Ashok R. Kelkar





The primary use of human speech is being the sign vehicle for language, man’s dominant medium of understanding his environment, indeed reality to which man has access and at the same time man’s dominant means of communication through which to gain access to his social environment.  But the primary use is not the sole use.  Loss of speech or absence of speech robs man not only of the prime sign vehicle of language but also of a means of expressing his feeling and of a vehicle of musical art.  Even in respect of the arts of language like poetry of drama, speech in an important element in the artistic vehicle.  In these days of writing and printing we are apt to lose sight of this: we only need to read a poem or a dialogue in a play ‘silently’ to realize that the text demands to the ‘heard’ even if only heard by the mind’s ear.  Any written or printed text (short of statistical tables or formulas of organic chemistry) when read out well is read effectively; coming from a poor reader the text becomes virtually unintelligible.  Singing, speaking eloquently, effective loud reading are all arts of speech.


            Sanskrit distinguishes between three modes of vai (speech: ordinary  speech

(bhāaa), reciting of a text from memory or from written record (pahana), and singing (gāna or gāyana).  Modern Indian languages tend to follow suit.  Marathi, for instance, disitingusies correspondingly between bolae, mahae, and e; the distinction is obscured slightly because mhanae a has also a second sense-saying something, especially through speech. It will be recalled that the Sanskrit etymon of mhanae, namely, bhaiti has he same sense.  Printer! print-as or-dahs and print as hyphen throughout namely, saying something.  There are occasions when an identidal text could be preseted through ordinary speech, reciting, or singling. A written poem, for example, can be read aloud, recited, or sung-kavitā-vācana, kavitā-pahana, and kavitā-gāyan are regularly distinguished Marathi.


One could look at these three arts of speech, which range from the most language-oriented (and least art-oriented) to the most art-oriented (and least language-oriented), from two distinct points of view, namely, the language-oriented and the sort-oriented.





            The very first question that any language-oriented investigation of the arts of speech needs to take up is-how does one go about distinguishing between the three categories of *** or speech, namely, ordinary speech, recitative, and song?  While the distinction is well recognized traditionally, there are no traditional definitions (as far as I am aware).


            To begin with, one could observe that ordinary speech and song are the most different from each other while recitative has some resemblance to either of them.  In other words, recitative could be thought of as a middle term with ordinary speech and song figuring as the two extremes.  But them, it soon becomes evident that the three constitute not a chain so much as a continuous spectrum with two poles.  In other words, there are some cases halfway between ordinary speech and recitative and other halfway between recitative and song.  Al l this indicates that the arts of speech all share certain variables such that observing how the variables are handled enables us to position a given specimen of speech within the spectrum.  It will be illuminating then to examine some specimens that a native speaker of the language concerned unhesitatingly identifies as ordinary speech, recitative, or song.


            Given that I am a native speaker of Marathi and an analysis of the Marathi speech system is readily available (kelkar 1993), a scrutiny of some Marathi specimens will be useful.


            Marathi operates with the following speech segments and prosodies.

(1)   (a)  Speech segments: p t b g, ph th h kh, bh dh, h gh, f; c č , j ǰ, čh, ǰ h, h, s š; m n  M; l , r, mh, nh, h, lh, h, rh; h, y v, vh; i u, e o, əe  ə, a,

I u , ə, a ī ū ; n


(b)               Speech prosodies

Junctures: hyphen, word space, single bar, double bar

Accents: phrase-nucleus, tone-nucleus (both marked the same, since the last nucleus in a tone group is automatically the tone-nucleus; the word-nucleus is fixed, typically penultimate)


Accent modifiers: single underline (making exclusivity, double underline (marking intensity) consisting of progressive tension loss and tension gain respectively

Tones: tone-rise and tone-fall either gradual, steep, extra-steep, or reversed

Tone modifiers: pitch-squeeze, pitch-stretch

Peripheral speech prosodies: paid & slow tempo, high & low pitch register, loud & soft volume, abrupt & smooth prosodic transitions, clip & drawl in sallable-balance

Note: the remaining peripheral phenomena (special speech segments and articulations, voice modifiers and qualities) need not detain us here.


First, a couple of specimens of ordinary speech. 


(2)    A dialogue by way of question and answer.

Tu-‘tyača-ši bollas // ||

‘ho \\ | ‘bollo-tyača-nokri-bəddəl \\ ||

kai-bollas /||

t ə’s ə-kaikhas-nai \||

‘you spoke to him? –Yes, I did, about his job.

--what was it?—Nothing special.’

(3)    A monologue by way of depiction

‘hosel-či ‘rum || || su’ s əjj  ə nsli /| t ‘ ri bə’ rya-pəi ki \\||

‘te-hotə hill-sen \ | ‘vara k sa ‘bhəṇṇ -bhəṇṇ sula-uta \\ ||


A room in a hotel.  Not well-appointed, but fairly good.  It was a hill station.  The wind was blowing hard, relectlessly.’


These specimens show a handling of rhythm features (namely, juncture, accent,

and their timing) an melody features (namely, accent modification, tone, tone modification, pitch register, and their timing) in accordance with the meaning being formulated and the speaker’s communicative intent (that is, speech sense together with speech address).  Note, for example, the abbreviations in the dialogue, the loose-knit depiction adding a detail at a time, the exclusive accent on ‘special’ and the intensive accent on the wind-depicting onomatope, and so on.  In ordinary speech the speaker’s art in respect of rhythm and melody is narrowly governed by sense and address.


            Before we move to specimens of recitative and song, the notation of rhythm and melody needs to be amplified.  Marathi recitative and song operate with the following.


(4)    (a) Rhythm

Pauses: comma, semicolon

Bars: single bar, dotted single bar, double bar (corresponding to tālī, kāla or kālī, sama in Indian music, roughly ordinary, elided, cyclic beats)—the time interval between successive bars being constant


Mora group: hyphen (the number of moras in each group being constant-typically, two)


(b)  Melody

Notes in the scale: S R G M P D N (being the sliding sol-fa scale of Indian music-corresponding to C D E F G A B in Western music), out of which M may be marked ‘sharp’ (tīvra) (Ro Go  Do No)

Octaves: S2 S1 S s S1 (corresponding to  atimandra, mandra, madhya, tāra, atitāra octaves of Indian music-ranging very low to very high)


Note: The tie line joins sequences of speech segments across rhythmic groups.


Specimens of recitative now allow.

(5)    The traditional recitative of multiplication tables or syllabaries or the like designed for young children (paravacā), of which some modes are being illustrated


|V                     | V                    |V                     |V


i’ka-vər ‘ek ‘a:kra: |       I’ka-vər ‘don ‘ba:ra ||

soīa-sətte baro ‘dərse | ‘soīa- əṭṭthe  əhavi’ sase |


(Mora count |8|8|8|8; Note the draw marked; to fill out. Moras.) ‘One-and-two, twelve.’ ‘Sixteen seven-times, twelve-and-hundred.  Sixteen eight-times, twenty-eight-and-hundred.


| /             |/            |//            |/            | /             | /\

ə’a            ‘i’i            ‘u’u            ‘eəi            ‘oəu            ‘əmm- əhha ||


‘kəka            ‘kikki             ‘kukku            ‘keki             ‘koku            ‘kəmm-kəhha ||


(Mora count |4|4|4|4|4|4; Note the clip marked with apostrophe, to fill out moras.)


‘a ā i  ī  u ū  e ai o au  a a. ‘

‘ka kā ki kī ku kū ke kai ko kau ka kh.


(6)    The traditional recitative of the divine names, often accompanied by a telling of rosary beads (japa), of which one of the modes is being illustrated.

 |N1-N2-S-S                      |N1-N1-S-S


ǰəyy-vihl       ǰəy-jəy-vihə: lə

(Mora count | 2-2-2-2| 2-2-2-2;)


Marathi verse prosody recognizes four kinds:

(a)     vtta, metre based on heavy and light syllable sequences, of Sanskrit origin

(b)    ǰəti, metre based on mora count (2) for heavy, 1 for light syllable), of Sanskrit, Prakrit, or native origin; (c) chanda, metre based on syllable count,  of Sanskrit of Prakrit or native origin (d) Muktapadya, non-metrical free rhythm verse, of late Modern origin.  The fist three admit of recitative rendition with traditional verse melodies.  For any given metre, more than one verse melody are sometimes variable.  Specimens of these three kinds of metre follow.  Any given metre may or may not feature recurrence of syllable or mora patterns.


(7)               verse in vtta metre (bhu jagaprayāta metre, which features recurrence is being illustrated, the poet being Rāmdās)

S | GRG   |G,             GR| GMP | MG

S’əda ‘       sərvə-da:          ‘ yogə     tūjha                    gh’əava ||


LH                   HLH                HL            HH                   LHH;


(Mora count 1|5|5|5|4 Syllable count 5, 7; Metrical count the foot Light Heavy recurs four times.)


A vtta metre like mālinī has a pattern without such recurrence: LLLLLLHH, HLHHLHH;


(8)   Verse in jāti metre (āryāīgi) metre, which is without recurrence is being illustrated, the poet being is Rmachandra sadave)


SSSS, SSMM;                                   RMGMP, PMG, RN1  SRG;


Su1 lok ‘vmna-ca | ‘bhMg-vai pr ‘siddh ‘tukya-I ||

GGGGMG, Ro Ro S;                     N1N1Ro G, MGR, N1 SS;

‘ovi dnya’neša-či |               ‘təisi ‘ariya mə’ yr-p Mta-I ||

(mora count 8 ,4; 8 ,10; 8, 4; 8, 10;)


(9)   Verse in chanda metre (ov  metre, which features recurrence, is being illustrated, in one of its many varities.


N1-S-R-MGo, GoR-SN1,

‘ye-gə-tu-gə:  ‘ga:i: |

vai-mədhe || ‘əm: -gai ||

            ‘tanhya ‘baa-la ‘dudu:             ‘dei |



(This is a lullaby stanza with 4 lines and a refrain.  Mora count 2-2-2-2, 2-2, (first, second line), 2-2-2-2-2-2-2, 2-2; up to 9 (third line), Syllable count upto 6 (first, second line), up to 4 (fourth line refrain)


A chanda metre like anuṣṭubh metre has a pattern without recurrence 2-2-2-2-2-2-2-2, and syllables 8 in alline.


            These different specimens of recitative illustrate varying adjustments between speech rhythm and speech melody on the one hand and song rhythm and song melody on the other had.  In the school child recitative (5), speech accent and juncture combines with song measure in respect of rhythm; speech tone (pitch contour) combines with song pitch level and range in respect of melody.  In the religious recitative (6), speech accent and juncture combines with song melody.  In verse recitative (7, 8, 9), relatively simpler patterns of song rhythm and melody are imposed on speech rhythm in various ways.


            Finally, a couple of specimens of song follow.


(10)           A popular song with druta kaharva song measure by was of rhythm and a popular song melody is being illustrated










go’vində :ye


to : | məjə-vəri


gu’ la: lə udhə’

||S-:    :SR-SN1



Go’vind : ye

To: m-vri

Gu’la: l udhəḷī

To: | |



Pikari ghe

To:| maja-vari

Rm: g u ‘v

To: || ‘gop-




Ge: |’ m-vri

RM: g




Gen: Mja


To: ||



(11)           A stage (from G.B. Deval’s play šāradā 1899) with jhaptāl song measure by way of rhythm and a popular song melody of being illustrated with its refrain


|| D-D

| r-s


| D-G-R

|| R-R









rə vədhu



|| No-No


: No










Ci pəri





                                                            Song measure count ||2|3:2|3

                                                            Verse rhythm count |5|5,|5|2; |5|5, |5|5 ;


If the recitative shows varying adjustments between speech rhythm and speech melody and song rhythm and song melody; in the song, the song rhythm an d melody takes over completely.  Speech juncture and accent, so submerged, contribute only to the intelligibility of speech meaning.  While one continues to handle rhythm and melody as one moves from ordinary speech to song, the whole orientation to art orientation.





            What is art?  As soon as man’s engagement with the world that he has made a home in by way of understanding that world and coping with it moves beyond a hand-to-mouth animal existence, the engagement tends to take either of two forms-the form of science and the form of art.  (Greek epistēmē and tekhnē, Latin scientia an šarstra, Sanskrit šātra/vidyā an ilpa/kalā convey basically the same insight into this duality.) Science tends to operate with logical form, reason, abstraction from raw existence, acceptance of what there obviously is, rule-conformity, measurement.  Art on the other hand tends to operate with imagined form, intuition, concretion from raw existence, transcendence of what there obviously is, sense of tact or finesse (tāratmya buddi) (as in approximating/pursuing an ideal or improvisation), guestimate.  The pursuit of science culminates in Science with a capital s; the pursuit of art culminates in Art with a capital A.


            In moving from the art of ordinary speech to the art of song, we appear to be moving from art in the  broader sense just sketched to Art in the culminative sense.  A useful point of entry in the understanding of Art in this narrower sense for our present purposes is characterizing an Art work or object or performance.


            How does a work of art exist at all?  What is it made of?  A work art exists at two levels-at the level of material and at the level of medium.  (This distinction is prefigured in Kant.)  Thus, a painting can be thought of as pigments mixed with oil smeared onto a stretched piece of canvas; it is it this level that it can be said to have been produced and be bought or insured against fire or theft.  But then, clearly, not every paint-dabed canvas-piece is a painting.  That is why nobody would call paint an canvas the medium of painting-at best they constitute the vehicle material of painting.  The vehicle material is to be distinguished from the other kind of material.  A painting can also be thought of as made out of its content material or experimental material.  The question, What does the painting show?  What is it about?, can be meaningfully asked about any painting, whether representational or non-representational.  Again, however, the painting the eluded us.  Not every painted every painted map will be a painting.  The painting as such exists at the level of medium-at that level the painting presents line and shape, colour, light an d shade as they operate within painting space.  These constitute the medium of painting.  It is at this level that the painting has been created (not merely produced); interpreted (not merely annotated or figured out); and appraised (not merely priced).  The medium is what imparts form or shape to the vehicle material as well as to the content material, what brings the material home or projects it to the recipient, what enables one to assign a work of art to certain art form to be received as painting or sculpture or music or poetry or whatever.  The medium of any art form operates at two levels-the level of technique and the level of style.  At the level of technique, the medium is the body of devices through which vehicle coveys content in the course of an act.  At the level of style, the medium is the fusion of qualities through which vehicle embodies content in the forming of an object, the line between vehicle and content getting blurred.


            The site of existence of a work of art way thus be set out as follows:


(12)(a) Material

Vehicle material

Content material

      (b) Medium

   Technique: devices: act: vehicle conveys content

   Style : qualities: object: vehicle embodies content


            Our present concern is with an art in which speech figures as vehicle material.  When language-oriented, speech subserves meaning consisting of speech sense and address. When art-oriented, speech fulfils the demands of delectation (āsvāda) and projection (kepaa/samrapaa).  In the recitative, there is an attempt to reconcile the two orientations.


            At the level of technique, one considers the musical possibilities of speech.  Speech and music share syllabication, rhythm, and melody.  (Harmony can appear in musical speech or song when there is more than one voice and need not detain us here.)


            Compare syllables in speech (as: nok and ri in the speech form nokri ‘job’) and in music (as: sa, re in sargam, a:, in ālāpī, na, tom in tarānā -all forms of non-verbal spoken music; dha, dhin in percussions, di, da in strings in instrumental music).  The syllables can be checked or free, clipped or drawled (nok, reg, tom, dhin, did as against ri, re, dha, da).  Some languages use checked syllables sparingly (Italian, Japanese), others use them in abundance (German, Chinese).


            Compare rhythm in speech in music.  Rhythm is the perception of a certain handling of time intervals between auditory figures auditory ground.  The time intervals may appear bound my measure or free from measure.  Rhythm in ordinary speech and recitative may be free or bound.  Bound rhythm typically involves the recurrence of some auditory property at measured intervals.  In ordinary speech the rhythm is ordinary free.  In some languages (like English or German) with a noticeable word accent, there is a tendency to accentual bound rhythm.


(13)           Compare—

(a)    ‘sat’ ‘ah’ ‘nəu dəha ‘ăkra’ ‘bara’ ‘tera (in Marathi counting, free rhythm)

(b)    ‘seven’ ‘eight’ ‘nine ‘ten’ ‘level ‘twelve thirteen (in English counting, accentual bound rhythm)


In recitative, bound rhythm is fairly common—as in (5, 6, 7 bhujagapryāta, 8 pādākulaka, 9 ovī).  This is especially seen with accentual meters in languages with accentual bound rhythm.


(14)           Compare


(a)    The-‘curfew’ tls the –‘knell-of ‘parting ‘day The-‘lowing ‘herd wind—‘slowly ‘o’ er the –‘lea (Thomas Grey, ‘Elegy Written in a Country churchyard’)

(b)    The –‘one red—‘leaf, the ‘last of –its-‘clan, That-‘dances as-‘often as-‘dance it-‘can.  ‘Hanging so-‘light and ‘hanging so-‘high On-the ‘topmost ‘twig that looks-‘up at-the –‘sky. (coleridge)


While both (14a) and (14b) illustrate accentual bound rhythm, there is an important difference. In the former, the measure is regular in that there is a simple alternation between unaccented and accented syllables.  In the latter, the measure is bound, but neither regular nor simple; perhaps it may be described as pleasingly and adequately variable-ornamented, in short.  In song, free rhythm of the ornamented kind can be seen in ālāpī.   Bound rhythm is much more common-relatively simple in (10, 11) ; more ornamented bound song rhythm can be seen in pieces that are far more music-oriented than language-oriented.  In sum, rhythm can be more and more music oriented.


(15)           Rhythm can be progressively more music-oriented:

(a)    free and simple

(b)    free and ornamented

(c)    bound and simple

(d)    bound and ornamented


To take purely musical example of the distinction between (15c) and (15d), the introduction of kāla and sama makes rhythm ornamented (alaṁkta); the tāla cycle may be less ornamented as in hekā or more so as in qāidah.  The tālī by itself and without will make for bound and simple rhythm.


            Finally, compare melody in speech and in music.  In Marathi  speech melody, for example, one operates with four pitch levels, namely, low, mid, high, extra high; four pitch contours, rise, rise prefixed with fall, fall, fall prefixed with raise; and three pitch ranges, normal, squeezed, stretched.  But in recitative of school lessons (5), the fall-rise tends to come off not as the usual speech melody high-mid-extra-high or extra-high-mid-high but as SN1S.  The melody thus tends to be bound by scale.  Languages differ in this respect.  Some (like Telugu, Marathi) have a ‘sing-song quality’: that is, their speech melody is incipiently bound and the contour depends almost wholly on pitch.  Others (like English, Hindi) have a ‘crisp quality’: that is, their speech melody is wholly free and the contour partakes of pitch, intensity, muscular tension.


(16)           Melody can be progressively more music-oriented:

(a)    free and simple

(b)    free and ornamented

(c)    bound and simple

(d)    bound and ornamented


Ordinary speech is free and simple in its melody.  Recitative tends to be free and ornamented as in (5), bound and simple as in (6,7), bound and ornamented as in (8,9).  Song is bound and ornamented as in (10,11).  In bound melody, pitch level, pitch contour, and pitch range operate with fixed intervals within a scale.  The pitch level of the middle c(ādhāra-aja) may be variable as in Indian and early European music, or fixed as in late European music.  Ornamentation in melody begins with mini-contours within contours.


            At the level of style, one considers the artistic possibilities of speech.  Speech in this context is inclusive of the musical possibilities of speech in respect of syllabication, rhythm, and melody (as we have just seen).  The questions to ask are as follows.


(17)           At the level of style—

(a)    How is speech as vehicle material conducive to insight into content material?

(b)    What qualities can speech achieve and so help define a style-

       (b1)    In respect of delectation?

       (b2)    In respect of projection?

Note: (a), (b1), (b2) correspond to abhinaya, sāvāda, and kepaa/sapara in the large sense.


In respect of conduciveness to insight into content material, speech enriched musically can offer certain possibilities.  It can offer iconicity—consider the depiction of relentless wind at (3).  It can offer mimicry—the peripheral feature of lax larynx quality can suggest drunkenness.  It can evince feeling—the peripheral features of high pitch register, loud volume, sharp voice quality (with energy concentration in higher overtones and high damping leading to ‘ringing’ tone) can evince anger.  It can assuage feeling—pitch squeeze, soft volume, light or mellow voice quality (with low damping leading to non-resonant tone) can bring reassurance to disturbed or fearful addressee.  It can  evoke or excite feeling-overall labiovelatization and nasalization, soft volume tend to call forth protectiveness from the addressee.  Iconicity, mimicry, evincing or assuaging or evoking feeling are all aspects of expressivity—to give definiteness to a much abused word.


In respect of delectation, speech enriched musically can contributed to certain aesthetic qualities.  Sparing use of junctures and fewer accentual nuclei promote aesthetic unification; more junctures and nuclei promote aesthetic dispersal.  Constant duration of intervals, mirror-imaging of pitch contours, appropriate selection of speech feature for desired quality promote aesthetic harmony; irregularity, unpredictability, frustrating of expectation promote aesthetic disharmony.  Abundance of detail, variation, novelty, liveliness in speech promote aesthetic richness; their lack promote aesthetic sparseness.  Low pitch register, smooth prosodic transition, low unrounded vowels, near-level pitch contour, mellow voice quality (with energy concentration in lower overtones and low damping) promote aesthetic infintitude;  precise articulation, abrupt prosodic transition, high unrounded vowels, contribute to a ‘mincing’ tone and aesthetic finitude.


Another way of exploiting  and exhibiting the possibilities of the vehicle material is achieving a certain quality of projection.  The Indian tradition offers certain useful, indeed illuminating, categories in this area.


(18)           Getting the addressee involved in this  work of art in one or more of these ways (in simultaneity or sequence)—

(a)    Respective mood (citta-vistāra, prasāda)

(b)    Kindled mood (citta-dīpti, pāruya)

(c)    Melting mood (citta-druti, saukumārya)

(d)    Restful mood (Citta-Višrānti, šānti)


Perhaps, the middle register, medium volume, steady and distinct, articulation, mellow voice quality induce receptivity by virtue of offering ample space; sharp or heavy voice quality (with high damping resulting in ‘ringing’ tone), loud voice, abrupt prosodic transition, retroflex and other ‘dark’ quality segments induce kindled mood by virtue of offering ‘tough’ or ‘luminous’ effects; soft voice, smooth prosodic transition, slurred articulation, feeling-laden voice (husky with pharynx air friction or glottis air friction), light voice quality (energy concentration in higher overtones and high damping) induce melting mood by virtue of ‘tender’ or ‘fluid’ effects; slow tempo, low pitch register, soft volume, smooth prosodic and artificulatory transition induce restful mood by virtue of effects of stillness.


It is obvious that there are properties of song that are purely musical (and thus shared with non-speech music, that is instrumental or non-verbal vocal music) and that the account just given of the expressivity function, the delectation function, and the projection function of song prosody fails to take note of these purely musical properties. These certainly have their own functions of expressivity, delectation, and projection, but they have been poorly understood and call a for a study across historical periods, cultural, groups, and laboratory subjects.


The Indian traditional thought about the performing arts applies the conceptions of  āvāda and kepana to singing and instrument-playing (gāyana and vādana) as well as dancing and paly-acting (narthana and ya) but the notion of abhinatya is applied only representational dancing (ntya) and play-acting, and not to non-representational or abstract dancing (ntta) nor to singing and instrument-playing.  But the element of expressivity (abhinaya in a large sense) does appear to be present in these letter art forms.  Again, while Indian traditional thought on the performing arts, perhaps because of it šāstra-orientation, is preoccupied more with technique than with style. It is high time that we pick up the thread from where Indian traditional thought leaves of.  This may help us to ask interesting questions such as the following.


(19)           The characteristic sequences (paka) of rāga pūriā are the following:


Gm Do Gm’G

M’N Ro S


Traditionally, singing or instrument-playing in this rāga is deemed to be conducive to karua rasa (the pathetic). Presumably, the presence of the pathos in the words helps, but is not critical.  For testing this out with listeners within and outside the tradition, wouldn’t a weaker’ hypothesis be more plausible?  The hypothesis namely that the tonal sequence is conducive to citta-druti (melting mood)?


            Given the present state of near-ignorance in this area, it will be more sensible to return to the speech or language end of the spectrum.



            First, some specimens intermediate between ordinary speech and recitative, between (2, 3) and (5, 6, 7, 8, 9).


(20)           Non-poetic drama: dialogue: non-rhetorical style

Sarasvati : u’gic kəlpəna-kərit-bsu-nəkos \\ |

                       To           : ‘hya- či ‘jəruri-nahi \\|| ‘tu svə’təMtrə-ashes /| pə-‘mi svəyəMbhu-ahe \\|| ‘ svə’t Mtryə \\||ai-svəyMbhūtvə/| hyaM-či-l’hai ‘laun-deya-t ‘mətləb-nahi \\ ||  svyMbhutv ‘imkel \\|| nehmi ‘ǰiMkel \| ‘aj-hi ‘ǰiMkel \\||


Sarasvati : pə-šīrdhər /|

            To          : ‘əikun-ghe || ‘mi-te ‘bhə ‘mage-ghear-nahi \\ || ‘tya-ca ‘lekəh-   krun/| mi-to-‘ čhapar-ahe \\|| hya-t- ‘təjo-nahi \\| ata-kə ša-t-hi ‘təjo- nahi \\ ||



(G.P. Deshapande, Uddhavasta dharma šāa, 1974, act, I scene 4: end)


‘Saraswathi: Don’t you keep imaging things to no purpose…

He        :  That was unnecessary.  You are independent, but I’m self-      dependent.   Little point in setting independence and self-dependence against each other.  Self-dependence will win.  Always win.  Win even today.


‘Saraswathi   :  But, Sridhar. . .

He                 :  Just listen! I’m not going to withdraw that speech. Indeed, turning it into an article, I’ll print it.  No compromise in this. Now, no com-pro-mise, no way !


(Note the effect of interruptions on junctures, and the tone modifiers, pitch squeeze marked and pitch stretch marked.  Sharp voice quality should suit Srdhar.  The three final tone gous, all tone-falls in loud volume are reminiscent of the terminal tihaiy in tabl-playing.  Free rhythm and melody.  Speech segments are typical of deliberate rather than informal delivery.)


(21)           Speech-making : monologue : non-rhetorical style mah’rašra-ča bhəvi’tə vvya- či /| hi-mohi-c mohi ‘laMb-əc-lamm s ə’phər-ahe \\| tya-‘səphri- til \| amhi-prə ‘vasi-ahot /| hi-‘bhavəna ‘nəmrəte-


nə-aplə -‘prə ttyek-paul-aktana \| apə-aplya-mə’na-ši ‘uhe li-pahi ǰ  e\\ ||


(Yashwant Rao chauhan)

The future story of  Maharashtra is a tremendous voyage of a long duration. In that voyage, we are but passengers.  This humble sentiment should keep us company at each step we take.’


 (Speech segments are again as in formal delivery.  Long breath groups segmented by non-final pauses.  Restrained in feeling. Peroratory low pitch register.  Texture of pitch-rises and pitch-falls.  Ornamented free rhythm.  Simple free melody).


(22)           Verse in mukapadya in kāvaya-vācana (the poet being A.R. Deshpande ‘Anil’), slightly rhetorical cə ‘kakəli ‘vij \\|| ‘gəMbhir ‘ghməuli ‘ghənə-gərǰna \\| ni’nadǰle ti’ce ‘sad-pəsadə \\|| su’ru i -jhali ‘vruši \\|

Mru ‘ga-ca-pa’us /| ‘bərsu-lage \\| thə’ rare ‘sruši \\||





‘flashed the lighting, deep the roar of the clouds sounded and resounded, comes the rain, the rain brought by the Mriga constellation begins pouring, the creation gets thrilled.’


(Note: the partial restoration of the elided to fit the ornamented free rhythm.  The adjustment of the accent groups likewise.  The melody is simple and free.  Kindled mood).


(23)           Speech –making: monolugue: rhetorical style

deš-MdhuM-no /| vi’ čar-kəra \\ || deša-la ‘ag-lagli- əstana \/| tumhi-kai ‘aplya-c su’kha- či ‘kaǰi-vahar /\ ||



‘Fellow countrymen, give a thought.  When country burns, are you just   going to care for your own little happiness?’


(The first and third tone groups in high pitch register and lax pharynx ‘booming’ voice, the other tow tone groups low pitch register.  The fourth tone group in tense pharynx voice.  All the four tone groups in loud volume, smooth prosodic transition, precise articulation.  Ornamented free rhythm.  Simple free melody.)

(24)    Poetic darama: monologue: rhetorical style Lakh mərot  /| pə-lakhaM-ca pš iMda-nə-məro /\ || raǰa /| tu ǰha ‘hukma- ča ek-eka bola –bərobər /| mohr’lelya ‘jhaaM-na sabǰi ‘lagtil \\ ||

(Ram Ganesh Gadkari, Rājasanyāsa, act 5, sene 5, Sābāji’s speech addressed to the king)



Let a hundred thousand die, but not the one who is their mainstay, O king, you have  only to say a word of command, and the trees in bloom will yield a harvest of Sābājīs like me.’ 

(Here also belongs the ceremonial calling our by attendants at law courts and darbār courts.)


(25)    Crying out in various moods

d’əhi: ‘ghyai –cə-ka də’hi:: \\||   

dahi on sale !’

(Hawker’s cry)

‘gəpəti-bappa \/ |’ mo: rəya /\ ||’ puhlya-vərši: \/| ləkr-ya /| ||


‘Father Ganpati, next year come back soon!

(Cry at the religious visarǰana procession of Ganapati, also called Morayā.)

‘ysvMt-ravaM-ca \/ | vi’ǰəy-əso /\ ||

hail Yashvant Rao ! Victory to him!.


(Slogan of political support.  (Likewise, slogans of political protest or rejection.)

 ǰi: rli /| | ‘ji: rli’ /\ | I ka masa- či \/ | ǰi: rli /\ ||

‘somebody got discomfited !’


(Children’s jeering cry.  Clips and drawls, reversed pitch contours, simple bound rhythm, ornamented free melody.

(26)    Narrative monologue: rhetorical style

a-pa nəgər-hotə \\| ti’thə -ek gərib ‘brahmhə-hota \\||

‘There was a city, broad and wide.  There lived a poor brahman.’

(Familiar way of opening a kahāī, a folk genre favoured by upper caste women.

hi-saha-uttəraM-či ‘kahi //| ‘paca-uttəri suphəl səMpru \\\||


Thus endeth the story of sixty pharases in five, yielding well and fully (saūpra).’

(Familiar way of closing a kahāī.)

ek-hota ‘raǰa \/ || ‘ek-hoti rai :I /\ ||

There was a king.  There was a queen too.’

Rai-mhəali /\ | ‘nəkko-nkko /\ | mi-raǰa-v ər \/ | ‘ru: ste /\||

The queen said, please don’t ! I agree to throw a tantrum at the king!’

(Both are examples from a nursery tale. Recurrence and mirroring of reversed tones.  Ornamented free rhythm and ornamented free melody.)


(27)    Children’s chants

|    S-D1      |S-D


ye-re    ye-re    pa: usa /\ |

tu:-la    deto    pi:sa /\ |

pəi: sa    jhala     kho: a /\|

pa: us    ala    ‘mo:ha /\ |



Do come, O rain !

Here I give a pice to you.

The pice turns out false!

The rain comes big!’

(Nursery rhyme.  Pitch-strentch, drawl, reversed tone, simple bound rhythm and melody.)

| N1-N1-S-S                 | N1-N1-S-S                        | N1-N1-S-S                        | N1-N1-S-S           

I’ riMg mi’ riMg     iəuMg  triMg\|             ləu’ Mga-triMg-ca            əbəb-baǰa \|


            | S-D1=S-D1                                             | G-R-RG-S

            ‘giya ‘gopi                                        utria ‘raa \\\ ||


(Nursery spell for identifying the ‘it’ in a game-opening and close in the course of the nursery rite of cake.)


            Now, some specimens intermediate between recitative and song, between (5, 6, 7, 8, 9) and (10, 11).


(28)    Verse in chanda metre (abhaga) metre, which features recurrence, is being illustrated in one of its many variations, the poet being Tukārām)


|| D1D-D1-s-s-s-ss               || RR-RS-S-R


Stanza ‘deva-ce ghərin’deven-keli ‘cori \| ‘deven ‘devə nagə ‘vuni ‘kela bhikai \\ ||

            | |  RR-RS-S-R                      | | R-S-S-S

Refrain dhan ‘viyan dhanva               ‘mag-ci-nahin ‘javen

            Jhani viya ‘dhanva \ |                         k viya ‘gan va / ||


            D1D1-D1 S-S –SS                 RR-RS-S-R


Stanza  səven- či-hota ‘cor                  əvghen-kelen va’len   

                        Ghə’ ričiya ghə rin |                                        Phan’vəliya-vəri \\ ||

Thrid     tuka-mhəne ‘yethe koi-cə                         na’ gəvə len-koṇa ‘gelen koa-cen


                        nahin \|                                                                          ‘kain /||

(Tukaram, the Government edition, no. 1840, line 2b slightly emended to fit the rhyme scheme)


(29)    Another poem of Tukrm, also in abhanga metre (no.3903 in the same collection), is presented here without the rhythm and melody statements.


First stanza            avə ’deəl ‘tisen ‘tujə aə’vinə \| vae səma’ dhanə’ ǰva

Refrain             ‘təi isen \ ||

                        nahin-yethen’kanhin  \ ləukika-či caə\| tujə-vṇa ‘godə

Second stanza  purəvi mə’ norəthəəMətrin-cen ‘artə \| dhaye-vəri

                        Gt ‘gain tu jhen \\ ||

Third stanza            tu’ka mhə’ e ‘lenkin aəvi ma’hera \| ‘gaun-ya səun’ sara ‘tujə

                        təi sen \ ||



While both (28) and (29) occupy the same position in the spectrum, namely intermediate between recitative and song, there is an important difference in that  (29) is also available for a clearly songlike rendering. The two poems represent two distinct veins in Tukārām, though there are no conventional terms for these.  (The  two comparable veins in Kabīr are recognized genres, namely, dohā (or saloku) and pada (or šabad) corresponding to (28) and (29) respectively.  Let us call these the cognitive and the cantatory modes.


            The first mode lends itself to sermonizing (pravacana).  The second mode to hym-singing (bhajana).  The first appeals to the intellect bearing the heat of sentiment, the second speaks to the sentiment bearing the heat of intellect.  In the examples chosen here, (28) presents a spiritual conundrum and offers a spiritually satisfying answer; (29) represents a throwing of oneself to divine compassion in the mood of a young bride’s aching for her parental home.  The English renderings should bring this out.


            First, (28) in a ‘modernized’ English rendering by Arun Kolatkar:


            It was a case Of God rob god.

            No cleaner job Was ever done.

            God left God’ Without a bean.

            God left no trace/No trail no track.

            The thief was lying/Low in His flat.

            When he moved/He moved fast.


                        Tuka says:


            Nobody was/Nowhere.  None


            Was plunderd/And Lost nothing.


            The experience of non-duality (advaita) between Me and Him is presented here with a certain impish humour characteristic of Tukārām (who shares it with Kabir in some of his moments).


            Next, (29) which turns on the notion of āavaī that means both pacifying or conciliating a child, a beloved, God through soothing gestures and also bringing out the possibilities of a melody or song by slowly teasing them out.  (The verb appears in the first and the last stanzas.)


Let me persuade as I love how to

Certainly that soothes my restive soul (1)

No caring here for the worldly

Nothing pleases without you dear God (R)

Fulfill my heart’s desire, my inner longing

Here I break into loud plaint of a song (2)

Tuka says: Even as the (married) daughter sings for maternal home 

    So left us sing to you in our worldly home


What is common to both these lyric modes of Tukaram, the cognitive and the cantatory, is, however, the primacy of expressivity (abhinaya) and projection (kepaa) over delectation (āsvāda)-recall (17a, b2, b1) respectively.


            At the level of technique, one considers the availability of various speech-art devices and their selection.  At the level of style, one takes up two questions.  First, which alternative  quality (achieved through the selection of this or that speech-device) is conductive (or non-conductive in a specific work of art, a specific context, a specific artist, a specific place and period ambience?  Secondly, conducive (or non-conducive) to which goal expressivity, delectation, or projection?


            The veil of near-ignorance in the area of the art-oriented consideration of speech can be lifted only through a resolute, precise, methodical, and sensitively alert inquiry.  What we hope to have provided in the present effort are some of the needed conceptual tools.


            This is a reworking of an earlier Marathi article ‘Kalāvatī vānī, to appear in my book Madhyamā (Pune, late 1996).


            The phonological apparatus was elucidated in Kelkar 1993 (Prosodies and their functions in Marathi, Bulletin of the Deccan College Postgraduate & Research Institute 51-52 (1991-92): 239-302, published 1993 (Katre felicitation volume).


            The present enquiry into the arts of speech is indebted to: G.H. Ranade, Hindustani music: An outline of its physics and aesthetics, 2nd edition, Pune: The Author, 1951.  Baburao Joshi, Chandašāstra va sagīta, Kolhapur: Ajab Pustakalaya, 1980, coming with an audiocassette.  Discussions with Dr. Veena Saharabuddhe (vocalist), Dr. Hari Sahasrabuddhe (computer scientist), and Mr. Prabhakar Jog (music arranger).  The debt extends both to the discussion and the illustrations.


            It is a pleasure to offer this to my friend Biligiri, who is equally entranced by language and music.

            Postscript:  It was a shock to learn of is sudden death after this place was ready.




            The Marathi original ‘Kalavatīi: Bole,mhanne gāe appeared in Ashok R. Kelkar’s Mdhyamā: Bhāā āni,  bhāāvyavahar, Pune: Mehta, 1996, 231-52.  The English version was intended for a Fesric for H. S. Biligiri, but has remained unpublished.