To the memory of Godard Schokker (1929–2009), who introduced me to Prākrit and Apabhraṃśa literature and thus gave me the key to understanding Caṅkam poetry

1. Introduction1

1The earliest śāstric text in Tamil is Tolkāppiyam.2 It is part of the so-called Caṅkam corpus, which otherwise includes eight anthologies of poems. Tolkāppiyam describes, or prescribes, the grammar and poetics of this literary tradition. The indebtedness of its grammatical part in Books 1 and 2 to the Sanskrit tradition, and in particular the Aindra or Kātantra school of grammar, has long been recognized (Burnell 1875: 8–20; Scharfe 1973). The indebtedness involves modes of analysis as well as terminology and shows that the Tamil tradition did not start from scratch but was in its origin inspired by earlier North Indian examples.

2From the poetic part in Book 3 of Tolkāppiyam, known as Poruḷatikāram, a similar picture emerges (Marr 1985: 52–68). Some topics were borrowed directly from the Sanskrit literary tradition. This applies, for instance, to the treatment of similes (uvamai, Skt. upamā) and moods (meyppāṭu, Skt. bhāva) in Chapters 7 and 8 respectively.3 The borrowings may also concern details, such as the distinction between nāṭaka vaḻakku and ulakiyal vaḻakku introduced in 1.56/1003, which corresponds to the one between the pair nāṭyadharmī and lokadharmī of the Nāṭyaśāstra (13.70 ff.; Takahashi 1989: 21; Tieken 2001: 163–164); or parts of chapters, such as the enumeration of the 32 uttikaḷ in 9.112/1614, which correspond to the 32 tantrayuktis mentioned in, among other texts, the Arthaśāstra (Chapter 15, Chevillard 2009).

3In this process of peeling offlayers of Sanskritic material from the Tolkāppiyam, one usually stops short of the classification of poetry into seven so-called tiṇai, or types, laid down in Poruḷatikāram, Chapters 1–5. This part of the literary theory is claimed to be a purely indigenous invention independent of any influence from the North (Zvelebil 1973: 85; Marr 1985: 15). In what follows, however, I will try to show that the literary theory is based on categories derived from a theory of music of North Indian, i. e. Sanskrit, origin. The earliest textual source for this theory is the Nāṭyaśāstra (before the 5^th century CE). However, the musical theory underlying Tamil poetics, if indeed related to the one outlined in the Nāṭyaśāstra, is already far removed from it. The musical theory appears to have followed the same trajectory as the later North Indian Rāga and Rāgamālā systems. The aim of this article is, among other things, to try and situate the Tamil literary theory more exactly in relation to these various North Indian musical theories.

2. The Tamil literary theory

4Old Tamil Caṅkam poetry is divided into two genres.4 On the one hand, we have poems depicting life in small villages in the countryside and, on the other, there are poems evoking a so-called heroic society. The heroic poems feature kings doing battle and bards trying to make a living by praising these kings’ heroic deeds. In the village poems farmers try to make the best of their love lives against all odds. For the love poems Poruḷatikāram has coined the term akam, or “interior”, and for the heroic poems the term puṟam, or “exterior”. In Poruḷatikāram the scenes of the poems, which are theoretically infinite in number, were brought down to distinct “types”, or tiṇais, that is, seven for the village poems and (twice) seven for the heroic poems. The various types were indicated by plant names. In what follows I will deal with the names given to the different types of Akam, which besides referring to plants and types of poetic scenes appear to designate specific melodic types as well. In the case of the names for the Puṟam types, though these are similar in nature, there does not appear to be any connection with melody. Most likely the names were merely invented to create a system identical to the one of Akam.

5Of the seven Akam types, the middle five are assigned to specific kinds of landscapes and are named after plants or flowers typical of those landscapes. Thus, scenes involving secret meetings of lovers (puṇartal) are set in hills and mountains (maivarai), and are referred to by the mountain flower kuṟiñci. Those describing a sulking wife (ūṭal) are found in a village amidst rice fields (tīmpuṉal). This type of scene is named after the marutam tree. A woman waiting (iruttal) for her lover to show up is portrayed against a background of a forest (kāṭu) in spring, the season of love, when the mullai-jasmine blooms. Situations involving women anxiously waiting for their lovers are set on the sea front (perumaṇal) among the fishermen community and are called neytal, after the white Indian water lily. In one way or another all poems deal with the impossibility of either lovers or husband and wife to be together and make love undisturbedly. A specific variation on the theme of separation is that of the man travelling (pirital) in summer at midday through a desiccated jungle and longing for his wife. These poems are named after the pālai flower typical of such dried-up tracts. Apart from the above-mentioned love situations (uri), like the secret meeting and the sulking wife, each landscape is also characterized by its own season and time of the day (peru- and ciṟupoḻutu, or “big” and “little” time respectively). For instance, the kuṟiñci scenes take place in winter in the middle of the night, and those of pālai in summer at midday. In addition, each landscape has its own native elements (karu), which include gods, local food, animals, plants, birds, musical instruments (or melodies) and occupations. Thus, a typical occupation of the mountain people is hunting and slash-and-burn agriculture, and their staple food is millet, while the god appearing among them is Murukaṉ. Typical animals are monkey, tiger and peacock.

6Beside these five types there are two more, the first and the seventh, which are, however, not associated with a particular landscape or named after a plant or flower. They are called kaikkiḷai and peruntiṇai respectively, two terms the meanings of which I will come back to later. The definition of kaikkiḷai in Poruḷatikāram 53/1000 consists of the description of a scene involving a young girl and a man who takes advantage of her shyness by interpreting her silence as a sign of agreement:

kaikkiḷai consists of a scene in which a man, frustrated and unhappy in the company of a young girl who is not (yet) ready for love, tries to overwhelm her by, alternatively, kind words and threats, and, when he gets no response, is not held back but goes on speaking to her in that way and enjoys himself making love to her.5

7While in kaikkiḷai the thwarted lover is keeping a clear head, in peruntiṇai (54/1001) he is making more desperate moves. One of these is riding the rugged trunk of the palmyra palm (ēṟiya maṭal) in front of the house of the girl who has rejected him. In another he has completely lost his mind from an excess of desire (tēṟutal oḻinta kāmattu mikutiṟa), and in the third his passions seem to have completely incapacitated him (mikka kāmattu miṭal). In this respect the fourth situation (actually the second in the enumeration, coming after riding the palmyra) is somewhat out of tune: it presents a lover who is too old (iḷamaitīrtiṟam).6 In both situations we are dealing with relationships between unequal partners: in kaikkiḷai the girl is helpless and the man too strong for her, in peruntiṇai it is the other way around. As I will argue below, these aspects turn up in the names of the two tiṇais. By placing these two situations at both ends of the series of tiṇais, the theory seems to demarcate the radius of action of the middle five, which would thus concern relationships between equal partners. If this is indeed what the literary theory wants to imply, which I think it does, it should immediately be added that on this point it also willfully ignores the poetic practice. For, in the poems the lovers are continuously frustrated — that is what the love poems are about and the men do ride palmyra trees and do try to win over very young girls.7 What is strange, though, is that only for peruntiṇai allowances are made: in 55/1002, as a kind of afterthought, it is stated that the four scenes of peruntiṇai may also be found in the type(s) mentioned before.8

8I am not concerned here with the question of whether there is any overlap between the various types distinguished in Poruḷatikāram or what the exact relationship is between this classification and the poems. Interestingly, the theory itself, by introducing the concept of tiṇaimayakkam, or confusion of types (14/961), allows for a considerable degree of overlap. A typical case of tiṇaimayakkam is when animals or plants of one type are found in poems of another type. Instead I am concerned with the origin of the literary theory. In this connection I want to draw attention to the fact that the names of the seven Akam types are also found in poetry as technical terms of music. While kaikkiḷai is originally a musical term — and, as we shall see, the same applies to tiṇai in peruntiṇai — the use of the plant names as technical terms of music seems to have a more complicated history. In what follows some of the instances of the plants names as musical terms in literature will be considered, beginning with those in the Cilappatikāram (Cil.). This text abounds in descriptions of performances of music, song and dance, and the terms in question are often found in contexts which are to a large extent self-explanatory.9 As such, these instances will be helpful in understanding the short and therefore vague instances of the terms in Caṅkam poetry.

3. Plant names as technical terms from music

9The Cilappatikāram contains several instances showing the use of the plant names kuṟiñci, pālai, mullai and marutam as musical terms. (In the Cilappatikāram the word neytal is not found as such.) Thus, in Cil. 3.149 we read about a pāṟpaṭa niṉṟa pālaippaṇ. The word paṇ in pālaippaṇ suggests that pālai is a melody type with a musical scale of its own. In this respect 28.33 contains more specific information. The passage refers to a pālai (melody) played on the harp in which the kural note is taken as the kural, or tonic.10 The following two lines (34–35) speak of a sweet kuṟiñci, which is produced by playing the tuttam note as the kural.11 Furthermore, in the so-called cevvaḻippālai the kural is “joined” with (cēr) the kaikkiḷai, or the third note, to form a chord (7.47),12 and in the viḷarippālai for that same purpose the first note (kural) is joined to the fifth (iḷi) (7.48).13

10Both harp and flute can be tuned according to these particular scales. Cil. 3.70–94 gives a description of this process of tuning for the pālai mode:

11Puis venait le maître du luth aux quatorze cordes. Pour établir les sept notes du mode il faisait résonner ensemble les deux cordes donnant la tonique et la note aiguë qui sont au centre de l’instrument, et sur lesquelles on peut accorder la tierce. Puis accordant la corde grave du côté de l’aigu et deux cordes aiguës du côté de la tonique, il accordait la sixte. De la quarte [de l’octave grave] jusqu’à la tierce [de l’octave aiguë] il jouait les quatorze cordes montrant ainsi l’ampleur du mode. Les séquences [qui forment les modes] apparaissaient dans un ordre prévu: en partant de la tierce on obtenait la gamme appelée paṭumalai. De la seconde commençait le mode cevvaḻi; de la septième le mode kōṭi; et de la sixte le mode viḷari; de la quinte le mode mēṟcem. Ainsi s’effectuaient les divers arrangements des intervalles. Sur le luth, les sons sont de plus en plus graves en allant vers la gauche; sur la flûte c’est le contraire. L’expert joueur de luth sait mêler les sons graves et les sons aigus avec les sons moyens d’une manière douce pour l’oreille.14

12The expression mullaiyantīṅkuḻal in Cil. 17.21.3 may accordingly be taken to refer to a flute tuned in the mullai mode.15

13It is tempting to connect these paṇs, or melodic types, with the North Indian rāgas. The Cilappatikāram, however, uses the term jāti, which, while not the direct precursor of the rāga, refers to an earlier, closely related concept.16 See in this connection Cil. 8.39–41, in which the marutam cāti (Skt. jāti) is said to consist of four types, called akanilaimarutam, puṟanilaimarutam, arukiyaṉmarutam and perukiyaṉmarutam respectively.17 As must be noted, the first two varieties have names, akam and puṟam, which are also found in the literary theory.

14Thus, while the words kuṟiñci, mullai, pālai and marutam (as indicated, in the Cilappatikāram no instances of neytal as a term from music are found) all refer to melodic types, kaikkiḷai is somewhat out of tune. It does not refer to a melodic type but, as in e.g. Cil. 7.47, it is the term for an individual note. The term tiṇai as in peruntiṇai is extremely rare in the Cilappatikāram and is not used in that text as a musical term.

15If we now turn to the Caṅkam poems proper, there the word kaikkiḷai is found only once, namely in Paripāṭal 11.136 in the expression kaikkiḷaikkāmam. kaikkiḷai clearly does not refer to a note here but it is used in its technical, literary meaning of “love between a young, innocent girl and an experienced womanizer”.18 I will return below to the use of kaikkiḷai in this meaning in a Caṅkam text. The word tiṇai in Caṅkam poetry has nothing to do with music. In the few known instances of the word it seems to mean “family, clan”, as in Puṟanāṉūṟu 373.28, viḷaṅkutiṇai vēntar “kings belonging to illustrious families or clans”. In the majority of instances the words kuṟiñci, mullai, neytal, marutam and pālai refer to the respective plants. In a few instances, however, they refer to pieces of music. Cases in point are Akanāṉūṟu 102.6 (kuṟiñci pāṭa, “while they sing a/the kuṟiñci”), Naṟṟiṇai 255.2 (kuṟiñci pāṭi ... kāṉavar, “the forest people ... sing a/the kuṟiñci”), Aiṅkuṟunūṟu 408.1 (pāṇar mullai pāṭa “while singers sing a/the mullai”) and Puṟanāṉūṟu 389.16–17 (makaḷir neytal kēḷaṉmār, “may your women never hear a/the neytal”). The first impression from these cases is that we might be dealing with songs typical of the respective regions, kuṟiñci, for instance, as a song typical of the mountains.19 However, the instance of mullai in the phrase mullai nal yāḻ pāṇa in Aiṅkuṟunūṟu 478.5, which may be compared with mullaiyantīṅkuḻal in Cil. 17.21.3, seems to suggest that these songs are also characterized by their own melodic mode or musical scale: “O singer with the good harp (tuned in) mullai”.

16As such these melodies seem to be capable of expressing certain moods. Thus, neytal in Puṟanāṉūṟu 194 is a song or a piece of music expressing sadness, in casu the sadness of widows:20

From one house the sound of a neytal emerges,
from the other a melody (pāṇi) on a drum (smeared) with cool mud.
Women who have their men with them (puṇarntōr),
don ornaments offlowers!
Women whose men have gone off (pirintōr),
pour down tears from grieving, kohl-rimmed eyes.
Surely, the one who created this world has no virtues.
This world is an unhappy place,
Only those who know its real nature
will know happiness.21

17In Puṟanāṉūṟu 149, marutam refers to a melody particularly suited to the morning. As such it is mentioned here alongside one particularly suited to the evening:

Naḷḷi! May you live long! Naḷḷi!
In the evening my men perform a/the marutam
on their lutes
and in the morning a cevvaḻi.
They have forgotten the proper time for each
because of your generosity in the giving of gifts.22

4. jāti, rāga and Rāgamālā

18In these two Puṟanāṉūṟu poems the melodic types of neytal and marutam are associated with a particular mood and a particular time of day respectively.23 These same associations are found in the poetical theory of Poruḷatikāram under the headings of uri and ciṟupoḻutu respectively. However, they are also found in the North Indian rāga system (and the so-called Rāgamālā paintings), and something similar is also found in the jāti system as outlined in the Nāṭyaśāstra. Indian music theory, from the Nāṭyaśāstra onwards, has been interested in notes and scales and their effects on, and power to produce, moods and sentiments. Traditionally, the notes and scales have served as focal points around which are brought together all kinds of situations which contributed to the production of the different moods. Thus, in the Nāṭyaśāstra the jātis were connected with the rasas. Having dealt with the production and composition of the various jātis in 28.37–141, Nātyaśāstra 29.1–13 continues with assigning rasas to the various jātis with reference to their dominant notes. For instance, 29.1:

The ṣaḍjodīcyavatī and ṣaḍjamadhyamā jātis can be used for the śr̥ṅgāra (erotic) and hāsya (comic) moods as the madhyama and pañcama notes abound (are amplified) in them.24

19It is possible to draw up a scheme like the following (the numbers after the jātis refer to the ślokas in Nāṭyaśāstra, Chapter 29):25

20The jātis gradually lost their importance and came to be subsumed under the rāgas. Nārada’s Saṅgītamakaranda (between 700 and 1100) introduced a classification into male, female and neuter rāgas. This classification is based on the attribution of the rāgas to the rasas. Those producing the vīra, adbhuta and raudra rasas are male, those producing the śr̥ṇgāra, hāsya and karuṇa rasas are female, and those producing the bībhatsa, bhayānaka and śānta rasas are neuter (Koch 1995: 30–31). In the same text the rāgas are also assigned to different seasons and times of the day (Koch 1995: 32, n. 15). A more elaborate system has been attested in Śārṅgadeva’s Saṅgītaratnākara (first half of the 13^th century). The seven notes are connected with the rasas, with animal cries (e.g. ṣaḍja with the cry of the peacock), colors, gods, ancestors, seers, demons, the four castes and regions. In the same treatise the rāgas were subjected to a similar categorization as well. As in the Saṅgītamakaranda the rāgas were also linked to the seasons, the times of the day and the rasas (Koch 1995: 31–33). This whole edifice has culminated in the Rāgamālā miniature painting, in which the male rāgas, the female rāgiṇīs and their sons, or rāgaputras, were depicted as gods or goddesses in settings appropriate to the moods of the rāga. The paintings were meant as objects of meditation (dhyāna). One of the earliest examples is the Rāgamālā by Kṣemakarṇa (16^th century). In Rāgamālā the great variety of rāgas is grouped around six main male rāgas (Bautze 2001).

5. Poetry and singing

21The developments seen in the Saṅgītaratnākara and Rāgamālā seem to reflect the meeting of music, (song) text and image. As I will try to show, Tamil literary theory is the product of such a combination as well. In order to make this clear, I will turn to the first three nūṟpās (nūṟpā is a loan translation of Sanskrit sūtra) of the Poruḷatikāram. While enumerating the seven tiṇais, the first nūṟpā appears to refer to an earlier work, which on the basis of the third nūṟpā may be identified as a treatise on singing and song texts.

22In the very first nūṟpā the number of tiṇais, beginning with kaikkiḷai and ending with peruntiṇai, is said to be seven. These same seven tiṇais are also said to have been treated before (muṟpaṭa kiḷanta eḻutiṇai). This phrase raises a problem, for this is the first time the seven tiṇais are mentioned in the Tolkāppiyam. There is no evidence of a lost part of the Poruḷatikāram preceding the present first chapter either.26 For a possible answer to what source the text is referring to here we may turn to the third nūṟpā. First, however, in the second nūṟpā, the middle five tiṇais are set apart. Five or four of these, if pālai is excluded, stand for different ecotypes (mountains, forests, rice fields and beach). Pālai, like Sanskrit aṭavī, is the wild jungle lying around or in between these areas. The third nūṟpā, then, turns to the categories of landscape (nilam) and time (poḻutu, the mutals, or “first or basic elements”), things native to the landscapes (karu) and love situations (uri) peculiar to the middle five tiṇais. The text reads:

mutal karuv uripporuḷ eṉṟa mūṉṟē
nuvaluṅ kālai muṟai ciṟantaṉavē
pāṭaluṭ payiṉṟavai nāṭuṅ kālai.

23The absence of any marker of coordination in the two parallel subordinate clauses (... nuvaluṅ kālai and...nāṭuṅ kālai) suggests that the two clauses each have a different relationship to the main sentence. It may be assumed that the second one, found after the main verb of the sentence, supplies a kind of afterthought. This would give a translation like the following:

Upon examination of the set of three things, namely the basic elements, the things native to the tracts and the love situations, that ordering makes sense, (in particular) when one (also) takes into consideration their use in song.27

24It should be noted that the word pāṭal in the third line refers to songs, not to poems like those of, for instance, Kuṟuntokai. Thus, in nūṟpā 56/1003 the word pāṭal refers to the songs embedded in the Kalittokai and Paripāṭal poems (Tieken 2001: 163; Tieken 2009; Takahashi 1989: 21). It seems that in presenting the tiṇai system the author of Poruḷatikāram looked from the corner of his eye at songs or singing, and apparently that song tradition was characterized by seven tiṇais as well.

25In this connection it is important to note that the key word of the literary theory, namely tiṇai, “class, type”, is a loan translation of the term for the melody types, jāti.28 It may therefore be assumed that the author of the Tamil literary theory was inspired by a theory of music and songs, if not by an actual treatise on these topics, which, whether theory or treatise, may well have been the “earlier” source referred to in the first nūṟpā. Apart from all this, it may be noted that, although the compound peruntiṇai is otherwise unknown, even the seventh tiṇai, like the other six, now appears to have been named by a term from music, or rather to include a term from music, tiṇai.

6. The numbers seven and five

26The term tiṇai suggests that the music theory concerned was formulated in Sanskrit or goes back to one formulated in that language. This theory must have resembled more closely the fully elaborated systems we find in the Saṅgītamakaranda or Saṅgītaratnākara than the one in the Nāṭyaśāstra. However, as to the relatively “meager” music theory found in the Nāṭyaśāstra, it should be noted that this text was only concerned with those aspects of the contemporary music theory which could contribute to the realization of the rasas. Furthermore, the Nāṭyaśāstra is only concerned with poetic conventions in a very general way. On the other hand, most Tamil scholars will consider the Saṅgītamakaranda and Saṅgītaratnākara too late for comparison with the Tamil literary theory. Both texts, however, may well refer to aspects of music which already had long traditions at the time of their composition. Apart from that, the question is: how old is Caṅkam poetry anyway? The dates assigned to Tamil Caṅkam poetry differ widely, ranging from the 3^rd century BCE to the 8^th or 9^th century CE. However, as I have tried to show on various other occasions, the early date has never been backed up with verifiable evidence (Tieken 2001, 2003, 2004, 2008^a, 2008^b, 2009). The evidence is mostly circular, that is to say, once it is accepted that the early date is not established as firmly as one wants to believe, most of the evidence put forward in its support collapses. In addition to that, if the Tamil literary theory has indeed been based, directly or indirectly, on a music theory formulated in Sanskrit, here we have another piece of evidence suggesting that the tradition is not as early or as independent as it is often claimed to be. As suggested already elsewhere, we should start from a post-Pallava date for Caṅkam poetry (Tieken 2001).

27With this relatively late date, the distance in time between the Tamil literary theory and the two Sanskrit treatises on music mentioned above becomes considerably smaller. In what follows, however, I do not intend to undertake a detailed comparison between the two systems. Instead I want to draw attention to the existence side by side in the Tamil literary theory of seven and five tiṇais, which raises several questions regarding the ways and means by which the literary theory was put together.

28As we have seen, the number of tiṇais in the Tamil literary theory is seven. Seven is also the number of the scales formed by the jātis, mentioned in the Kuḍimiyamalai inscription of the South Indian Pallava king Mahendravarman I (7 th century, Nijenhuis 1992: 38). It should be noted that in the Nāṭyaśāstra the number of the jāti scales (grāmas) is not clear. Two have been explicitly mentioned in Chapter 28 on the jātis, namely the ṣaḍjagrāma and madhyamagrāma, so called after their first notes, ṣaḍja and madhyama respectively. Beside these two, the text mentions four more. Thus, in Chapter 32, 428–429 it speaks of altogether six different scales to be used in the six phases, or sandhis, in a play, that is, beside the ṣaḍjagrāma (in the pratimukha) and madhyamagrāma (in the mukha), the ṣāḍava (in the pūrvaraṅga), sādhārita (in the garbha), pañcama (in the avamarśa) and kaiśika (in the nirvahaṇa). In Tamil poetics, however, the number seven is reached only by changing the criteria halfway as well as the terminology used in the classification. Thus, only the middle five, the so-called aintiṇai, are associated with eco-types, seasons, times of the day, and things native to the eco-types, and are named after flowers or plants. The other two types by contrast are not connected with such things as landscapes (and therefore could not be given plant names). They are just headings for a number of (more or less reprehensible) love scenes, for which appropriate names were selected, or coined, drawing from music terminology proper: kaikkiḷai, which along with kai, “smallness, (something in need of) a helping hand”, could possibly serve to express the young girl’s helplessness, and peruntiṇai, which along with peru “big” suggested the woman’s aloofness.

29Given this situation it is tempting to conclude that the plant names were not part of the original music theory. That is to say, the seven tiṇais (or jātis) of the music theory were not, or not yet, named after plants. In any case, if they were, it is difficult to see why such names were not used for the first and the seventh poetic tiṇais as well. It may be assumed that the plant names were introduced only at the moment when music and poetry came to be considered together: they were suggested by the poetic scenes. This assumption raises several questions. For one thing, it is difficult to decide if at that moment scholars were working on the basis of an existing classification of poetry through five landscapes or not, or if the five landscapes constituted an earlier poetic convention or a category created only on that moment. I will return below to this point. Whatever had been the case, in either scenario a strange light is cast on the use in the poems of the plant names for melodic types expressive of specific moods. These occurrences suggest either that the supposed earlier poetic conventions with the five tiṇais had been based on a music theory as well, which already had become part of the poetic tradition itself, or that the poetry which we now have was a scholarly affair dependent on the literary theory as formulated in the Poruḷatikāram. In connection with the latter point I want to mention the possibility that the Caṅkam corpus as we now have it may well be a fairly late version produced under the patronage of the Cēras, who through this literature tried to revive an earlier literary culture of the Pāṇṭiyas of the east coast (Tieken 2001: 230–234). Such a “revival” presupposes the availability of a scholarly tradition regarding the poetry in question.

30In this connection I want to draw attention to the fact that the instances in the Caṅkam poems of the words kuṟiñci, neytal, mullai, marutam and pālai, if borrowed from the literary theory, do not stand on their own. They are matched by some other such borrowings. One of these is the expression kaikkiḷaikkāmam, “mismatched love”, in Paripāṭal 11.136, already referred to above. Some other instances are found in the Porunarāṟṟuppaṭai and Maturaikkāñci, both from the Pattuppāṭṭu. Thus, in Porunarāṟṟuppaṭai 218–221 we find tiṇai, meaning “melodic type”, as well as mullai in the meaning of “forest” (this meaning is vouchsafed by the preceding expression nīṉiṟa, which cannot possibly refer to the jasmine):

kuṟiñci paratavar pāṭa neytal
naṟumpūṅ kaṇṇi kuṟavar cūṭak
kāṉavar marutam pāṭav akavar
nīṉiṟa mullaip paẖṟiṇai nuvala,

While the fisher folks from the sea coast sing mountain songs (kuṟiñci), while the mountain people wear garlands made of fragrant flowers from the sea coast (neytal), while the people from the forests sing songs from the rice fields, and the cultivators (akavar)29 from the rice fields sing (the) various melodies (paẖṟiṇai) from the dark (nīṉiṟa) forest...

31Maturaikkāñci 238–326 provides stereotyped descriptions of the five types of landscapes (marutam in 238–270, mullai in 271–285, kuṟiñci in 286–301, pālai in 302–314 and neytal in 315–325). These descriptions are concluded in line 326 with the phrase aimpāl tiṇai, “so far the five tiṇais”. Here we see the words marutam, etc. and tiṇai in their technical meanings of landscapes and landscape types respectively. I have already dealt with the occurrences of this set of words in Cilappatikāram. As to the implication of the occurrence of these technical terms in the other two collections, Paripāṭal and Pattuppāṭṭu, it should be noted that scholars disagree on the position of precisely these two texts vis-à-vis “Caṅkam” poetry. Thus, while the Paripāṭal is mentioned in the traditional list of Caṅkam works, modern scholars tend to place it apart as a late text. The case of the Pattuppāṭṭu is the opposite: the text is not found in the list but scholars like nevertheless to include it in the Caṅkam corpus. As I have argued elsewhere, however, both the Paripāṭal and Pattuppāṭṭu represent “merely” genres different from that of the so-called short Caṅkam poems, but are as such not necessarily later than these. Or, to put it differently, the short Caṅkam poems are not demonstrably earlier than the Paripāṭal and Pattuppāṭṭu (Tieken 2001: 170–182 and 196–200). Most likely, the latter two texts had their origin in the very same milieu as the other Caṅkam poems. The Pattuppāṭṭu is the result of the attempt to adapt the style and conventions of stanzaic Caṅkam poetry to the Mahākāvya genre (Tieken 2001: 196–200). In the Mahākāvya genre, showing offone’s learning is a common poetic device. In addition, it offers simply more space for literary games of this sort than do the relatively short poems of Caṅkam proper. The rareness of technical terms borrowed from the literary theory in the Caṅkam poems could after all well be a matter of genre.

32Above, it has been assumed that the number seven has its origin in music, with its seven jātis (and its seven notes). As far as poetry is concerned, however, seven must have proved problematic from the very beginning. The number of recognized seasons is six and, whatever way day and night are divided, a division in seven is not found.30 Of course almost anything can, if necessary, be forced into a division of seven, but in the case of, for instance, the seasons and the day and night, in order to be convincing one had to stick to the traditional divisions. Furthermore, the division of the landscapes has resulted in five distinct types, not seven. Above I have already noted that it is unclear if this division into five represented an earlier literary tradition or was one newly introduced. In this connection it might be noted, though, that, if the final result was to be seven, starting with five has resulted in a strikingly neat system, neater than would have been possible with, for instance, four or six, with a middle group of five and one placed at either end. An entirely different question is why landscape with its native plants or flowers was made the primary category in the classification, rather than, for instance, season. A reason might be precisely that in this case the number had not been fixed yet, contrary to that of season or time of day, so that it could be adapted to the circumstances. In addition to that, however, it is also possible to read nationalistic motives in the classification. The Tamil speaking world is defined as a world in itself. Every type of landscape is represented in it. “Tamilnadu” may be smaller than India but it is just as varied and complete.31 Taken in this way the Tamil practice of distinguishing landscapes might be compared with the identifications of the individual rāgas with different regions of India (Jaunpurī, Multānī, etc.) in, for instance, the Saṅgītaratnākara.32

7. Conclusion

33Above, I have argued that the Tamil literary theory was inspired by a theory of music that originally must have been formulated in Sanskrit. The latter theory distinguished seven jātis, which we meet again in the seven tiṇais. The Tamil literary theory itself seems to expressly acknowledge its indebtedness to a song tradition. All this raises more questions than can be answered here. The questions concern the way the music theory was actually implemented and the relationship between the literary theory and the poems that we now have. The latter points need a fresh, thorough investigation of Tamil literature, in which all options are leftopen. This applies in particular to the question of the dating of Caṅkam literature, which has unfortunately developed into a trench war. In this connection I should, however, note once more that the fact that the key term of the Tamil literary theory, tiṇai, is a loan translation of a Sanskrit theory of music seems one more piece of evidence for a late date of the Tamil theory.