The History of the Bengali Language/Lecture 9
SOME FACTORS RELATING TO PHONOLOGY AND ACCENT
LECTURE IX
Section 3
Accent traced in Sandhi and Compound Formations
Sandhi.—The phenomenon of euphonic combination or Sandhi should engage our attention next, as by a study of it we can partly ascertain many phonetic and accentual peculiarities. How some stiff Sanskrit rules of Sandhi can be simplified on reference to the original Vedic sound-value of some letters, has been discussed separately, and this discussion has been relegated to an appendix to this lecture. It is true, that unlike what is noticeable in Greek, Chāndasa does not allow any hiatus to exist in a word, but the rigid Sanskrit Sandhi-rules by virtue of which two or more independent words are linked together in an agglutinated unit, do not appear to have obtained in Chāndasa. I need hardly assert, that in a living speech, in which ease and fluency in the matter of articulation of sentences can never be disregarded, and in which words must be uttered in an intelligible manner, Sanskrit rules of Sandhi cannot be strictly enforced. The component parts of Purohita for example, may not be allowed to stand separate, since the newly-formed word, has a distinct signification of its own, but the force of the word Adya, (অদ্য) for example, disappears, if Adyendrasya অদ্যেন্দ্রস্য be substituted for অদ্য ইন্দ্রস্য. We get such a line as বয়ম্ অদ্য ইন্দরস্য প্র ইষ্টা in the Vedic Padapātha, while the Sanskritic form of the text gives us বয়মদ্যেন্দ্রস্য প্রেষ্টা, which involves the loss of three syllables required by the metre. I cannot speak here anything regarding what is called স্বরভক্তি as noticed in the Vedic pronunciation of ইন্দর for ইন্দ্র, but the example will fully show, how the rigid Sandhi rules of Sanskrit Grammar make a simple speech, unintelligible. We find in the Vedic verses, as is natural in a living speech, that each foot, nay each hemistich, stands apart, without being united in Sandhi with a succeeding foot or hemistich. We observe only in some rare cases in Sanskrit, that one hemistich is not united with another in a verse, where Sandhi combination is possible; the following is a couplet which illustrates this sort of deviation from the usual rule:
যো ধ্রুবাণি পরিত্যজ্য অধ্রুবাণি নিসেবতে
ধ্রুবানি তস্য নশ্যন্তি অধ্রুবম্ নষ্টমেবহি।
It becomes perfectly clear, that once when the Vedic language became obsolete, various cases of euphonic combination occurring in that language were studied very carefully, and a good number of generalized artificial সন্ধি (Sandhi) rules were framed for their rigid application in what is called the Sanskrit language. This is why the processes noticeable in Chāndasa in such euphonic combinations, as are due to the influence of accent (which is a living factor in a living speech), are not at work in Sanskrit; instances of lengthening the accented vowel, as in বিশ্বামিত্র (বিশ্ব + মিত্র) or অনূপ (অনু + আপ) or of dropping the unaccented vowel as in চার্বাক (চারু + বাক্), are not obtainable in Sanskrit.
It is a fact that the rules of Sandhi, as are noticeable in Pāli, are not wholly in agreement with the Vedic rules; that this very deviation shows the living character of that earliest-known Māgadhi Prākṛta, is what we should duly appreciate. How because of the natural accent of the speakers, and owing to the changed value of the sound of some letters, such euphonic combinations occurred in Pāli, as দেবাতি (দেব + ইতি), পন্নরসেরিব (পন্নরসে + ইব), পাদন্তি (পাদং + ইতি), etc., should be a subject of special study. That the Sanskrit Sandhi-system does not support this claim of Sanskrit, that it was naturally evolved out of Chāndasa should be duly noted; we can very clearly see, that the natural Vedic rules of euphonic combination have only been artificially extended in Sanskrit, to cases where combination brings about stiffening of the speech and unintelligibility of meaning.
It is certainly very true, that our Bengali Sandhi system is not worked by the rules of the old time speeches, but we proceed to show that the very principle which governed the phenomenon of euphonic combination, in the speeches of ancient times, governs to-day the Sandhi system of ours. Before I cite examples to substantiate my proposition, I should notice an objection which is raised by some in this direction. On the basis of a superficial and unscientific observation of the fact, that the rules which govern the formation of such combinations as ইত্যাদি, মনোহর, etc., do not prevail in Bengali, some scholars have gone the length of asserting that the natural phenomenon of euphonic combination does not at all exist in Bengali. We have certainly borrowed the words ইত্যাদি and মনোহর in their entirety, and cannot disjoin them in our language; no doubt our ইতি means finish, আদি means original, and ইত্যাদি means etcetera; again there is no such word as মনস্ in Bengali, nor the word হর, if not a name of a man, can have any meaning independently. It is also true that ই and আ do not combine according to Sanskrit rule to form য়া, but we have noticed previously, that they coalesce in Bengali quite in another fashion. Non-observance of Sanskrit rules does not however justify us to formulate, that euphonic combination is unknown in Bengali.
In Bengali, we do not and cannot combine different words into one agglutinated whole, for we utter our words one after the other, to convey distinct meaning of them to other ears; but different vowel sounds and allied consonants do combine to form one word. With the অপভ্রংশ word মশা, the Sanskrit word অরি (enemy) being joined in sandhi, we have got one word to mean the distinct article মশারি (mosquito curtain); to signify a special sort of আলু (bulb) the adjective গোল and আলু have been joined to form the word গোল আলু; the word কাঁচা as an adjective of কলা does not and cannot change its form, but when the words are combined together to signify the sort of কলা which is used as vegetable, the final আ of the adjective being dropped, the word কাঁচকলা has been formed; we may also get the examples ঘোড়া + গাড়ি = ঘোড়গাড়ি, যোড়া + বাঁধা = যোড়বাঁধা, বোকা + চন্দ্র = বোকচন্দ্র (fool), etc. In such examples as কাঁচ কলা, যোড়বাঁধা, etc., we notice the loss of unaccented আ finals of the first component parts of the words; similarly we find the loss of ই of ইচ্ছা in the phrase যাচ্ছেতাই, where the accented যাহা in the form of যা occurs as the first component. We have seen that in the Vedic language, there were elision of the final vowel sounds of the unaccented syllables in their euphonic combination with unaccented syllables, and that is why the final উ of চারু dropped in euphonic combination with accented বাক্, to form the word চার্বাক to signify a man of persuasive speech.
That the lengthening of vowel sound as noticeable in the Vedic words বিশ্বামিত্র, বৃসা-কপি, etc., is also noticeable in the old Prākṛtas as well as in modern Vernaculars, requires to be pointed out. We may notice, for example, the Pāli idiomatic expression ফলাফল (various sorts of fruits) in such a sentence as ফলাফল স্স অন্তো ন অত্থি, to see unmistakeably that the word has not been formed by the combination of ফল + অফল. This sort of duplication to indicate either variety or etcetera is very much current in Bengali; it is also the rule in Bengali, that in the process of duplication, an আ comes in as a joining link. We must clearly see, that the Bengali words ফলাফল, হিতাহিত, মতামত, চলাচল, etc., do not combine good and bad ideas together; চলাচলের রাস্তা does not signify the path for going and not going. The 'আ' that comes in here as a joining link, indicates emphasis only, when occurs in the formation of compound words; compare the emphasized forms কপাকপ, চটাপট, টপাটপ্, etc., with the ordinary forms কপকপ, চটপট, টপটপ, etc. I feel tempted to notice, that in common parlance it is difficult to many of us to keep the purity of the Sanskrit word দুরবস্থা; because of accent on অবস্থা the word is pronounced usually as দুরাবস্থা.
The natural rule by which one consonant is changed into another, because of the genetic affinity between the consonants, is also at work in Bengali; এক + গঙ্গা = এগ্গঙ্গা পাঁচ + জন = পাঁজ্জন, ছোট্ (ছোট = younger) + দাদা = ছোড়দা, যত + দিন = যদ্দিন are examples. The Sandhi rules of Prākṛta Grammar as are still at work, should be carefully studied by the students in this connection; I point out here a few cases only, where Bengali is in agreement with the old Prākṛtas, in the matter of euphonic combination. From চড়্ + চড়্ we get চচ্চড় and from জগৎ + বন্ধু we get জগবন্ধু; final ৎ is very often dropped in Pāli even though there is no euphonic combination with the initial letter of a succeeding word, for example, কিঞ্চি is the form for কিঞ্চিৎ, and আসী is the representative of the fuller form আসীৎ. I may remark in conclusion, that the Dravidian method of Sandhi combination, is noticeable in some rare cases only; in Tamil 'ম + কাই' and 'তে + কাই' for example, will be মাঙ্গাই (mango) and তেঙ্গাই (cocoanut) respectively; this growth of nasal sound in Sandhi has only been noticed by me in খোলা + কুচি = খোলাম্ কুচি.
Samāsa—সমাস.—I have spoken above, that change of vowel as well as of consonant takes place in the formation of compounds called samāsa (সমাস); but as many noted scholars are of opinion, that barring a few stray examples, we cannot get samāsa compounds of genuine Bengali words, I must show that compounds or samāsa of all sorts exist in Bengali. I consider this question to be important, for it is to be seen, whether the old mode of thinking which brought about samāsas in particular forms, is still our inheritance or not; it must be borne in mind, that the racial peculiarity in the matter of thinking, governs the style and structure of a language. I cite below the Bengali samāsa forms exactly in that classified order which is maintained in authoritative Sanskrit Grammars.
অব্যয়ীভাব—Adverbial Compounds.— 1. In the following examples, অব্যয় words do not occur as in Sanskrit, but the compound forms indicate the sense of the অব্যয়ীভাব সমাস—(a) indicating বীপ্সা—গলি গলি, বাড়ি বাড়ি, রাস্তায় রাস্তায়, রোজ রোজ, etc; (b) ক্রম—পরপর, পিছুপিছু, etc.; (c) অনতিক্রম—রাতারাত, যত পারি, যা পারি etc. The following examples may be contrasted which are not compound forms, viz.—দেখ্তে দেখ্তে (quickly), চল্তে চল্তে (by excessive walking), গাছ্টি বাড়্তে বাড়্তে (in the course of growth) শুকিয়ে গেল, ফল পড়্তে পড়্তে (just on falling) কুড়িয়ে নিল, etc.; in these cases, infinitives being doubled, the sense of repetition has been expressed, but the words do not form Samāsa compounds, (d) The following examples indicating "the whole of" are closer in relation with অব্যয়ীভাব forms than with others: মাঠ্কে মাঠ, বাজারকে বাজার, ঘর্কে ঘর, etc. (e) Where to indicate পর্য্যন্ত (up to) 'আ' occurs in Sanskrit, as in আকণ্ঠ, আকর্ণ etc., only the doubling of the word takes place in Bengali; e.g., 'গলায় গলায়' খাওয়া, 'কানায় কানায়' ভরা, etc., এক নৌকা লোক, ভরপেট, কুঁচ্কি কণ্ঠা may be compared with these forms, as indicating the sense of অব্যয়ীভাব৷
তৎপুরুষ—Determinative—2.—If the examples grouped under the following sub-he&ding (1) be regarded as I suggest, as of তৎপুরুষ class (being Determinative, or rather Dependent) wherein the nominative case predominates, we may hold that we have তৎপুরুষ with vengeance in Bengali. I may then classify the তৎপুরুষ forms as কর্ত্তা প্রধান, কর্ম্ম প্রধান and so forth, looking to the sense which the forms convey.
(1) The তৎপুরুষ of nominative prominence, or কর্ত্তা প্রধান তৎপুরুষ:—For this entirely new class of Bengali compounds, my examples are,—দাগলাগা, বাজপড়া, etc. We have to note that such Sanskrit forms as তৈললিপ্ত, বজ্রাহত, etc., are construed as তৃতীয়া তৎপুরুষ; the form কাদামাখা may be construed as করণপ্রধান তৎপুরুষ, but কাদালাগা cannot be so construed; we have also to notice that the forms of my example cannot be classed under বহুব্রীহি, for, in বহুব্রীহি forms a person or thing must be indicated irrespective of the meaning of the component parts. This is why I have suggested this new nomenclature for a class of compound words. Mine is a suggestion merely, and not an authoritative statement. Compare all the compound forms occurring in the sentence আমার লাল-পেড়ে তেল-ধুতি খানার এই খানটা বাক্চী-বাড়ীর বাদুড়-চোষা গাছ-পাকা জামের রসের দাগ-লাগা; the 1st is বহুব্রীহি, the 2nd is তৎপুরুষ of 4th class, the 3rd is ষষ্ঠী তৎপুরুষ, the 4th is তৃতীয়া তৎপুরুষ, the 5th is ৭মী তৎপুরুষ; and so now the character of দাগ-লাগা may be appreciated.
(2) কর্ম্মপ্রধান তৎপুরুষ (object-indicating):—মানুষ খেকো (as a tiger), মন ভুলান, ইঁদুর ধরা (as a কল or machine), বৈ পড়া (as in বৈ পড়া বিদ্যা), etc., are examples.
(3) করণ-প্রধান or agency-indicating:—হাত-গড়া, কাপড়-ঘেরা, এক-কম-কুড়ি (signifying ঊণ or less by one); at times 'এ' is added to the final letter of the first part of the compound as আরশুলায়-চাটা (আরশুলা-চাটা is another form), মেঘে ভরা (also মেঘ-ভরা), etc. In বাদুড় চোসা, the final letter does not take an 'এ'৷
(4) Purpose-indicating or উদ্দেশ্য বাচক:—তেল-ধুতি (cloth worn for besmearing the person with oil), পা-জামা (for পা, i.e., trousers), মরা-কান্না (wailing befitting the occasion of death in the family), বসৎ-বাড়ী (house intended for dwelling), etc.
(5) অপাদান-বাচক, to signify 'away from':—পাল-ছাড়া (straying away from the flock or herd), সৃষ্টি-ছাড়া (different from what is usual), ঘর-পালান ছেলে (a run-away boy), পাহাড়-ঝরা (পাহাড় থেকে) জল, etc.
(6) Relation-indicating or সম্বন্ধ বাচক:—ধর্ম্ম ভয় (the word ভয় does not affect the character of the compound in Bengali), বিয়ে-বাড়ী, মাস-কাবার, বন-বেড়াল, etc.
(7) Locative or স্থান কাল বাচক:—গাছ-পাকা (ripened on the tree), নৌকা-ভরা লোক (i.e., নৌকায়—full up in the boat), ঘর পোষা, etc.
কর্ম্মধারয়—Descriptive—3.—(1) আধফোটা, কালপেঁচা (an inauspicious owl); (2) কাঁচা পাকা (ripe and unripe), তাজা মরা; (3) মিশ্-কাল (black like মিশি or black tooth powder, here the final ই of মিশি has been dropped), ছেলেবুদ্ধি, কাঁচ-পোকা (looking like glass), সোনা-মুগ (a lentil, like gold in colour), etc.
দ্বিগু—Numeral compounds—4.—পাঁচ-গজি (as a cloth), তে-হাতি, দু-পিঠে (lit. having two sides), এক চোখো (as a judgment, disclosing partiality).
বহুব্রীহি—Possessive compounds—5.—মা-মরা, হাঁড়িমুখো (long-faced), পাড়া বেড়ানী, বাইশ কর্ম্মা (one who does nothing), etc.
দ্বন্দ্ব—Copulative—6.—মানুষ-গরু, লেপ-কাঁথা, আম-কাঁঠাল, ভাল-মন্দ, উনিশ-বিশ, etc.
Duplicated Words.—The words which are duplicated on account of emphasis, to indicate repetition, or to express the idea of excessiveness, should be noticed and classified here, to distinguish them from the Samāsa compounds. I need hardly remind you that according to the যঙন্ত rule and by the rule of the ণমুল্প্রত্যয়, words are duplicated in Sanskrit to indicate repetition or excess. I refer you to the whole section of the Siddhanta Kaumudi entitled Dvirukta Prakaranam which begins with the rule সর্ব্বস্যদ্বে, wherein reduplication of various sorts has been illustrated.
1. (a) The adverbs সোজাসুজি (quite direct), পাশাপাশি (close by the side), মাঝামাঝি (right through the middle), etc. (indicating 'very much'), and হাঁটাহাঁটি, ছুটাছুটি, টানাটানি etc., (indicating 'repetition'), may be classed under one head. Such adverbs as মোটামুটি from মোট (total; taking the whole roughly into consideration) and গোড়াগুড়ি from গোড়া (beginning from the very beginning) come also under this head, as the idea involved in the words is that মোট or গোড়া is taken repeatedly or much into consideration. The vowel changes in this class of duplication must be noted.
{b) কাটাকাটি, মারামারি, চুলাচুলি, হাতাহাতি, etc., fall also under this head as a sub-class, as a slightly-differing sense of reciprocity in fight is indicated by them. মুখামুখি (tete-a-tete), চোখা চোখি (each seeing the other), কোলাকোলি (the act of embracing), etc., are also of this class.
2. Though the idea "very much" is in the following words, they differ from the first class in meaning as well as in form; the words are duplicated without undergoing any change. Thinking too much of, or having anxious solicitude for, or making too much of, will be found to be the idea involved in বাড়ি বাড়ি, দাদা দাদা, and কলিকাতা কলিকাতা, in the following sentences,—আমার মনটা বাড়ি বাড়ি করিতেছে, সে দাদা দাদা করিয়া মরিল, তুমি বড়ই কলিকাতা কলিকাতা কর৷
3. When duplication takes place to indicate 'almost like,' or 'similar to,' no change of vowel takes place. The forms কাদা কাদা, জ্বর জ্বর, ঠাণ্ডা ঠাণ্ডা, মানুষ মানুষ, etc., are examples.
4. In the class of duplication noticed below, there is this special peculiarity, that in the process of duplication, the original word without being repeated, is conjoined to a synonym of it. Agglutination of two seemingly different words, should not mislead us to consider the word as a Samāsa compound. The adverbs করেকম্মে, ভেবেচিন্তে, ফেলেছড়ে, বলেকয়ে, মেগেযেচে, etc., are fitting examples. The noun forms ঘরবাড়ি, লোকজন, সোরগোল, মাথামুণ্ডু, গাগতর, etc., are also similarly duplicated. Some duplicated words of this class, may elude detection of their character, as in either the first or the last augmented portion, some obsolete or unfamiliar words appear. I give a few examples. In the word আশেপাশে, the first portion is the Vedic word আশা, which has the same meaning as পাশ of Bengali; the adverb চলেবুলে and the noun form ছেলেপিলে may also be considered; the word পিলা is a Dravidian word for child, and the word বুলা meant walking in old Bengali, and in that sense the word is still in use in Oria; in the word খুঁজে পেতে, the last portion পেতে or পাতা comes very likely from Hindi পাত্তা (cf. পাত্তা নেহি মিলা, no trace is obtained); the word পাতি in পাতি পাতি করিয়া খোঁজ, seems to be also of the same origin. Let me adduce a few examples to show, that a word of foreign origin or of classical origin, though really a synonym of a word, is used either as an adjective or an adverb to its synonym, because the real import of the foreign or classical word is lost sight of, or is imperfectly understood; the word গন্ is a corruption or অপভ্রংশ of Pāli গিনি (Sanskrit অগ্নি); this অপভ্রংশ form is found retained in the phrase গন্গনে আগুন. The word or stem হন্ is of Kolarian origin, and it signifies walking; it is this হন্ which we meet with in our হন্হন্ করিয়া ছোটা. These words should not be confused with the words of onomatopoetic origin. In গালি-গোপ্তান, the second Persian word is a synonym of the first. I have heard school-boys saying আকাশ খুব clear পরিষ্কার হয়েছে; as a translation of police investigation we at times meet with তদারকের অনুসন্ধান in our Bengali newspapers. Whatever that may be, let me add a few more examples as may fall under this class. They are: ঘুরেফিরে, ঠেঙ্গিয়ে পিটিয়ে, ডেকে হেঁকে, তেতে পুড়ে, বন জঙ্গল, বন বাদাড়, বেঁচে বর্ত্তে, ভিজে তিতে, ভেঙ্গে চুরে, হাট বাজার, etc.
5. Almost connected with the fourth class is the class I now describe. To give special emphasis to an idea, two words are so joined together, as the second portions may indicate the consequence or completion of the action, indicated by the first portions of the compounds. A few examples are: আণ্ডা বাচ্ছা, উলট্ পালট্, কেঁদে কোকিয়ে, ঘসে মেজে, জ্বলে পুড়ে, টেনে হেঁচ্ড়ে, দেখে শুনে, ধুয়ে মুছে, নেয়ে ধুয়ে, নেচে কুঁদে, বুঝে সুঝে (সুঝে = to see), মেরে কেটে, রেঁধে বেড়ে, লড়ে ভিড়ে.
6. Such duplications as মানে মানে, প্রাণে প্রাণে should perhaps be classed separately, as they indicate—anyhow saving or protecting মান (honour) and প্রাণ (life).
7. It is difficult to say whether the second portions of the following compounds are meaningless additions, or that they once had some significance, and as such should be grouped under the fifth class. The words are:—কুড়িয়ে বাড়িয়ে, জুড়ে তেড়ে, কাপড় চোপড়. If the last named example is the representative of the প্রাকৃত idiom বত্থ-চিবর, চোপড় may be easily explained.
8. To indicate etcetera or 'the like,' the words are generally duplicated with the loss of the initial letters and by the substitution of ট for the initial letters. কাজ টাজ, ভাত টাত, মাছ টাছ are very familiar examples. When disgust is sought to be expressed, the initial letter of the duplicated portion is usually changed into ফ, as—এ গরমে কাজ ফাজ করা চলে না, রাত্রে ভাত ফাত খাওয়া দায়, মাছ ফাছে কাজ কি? etc. It is to be noticed, that in some cases, duplication is made not with the ট initial but with some other letters. চাকর বাকর, বাসন কোসন, রকম সকম, etc. are examples; it is rather difficult to enunciate any general rule for these irregular forms. It may be, that in the cases of these exceptions, the augmented portions are but representatives of some obsolete words, and if so, must be grouped under class four. As in পূজা আর্চা the augmented portion is a contracted form of অর্চনা, so there may be many augmented forms, the meaning of which may be traced.
9. Such onomatopoetic words as কচ্, খঙ্, পট্, মড়্ are generally duplicated in their use. It is worth noting here, that many words simulate onomatopoetic origin, though they are really but অপভ্রংশ forms; ধব্ধবে is from ধবল (white) or from ধোয়া (washed clean), মিশ্মিশে is from মিশি (black tooth powder), ফুট্ফুটে is from Sanskrit স্ফুট্, ফুরফুরে is from Sanskrit স্ফুর্, ভুরভুর is from ভূরি (much). A special class of onomatopoetic words as ঠঙ্ ঠঙ্, ঝন্ জন্, খট্ খট্, etc. is of special interest in the Prākṛta dialects; in olden times the use of such words as well as of Deśī words of all sorts was prohibited by the Sanskrit Grammarians because of their vulgar origin. This is exactly why they are of importance in a history of language. I reproduce in Appendix I, my paper on onomatopoetic words which was published in the Journal of the Royal Asiatic Society in 1905.