিকায় পড়ে রণে রাবণ চিন্তিত
যোড়-হাতে বাপের আগে কহে ইন্দ্রজিৎ।
The word চিন্তিত rhyming with ইন্দ্রজিৎ leaves no room for even a careless reciter to articulate its final ত with a vowel sound; that যোড় and বাপের carry normal pronunciation, is absolutely clear. The supporters of the contrary opinion are found even to name অন্নদামঙ্গল of Bhāratchandra in the list of the delinquents, without noticing that if we leave a few artificial Sanskritic verses out of consideration, Bhāratchandra must be given the credit of having composed verses with words of common use with their natural accent. অন্নদামঙ্গল as well as his other works, abounds with verses like:
"ঝাঁকড় মাকড় চুল্, নাহি আঁদি সাঁদি,
হাত্ দিলে ধূলা ওড়ে যেন কেয়া-কাঁদি।"
When, more than fifty years ago, we read Madanmohan's poem পাখী সব, etc., we were not asked by our village teachers to deviate from the usual pronunciation, and did never read সব, রব, কুসুম, রাখাল, গরুর পাল, শিশুগণ দেয় মন, etc., with final অ sound.
Despite the fact that Madhusudan has drawn largely upon the Sanskrit vocabulary for some effect in the blank verse, his verses have to be read by putting proper accents upon the words. Is it not true, that far from imposing an adventitious artificial jingle upon our verse, and far from throttling our natural accented words to death, Hemchandra has made the Bengali verse to move with natural accent, with uncommon vigour and rapidity? I think I have made it clear, that Bengali words with their accent have always been used in our Bengali verse, and the exception to the rule has only been occasional, where there has been an imitation of the forms of some Vaiṣṇava poets. The imported metrical system of some Vaiṣṇava poets could never take root in Bengal.
I proceed now to show from the history of evolution of our metrical system, that the accent system which now prevails, has been the accent system of the Bengali-speaking people, from the time about which faint suggestions can only be made, by taking a stand at the shadowy threshold of dim past. The history of the evolution of human institutions has made us familiar with this phenomenon, that our verse with our metrical system owes its origin to tribal festive songs, which in their turn originated from primitive expressions of emotions. In its normal condition, therefore, no metrical system of any race, can have any other basis than an accentual one.
To ensure convenience, I refer first of all to such Sanskrit metres as are of undoubted late origin. মানবকা-ক্রীড় is a ছন্দ of late origin and its origin in Magadha-Gauḍa cannot be seriously doubted. The hemistiches of one line of this verse are divided below by partition lines:
আদি গতং । তুর্ষ গতং ।পঞ্চ মকং । চান্ত্য গতং।
Each hemistich is really a complete foot, and the characteristics of it are repeated in subsequent hemistiches. A portion of our nursery rhyme will be seen to be exactly in accord with it. It is as follows:
হাত ঘুরুলে । নাড়ু দেব ।নৈলে নাড়ু । কোথা পাব ।
The apparent inaccuracy in the second hemistich of the first foot disappears if the সুর or tune underlying the Sanskrit as well as the Bengali verse is rightly caught. Uniformity in Sanskrit metre is maintained by the fixity of long and short sounds, while without following the Sanskrit rule, mere tune may maintain the purity of the metre with natural accent in the Bengali verse. Compare the same ছন্দ in another nursery rhyme:
ডালিম গাছে, ফিলপু নাচে;তাক্ ধিনাধিন্ বাদ্যি বাজে।
It will be seen that how the four akṣaras required in each hemistich in Sanskrit, correspond exactly to the four syllable-unit of Bengali. No one will venture to say, that our village girls or matrons imitated the মানবকাক্রীড়; that the Paṇḍits utilised the indigenous chhanda for a Sanskrit metre verging upon মাত্রাবৃত্ত, cannot be doubted. When songs were composed with mātrās, numerous chhandas cropped up in Sanskrit, and the verses were set in indigenous tunes. To illustrate this properly, I take a verse of a very familiar song from the গীতগোবিন্দ. I divide the lines for the purpose of my analysis, and put the fag portions in brackets.
বদসি যদি । কিঞ্চিদপি । দন্তরুচি । (কৌমুদী)
হরতি দর । তিমিরমতি । (ঘোরং) ।
স্ফুরদধর সীধবে । তব বদন । (চন্দ্রমা) ।
রোচয়তি । লোচন । (চকোরং) ॥
প্রিয়ে চারুশীলে etc., as ধূয়া or refrain.
It is first to be noticed, that each portion divided by partition lines consists of either five letters or five mātrās. The beginning of the refrain portion if divided similarly, a great artificiality will be noticed, since the first division will take in only the first three letters প্রিয়েচা; but if sung according to the tune, this unnaturalness will disappear, and the whole refrain will be found to be set in music with all regularities. Compare with it the line composed in Bengali রেক্তাছন্দ, already quoted above, and is quoted again, for facility of reference:
নানারকি নানাকেলে—
নানারকি নানাকেলে । রাজ্য পেলে ।
তাতেই এত । জারি ।
If we exclude the introductory নানারকি নানাকেলে which is prouounced as অনুদাত্ত, and if we set apart the word জারি as a tag, the essential agreement between the Bengali metre and the Sanskrit metre, will be obvious; the word জারি if pronounced with lengthening sound as is done in reciting a verse of the রেক্তাছন্দ, its agreement with ঘোরং, will also become clear. In রেক্তাছন্দ there is an introductory portion which is of peculiar nature; the first portion of the first line becomes the independent introductory portion of the verse. The introductory line সে যে গো পুষ্যি এঁড়ে, must first be articulated as অনুদাত্ত, and then it is to be repeated as উদাত্ত with the other portions of the verse. The verse then will stand thus—
সে যে গো পুষ্যি এঁড়ে—
সে যে গো পুষ্যি এঁড়েদস্যি ভেড়ে
নস্যি কর তাকে ।
That the refrain portion, therefore of the Sanskrit sons, originates from the introductory tag, becomes rather clear. No doubt there has been some lengthening of the tune in Sanskrit in the refrain portion, but this is because a uniformity of the metre has been maintained. To explain the matter more clearly, let me notice here a verse which has been composed by exactly adopting the metre of Jayadeva's song বদসি যদি, etc.; it will be observed in this verse that the essential character of the chhanda has not been affected, even though there has been either a little lengthening or a little shortening of some mātrās, in conformity with the genius of the Bengali metre. The Bengalicised Sanskrit verse runs as:—
তেমন-ইকি আসে ঊষা সে সোণালি সুষমায়,
- সাজায়ে শ্যামল দেহ শরতের?
শুনি যবে পাখীদের আনন্দের ঘোষণায়
- ভেঙ্গে যায় নীরবতা জগতের?
If the portions হরতি, etc., and রোচয়তি, etc., of Jayadeva's song, and the corresponding lines of the Bengali verse be put aside, the following Bengali verse composed after an indigenous Bengali metre, will be found to be in perfect agreement with the chhanda in question:
ঠাকুরমা, সেই ছেলে বেলায়, ঘুম পাড়াবার ফন্দিতে,
এক যে রাজার মজার গল্পের হুঁ হুঁ জোড়া সন্ধিতে
এম্নি ক'রে ঢেলে দিতেন নিদ্রালসের আবল্লি,
নেতিয়ে পড়্তে হ'তই ঘুমে, রাজারাণী যা বল্লেই ।
For an example of a Sanskrit metre, corresponding to or agreeing with the Bengali লঘুত্রিপদী, which may easily be conceived to have evolved the longer or দীর্ঘত্রিপদী, I lay again the poet Jayadeva under contribution. In his
চল সখি কুঞ্জংসতিমির পুঞ্জং
শীলয় নীল নিচোলং ।
The Paṇḍits who look for our পয়ার, a respectable origin, make the Sanskrit অনুষ্টুভ, the forbear of our humble পয়ার, though the tune and the form of the one do not agree with those of the other. That our nursery rhyme বিষ্টি পড়ে টাপুর টুপুর, etc., is wholly in accord with our পয়ার, cannot even for a moment be doubted. Though the lines move on keeping time with the note of a tune imbedded in syllabic accent, they contain fourteen letters, and at the end of the first foot of a line consisting of eight letters, we get the requisite cæsura. It will be noticed in the last line of the verse quoted below, that রা of রাগ being lengthened by the stress of an accent, the loss of one letter has been made up for. The verse is:
বিষ্টি পড়ে টাপুর টুপুর নদী এল বান,
শিব ঠাকুরের বিয়ে হবে তিনটি কন্যাদান;
একটি কন্যা রাঁধেন বাড়েন, একটি কন্যা খান,
আর একটি রাগ ক'রে বাপের বাড়ি যান।
The purpose I have in view, does not allow me to write elaborately on the genesis of our metrical system; a separate treatise should be devoted to the execution of the work. The nature of our accent, the accentual basis of our metrical system, and the fact that our accent has been retaining its peculiar character since long, are factors which should principally engage our attention. That in respect of our accent, a long continuous current may be observed to have flown through ages, will be clearer when we consider other facts, and the readers will have to form their opinion, by considering the effects of what may be called cumulative evidence.
As to the origin of many Sanskrit metres from popular rhymes, such non-Sanskritie names as তোটক (indicating special emphasis on the third letter—তিয়ট + ক), দোধক (of equal length with তোটক, leaving out the initial and the final accented syllables; the accent falls regularly after two intermediate unaccented syllables—i. e., ধক্ or ধাক্কা or 'push' comes after 'দো' or two) লীলা খেল (the word খেল indicates the origin), etc., may be referred to.
I cannot certainly deal with our metrical system at a greater length here, but I cannot at the same time conclude this section of my lecture without referring you to the Hindi and the Oriya modes of reciting poetry, in contrast with our mode. I speak of the mode of recitation only, as it is not possible to analyse here the metrical system of Oriya and Hindi. I doubt not, that you have heard in this city the Oriyas and the up-country men to read aloud their verses. It must be a familiar experience, that from the sound alone from a good distance, a Bengali can know whether an Oriya poem or some Hindi দোঁহা's are being recited. Even where there is no musical chanting, the character of the metres will indicate the characteristic difference. All this is due wholly to different accent systems. As the style of a language is the expression of the thought of the speakers, so is the metrical system in a language, due to the special accent system of the people. The Oriya verse of lines of 9 akṣaras, if read in Bengali fashion, the composition will sound like a disjointed prose piece; so also it will be with such ত্রিপদী lines of Upendra Bhanja, as:
ব্যধি বন্তর বাঞ্ছিত উপচার
বৈদ্য বিহিলার পরিযে
Similarly if the lines of a Hindi দোঁহা be not read in the Hindi fashion, the music of the lines will fade away. Without pronouncing any definite opinion as to whether the lines quoted below, were composed in old Bengali or Hindi, I may bring to your notice, the basic Hindi character of the metre of the lines; the non-hasanta sound of the final syllable and the long sound of the penultimate, as have to be maintained in rightly reciting the lines, are to be duly noted. The lines are:
কাআ তরুবর পঞ্চবিডাল
চঞ্চল চীএ পইঠো কাল॥
I do not mean to be exhaustive here; I want however to impress upon you, that we can solve many linguistic problems at least partly, if we take the factor of accent, deeply into our consideration.