And how to put up with these stupid sectarians Sullying the holy books with murderous commentaries?
Victor Hugo

1The Book of Love by Tiruvaḷḷuvar, circa 5^th c. CE, is the third part of a triptych, the Tirukkuṟaḷ, the most popular and widely diffused text of the early Tamil literature known as the Caṅkam, its other aphorisms on wisdom and fortune reflecting the ideal of the ancient Tamils.1 In twenty-five chapters of ten distiches each, it is the quintessence of the Tamil art of love, and the medieval commentaries that go with it express in exemplary manner all the approaches elaborated around the akam texts, that is: interior literature on amorous themes. A seminar devoted to literary genres within the Indian domain is therefore the ideal occasion on which to propose, from a privileged position, variations on a wider vision of the commentary as the particular mode of expression of a culture.

2The great mass of literary, grammatical and religious commentaries in the literature of India terrify or fascinate, but, before attempting to order the apparently rather contradictory visions and to justify our own choice, the first thing to notice is that the phenomenon is universal and has a special place in occidental literature too.

3Before the least methodological reflection on the nature and range of a commentary is undertaken, it is the proliferation and the universality of the genre which is striking. Confining ourselves to the domain of French, and given that pastiche consecrates a genre, we recall, for example, Le Chef-d’oeuvre d’un inconnu, of Thémiseul de Saint Hyacinthe, presented as the edition of a learned dissertation by Doctor Chrysostome Matanasius on the five stanzas of an anonymous, popular, short poem which sings the love of Colin and Catos.

4Published in 1714, this erudite hoax, which is also a wink at libertine thought, went through a dozen editions in half a century. The charges against the uncontrolled dizziness of the erudition (ornithological for example), against the delirious interpretation applied to the punctuation, the onomatopoeia and the truisms, and against the elusive learning that never gives its sources and through a rare dithyrambe exposes a weakness in the metre as though it were a literary merit, are so familiar, and so well identify the breadth of a genre that this humorous piece, meant for initiates, has taken on universal value and could just as well serve as a caricature of the Indian commentators too.

5The popularity of Le Chef-d’oeuvre was so great that it often appears in the very classic study of private libraries of the 18^th c. by Daniel Mornet: “129 times, which puts it in the 14^th row of frequency, out of all the genres together, outnumbering the novels”.2 This does not surprise us. If this text which had been temporarily forgotten is being reissued today, it is because the importance of rhetoric at the heart of French literature and the systematic exploration of the classical grammatical traditions have already been, for many years, the subject of research and of important works which contribute to the rehabilitation, supposing it to be necessary, of the literature of commentaries. These are, consequently, an invitation to integrate Indian studies in this area into an ensemble of epistemological approaches, quite various in origin, which tend to converge in their analyses and their results.

6Whilst, in India, the works of Pāṇini in Sanskrit and the tradition of the Tolkāppiyam in Tamil are the subject of research, of new editions and translations, we witnessed an analogous re-launching in France when, in 1980, Marc Fumaroli published his monumental thesis on L’Âge de l’Éloquence, Rhétorique et “res literaria” de la Renaissance au seuil de l’époque classique, and when, the following year, a no less enormous study on the Ars Donati and its diffusion, with a critical edition of the Latin text by Louis Holtz,3 established itself, with reference to the grammatical teachings symbolised by Donat for Latin scholars, on a route that numerous works have been ceaselessly marking out ever since. We thus have no hesitation in resorting to the analogy of an occidental classical world, more and more thoroughly explored, to evoke for the benefit of non-Tamil scholars what represents, in form as in substance, the literature of the Tamil commentators.

7Behind a preliminary formal question: commentaries or scholia, marginal notes or technical treatises, looms the fundamental problem of the relationship between the work and its exegeses, between the literature and the grammatical tradition, between written monuments of erudition and the living milieu that animates them within a learned oral tradition which has nothing exclusively Indian about it.

8The better to understand the history of our own rhetoric, we simultaneously refine our vision of the pandit and his role. As we examine more thoroughly all the grammatical traditions specifically in each area, and therefore, at the heart of the Tamil or Sanskrit tradition, the parallel between two different worlds stands out more clearly, the erudite scholars and the school teachers, and between the great masters and the more modest connective tissue of minor grammarians, with, always and everywhere, the terminological stiffness which characterises the latter. An urgent task for research, therefore, is the cleansing of the phraseologies by a return to the sources and the quest for successive chronological layers.

9We shall say no more than a word about another striking analogy. The terms of Alexandrinism or of Hellenistic or Byzantine culture often suggest themselves in our approach to India. In a certain way, from the time when to be “Greek” at Rome signified belonging to a culture, the book itself is thereby privileged, and grammar reserved first and foremost for literary work rather than for the science of the current language. The Tamil grammatical tradition is also fitted to a particular mode of literary expression, that of the great works recognised as such. In contrast to modern linguistic research, the early grammatical tradition puts first, in Tamil as in Latin or Greek, the analysis and interpretation of texts. This axiom has two practical consequences both of which are obstacles to linguistic generalisation. The first is that during the process there is a tendency for the systematic character to be lost and to have to be recuperated all the time, and the second is that normative observation is invariably tempered by the principle, known since Quntilian at least, that poetic utterance transcends linguistic correctness. It is in this way that the lexicographic work of the Tamil commentators focuses, as that of the rhetoricians, on the study of homonyms and rare words, the terms called uriccol, “specific”.

10The reading of early Tamil literature is therefore almost always doubled, in the manuscripts as in the printed editions, by systematic annotations which it is convenient to call “commentaries”. More than a series of “glosses”, or remarks on the terms and proceeding by synonyms, these notes amount to a “perpetual commentary” which follows the texts step by step, to develop the meaning or specify articulations; this commentary is accompanied by scholia, marginal notes consisting of collations of variants, and literary extrapolations which reveal, explicitly or, more often, by veiled allusion, the independent existence of a theoretical debate about the texts as well as the presence, beneath the surface, of another corpus devoted to the theorisation of this debate. Just as, regarding Homer or Pindar, we may hesitate over the most pertinent terminology: the mass of scholia tend to expand into commentaries, and those preserved reflect the independent existence of more technical treatises, often by the same authors.

11We have therefore, in Tamil, on the one hand, a very fine tradition of grammatical treatises augmented by technical commentaries with grammar as subject and literature for reference, centred around morphological debate, the aesthetic of the language, and certainly the poetics and on the other hand, one, or several, outstanding commentaries attached to each great literary text. We should guess that there were countless inserts and cross checkings between the first series, which tends to be spread out, in meaning more than through form, in treatise on grammar, prosody or rhetoric, and the second, which seems more like a selective exercise of memory and reflection applied to texts of another kind.

12We may speak in terms of “literary” and “grammatical” commentaries,4 but there is, in actual fact, a single corpus, in which the same tools have various different applications and where the first concern of the authors is not to enlighten us in advance as to the nature of their tools. The great commentators have clearly created, in the margins of literary texts, an actual work as well as linguistic and rhetorical systems but it is, all the same, through those systems that we may today embark upon the clarified reading of the works. At the same time, the link of allegiance to the work of good authors remains the claim and the first exigency of the commentators: these refute or complement each other in the name of the very texts they cite for their various divergent readings.

13This is a long standing debate in which no commentator’s effort at discrimination is without polemic and vanity, though they all advocate perfect modesty and oppose controversy as well as panegyric. All commentators are theoretically supposed to subscribe to the remark of Saint Jerome for whom a commentary meant: “that I do not seek to have my own words praised, but that what has been well said by another be understood in the sense in which he has said it”.5

14Five or six centuries after Saint Jerome, a very beautiful Tamil kāvya (Tam. kāppiyam), the Cīvakacintāmaṇi, would provide that idea of the appropriateness of the commentary to the text with a remarkable image of which it was the genesis. Evoking the perfection of a lyric performance, the poet compares the accord between the song and its accompaniment on the lute with the kite and its shadow (III, 143). The commentator, Nacciṉārkkiṉiyar would later explain that the song “climbs” the scale as the kite rises in immobile flight, to the exact place of its choice, while no trembling of its wings alters its shadow projected on the earth. But even before him, both the commentators on the Cilappatikāram (III, 143), repeating each other word for word, had employed this image for the union of two groups of various musical instruments; the anonymous verse preface on the glory of Aṭiyārkkunallār, (one of these two commentators), hints at it to celebrate the richness this latter put into his work in the service “of the poem and of its commentary which are like the kite and its shadow”:

Paruntu niḻalumenap pāvu muraiyum.

15The commentary and the text, as the kite and its shadow are indissoluble, covering one another exactly with neither dissonance nor misjoin. This is antithetical to the shifts and slips, vaḻu, denounced in the work of others by those who make every effort to avoid them in their own texts. The metaphor of the kite in the preface to Naṉṉūl is again used to characterise one of the steps in the progression from sūtra to sūtra of a text: unconcerned with sequences, the kite drops from a great height straight down upon its prey.6 But in this context the original image has lost something of its admirable strength and uniqueness.

16If the literary and grammatical commentaries form a coherent unit, the only one, moreover, which is referred to as a matter of course, we shall benefit by remembering that any work of technical literature (the śāstra) will usually have its retinue of commentaries; the Sanskrit Dharmaśāstra(s) are an impressive example but there are others, and in Tamil itself. We mention here a single case, that of the religious commentaries grafted onto the corpus of Vaishnava Bhakti, the hymns of the Āḻvār, since an essential characteristic of these texts is of immediate interest with regard to an original aspect of the most celebrated commentator on the Book of Love (Kāmattuppāl), Parimēlaḻakar. The important place of Sanskrit vocabulary in these medieval theological exegeses is not always properly understood. It has given birth to a particular jargon, the maṇipravāḷam, a mixture of Sanskrit and Tamil (also to be found in various Saivite and, more often, Jain texts) which some have unfortunately chosen to see as the esoteric expression of the sectarian tendencies of later vaishnavism, and not as a positive contribution to Tamil culture. But, as some recent works7 make evident, the historical reality is more gratifying for the Vaishnava. On the one hand, it was they who made of the Tamil texts of their hymnists veritable sacred texts of a theological authority essential to the doctrine: for the first time in India, in this current of ubhayavedānta, a vernacular language, Tamil, is placed on a par with Sanskrit, not only for the singing of hymns, but also for defining and specifying the revelation. This is a phenomenon which must not be underestimated. On the other hand, at the origins of the sectarian disagreements between Tamil Vaishnavite “southernists” and “northernists” there is, first of all, the spreading out of the same current in two different contexts. In the South, which is to say, at Sriraṅkam, the Tamil milieu showed itself more co-operative and convivial, the most popular predication there being permeable to the vernacular culture to a greater degree. In the North, that is, at Kanchipuram, an environment more cosmopolitan, philosophically diverse and critical, imposed a more polemic logic and an argumentation of external usage whose elements and vocabulary must have been taken from its universal storehouse and supported, beyond Tamil, by Sanskrit, which had greater prestige as a language of communication was better adapted to conveying conviction. When we wonder about the reasons for this choice and remember that the commentator of the Book of Love, whose approach was the most sanskritized, Parimēlaḻakar, belonged to the Kanchipuram of the 13^th c., and was thus immersed in a cultural milieu which naturally led him to see in Sanskrit the most universal expression for the propagating of his views, we are no longer surprised to see him using a more capacious terminology in the service of the Book of Love. Today, with all the necessary transpositions in place, modern commentators cannot do any differently.

17The series of commentaries attached to the classical texts is of very variable extension. We have spoken of the considerable proliferation of Vaishnava religious commentaries. In the profane corpus, it is the fundamental grammatical text, the Tolkāppiyam, which has engendered the greatest number of commentaries; the habit of giving to its gloss the character of an all-encompassing teaching still survives. The latest exegesis to date preens itself on a brief exposition of the haiku and another on contemporary vers libre (putuk kavitai); no matter if the former is not at all technical and the latter purely negative and polemical, the essential is to show that the tradition is quite able to accommodate them both. In contrast, the classical texts of Tamil literature generally have no more than one commentary each, and that not always complete. The Cilippatikāram and the Tirukkōvaiyār have two. By this token, the Book of Love is very privileged, since stanza 41 of the Toṇṭai maṇṭala catakam gives ten commentators for the Tirukkuṟaḷ. Five, it is true, are nothing but names, Tarumar, Tāmattar, Naccar, Tirumalaiyār and Maḷḷar, but the other five texts survive. In a chronological, and rather conjectural and uncertain, order they are: Maṇakkuṭavar (c. 950-1050), Paripperumāḷ (c. 1000-1100), Kāḷiṅkar (c. 1100-1200), Parimēlaḻakar (1250-1300), and Paritiyār (c. 1450). An old anonymous commentary, also edited, and the modern gloss of Rāmānuja Kavirāyar, considered, however, as classical, puts the number of known commentaries at more than half a dozen, not counting contemporary editors. Only the Tirumurukāṟṟuppaṭai, which owes its popularity to its religious character, approaches this abundance, acknowledging five commentators: Nacciṉārkkiṉiyar, Parimēlaḻakar, Kavipperumāḷ (often identified as Paripperumāḷ), Pariti, and a “Commentator” (uraiyāciriyar), who has been tentatively identified as Pērāciriyar, a case of giving to the rich. We see from this enumeration, in fact, that the great names are often the same and that we are dealing less with homonyms than with a relatively narrow milieu where cross references are frequent. The detail of the grammatical or literary theorisation underlying this corpus is also the fruit of the progressive elaboration of systematic expositions to which the commentators always refer implicitly. The reading of the Book of Love, therefore, goes through the succession of those works which, properly speaking, intermingle with the chronology of its exegetists. We should briefly remember these stages.

18The Book of Love itself belongs to the last wave of the literature of what is called the Caṅkam which is still seen as extending from the beginning of the common era to the 6^th c. or even 7^th c. To that ensemble already belonged the brief and sibylline sūtra(s) of the Tolkāppiyam, which is the linguistic and rhetorical conscience of the Caṅkam poets, given by the tradition a date anterior to that given to the poems; a hypercritical modern diorthontic movement rips it up into successive strata without, however, carrying full conviction.

19Following the Caṅkam poems, comes the period of the colophons that the manuscripts have also preserved for us, the tacit assumption being that they are more or less contemporary with the final arrangement of the texts into Anthologies, giving evidence thereby of a first effort at systematic application of the theory to the existing poems. We have already noted of the waverings, if not downright contradictions, in this game of labels, the indicative value of which has subsequently taken on uncontested authority.

20The first systematic erotic treatise seems to have been composed towards the end of the same period, and most probably in the Pāṇṭiya country around Madurai. This materia amatoria (akapporuḷ) in sixty sūtra is attributed to the “Lord” (Iṟaiyaṉār), Siva himself, and its commentary by a certain Nakkīrar takes its terminus ad quem from the generally accepted identification of the Pāṇṭiya king celebrated by the series of stanzas used as illustrations to the aphorisms and constituting, what is called the Pāṇṭikkōvai, the first model of a genre.

21In the 9^th c. the Tirukkōvaiyār of Māṇikkavācakar gives us, in 400 stanzas “the sacred string” of 400 stated situations (tuṟai) of conventional utterances (kiḷavi) which imposed themselves, thanks to the authority of the commentary of Pērāciriyar in the 13^th c., as the quasi immutable archetype of the sequential unfolding of a love story. Compared with this monumental poem on “religious” eroticism devoted to the praises of Siva at Cidambaram (Tam. Tillai), the 25 sūtra on love in the Tamiḻneṟiviḷakkam are recalled here for the sake, simply, of the memory. The commentary of Maṇakkuṭavar on the Book of Love is dated from towards the end of the 10^th c., which makes the text unquestionably the first Tamil commentary devoted to an exclusively literary work (as opposed to a technical treatise such as the Iṟaiyaṉār akapporuḷ.).

22The 11^th c., in fact, saw the flourishing of the first great commentaries: that of Iḷampūraṇar on the Tolkāppiyam, the anonymous commentary to the Cilappatikāram, and a second commentary to the Book of Love, by Paripperumāl, who, though he copied fairly faithfully that of Maṇakkuṭavar, left the memory of a subtle scholar and a good Sanskritist. The Vīracōḻiyam appeared in the same century with a commentary probably of the same date.

23Here, for the first time, Tamil grammar is systematically presented in the terms of Sanskrit grammar; but the dozen stanzas allotted to the subject of love, which are also the most poorly edited, have exceptionally conserved the substance and terms of the Tamil vision. The latest edition, by T. V. Gopal Iyer, is accompanied with his own commentary, but, is nevertheless not in the format of a critical edition (Srimat Antavan Asramam, Chennai, 2005).

24With the 12^th, and certainly the 13^th c., the age of the commentators reaches its apogee, when Naṟkavirāca Nampi compiled the Akapporuḷ viḷakkam (Light on the matter of love), the perfect, complete, classical compendium. The next two commentators on the Book of Love were the Saivite, Kāḷiṅkar, and the Vaishnavite, Parimēlaḻakar, who belonged respectively to the beginning and to the end of the 13^th c., which gave the latter who was in a position to quote from and correct the former the reputation he still has of being the definitive authority.

25The 14^th and 15^th centuries, however, maintained the tradition of the commentaries at a high level: Nacciṉārkkiṉiyar is without doubt the most brilliant interpreter of the Tolkāppiyam, the Kaḷaviyaṟ Kārikai and its commentary representing the last important classic on the art of love. Paritiyār, around 1450, writes the last classic commentary of the Book of Love to survive, his text never succeeding, however, in causing that of Parimēlaḻakar to be superseded.

26The grammarian tradition has been continuous, and we shall pause here, rather arbitrarily, since the moment has come for us to assess that tradition in the diversity of its approaches which are almost all represented in their essentials in the works of the authors already cited.

27The presentations accessible in English are indispensable for a view of the entirety. But the general studies, since the pioneering work of John Marr and V. Sp. Manickam up to the popular handbook by K. Zvelebil,8 finding it difficult to be comprehensive, have practised a technique of zooming, which does not always manage to make clear exactly which level it is operating on: certainly not that of a particular text; the style of exposition of the doctrine also precludes the inclusion of a chronological dimension. By contrast, the very systematic monograph of Takanobu Takahashi9 creates much more awareness of the intervals and possible gaps between different states of the theorisation which are separated by durations doubtless difficult to estimate with precision. He has a tendency to forget, due perhaps to being too much concerned with the comparison between the rhetorical theory and the “reality” of the known poems, that this is the preoccupation of a critic or a theoretician and not the most important aspect for the authors themselves. His work, however, brings out the contradictions between the different commentaries on the same text and, moreover, dares to question the verdict of the colophons which, as we have already said, has been too eagerly accepted. We meet with a number of cases where the colophons themselves offer a choice of two, or even three, possible labels for the same poem. Might not this be interpreted as a proper flexibility as much as the reflection of debates that were left open by the commentators themselves? The truth is indubitably to be found in a delicate appreciation of the margins of deviation available to the authors for the expression of their personality in a conventional framework which has its limitations, but we know so many cultures which offer comparable models. Be that as it may, the work of M. Takahashi is certainly one of the best and the one most to be recommended to those who wish to enter into the heart of the matter. We must mention too a very solid thesis by Indira Manuel submitted at the University of Trivandrum in 1984 but published only in 1997.10 This is a very minute and notably complete ordering of the theory, but without any comparison with the literature.

28When we pore over the existing commentaries to the Book of Love, we immediately notice, with regard to the economy of the work, an almost total consensus on the number, title, and sequence of chapters, and the choice of distiches that each groups together. On the other hand, the arrangement of ten distiches within each chapter is astonishingly variable; it continues to be surprising, that in a text so much studied that arrangement has not sparked off more interest amongst critics. Parimēlaḻakar, however, whose text has so much authority that it is practically the only one on sale, and, too, is the only one whose numeration, and therefore arrangement, is followed by all the collective editions which gather together the texts of other commentators, but always according to his, that is: in the inverse order of chronological appearance. Parimēlaḻakar is thus often alone in his opinion. Of 250 distiches, only five, that is, one out of fifty (of which three are at the beginning of chapters), are in the same place within a chapter according to all the commentators. Not more than a dozen follow the order advocated by Parimēlaḻakar in four commentaries out of five, while thirteen, and among them the last chapter in its entirety, are unanimously placed in the same rank, distinct from the order prescribed by Parimēlaḻakar.

29The distribution of distiches in each chapter is probably largely arbitrary (even this is not much remarked!), but this rather futile collating still provides evidence to show that the authority conferred on Parimēlaḻakar has too easily devalued the textual tradition anterior to him: a tradition that no one has bothered to know or to research. For example, the disruption he introduced in the last chapter, whose sequence was certainly unanimous prior to him, is obviously intended to allow for his version to conclude with the last letter of the Tamil alphabet, as the Kuṟaḷ opened with the first letter, though, to be sure in an indicative manner. Must we then rejoice to see a grid for such apocalyptic reading taking the place of a tradition with three centuries of different exegeses to its credit, of whose diversity we have lost all idea?

30It must certainly be remembered that each distich must be, in fact, treated as an isolated and independent poem and that the important thing is the grouping by chapter, and not the progression within each chapter where nothing happens, neither development nor dramatic action. The Book of Love, like all the Caṅkam anthologies, is, first of all, a collection whose significant structure and arrangement are on another level of interpretation than that of the poems taken individually. The few lines which in the commentaries connect the chapters are, in fact, the only place where we find the rare sequential texts which reveal to us, along with the general introductions and the closing remarks, the particular view of each commentator. Added up, these summarising visions help us to go from puzzle to structure with the support, naturally, of the theoretical representations, elaborated independently of the texts but essential to their comprehension.

31Maṇakkuṭavar, for example, the earliest exegetist, organises the Book of Love according to three successive types of union of the hero and the heroine. The first type is founded on the miracle of the first encounter, exceptional (arumai) union which reveals the delights of sexual joy. This takes up three chapters (109-111). The second type, detailed in eighteen chapters (112-129), concerns union after separation and thus mostly sing of waiting and the longing for union. The third type, celebrated in four chapters (130-133), is the union which takes place after a dispute and consummates the happiness of the reconciliation, refined with voluptuousness.

32Maṇakkuṭavar is taken up and developed by Paripperumāḷ, often in analogous terms, but with a supply of Sanskrit references (we recognise the Daśarūpaka and Vātsyāyana), and the Tamil copying of the three elements of the erotic rasa: ayoga, viprayoga and sambhoga, and also of Tamil ones (Tolkāppiyam and the Akapporuḷ of Iṟaiyaṉār, that is, the most classical and fundamental). This balance betrays the author’s concern with conciliating the tripartite division adopted for the ensemble, which is most often of Sanskrit obedience with a conception of five types of behaviour (or tiṇai) linked with five regions or landscapes, which are presented in his Tamil sources and to which we shall come back, but not without firmly underlining, from now on, the presence of that vision as early as the first exegetists. Five Tamil elements, joy, separation, waiting, lamentations and disputes, may be reduced to three Sanskrit terms if we postulate that waiting and lamenting are part of separation and simply particular aspects of it; in the same way we may rally to the first view of the three unions as expressed by Maṇakkuṭavar: 1) on the occasion of the first meeting, 2) after separation, and 3) at the end of a quarrel.

33Paripperumāḷ, a subtle conciliator, also practises the art of transitions thus integrating the entire global structure of the Tirukkuṟaḷ into the fundamental, and pan-Indian, grid of the aims of man (puruṣārtha). Introducing the Book of Love after the two previous Books, he states in substance that it is devoted to the pleasure born of the union with a young woman, as may take place with a virgin, a prostitute or the wife of another. Now, as the Book of Wisdom (Aṟattuppāl) rejects the wife of another as a sin, whilst the Book of Fortune (Poruṭpāl) advises against wasting money with a prostitute, the Book of Love thus shows how to discover, with a virgin, the happiness reserved for the gandharvas, the only possible one of eight types of marriage which it makes sense to see as ideal given that three of the other seven are founded upon unrequited love and four are ill assorted.

34This is how, in a few anodyne lines, the learned commentator quite casually introduces us into the core of the theoretical models of the Tamil art of love, according to the first sūtra(s) on the subject of love in the Tolkāppiyam, or the beginning of the Akapporuḷ of Iṟaiyaṉār, which are used, with various Sanskrit antecedents, as the overall framework for the reading of the Book of Love. We shall see that Parimēlaḻakar did not do much better two centuries later. We shall also bear in mind the mention of the three aims of man (aṟam, poruḷ, iṉpam, or as they are better known in Sanskrit, dharma, artha, kāma, order, profit, pleasure), that Parimēlaḻakar too would take care not to forget at the end of his commentary.

35The commentator Kāḷiṅkar would, in his turn, borrow from the Tolkāppiyam another concept rich in ulterior development. From here on we turn to the statements, or utterances, and the speakers; what is important is to know who is speaking: the lover (the first seven chapters), the beloved (the following twelve chapters) or the two alternately (the last six chapters). A poem attributed in the Garland of Tiruvaḷḷuvar to Mōcikīraṉār, a Caṅkam poet, presents the same distribution and might therefore be at the origin of Kāḷiṅkar’s thesis. This thesis is hard to accept without qualification, however, even if accord on three fifths of the text is reached without discussion. Each distich is, in fact, an isolated utterance which could be put into more than one mouth and, aware of the problem, most of the exegetists having posed it and resolved it in various different ways, often prudently avoiding taking the high ground and putting forward an overly general proposition. Parimēlaḻakar also seems to attribute to the lover, chapters 109-112, to the heroine, chapters 116-127, and to share chapters 113-115 and 128-133 between the two. Maṇakkuṭavar gives three chapters to the hero (110-112) as compared with six for the heroine (115-127) and six chapters are shared out in varying proportions (109, 113, 114, 128, 133).

36The choices are delicate in fact in the detail. Thus the use of “we”, the first person plural, prevents us from making a clear distinction between the heroine and her companion who often speaks in her name using “we” as the indefinite “on” of our European duennas. The Tamil language has plural or respectful forms which are epicene and this helps to increase the ambiguity. In distich 1151, a variant has to be introduced, as Kāḷiṅkar, and Iḷampūraṇar the first commentator of the Tolkāppiyam, have done, to align the distich with the rest of the chapter, that is: the mistress addressing her attendant; otherwise we must suppose that the young woman is directly addressing her lord or that it is the attendant who has been delegated to do that: three possibilities altogether, just for a simple possessive adjective: “your” (niṉ) return or “his” return.11

37The problem which arises here is of importance for all Caṅkam literature. On the one hand, there must not be any mistake: if the distiches of the Book of Love sometimes look like repartee, this does not mean that a consistent dialogue can be made out; these are rather the words of the author juxtaposed than an exchange of retorts. On the other hand, the dramatic element is often subjacent in this literature and the Tolkāppiyam recognises no fewer than a dozen possible interveners.12 A major part of its exposition of the amorous rhetoric consists in defining who is addressing whom and with what type of intention.

38This brings us close to another form of particularly elaborated theory of literature, but this time in Sanskrit: dramaturgy. The analogies between it and the situations evoked by the Tamil poems are, naturally, multiple, up to the terminological parallelism. We need to emphasise, however, that the Tamil commentators, though passionately interested in Sanskrit rhetoric, were not engaged in these kinds of comparison whereas to indologists, by contrast, this seems one of the most urgent tasks incumbent upon research.

39The commentators do not, for all that, seem to have been afraid of keeping company with Sanskrit, and Parimēlaḻakar less than others. We have seen above why the cultural, and also religious, milieu which was his own, at Kanchipuram at the end of the 13^th c., led him into spreading the glory of the Book of Love by unhesitatingly demonstrating the way it was speaking Sanskrit in Tamil, and the reverse. For this, as we shall see now, is certainly what the two short paragraphs of his presentation are about.

40In the first stage, Parimēlaḻakar has it that that part of the Tirukkuṟaḷ speaks of pleasure (iṉpam), defined by amorous pleasure (kāma iṉpam), which is enjoyed with the five senses and whose excellence (ciṟappu) is presented in a Sanskrit work by the king Bhoja, which proclaims this savour (cuvai) to be the one which contains all the others. It is to be understood that it is a matter of śṛṅgāra rasa, object of the śṛṅgāra prakāśa, whose definition consists essentially of two parts, enjoyment and separation, sambhoga and vipralambha in Sanskrit, and puṇarcci and pirivu in Tamil.

41If we now accept as evident, with regard to the Tamil term pirivu, separation, the suggestion, guilelessly put forward by Parimēlaḻakar that the Book of Love also brings together waiting, lamenting and dispute, taking up the procedure of Paripperumāl described above and developing it, we shall bring together under the two terms, which cover the two fundamental concepts of the Sanskrit śṛṅgāra rasa, the five categories which are those described in the Tamil Tolkāppiyam under the vocable tiṇai, “behaviour or state of soul”. This is but the first step in a syncretic progression, a daring one, to say the least.

42In the following stage (note that everything is said in only three sentences!) Parimēlaḻakar hints that, in conformity with the Tamil works, the author has reclassified puṇarcci and pirivu, enjoyment and separation, under the designations kaḷavu and kaṟpu, spontaneous love and domestic love, which constitute the key notes of classical Tamil erotic art, corresponding to two basic kinds of conduct, called kaikōḷ, and, as such, made each the subject of a separate chapter of the Tolkāppiyam. Under cover of a simple terminological convenience, that sided in a way with majority usage, Parimēlaḻakar succeeded, around the two words, puṇarcci and pirivu used as interface, in switching the entire system over, to the Sanskrit (sambhoga/ vipralambha) or Tamil (kaḷavu/kaṟpu) point of view, according to his aims, and letting it drift through the centuries under the double banner of Tolkāppiyam and Bhoja. We could even say, today, that Parimēlaḻakar conceived a forerunner of the modem.

43The rest is mere child’s play and it was possible thenceforth for Parimēlaḻakar to describe in a few lines the subject of kaḷavu which constitutes the material of the first seven chapters. This is of course the natural and spontaneous union between two beings who meet without anyone’s intervention and come together, noticing their complete equality of body, quality, age, condition etc., a union, in brief, worthy of the gandharva (s), sheltered from illness, old age and death. As for the love defined by kaṟpu, which takes up eighteen chapters, a last colophon of Parimēlaḻakar specifies it in conformity with the Sanskrit works, as bringing together every kind of absence: that which provokes leaving on a journey (celavu, Skt. pravāsa, ch. 116), the unbearable sorrow of separation (āṟṟāmai, Skt. viraha, ch. 117-126), the wait riddled with the desire for the absent one (viṭuppu, Skt. ayoga, ch. 127-129), and the quarrel (pulavi, Skt. manaspardhā, ch. 130-133). Whether the ultimate in precision or in stylishness of the commentator, there is no place here for absence in the form of ascetic renunciation, as in the Sakskrit texts, since the fourth of the aims of man, deliverance (mokṣa) does not come into the Tirukkuṟaḷ. Parimēlaḻakar had remembered that, in the Tolkāppiyam, as in the works of the 8^th c. Tirumaṅkai Āḻvār, kāma contains in itself the concept of mokṣa, but he refrained from referring more explicitly to this idea, which was taken over by Bhakti literature in which it plays an important role.

44Parimēlaḻakar evidently did not invent this division of the Book of Love into two parts since reference is made to it by an earlier commentator, Kāḷiṅkar; but he has given it such an outline, making of it the key to his entire vision of the text in the face of all other possible poetics, that there has been an insistence since then on seeing it as the only acceptable structuring of the work, the mark, so to speak, of Tamil specificity; which in view of the ambivalence of Parimēlaḻakar’s reasoning, is paradoxical to say the least.

45There is no point, however, in challenging the power of a myth, so we leave to the historians of mentalities to decide whether or not kaḷavu and kaṟpu remain two cardinal concepts of the Tamil psyche and that they share the Book of Love according to the two edged sword of Parimēlaḻakar.

46It is therefore especially presumptuous and perilous to attempt to disassociate oneself from a doctrine which has been unanimously accepted for seven centuries. We have, all the same, aimed at showing that there exists, for deciphering the Book of Love, at least one other grid of traditional reading which continues to have in its favour some very forcible presumptions; we mean the concept of tiṇai, region or state of soul, to which the Tolkāppiyam has devoted its first sūtra(s) allotted to amorous subjects.

47Our edition has deliberately ignored the tradition of the known editions and translations which all present the Book of Love in two parts, kaḷavu and kaṟpu, according to the schema of Parimēlaḻakar. We have, tentatively, chosen to redistribute it according to the five tiṇai,

• because that idea is the first to come to mind as regards the love poetry of the Caṅkam,
• because that distribution is overwhelmingly in the majority, in the arrangement of the Anthologies, and
• because it existed for the Book of Love itself in a tradition anterior to Parimēlaḻakar, of which some traces remain and which is worthy of more attention than it has so far received.

48The first chapter that the Tolkāppiyam dedicated to matters of love opens with the idea of tiṇai, and the rest of the chapter is devoted to clarifying the meaning and the field. Tiṇai is a one word illustration of Amiel’s phrase: “A landscape is a state of the soul.” But there are seven tiṇai to be fitted in and only five landscapes on the “carte du Tendre”, the love itinerary of Tamil. This is problematical only with respect to the interpretation by Iḷampūraṇar, who holds to the equation of tiṇai with a substance which is in some way material and argues that the designations of tiṇai are derived from the names of the flowers characteristic to each region, which we would naturally call emblematic. It is there that accident and substance are confused and, on this point, we prefer the, perhaps more essentialist, reading of Nacciṉārkkiṉiyar, in which the idealism corresponds more closely to the literary situation which is, in the final analysis, the real subject of the debate. According to this commentator, tiṇai is taken in the sense of oḻukkam, a type of conduct or behaviour, a state of the soul. After that, it is no longer surprising that two particular tiṇai, two extreme types of behaviour, cannot in any sense be part of the amorous Tamil landscape. Properly speaking these are fringe attitudes between which the other five types make up a balanced environment. There is in fact no actual landscape which is not socialised, and the two extreme types of behaviour involved (love which is unilateral and thus with no possible reciprocity, and love between an ill assorted pair who are not made to agree), are naturally rejected as anomalies by a social group which assumes two partners in accord and in harmony, qualified to enjoy their discordance. The hard reality, both in literature and in life, which is that such kinds of behaviour do exist and do not enter the Tamil landscape which refuses to integrate them, conforms perfectly to the thesis of Nacciṉārkkiṉiyar even if it constitutes, on the other hand, a very serious obstacle to that of Iḷampūraṇar.

49In the other five tiṇai, by contrast, five regions harmoniously harbour five soul states, but the meticulousness shown by Nacciṉārkkiṉiyar remains an essential warning if we seek better to understand the interior dynamic of this classificatory vision rather than being satisfied with a nomenclature. If the tiṇai are first of all, really, as he believes, kinds of comportment, or of states of the soul, it is these states which constitute the specific material (uri-p-poruḷ) and the land which shelters them is then no more than the containing or supporting element. The use of one and the same word for these two things emphasises a certain co-substantiality, but no more nor less than does the very beautiful metaphor the commentator suggests: a single word, viḷakku, expresses both the lamp and the light, the support and its raison d’être. Everything is said through that image, and we may thereby allow free range to the descriptive and classificatory mania, (demanding patience from the reader certainly), the details of which are easy to find in the general works and histories of the literature to which we refer that reader.13

50For us to understand the essential, we must emphasise the hierarchy of the constituent elements of the specific meaning or material (poruḷ) within each tiṇai. The theory names three types, 1) specific, (uri-), 2) primordial (mutal-, which refers to the supreme deity as compared to the letter A in the Tamil alphabet which is both first letter and first driving principle), and 3) seminal (karu, the name of the seed or the embryo). To clarify their interrelations, the theory, which then occupies itself with the possible permutations, provides an excellent criterion of relative significance, since the most mobile are revealed as the least fundamental. We find them thus in the reverse order:

1. It is that the karu-p-poruḷ, the embryos of development, which are the most numerous and picturesque, amount finally to a showroom of accessories from which it is easy to take, for each region, details of the flora and fauna, work and days, the principal deity and certain specific traits which have led modern interpreters to a rather anachronistic, ecologically oriented reading of the literature, whilst anthropologists have been able to collect names of functions and eating habits.
2. By contrast, the primordial elements (mutaṟporuḷ) come nearer to the heart of the problem and, in point of fact, we see the commentators perplexed: of two primordial series, one concerning time and the other space, the second is more primordial than the first, or so it appears. There are, in actual fact, two timespans, a short one limited to the cycle of the day (for example, the lovers’ clandestine midnight rendezvous, the flight across the desert under the noonday sun, the returns in the early morning after a night’s absence that unleashes conjugal rage etc.), and a long one extended over the annual cycle of seasons, the dryness of summer being most appropriately the symbol of absence whilst the rains promise reunions, whether at home or in secret meetings. We notice how, in the texts of the poems, these conventions are still able to protect the flexibility without endangering the system of the rhetoric.
3. Primordial too, and much more so, is the distribution in space. We know that the five regions, which are the mutaṟporuḷ, may be called tiṇai as much as may the five types of conduct or states of soul which are the uri-p-poruḷ, the proper or specific elements, the only ones, finally, that seem to be considered as truly immutable, from the first time they appear in the text of the Tolkāppiyam itself, but which are, simultaneously inseparable from the spatial primordial elements.

51Everything has already been said by the apologue of the word, viḷakku, light and lamp: it is difficult to conceive of the light without the support of the lamp but it is, all the same, the light which is the raison d’être of the lamp. In parallel, each type of conduct must have its frame and an inanimate, extinguished landscape vanishes. We have, thus, five landscapes: mountain, desert, forest, shore and plains to which correspond, as a light to its lamp, five states of soul, most often expressed by verbs: to enjoy, to be separated, to wait, to ravage oneself, to quarrel. We note that the desert is the subject of separate treatment; as an “intermediate zone”, it is simply, in some manner, the zero degree of the other regions and, for example, has no tutelary deity at its origins; the obsession with exhaustive classifications does ultimately however, cause the lacunae to disappear. A transit area, a promise in some way of extra-territoriality, the desert is the domain par excellence of separation, a state with no meaning either, except with reference to the other states or to two other moments of time, one before and one after. In order not to be misled by this entirely relative indigence, we note that the desert is, in the existing literature, the region most sung, before even that of the mountains which is, however, the most significant in the Tamil literary imagination.14

52The most daring poetic licence, that is, confusion of tiṇai (tiṇai mayakkam) consists in knowingly, violating those correspondences which are at the heart of the Tamil poetic universe as its second nature.

53The habit has taken hold of speaking of an “interior landscape” with reference to the imaginary world of the Anthologies of the Caṅkam, and all the phraseologies do in fact contribute to the same overall perception: at the centre of the system is the essential relationship of a landscape to a state of the soul with, on the periphery, the circle of hours and seasons and the allusive glimpses of accessories and associates. More than any other, the concept of tiṇai is omnipresent in the love literature of the Caṅkam and in the particular case of the Book of Love the attention converges, not on the picturesque but on interiority. That is to say that the poetic progression towards the archetype imposes, even, curiously enough, on Parimēlaḻakar himself, the psychological vocabulary of the tiṇai as an ineluctable reference. It has not escaped anyone’s notice that these songs of amorous lyricism have the Tamil map of the “Tendre” as framework. The conventions of the vocabulary have compelled us to designate by the names of the regions the five sections of the Book of Love, but we might have been even closer to the truth had we spoken of songs of enjoyment, separation, waiting, lamentation and quarrels.

54It is thus accepted that the approach to the amorous material by way of the landscape is the first and most basic to be taught and practised in ancient Tamil and that, in the order of exposition of the Tolkāppiyam, the concepts of kaḷavu and kaṟpu come only later. But it is also this system of classification that dominates by far the greater part of the formal classification of the Caṅkam Anthologies. Of the Eight (great) Anthologies, five are devoted to the love lyric, of which four are classified according to the five tiṇai. The Aiṅkuṟunūṟu is made up of five brief “centuries” or “hundreds” each of which celebrates a different tiṇai. The Kalittokai contains a hundred and fifty poems, around thirty for each tiṇai, attributed either to a single author or to five different authors, or even more according to various colophons, in the order: desert, mountain, plains, forest, shore. The distribution of the five regions within the four hundred poems of the Akanāṉūṟu is of an astonishing refinement which gives remarkable value to the division by tiṇai and the care it has brought to it. We count in fact:

55200 desert poems: all the unevenly numbered ones: 1, 3, 5, 7, etc.

5680 mountain poems: all the poems with ratio 2 and 8: 22 and 28, 32 and 38 etc.

5740 forest poems: all those with ratio 4: 4, 14, 24 etc.

5840 poems of the plains; all those with ratio 6: 6, 16, 26 etc.

5940 poems of the shore: all those with ratio 10: 10, 20, 30, 40 etc.

60The anthology of Naṟṟiṇai is given over to disorder, but each time that there is a title it is that of a tiṇai and the four hundred poems (398 extant) are counted thus: 132 for the mountain, 106 for the desert, 102 for the shore, 32 for the plains and 28 for the forest. Only the anthology of Kuṟuntokai is classified according to speaker.

61The ten texts collected together in the Pattuppāṭṭu are of disparate structures and six are not related to the love literature. But of the other four, three have the name of a tiṇai in the title itself: Song of the mountain (Kuṟiñcippāṭṭṭu), Song of the forest (Mullaippāṭṭṭu), Haven–desert (Paṭṭiṉappālai).

62The Tirukkuṟaḷ itself belongs to an ensemble of Eighteen Minor Collections (Patiṉeṇkīḻkkaṇakku), of which only six are clearly related to the love lyric as in the Book of Love. Five of the six are explicitly, even in their titles, collections arranged according to the principle of five tiṇai. Their names are respectively: Fifty stanzas on the five tiṇai (Aintiṇai Aimpatu); Seventy stanzas on the five tiṇai (Aintiṇai Eḻupatu); Fifty stanzas on the language of the tiṇai (Tiṇaimoḻi Aimpatu); A hundred and fifty stanzas on the garland of the tiṇai (Tiṇaimālai nūṟṟimpatu); The way to behave [according to the tiṇai] (Kai-n-nilai, or kai in the sense of conduct, o ḻukkam and tiṇai; it consists of five times twelve poems but the text is to date incomplete). The sixth text does not contravene this tedious monotony since it is a matter of Forty stanzas on the season of rains (Kārnāṟpatu), in which the action takes place in a single tiṇai, the forest.

63Is any further emphasis needed? The arrangement of the Book of Love according to five tiṇai appears so natural and ordinary in this context that we dream of crediting Parimēlaḻakar with an honourable concern for originality in looking for something else. What actually needs to be explained is not why the Book of Love must be cut up into five by five chapters according to the five tiṇai, but rather why this banal and obvious division has not prevailed. This is a real question and not a paradox, but we have never seen it asked in these terms.

64The view of the literature of love according to the grid of five regions is the most classical and the most “Tamil” in all the earliest literature. But we have a few indications that, from the slightly more recent epoch of the Tirukkuṟaḷ, this type of approach might however, have increasingly become repetitive and its resilience a little scholastic. The six Minor Collections we have mentioned specifically represent a quite conscious and somewhat laborious effort to focus again on the normative regular framework of the system. We sense in looking at the way they are composed some kind of an effort, still elegant but perhaps a little artificial, in the gathering together and revitalising of a patrimony: the Book of Love may be inscribed in that moment of time as an apogee.

65We also consider the fact that the Vīracōḻiyam, like its commentary of the 11^th c, sanskritised the Tamil grammar in its totality, but, strangely enough, the essential of the five Tamil tiṇai has been preserved for the rhetorical exposition of the amorous material. It seems, however, to have failed to convey the message on that exact point, as it is this part of the commentary of which least has survived and which looks today to be the most confused.

66Are we then to be surprised that Parimēlaḻakar, eager to serve the Book of Love by valorising its most universal scope, did not see the opportunity to put the accent on a framework which brought with it the first stigmata of an obsolescent poetic? On the other hand, the forceful idea of the binomial kaḷavu and kaṟpu, the two basic types of conduct, without doubt sufficiently lively to be brought to the fore, certainly as a copy of Sanskrit binomial sambhoga and vipralambha, was obviously triumphant. In his time and in his milieu, Parimēlaḻakar no doubt better served the glory and brilliance of the Book of Love by proposing, out of the preoccupation with perpetual youthfulness, the very reading against which we are now rebelling. But if he was right then, it is probably not artificial or abnormal on our part today to bring about the re-emergence of that which the 13^th c. could not but strive to occlude.

67As do all the theses, ours meets with resistance, starting with the enormous weight of acquired habit and of universally diffused half truths. We have, even so, suggested our grid with all the discretion of the early commentators, by way of colophons or sub-titles, so as to open to the attentive reader a slightly fresher perspective. In so doing, we have assuredly sought to remember which level the Book of Love is on, much less that of intrigue or suggestive description than of the archetypes of the Tamil world of the imagination. At this archetypal level then we again inescapably find the five tiṇai as the nucleus of the atom in which landscape and soul state endlessly enthral one another.

68Finally, we shall be happy if the references we have collected, thanks to R. S. Sambasiva Sarma, confirming the suggestion, and the evidence, of T. P. Meenakshisundaram and of Chelvakesavaraya Mudaliar as to the existence of a manuscript tradition that supports this presentation,15 prove an incitement to young researchers to rediscover the too long neglected path of the genuine critical editions.

69To conclude, the lesson of all these avatars of the Book of Love in the prism of its commentaries, is that more distance must be taken with regard to any immediate and particular exegeses. It must be remembered too that, whether it be a case of a simple label or of a colophon, the complex heritage of the whole grammar is carried within it. This means, in the case of the love lyric, that the theoretical view oscillates perpetually, and without giving warning of changes in perspective, between the possible approaches and presentations we have defined, quite distinct in theory but constantly overlapping in practice. It also means that between the poems, the commentaries and our present reading, many centuries passed carrying their load of language and cultural changes; this implies especially the necessity to do full justice to the bilingual culture of Tamil medieval elite. This interference of history increases the complexity of possible readings of a text through the works of the commentators as this implies essentially the quest for a milieu and its representations, from the most basic to the most refined.

70To approach the exegesis of a text through its scholarly interpreters (a cultural exercise par excellence) is, in one respect, to discover new harmonics and to root that text in the history and in the revealed imaginary world of its original public. But, in another perspective, this exercise sharpens, by the contact it involves with a different world of learning, the tools of our own scholarship and it makes us better able to measure their limitations and efficacy. Are we not also, in fact, modern commentators whose readings and glosses are the only way by which the texts may survive? This frightening responsibility is what Vladimir Nabakov is subtly charging us with in Pale Fire, if he is justified in writings, “for better or for worse, it is the commentator who has the last word.”

71Text of a lecture delivered on 1^st of April 1992 at a Seminar on “Literary Genres in India”, Paris III University, and published in Genres Littéraires en Inde, publication de la Sorbonne Nouvelle sous la responsabilité de Nalini Balbir, 1994.