8/9/2019 DeRomanis Topoi Suppl 11 Libre http://slidepdf.com/reader/full/deromanis-topoi-suppl-11-libre 1/14 Topoi Suppl. 11 (2012) p. 3-4 SOMMaiRe topo, Supplém11 auour u PépldlÉét xs éés pr M.-Fr. B oussac , J.-Fr. s alles J.-B. Y on Sommaire 3-4 Présentation 5-6 Grlcurs D. m arCotte , « LePériple de la mer Érythrée dans son genre et sa tradition textuelle » 7-25 P. a rnaud , « LePeriplus Maris Erythraei : une œuvre de compilation aux préoccupations géographiques » 27-61 J. d esanGes , « L’excursus de Pline l’Ancien sur la navigation de mousson et la datation de ses sources » 63-73 dl’Égypà l’iB. f auConnier , « Greco-Roman merchants in the Indian Ocean : Revealing a multicultural trade » 75-109 Mr RougafrquP. Pomey , « À propos des navires de la mer Érythrée : découvertes récentes et nouveaux aspects de la question » 111-132 S. s ideBotham & i. Z yCh , « Results of Fieldwork at Berenike: A Ptolemaic-Roman Port on the Red Sea Coast of Egypt, 2008-2010 » 133-157
t x s eacute eacutes p r M-FrBoussac J-Fr s alles J-BYon
Sommaire 3-4
Preacutesentation 5-6
G r l c urs
D m arCotte laquo LePeacuteriple de la mer Eacuterythreacutee dans son genre et sa tradition textuelle raquo 7-25Pa rnaud laquo LePeriplus Maris Erythraei une œuvre de compilation aux preacuteoccupations geacuteographiques raquo 27-61J d esanGes laquo Lrsquoexcursus de Pline lrsquoAncien sur la navigation de mousson et la datation de ses sources raquo 63-73
d lrsquoEacutegyp agrave lrsquoiB f auConnier laquo Greco-Roman merchants in the Indian Ocean
Revealing a multicultural trade raquo 75-109
M r Roug afr quP Pomey laquo Agrave propos des navires de la mer Eacuterythreacutee deacutecouvertes reacutecentes et nouveaux aspects de la question raquo 111-132S s ideBotham amp i Z yCh laquo Results of Fieldwork at Berenike A Ptolemaic-Roman Port on the Red Sea Coast of Egypt 2008-2010 raquo 133-157
V BuCCiantini laquo The Limits of Knowledge Explorations of and Information from the Horn of Africa to the East African Coast in the Greco-Roman Tradition raquo 159-176
ar bm B ukharin laquo The Coastal Arabia and the adjacent Sea-Basins in thePeriplus of the Erythrean Sea (Trade Geography and Navigation) raquo 177-236J s ChietteCatte laquo LrsquoArabie du Sud et la mer duiii e siegravecle av
au vi e siegravecle apr J-C raquo 237-273A r ouGeulle laquo Syagros et autres eacutetablissements cocirctiers du Hadramawt
GolfJ- Fr s alles laquo Le Golfe persique dans lePeacuteriple de la mer Eacuterythreacutee connaissances fondeacutees et ignorances reacuteelles raquo 293-328
if d e r omanis laquo On Dachinabades and Limyrike in thePeriplus Maris Erythraei raquo 329-340
heacuter g sCla lliBert laquo Les reacuteseaux de navigation du deacutebut de lrsquoegravere chreacutetienne auxvi e siegravecle Rencontre de populations eacutechanges commerciaux et matrimoniaux
concurrence agrave lrsquoouest et agrave lrsquoest de Madagascar raquo 341-357Eacutev allet laquo LePeacuteriple au miroir des sources arabes meacutedieacutevales Le cas des produits du commerce raquo 359-380
On DachinabaDes and Limyrikein the PeriPLus maris erythraei
It is a truism that the names of regions or choronyms are a far more interestingsubject of study from a historical point of view than the names of places ortoponyms While the rationale for the latter lies in too distant and obscure timesand tells very little about the subsequent history of a place choronyms are sensitiveto historical dynamics and their signi cance may uctuate either objectively inthe diachrony or subjectively in the synchrony They may therefore be highlysigni cant for the political economic and social history of both the regions towhich they refer and the surrounding areas When compared to their equivalentsin Indian sources the occurrences in thePeriplus Maris Erythraei (PME ) of thetwo Indian choronyms Dachinabades and Limyrike allow for interesting insightsabout the historical geography of the subcontinent and the perspective by which itsauthor interprets it
Barygaza its gulf and the Ariake region around it marks a sort of geopoliticalwatershed in the map of the author of thePME On one side it signals the startof the kingdom of Manbanes and the whole of India laquo Right after Barake is the
gulf of Barygaza and the land of the Ariake region beginning of the kingdomof Manbanes and the entire India raquo1 On the other side it adjoins the land called Dachinabades which stretches widely towards the south laquo Immediately afterBarygaza the adjoining land extends from north to south For this reason the regionis called Dachinabades the south indeed is calleddachanos in their language raquo2
This emphasis on Barygaza and its region is hardly surprising In thePME the Indian Ocean is mainly represented from the perspective of Roman trade and
Barygaza was speci ed as one of the three major Indian destinations of the Romanmerchants active in those seas ndash the other two being Barbarikon at the mouth of theIndus and the emporia of the Limyrike in present-day Kerala3 Much less obvioushowever is the awareness of the importance of Barygazarsquos inland connections with Dachinabades a choronym which perhaps not insigni cantly is not found in anyother Greek or Latin work
Dachinabades is the Greek transcript of a Prakrit form of the Sanskrit Dakṣiṇāpatha a notion which evokes ndash in different ways as we shall see ndash thesouthern parts of India From it the concept of Deccan arose4 Exceptionally andquite interestingly the author of thePMEoffers an interpretation of the expression
by referencing the local language5
Even more remarkable however is the fact thathis understanding of the term is strictly speaking both imprecise and incomplete Asa matter of fact the compound transcribed in Greek as Dachinabades is composedfrom two words the adverbdakṣiṇā which actually means laquo southwards raquoand the substantive patha laquo road route raquo which the author omitted to translateAgain strictly speaking his explanation of the term is inexact Since it meanslaquo southwards route raquo the expression Dakṣiṇāpatha cannot be logically explainedwith the circumstance that after Barygaza the land (ἤπειρος) extends from north tosouth
However the omission of patha and the imprecise rendition ofdakṣiṇā arenot casual mistakes At the time of thePME Dakṣiṇāpatha was not understoodanymore in its original sense of laquo southwards route raquo In the everyday languageof the locals the hodonym (name of a road) laquo southwards road raquo had become achoronym designating the region south of the Barygaza region As a consequencea translation of the second element of the compound ( patha) and a proper
Ἀρωμάτων ἀφιέντες οἱ μὲν εἰς Λιμυρικὴν πλέοντες ἐπὶ πλεῖον τραχηλίζοντεςοἱ δὲ εἰς Βαρύγαζαν οἵ τε (Muumlller οἱ δὲ) εἰς Σκυθίαν κτλ Consequently lists ofthe exchanged items are given only for Barbarikon (PME 39) Barygaza (PME 49)and the emporia of the Limyrike (PME 56) Ships and merchants sailing back fromeither Barygaza and the Ariake or Barygaza or Limyrike may call respectively at theemporia of the Somali coast (PME14) or Socotra (PME 31) and Moscha (PME 32)Barygaza is a major destination also for merchants from Muza (PME21) and Cane(PME27) its ships sail to Omana (PME36) it receives goods from Minnagara inAriake (PME41) and Ozene (PME48)
4 Cfr rG B handarkar Early History of the Dekkan Bombay 18952 p 1-2 dk C hakraBarti The Archaeology of the Deccan Routes The Ancient Routes fromthe Ganga Plain to the Deccan New Delhi (2005) p 1-21
5 Elsewhere African and Arabian toponyms are etymologically explained byreferencing the Greek cfrPME 16 [hellip]τὰ Ῥάπτα [τὰ] (L Stuck)λεγόμεναταύτην ἔχον τὴν προσωνυμίαν ἀπὸ τῶν προειρημένων ῥαπτῶν πλοιαρίων 26 Εὐδαίμων Ἀραβία [hellip] Εὐδαίμων δrsquo ἐπεκλήθη κτλ
understanding of the rst (dakṣiṇā) must have appeared respectively misleadingand meaningless to the author of thePME and his Indian informers
This evolution from hodonym to choronym is not unparalleled Similar casesare those of the ancient Italian choronyms Aemilia(modern Emilia) andFlaminiawhich come out from the hodonymsvia Aemiliathe former andvia Flaminiathelatter However such transformations take place only when a new trunk road deeplyaffects the material life of the surrounding areas The commercial importanceof what must have been the original Dakṣiṇāpatha 6 ndash the main trunk road thatwent from Barygaza to Paithana Tagara and probably the coastal regions of theBay of Bengal ndash is still clearly recognizable in the text of thePME In chapter 51
the location of the two major emporia of Dachinabades Paithana and Tagara isaccompanied by speci c information concerning the number of travel days fromBarygaza and Paithana respectively laquo Among those in the same Dachinabadestwo are the most conspicuous emporia Paithana which is twenty days travel southof Barygaza and ten days east from it Tagara another very big city raquo7 Extensionsof the laquo southwards route raquo from Tagara to the coastal regions of the Bay of Bengalare not explicitly mentioned Still they are implied by the merchandise laquo of thecoastal parts raquo rst exported to Tagara and then re-exported to Barygaza8
In its original hodonymic sense the expression Dakṣiṇāpatha occurs in apassage of the Arthaśāstra laquo In the case of a land-route lsquoThe route to the Himavatis preferable to the southwards route Commodities such as elephants horsesperfumes ivory skins silver and gold are of very high valuersquo So say the teacherslsquoNorsquo says Kauṭilya lsquoleaving out commodities such as blankets skins and horses(those items) and commodities such as conch shells diamonds rubies pearls andgold are more plentiful on the southwards routersquo raquo9
6 Cfr eghP r ay Monastery and Guild Commerce under the Sātavāhanas Delhi(1986)
7 PME 51 τῶν δὲ ἐν αὐτῇ τῇ Δαχιναβάδει δύο ἐστὶν τὰ διασημότατα ἐμπόριαΠαίθανα μὲν(Schwanbeck ἐπιφαινόμενα) ἀπὸ Βαρυγάζων ἔχουσα ὁδὸν ἡμερῶνεἴκοσι πρὸς νότον ἀπὸ ltδὲgt (edd)ταύτης ὡς ἡμερῶν δέκα πρὸς ἀνατολὴν ἑτέραπόλις μεγίστη Ταγάρα (Fabricius Ταγάρα μεγίστη) Hereἐν αὐτῇ τῇ =ἐν τῇ αὐτῇ cfr h f risk Le peacuteriple de la mer Eacuterythreacutee Suivi drsquoune eacutetude sur la tradition et lalangue Goumlteborg (1927) p 65-66
This passage is noteworthy in many ways Besides showing the term inits original meaning and con rming that it actually referred to a land-route(sthalapatha) and not to as would have been theoretically possible a water-route(vāripatha) it proves that Dakṣiṇāpatha was a commercial road used by merchantswho could choose to carry on business either along the laquo southwards raquo or the Haimavata route10 It makes therefore clear that the laquo southwards route raquo was acontrastive denomination suggested by the opposition (and competition) betweenthe laquo southwards raquo and the Haimavata routes For a long time trade with Himālayahad been considered preferable to trade along the laquo southwards route raquo but whenthose lines of the Arthaśāstra were conceived the latter turned out to be more
lucrative than the former thereby disproving the eldersrsquo opinions It is thereforeself-evident that the expression Dakṣiṇāpatha was forged in an area around thenorthern end of that route by a commercial milieu whose trading activities werepoised between the Himālaya and Deccan This circumstance explains why Dakṣiṇāpatha is so to speak a one-way hodonym oriented in a direction north-south
It has been observed that the contextual reference to Deccan and Himālayaland-routes implies a standpoint laquo of a person in the midlands neither a southernernor a northerner raquo11 The text of thePME may suggest a more precise localizationIn chapter 47 the list of tribes lying behind Barygaza ndash which includes theArachosians the Gandharians and the Bactrians12 ndash re ects Barygazarsquos commercialconnections with the Himālayan region The same conclusion can be drawn fromchapter 48 where import to Barygaza via Ozene (= Ujjayinī) of nard costus andbdellium from the laquo upper places raquo is mentioned13 Just mid-way between the
10 For archaeological evidence pointing to connections between Deccan and Gandharacfr P BranCaCCio laquo Close Encounters Multicultural Systems in Ancient India raquoin d s rinivasan (ed)On the Cusp of an Era Art in the Pre-Kuṣāṇa World Leiden(2007) p 385ndash398 e ad The Buddhist Cave at Aurangabad Transformationin Art and Religion Leiden (2010) p 67-68 69-70 e ad laquo Terracottas fromWestern Deccan An Exploration of Sources and Transmission of Models in theEarly Historic Period raquoin Ph Granoff (ed) A Pantheon Rediscovered ChangingPerceptions of Early Historic India forthcoming
11 rP k anGle TheKauṭilīya Arthaśāstra Part II An English Translation with Criticaland Explanatory Notes Bombay (19722) p 360 nt 22
Himālaya and Deccan regions the Barygaza-Ozene region was perhaps the onlyIndian area where the alternative between a Haimavata and a Dakṣiṇāpatha land-route made perfect sense
Incidentally I would like to point out another example in which a comparisonwith the text of the Arthaśāstra may improve our understanding of the text ofthe PME Again it relates to the Dachinabades land-route In chapter 51 itemsfrom Paithana and Tagara are brought to Barygazaπορείαις ἁμαξῶν καὶ ἀνοδίαιςμεγίσταις which Casson translates as laquo by conveyance in wagons over very greatroadless stretches raquo14 Muumlllerrsquos Latin rendition is similar laquo per loca maximeinvia plaustris merces devehuntur raquo15 Both these translations assume that such
long journey was made on stretches of land that were simultaneously roadless andaccessible by wagons which is contradictoryEvoked by the author of the Arthaśāstra in the very same chapter containing
the comparison between Dakṣiṇāpatha and Haimavata land-route is thedistinction betweencakrapatha laquo the wheel-track raquo pādapatha laquo the foot-path raquoand kharoṣṭrapatha laquo the ass- or camel-road raquo laquo [hellip] as between a wheel-trackand a foot-path the wheel-track is preferable as it enables undertakings on a largescale Alternatively path for donkeys and camels in conformity with place and time(is to be preferred) raquo16 This line suggests that a long land-route may not be entirelysuitable to wheeled vehicles and may therefore require changes in the mode oftransport such as shifting between wagons and donkeys It is consequently probablethat atPME51 theκαί which connectsπορείαις ἁμαξῶν andἀνοδίαις μεγίσταις is disjunctive from Paithana and Tagara the items are brought laquo with journeysof wagons and long roadless tracts raquo In other words laquo journeys of wagons raquo(πορεῖαι ἁμαξῶν) and laquo roadless tracts raquo (ἀνοδίαι μέγισται) refer to two distinctmodalities of transport each adopted as needed along different legs of the voyageThe long route from Tagara to Paithana and Barygaza was to be travelled withwagons on laquo wheel-track raquo (cakrapatha) only in part for long stretches lsquothe ass- orcamelroadsrsquo (kharoṣṭrapatha) were the only option available
While in the Arthaśāstra the term Dakṣiṇāpatha still retains its originalmeaning of laquo southwards route raquo its earliest inscriptional occurrences already show
Σκυθίας ὅ τε κόστος καὶ ἡ βδέλλα The same items are available at Barbarikon onthe mouth of the Indus PME39
14 l C asson ThePeriplus Maris Erythraei Text with Introduction Translation andCommentary Princeton (1989) p 83
15 C m uumlller Geographi Graeci Minores I Paris (1855) p 294 A more t translationis given byWh s Choff The Periplus of Erythraean Sea New York LondonBombay and Calcutta (1912) p 43 laquo [hellip] brought down to Barygaza from theseplaces by wagons and through great tracts without roads raquo
16 Arthaśāstra VII 12 27 tatrāpi cakrapādapathayoścakrapatho vipulārambhatvātcchreyāndeśakālasambhāvano vā kharoṣṭrapathaḥ
its evolution into a choronym Dakṣiṇāpatha was claimed by Sātavāhana rulers as thegeographical frame of their sovereignty Dakhiṇāpathapati and Dakhiṇāpathesara are the synonymous formulas (both meaning laquo Lord of Dakhiṇāpatha raquo) whichoccur in inscriptions from Nāṇeghāṭ Nāsik and Girnār17 The inscriptions of Nāsikand Girnār refer respectively to Vāsiṣṭhīputra Śrī Puḷumāvi and to his immediatesuccessor Vāsiṣṭhīputra Śrī Śātakarṇi The inscription from Nāṇeghāṭ probablyrefers to Simuka the founder of the dynasty
The lordship over Dakhiṇāpatha claimed by the Sātavāhanas explains howone of them called laquo the elder Saraganos raquo (= Sātakarṇi) by the author of thePME made Kalliena not far from Nāṇeghāṭ a laquo lawful emporion raquo18 It explains also
why Ptolemy locates at Baithana (= Paithana one of the two most conspicuousemporia ofPME rsquos Dachinabades) the royal residence of Sirip(t)olemaios (= SiriPuḷumāvi)19 Nonetheless however powerful (or charlatan) were the Sātavāhanakings20 their domain could have hardly coincided with the Dakṣiṇāpatha of somepassages of thePurāṇas where it includes Pāṇḍya Kerala and Cola in the southernend of the subcontinent21 Disconnected from any political or economical realitythis expanded Dakṣiṇāpatha is the result of a generalization elaborated by the
17 vv m irashi The History and Inscriptions of the Sātavāhanas and the WesternKshatrapas Bombay (1981) n 3 l 2 p 11 (dakhiṇāpathapatiin the Nāṇeghāṭinscription of Nāganikā) ibid n 18 l 11 p 46 (dakhiṇāpathesarain the Nāsikinscription of Vāsiṣṭhīputra Śrī Puḷumāvi) dC s irCar Select Inscriptions Bearingon Indian History and Civilization Volume I From the Sixth Century BC to theSixth Century AD Calcutta (19652) n 67 l 12 p 178 (Vāsiṣṭhīputra Śrī Sātakarṇistyled as dakhiṇāpathapatiin the Rudradāman inscription of Girnār)
18 PME 52 τοπικὰ δὲ ἐμπόρια κατὰ τὸ ἑξῆς κείμενα Ἀκαβαρου Σούππαρα (edd ἀπὸ Βαρυltγάζωνgt Σούππαρα Muumlller in comm Ἀκαβαρους οὔππαρα) Καλλίενα (καὶ ΚαλλίεναFrisk) πόλις ἡ ἐπὶ τῶν Σαραγάνου τοῦ πρεσβυτέρου χρόνωνἐμπόριον (Gelenius ἐμποριω) ἔνθεσμον γενομένη (Fabricius γενόμενον) Forthe identi cation of the laquo elder Saraganos raquo with the husband of queen Nāganikācfr am s hastri The Sātavāhanas and the Western Kshatrapas a historical framework Nagpur (1998) p 55-60
19 Ptol Geogr VII 1 82 Βαίθανα βασίλειον Σιριπτολεμαίου (UKX Σίριοςπολεμίου V Σίριος πτολεμαίου R Σιροπολεμίου A Σιριπολεμαίου Z) ριζ ιη LʹIt is to be noticed that in Bhāgavatapurāṇa IX 1 41-42 the title pratiṣṭhānapati (skt Pratiṣṭhāna = Gk Paithana) is given to a king whose three sons weredakṣiṇāpatharājānaḥ laquo kings ofdakṣiṇāpatha raquo
20 s hastri op cit p 7-12 C s inoPoli laquo On the Edge of the Empire Formand Substance in the Sātavāhana Dynasty raquoin se a lCoCk tn drsquoa ltroy kd m orrison and Cm s inoPoli (eds) Empires Perspectives from Archaeologyand History Cambridge (2001) p 162-178
21 Mātsyapurāṇa 114 46-49 Vāyupurāṇa45 124-128 Mārkandeyapurāṇa57 45-49 where the names of pāṇḍyāḥ keralāḥ andcolāḥ are corrupt in puṇḍrāḥ kevalāḥ andgolāṅgūlāḥ
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
V BuCCiantini laquo The Limits of Knowledge Explorations of and Information from the Horn of Africa to the East African Coast in the Greco-Roman Tradition raquo 159-176
ar bm B ukharin laquo The Coastal Arabia and the adjacent Sea-Basins in thePeriplus of the Erythrean Sea (Trade Geography and Navigation) raquo 177-236J s ChietteCatte laquo LrsquoArabie du Sud et la mer duiii e siegravecle av
au vi e siegravecle apr J-C raquo 237-273A r ouGeulle laquo Syagros et autres eacutetablissements cocirctiers du Hadramawt
GolfJ- Fr s alles laquo Le Golfe persique dans lePeacuteriple de la mer Eacuterythreacutee connaissances fondeacutees et ignorances reacuteelles raquo 293-328
if d e r omanis laquo On Dachinabades and Limyrike in thePeriplus Maris Erythraei raquo 329-340
heacuter g sCla lliBert laquo Les reacuteseaux de navigation du deacutebut de lrsquoegravere chreacutetienne auxvi e siegravecle Rencontre de populations eacutechanges commerciaux et matrimoniaux
concurrence agrave lrsquoouest et agrave lrsquoest de Madagascar raquo 341-357Eacutev allet laquo LePeacuteriple au miroir des sources arabes meacutedieacutevales Le cas des produits du commerce raquo 359-380
On DachinabaDes and Limyrikein the PeriPLus maris erythraei
It is a truism that the names of regions or choronyms are a far more interestingsubject of study from a historical point of view than the names of places ortoponyms While the rationale for the latter lies in too distant and obscure timesand tells very little about the subsequent history of a place choronyms are sensitiveto historical dynamics and their signi cance may uctuate either objectively inthe diachrony or subjectively in the synchrony They may therefore be highlysigni cant for the political economic and social history of both the regions towhich they refer and the surrounding areas When compared to their equivalentsin Indian sources the occurrences in thePeriplus Maris Erythraei (PME ) of thetwo Indian choronyms Dachinabades and Limyrike allow for interesting insightsabout the historical geography of the subcontinent and the perspective by which itsauthor interprets it
Barygaza its gulf and the Ariake region around it marks a sort of geopoliticalwatershed in the map of the author of thePME On one side it signals the startof the kingdom of Manbanes and the whole of India laquo Right after Barake is the
gulf of Barygaza and the land of the Ariake region beginning of the kingdomof Manbanes and the entire India raquo1 On the other side it adjoins the land called Dachinabades which stretches widely towards the south laquo Immediately afterBarygaza the adjoining land extends from north to south For this reason the regionis called Dachinabades the south indeed is calleddachanos in their language raquo2
This emphasis on Barygaza and its region is hardly surprising In thePME the Indian Ocean is mainly represented from the perspective of Roman trade and
Barygaza was speci ed as one of the three major Indian destinations of the Romanmerchants active in those seas ndash the other two being Barbarikon at the mouth of theIndus and the emporia of the Limyrike in present-day Kerala3 Much less obvioushowever is the awareness of the importance of Barygazarsquos inland connections with Dachinabades a choronym which perhaps not insigni cantly is not found in anyother Greek or Latin work
Dachinabades is the Greek transcript of a Prakrit form of the Sanskrit Dakṣiṇāpatha a notion which evokes ndash in different ways as we shall see ndash thesouthern parts of India From it the concept of Deccan arose4 Exceptionally andquite interestingly the author of thePMEoffers an interpretation of the expression
by referencing the local language5
Even more remarkable however is the fact thathis understanding of the term is strictly speaking both imprecise and incomplete Asa matter of fact the compound transcribed in Greek as Dachinabades is composedfrom two words the adverbdakṣiṇā which actually means laquo southwards raquoand the substantive patha laquo road route raquo which the author omitted to translateAgain strictly speaking his explanation of the term is inexact Since it meanslaquo southwards route raquo the expression Dakṣiṇāpatha cannot be logically explainedwith the circumstance that after Barygaza the land (ἤπειρος) extends from north tosouth
However the omission of patha and the imprecise rendition ofdakṣiṇā arenot casual mistakes At the time of thePME Dakṣiṇāpatha was not understoodanymore in its original sense of laquo southwards route raquo In the everyday languageof the locals the hodonym (name of a road) laquo southwards road raquo had become achoronym designating the region south of the Barygaza region As a consequencea translation of the second element of the compound ( patha) and a proper
Ἀρωμάτων ἀφιέντες οἱ μὲν εἰς Λιμυρικὴν πλέοντες ἐπὶ πλεῖον τραχηλίζοντεςοἱ δὲ εἰς Βαρύγαζαν οἵ τε (Muumlller οἱ δὲ) εἰς Σκυθίαν κτλ Consequently lists ofthe exchanged items are given only for Barbarikon (PME 39) Barygaza (PME 49)and the emporia of the Limyrike (PME 56) Ships and merchants sailing back fromeither Barygaza and the Ariake or Barygaza or Limyrike may call respectively at theemporia of the Somali coast (PME14) or Socotra (PME 31) and Moscha (PME 32)Barygaza is a major destination also for merchants from Muza (PME21) and Cane(PME27) its ships sail to Omana (PME36) it receives goods from Minnagara inAriake (PME41) and Ozene (PME48)
4 Cfr rG B handarkar Early History of the Dekkan Bombay 18952 p 1-2 dk C hakraBarti The Archaeology of the Deccan Routes The Ancient Routes fromthe Ganga Plain to the Deccan New Delhi (2005) p 1-21
5 Elsewhere African and Arabian toponyms are etymologically explained byreferencing the Greek cfrPME 16 [hellip]τὰ Ῥάπτα [τὰ] (L Stuck)λεγόμεναταύτην ἔχον τὴν προσωνυμίαν ἀπὸ τῶν προειρημένων ῥαπτῶν πλοιαρίων 26 Εὐδαίμων Ἀραβία [hellip] Εὐδαίμων δrsquo ἐπεκλήθη κτλ
understanding of the rst (dakṣiṇā) must have appeared respectively misleadingand meaningless to the author of thePME and his Indian informers
This evolution from hodonym to choronym is not unparalleled Similar casesare those of the ancient Italian choronyms Aemilia(modern Emilia) andFlaminiawhich come out from the hodonymsvia Aemiliathe former andvia Flaminiathelatter However such transformations take place only when a new trunk road deeplyaffects the material life of the surrounding areas The commercial importanceof what must have been the original Dakṣiṇāpatha 6 ndash the main trunk road thatwent from Barygaza to Paithana Tagara and probably the coastal regions of theBay of Bengal ndash is still clearly recognizable in the text of thePME In chapter 51
the location of the two major emporia of Dachinabades Paithana and Tagara isaccompanied by speci c information concerning the number of travel days fromBarygaza and Paithana respectively laquo Among those in the same Dachinabadestwo are the most conspicuous emporia Paithana which is twenty days travel southof Barygaza and ten days east from it Tagara another very big city raquo7 Extensionsof the laquo southwards route raquo from Tagara to the coastal regions of the Bay of Bengalare not explicitly mentioned Still they are implied by the merchandise laquo of thecoastal parts raquo rst exported to Tagara and then re-exported to Barygaza8
In its original hodonymic sense the expression Dakṣiṇāpatha occurs in apassage of the Arthaśāstra laquo In the case of a land-route lsquoThe route to the Himavatis preferable to the southwards route Commodities such as elephants horsesperfumes ivory skins silver and gold are of very high valuersquo So say the teacherslsquoNorsquo says Kauṭilya lsquoleaving out commodities such as blankets skins and horses(those items) and commodities such as conch shells diamonds rubies pearls andgold are more plentiful on the southwards routersquo raquo9
6 Cfr eghP r ay Monastery and Guild Commerce under the Sātavāhanas Delhi(1986)
7 PME 51 τῶν δὲ ἐν αὐτῇ τῇ Δαχιναβάδει δύο ἐστὶν τὰ διασημότατα ἐμπόριαΠαίθανα μὲν(Schwanbeck ἐπιφαινόμενα) ἀπὸ Βαρυγάζων ἔχουσα ὁδὸν ἡμερῶνεἴκοσι πρὸς νότον ἀπὸ ltδὲgt (edd)ταύτης ὡς ἡμερῶν δέκα πρὸς ἀνατολὴν ἑτέραπόλις μεγίστη Ταγάρα (Fabricius Ταγάρα μεγίστη) Hereἐν αὐτῇ τῇ =ἐν τῇ αὐτῇ cfr h f risk Le peacuteriple de la mer Eacuterythreacutee Suivi drsquoune eacutetude sur la tradition et lalangue Goumlteborg (1927) p 65-66
This passage is noteworthy in many ways Besides showing the term inits original meaning and con rming that it actually referred to a land-route(sthalapatha) and not to as would have been theoretically possible a water-route(vāripatha) it proves that Dakṣiṇāpatha was a commercial road used by merchantswho could choose to carry on business either along the laquo southwards raquo or the Haimavata route10 It makes therefore clear that the laquo southwards route raquo was acontrastive denomination suggested by the opposition (and competition) betweenthe laquo southwards raquo and the Haimavata routes For a long time trade with Himālayahad been considered preferable to trade along the laquo southwards route raquo but whenthose lines of the Arthaśāstra were conceived the latter turned out to be more
lucrative than the former thereby disproving the eldersrsquo opinions It is thereforeself-evident that the expression Dakṣiṇāpatha was forged in an area around thenorthern end of that route by a commercial milieu whose trading activities werepoised between the Himālaya and Deccan This circumstance explains why Dakṣiṇāpatha is so to speak a one-way hodonym oriented in a direction north-south
It has been observed that the contextual reference to Deccan and Himālayaland-routes implies a standpoint laquo of a person in the midlands neither a southernernor a northerner raquo11 The text of thePME may suggest a more precise localizationIn chapter 47 the list of tribes lying behind Barygaza ndash which includes theArachosians the Gandharians and the Bactrians12 ndash re ects Barygazarsquos commercialconnections with the Himālayan region The same conclusion can be drawn fromchapter 48 where import to Barygaza via Ozene (= Ujjayinī) of nard costus andbdellium from the laquo upper places raquo is mentioned13 Just mid-way between the
10 For archaeological evidence pointing to connections between Deccan and Gandharacfr P BranCaCCio laquo Close Encounters Multicultural Systems in Ancient India raquoin d s rinivasan (ed)On the Cusp of an Era Art in the Pre-Kuṣāṇa World Leiden(2007) p 385ndash398 e ad The Buddhist Cave at Aurangabad Transformationin Art and Religion Leiden (2010) p 67-68 69-70 e ad laquo Terracottas fromWestern Deccan An Exploration of Sources and Transmission of Models in theEarly Historic Period raquoin Ph Granoff (ed) A Pantheon Rediscovered ChangingPerceptions of Early Historic India forthcoming
11 rP k anGle TheKauṭilīya Arthaśāstra Part II An English Translation with Criticaland Explanatory Notes Bombay (19722) p 360 nt 22
Himālaya and Deccan regions the Barygaza-Ozene region was perhaps the onlyIndian area where the alternative between a Haimavata and a Dakṣiṇāpatha land-route made perfect sense
Incidentally I would like to point out another example in which a comparisonwith the text of the Arthaśāstra may improve our understanding of the text ofthe PME Again it relates to the Dachinabades land-route In chapter 51 itemsfrom Paithana and Tagara are brought to Barygazaπορείαις ἁμαξῶν καὶ ἀνοδίαιςμεγίσταις which Casson translates as laquo by conveyance in wagons over very greatroadless stretches raquo14 Muumlllerrsquos Latin rendition is similar laquo per loca maximeinvia plaustris merces devehuntur raquo15 Both these translations assume that such
long journey was made on stretches of land that were simultaneously roadless andaccessible by wagons which is contradictoryEvoked by the author of the Arthaśāstra in the very same chapter containing
the comparison between Dakṣiṇāpatha and Haimavata land-route is thedistinction betweencakrapatha laquo the wheel-track raquo pādapatha laquo the foot-path raquoand kharoṣṭrapatha laquo the ass- or camel-road raquo laquo [hellip] as between a wheel-trackand a foot-path the wheel-track is preferable as it enables undertakings on a largescale Alternatively path for donkeys and camels in conformity with place and time(is to be preferred) raquo16 This line suggests that a long land-route may not be entirelysuitable to wheeled vehicles and may therefore require changes in the mode oftransport such as shifting between wagons and donkeys It is consequently probablethat atPME51 theκαί which connectsπορείαις ἁμαξῶν andἀνοδίαις μεγίσταις is disjunctive from Paithana and Tagara the items are brought laquo with journeysof wagons and long roadless tracts raquo In other words laquo journeys of wagons raquo(πορεῖαι ἁμαξῶν) and laquo roadless tracts raquo (ἀνοδίαι μέγισται) refer to two distinctmodalities of transport each adopted as needed along different legs of the voyageThe long route from Tagara to Paithana and Barygaza was to be travelled withwagons on laquo wheel-track raquo (cakrapatha) only in part for long stretches lsquothe ass- orcamelroadsrsquo (kharoṣṭrapatha) were the only option available
While in the Arthaśāstra the term Dakṣiṇāpatha still retains its originalmeaning of laquo southwards route raquo its earliest inscriptional occurrences already show
Σκυθίας ὅ τε κόστος καὶ ἡ βδέλλα The same items are available at Barbarikon onthe mouth of the Indus PME39
14 l C asson ThePeriplus Maris Erythraei Text with Introduction Translation andCommentary Princeton (1989) p 83
15 C m uumlller Geographi Graeci Minores I Paris (1855) p 294 A more t translationis given byWh s Choff The Periplus of Erythraean Sea New York LondonBombay and Calcutta (1912) p 43 laquo [hellip] brought down to Barygaza from theseplaces by wagons and through great tracts without roads raquo
16 Arthaśāstra VII 12 27 tatrāpi cakrapādapathayoścakrapatho vipulārambhatvātcchreyāndeśakālasambhāvano vā kharoṣṭrapathaḥ
its evolution into a choronym Dakṣiṇāpatha was claimed by Sātavāhana rulers as thegeographical frame of their sovereignty Dakhiṇāpathapati and Dakhiṇāpathesara are the synonymous formulas (both meaning laquo Lord of Dakhiṇāpatha raquo) whichoccur in inscriptions from Nāṇeghāṭ Nāsik and Girnār17 The inscriptions of Nāsikand Girnār refer respectively to Vāsiṣṭhīputra Śrī Puḷumāvi and to his immediatesuccessor Vāsiṣṭhīputra Śrī Śātakarṇi The inscription from Nāṇeghāṭ probablyrefers to Simuka the founder of the dynasty
The lordship over Dakhiṇāpatha claimed by the Sātavāhanas explains howone of them called laquo the elder Saraganos raquo (= Sātakarṇi) by the author of thePME made Kalliena not far from Nāṇeghāṭ a laquo lawful emporion raquo18 It explains also
why Ptolemy locates at Baithana (= Paithana one of the two most conspicuousemporia ofPME rsquos Dachinabades) the royal residence of Sirip(t)olemaios (= SiriPuḷumāvi)19 Nonetheless however powerful (or charlatan) were the Sātavāhanakings20 their domain could have hardly coincided with the Dakṣiṇāpatha of somepassages of thePurāṇas where it includes Pāṇḍya Kerala and Cola in the southernend of the subcontinent21 Disconnected from any political or economical realitythis expanded Dakṣiṇāpatha is the result of a generalization elaborated by the
17 vv m irashi The History and Inscriptions of the Sātavāhanas and the WesternKshatrapas Bombay (1981) n 3 l 2 p 11 (dakhiṇāpathapatiin the Nāṇeghāṭinscription of Nāganikā) ibid n 18 l 11 p 46 (dakhiṇāpathesarain the Nāsikinscription of Vāsiṣṭhīputra Śrī Puḷumāvi) dC s irCar Select Inscriptions Bearingon Indian History and Civilization Volume I From the Sixth Century BC to theSixth Century AD Calcutta (19652) n 67 l 12 p 178 (Vāsiṣṭhīputra Śrī Sātakarṇistyled as dakhiṇāpathapatiin the Rudradāman inscription of Girnār)
18 PME 52 τοπικὰ δὲ ἐμπόρια κατὰ τὸ ἑξῆς κείμενα Ἀκαβαρου Σούππαρα (edd ἀπὸ Βαρυltγάζωνgt Σούππαρα Muumlller in comm Ἀκαβαρους οὔππαρα) Καλλίενα (καὶ ΚαλλίεναFrisk) πόλις ἡ ἐπὶ τῶν Σαραγάνου τοῦ πρεσβυτέρου χρόνωνἐμπόριον (Gelenius ἐμποριω) ἔνθεσμον γενομένη (Fabricius γενόμενον) Forthe identi cation of the laquo elder Saraganos raquo with the husband of queen Nāganikācfr am s hastri The Sātavāhanas and the Western Kshatrapas a historical framework Nagpur (1998) p 55-60
19 Ptol Geogr VII 1 82 Βαίθανα βασίλειον Σιριπτολεμαίου (UKX Σίριοςπολεμίου V Σίριος πτολεμαίου R Σιροπολεμίου A Σιριπολεμαίου Z) ριζ ιη LʹIt is to be noticed that in Bhāgavatapurāṇa IX 1 41-42 the title pratiṣṭhānapati (skt Pratiṣṭhāna = Gk Paithana) is given to a king whose three sons weredakṣiṇāpatharājānaḥ laquo kings ofdakṣiṇāpatha raquo
20 s hastri op cit p 7-12 C s inoPoli laquo On the Edge of the Empire Formand Substance in the Sātavāhana Dynasty raquoin se a lCoCk tn drsquoa ltroy kd m orrison and Cm s inoPoli (eds) Empires Perspectives from Archaeologyand History Cambridge (2001) p 162-178
21 Mātsyapurāṇa 114 46-49 Vāyupurāṇa45 124-128 Mārkandeyapurāṇa57 45-49 where the names of pāṇḍyāḥ keralāḥ andcolāḥ are corrupt in puṇḍrāḥ kevalāḥ andgolāṅgūlāḥ
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
On DachinabaDes and Limyrikein the PeriPLus maris erythraei
It is a truism that the names of regions or choronyms are a far more interestingsubject of study from a historical point of view than the names of places ortoponyms While the rationale for the latter lies in too distant and obscure timesand tells very little about the subsequent history of a place choronyms are sensitiveto historical dynamics and their signi cance may uctuate either objectively inthe diachrony or subjectively in the synchrony They may therefore be highlysigni cant for the political economic and social history of both the regions towhich they refer and the surrounding areas When compared to their equivalentsin Indian sources the occurrences in thePeriplus Maris Erythraei (PME ) of thetwo Indian choronyms Dachinabades and Limyrike allow for interesting insightsabout the historical geography of the subcontinent and the perspective by which itsauthor interprets it
Barygaza its gulf and the Ariake region around it marks a sort of geopoliticalwatershed in the map of the author of thePME On one side it signals the startof the kingdom of Manbanes and the whole of India laquo Right after Barake is the
gulf of Barygaza and the land of the Ariake region beginning of the kingdomof Manbanes and the entire India raquo1 On the other side it adjoins the land called Dachinabades which stretches widely towards the south laquo Immediately afterBarygaza the adjoining land extends from north to south For this reason the regionis called Dachinabades the south indeed is calleddachanos in their language raquo2
This emphasis on Barygaza and its region is hardly surprising In thePME the Indian Ocean is mainly represented from the perspective of Roman trade and
Barygaza was speci ed as one of the three major Indian destinations of the Romanmerchants active in those seas ndash the other two being Barbarikon at the mouth of theIndus and the emporia of the Limyrike in present-day Kerala3 Much less obvioushowever is the awareness of the importance of Barygazarsquos inland connections with Dachinabades a choronym which perhaps not insigni cantly is not found in anyother Greek or Latin work
Dachinabades is the Greek transcript of a Prakrit form of the Sanskrit Dakṣiṇāpatha a notion which evokes ndash in different ways as we shall see ndash thesouthern parts of India From it the concept of Deccan arose4 Exceptionally andquite interestingly the author of thePMEoffers an interpretation of the expression
by referencing the local language5
Even more remarkable however is the fact thathis understanding of the term is strictly speaking both imprecise and incomplete Asa matter of fact the compound transcribed in Greek as Dachinabades is composedfrom two words the adverbdakṣiṇā which actually means laquo southwards raquoand the substantive patha laquo road route raquo which the author omitted to translateAgain strictly speaking his explanation of the term is inexact Since it meanslaquo southwards route raquo the expression Dakṣiṇāpatha cannot be logically explainedwith the circumstance that after Barygaza the land (ἤπειρος) extends from north tosouth
However the omission of patha and the imprecise rendition ofdakṣiṇā arenot casual mistakes At the time of thePME Dakṣiṇāpatha was not understoodanymore in its original sense of laquo southwards route raquo In the everyday languageof the locals the hodonym (name of a road) laquo southwards road raquo had become achoronym designating the region south of the Barygaza region As a consequencea translation of the second element of the compound ( patha) and a proper
Ἀρωμάτων ἀφιέντες οἱ μὲν εἰς Λιμυρικὴν πλέοντες ἐπὶ πλεῖον τραχηλίζοντεςοἱ δὲ εἰς Βαρύγαζαν οἵ τε (Muumlller οἱ δὲ) εἰς Σκυθίαν κτλ Consequently lists ofthe exchanged items are given only for Barbarikon (PME 39) Barygaza (PME 49)and the emporia of the Limyrike (PME 56) Ships and merchants sailing back fromeither Barygaza and the Ariake or Barygaza or Limyrike may call respectively at theemporia of the Somali coast (PME14) or Socotra (PME 31) and Moscha (PME 32)Barygaza is a major destination also for merchants from Muza (PME21) and Cane(PME27) its ships sail to Omana (PME36) it receives goods from Minnagara inAriake (PME41) and Ozene (PME48)
4 Cfr rG B handarkar Early History of the Dekkan Bombay 18952 p 1-2 dk C hakraBarti The Archaeology of the Deccan Routes The Ancient Routes fromthe Ganga Plain to the Deccan New Delhi (2005) p 1-21
5 Elsewhere African and Arabian toponyms are etymologically explained byreferencing the Greek cfrPME 16 [hellip]τὰ Ῥάπτα [τὰ] (L Stuck)λεγόμεναταύτην ἔχον τὴν προσωνυμίαν ἀπὸ τῶν προειρημένων ῥαπτῶν πλοιαρίων 26 Εὐδαίμων Ἀραβία [hellip] Εὐδαίμων δrsquo ἐπεκλήθη κτλ
understanding of the rst (dakṣiṇā) must have appeared respectively misleadingand meaningless to the author of thePME and his Indian informers
This evolution from hodonym to choronym is not unparalleled Similar casesare those of the ancient Italian choronyms Aemilia(modern Emilia) andFlaminiawhich come out from the hodonymsvia Aemiliathe former andvia Flaminiathelatter However such transformations take place only when a new trunk road deeplyaffects the material life of the surrounding areas The commercial importanceof what must have been the original Dakṣiṇāpatha 6 ndash the main trunk road thatwent from Barygaza to Paithana Tagara and probably the coastal regions of theBay of Bengal ndash is still clearly recognizable in the text of thePME In chapter 51
the location of the two major emporia of Dachinabades Paithana and Tagara isaccompanied by speci c information concerning the number of travel days fromBarygaza and Paithana respectively laquo Among those in the same Dachinabadestwo are the most conspicuous emporia Paithana which is twenty days travel southof Barygaza and ten days east from it Tagara another very big city raquo7 Extensionsof the laquo southwards route raquo from Tagara to the coastal regions of the Bay of Bengalare not explicitly mentioned Still they are implied by the merchandise laquo of thecoastal parts raquo rst exported to Tagara and then re-exported to Barygaza8
In its original hodonymic sense the expression Dakṣiṇāpatha occurs in apassage of the Arthaśāstra laquo In the case of a land-route lsquoThe route to the Himavatis preferable to the southwards route Commodities such as elephants horsesperfumes ivory skins silver and gold are of very high valuersquo So say the teacherslsquoNorsquo says Kauṭilya lsquoleaving out commodities such as blankets skins and horses(those items) and commodities such as conch shells diamonds rubies pearls andgold are more plentiful on the southwards routersquo raquo9
6 Cfr eghP r ay Monastery and Guild Commerce under the Sātavāhanas Delhi(1986)
7 PME 51 τῶν δὲ ἐν αὐτῇ τῇ Δαχιναβάδει δύο ἐστὶν τὰ διασημότατα ἐμπόριαΠαίθανα μὲν(Schwanbeck ἐπιφαινόμενα) ἀπὸ Βαρυγάζων ἔχουσα ὁδὸν ἡμερῶνεἴκοσι πρὸς νότον ἀπὸ ltδὲgt (edd)ταύτης ὡς ἡμερῶν δέκα πρὸς ἀνατολὴν ἑτέραπόλις μεγίστη Ταγάρα (Fabricius Ταγάρα μεγίστη) Hereἐν αὐτῇ τῇ =ἐν τῇ αὐτῇ cfr h f risk Le peacuteriple de la mer Eacuterythreacutee Suivi drsquoune eacutetude sur la tradition et lalangue Goumlteborg (1927) p 65-66
This passage is noteworthy in many ways Besides showing the term inits original meaning and con rming that it actually referred to a land-route(sthalapatha) and not to as would have been theoretically possible a water-route(vāripatha) it proves that Dakṣiṇāpatha was a commercial road used by merchantswho could choose to carry on business either along the laquo southwards raquo or the Haimavata route10 It makes therefore clear that the laquo southwards route raquo was acontrastive denomination suggested by the opposition (and competition) betweenthe laquo southwards raquo and the Haimavata routes For a long time trade with Himālayahad been considered preferable to trade along the laquo southwards route raquo but whenthose lines of the Arthaśāstra were conceived the latter turned out to be more
lucrative than the former thereby disproving the eldersrsquo opinions It is thereforeself-evident that the expression Dakṣiṇāpatha was forged in an area around thenorthern end of that route by a commercial milieu whose trading activities werepoised between the Himālaya and Deccan This circumstance explains why Dakṣiṇāpatha is so to speak a one-way hodonym oriented in a direction north-south
It has been observed that the contextual reference to Deccan and Himālayaland-routes implies a standpoint laquo of a person in the midlands neither a southernernor a northerner raquo11 The text of thePME may suggest a more precise localizationIn chapter 47 the list of tribes lying behind Barygaza ndash which includes theArachosians the Gandharians and the Bactrians12 ndash re ects Barygazarsquos commercialconnections with the Himālayan region The same conclusion can be drawn fromchapter 48 where import to Barygaza via Ozene (= Ujjayinī) of nard costus andbdellium from the laquo upper places raquo is mentioned13 Just mid-way between the
10 For archaeological evidence pointing to connections between Deccan and Gandharacfr P BranCaCCio laquo Close Encounters Multicultural Systems in Ancient India raquoin d s rinivasan (ed)On the Cusp of an Era Art in the Pre-Kuṣāṇa World Leiden(2007) p 385ndash398 e ad The Buddhist Cave at Aurangabad Transformationin Art and Religion Leiden (2010) p 67-68 69-70 e ad laquo Terracottas fromWestern Deccan An Exploration of Sources and Transmission of Models in theEarly Historic Period raquoin Ph Granoff (ed) A Pantheon Rediscovered ChangingPerceptions of Early Historic India forthcoming
11 rP k anGle TheKauṭilīya Arthaśāstra Part II An English Translation with Criticaland Explanatory Notes Bombay (19722) p 360 nt 22
Himālaya and Deccan regions the Barygaza-Ozene region was perhaps the onlyIndian area where the alternative between a Haimavata and a Dakṣiṇāpatha land-route made perfect sense
Incidentally I would like to point out another example in which a comparisonwith the text of the Arthaśāstra may improve our understanding of the text ofthe PME Again it relates to the Dachinabades land-route In chapter 51 itemsfrom Paithana and Tagara are brought to Barygazaπορείαις ἁμαξῶν καὶ ἀνοδίαιςμεγίσταις which Casson translates as laquo by conveyance in wagons over very greatroadless stretches raquo14 Muumlllerrsquos Latin rendition is similar laquo per loca maximeinvia plaustris merces devehuntur raquo15 Both these translations assume that such
long journey was made on stretches of land that were simultaneously roadless andaccessible by wagons which is contradictoryEvoked by the author of the Arthaśāstra in the very same chapter containing
the comparison between Dakṣiṇāpatha and Haimavata land-route is thedistinction betweencakrapatha laquo the wheel-track raquo pādapatha laquo the foot-path raquoand kharoṣṭrapatha laquo the ass- or camel-road raquo laquo [hellip] as between a wheel-trackand a foot-path the wheel-track is preferable as it enables undertakings on a largescale Alternatively path for donkeys and camels in conformity with place and time(is to be preferred) raquo16 This line suggests that a long land-route may not be entirelysuitable to wheeled vehicles and may therefore require changes in the mode oftransport such as shifting between wagons and donkeys It is consequently probablethat atPME51 theκαί which connectsπορείαις ἁμαξῶν andἀνοδίαις μεγίσταις is disjunctive from Paithana and Tagara the items are brought laquo with journeysof wagons and long roadless tracts raquo In other words laquo journeys of wagons raquo(πορεῖαι ἁμαξῶν) and laquo roadless tracts raquo (ἀνοδίαι μέγισται) refer to two distinctmodalities of transport each adopted as needed along different legs of the voyageThe long route from Tagara to Paithana and Barygaza was to be travelled withwagons on laquo wheel-track raquo (cakrapatha) only in part for long stretches lsquothe ass- orcamelroadsrsquo (kharoṣṭrapatha) were the only option available
While in the Arthaśāstra the term Dakṣiṇāpatha still retains its originalmeaning of laquo southwards route raquo its earliest inscriptional occurrences already show
Σκυθίας ὅ τε κόστος καὶ ἡ βδέλλα The same items are available at Barbarikon onthe mouth of the Indus PME39
14 l C asson ThePeriplus Maris Erythraei Text with Introduction Translation andCommentary Princeton (1989) p 83
15 C m uumlller Geographi Graeci Minores I Paris (1855) p 294 A more t translationis given byWh s Choff The Periplus of Erythraean Sea New York LondonBombay and Calcutta (1912) p 43 laquo [hellip] brought down to Barygaza from theseplaces by wagons and through great tracts without roads raquo
16 Arthaśāstra VII 12 27 tatrāpi cakrapādapathayoścakrapatho vipulārambhatvātcchreyāndeśakālasambhāvano vā kharoṣṭrapathaḥ
its evolution into a choronym Dakṣiṇāpatha was claimed by Sātavāhana rulers as thegeographical frame of their sovereignty Dakhiṇāpathapati and Dakhiṇāpathesara are the synonymous formulas (both meaning laquo Lord of Dakhiṇāpatha raquo) whichoccur in inscriptions from Nāṇeghāṭ Nāsik and Girnār17 The inscriptions of Nāsikand Girnār refer respectively to Vāsiṣṭhīputra Śrī Puḷumāvi and to his immediatesuccessor Vāsiṣṭhīputra Śrī Śātakarṇi The inscription from Nāṇeghāṭ probablyrefers to Simuka the founder of the dynasty
The lordship over Dakhiṇāpatha claimed by the Sātavāhanas explains howone of them called laquo the elder Saraganos raquo (= Sātakarṇi) by the author of thePME made Kalliena not far from Nāṇeghāṭ a laquo lawful emporion raquo18 It explains also
why Ptolemy locates at Baithana (= Paithana one of the two most conspicuousemporia ofPME rsquos Dachinabades) the royal residence of Sirip(t)olemaios (= SiriPuḷumāvi)19 Nonetheless however powerful (or charlatan) were the Sātavāhanakings20 their domain could have hardly coincided with the Dakṣiṇāpatha of somepassages of thePurāṇas where it includes Pāṇḍya Kerala and Cola in the southernend of the subcontinent21 Disconnected from any political or economical realitythis expanded Dakṣiṇāpatha is the result of a generalization elaborated by the
17 vv m irashi The History and Inscriptions of the Sātavāhanas and the WesternKshatrapas Bombay (1981) n 3 l 2 p 11 (dakhiṇāpathapatiin the Nāṇeghāṭinscription of Nāganikā) ibid n 18 l 11 p 46 (dakhiṇāpathesarain the Nāsikinscription of Vāsiṣṭhīputra Śrī Puḷumāvi) dC s irCar Select Inscriptions Bearingon Indian History and Civilization Volume I From the Sixth Century BC to theSixth Century AD Calcutta (19652) n 67 l 12 p 178 (Vāsiṣṭhīputra Śrī Sātakarṇistyled as dakhiṇāpathapatiin the Rudradāman inscription of Girnār)
18 PME 52 τοπικὰ δὲ ἐμπόρια κατὰ τὸ ἑξῆς κείμενα Ἀκαβαρου Σούππαρα (edd ἀπὸ Βαρυltγάζωνgt Σούππαρα Muumlller in comm Ἀκαβαρους οὔππαρα) Καλλίενα (καὶ ΚαλλίεναFrisk) πόλις ἡ ἐπὶ τῶν Σαραγάνου τοῦ πρεσβυτέρου χρόνωνἐμπόριον (Gelenius ἐμποριω) ἔνθεσμον γενομένη (Fabricius γενόμενον) Forthe identi cation of the laquo elder Saraganos raquo with the husband of queen Nāganikācfr am s hastri The Sātavāhanas and the Western Kshatrapas a historical framework Nagpur (1998) p 55-60
19 Ptol Geogr VII 1 82 Βαίθανα βασίλειον Σιριπτολεμαίου (UKX Σίριοςπολεμίου V Σίριος πτολεμαίου R Σιροπολεμίου A Σιριπολεμαίου Z) ριζ ιη LʹIt is to be noticed that in Bhāgavatapurāṇa IX 1 41-42 the title pratiṣṭhānapati (skt Pratiṣṭhāna = Gk Paithana) is given to a king whose three sons weredakṣiṇāpatharājānaḥ laquo kings ofdakṣiṇāpatha raquo
20 s hastri op cit p 7-12 C s inoPoli laquo On the Edge of the Empire Formand Substance in the Sātavāhana Dynasty raquoin se a lCoCk tn drsquoa ltroy kd m orrison and Cm s inoPoli (eds) Empires Perspectives from Archaeologyand History Cambridge (2001) p 162-178
21 Mātsyapurāṇa 114 46-49 Vāyupurāṇa45 124-128 Mārkandeyapurāṇa57 45-49 where the names of pāṇḍyāḥ keralāḥ andcolāḥ are corrupt in puṇḍrāḥ kevalāḥ andgolāṅgūlāḥ
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
Barygaza was speci ed as one of the three major Indian destinations of the Romanmerchants active in those seas ndash the other two being Barbarikon at the mouth of theIndus and the emporia of the Limyrike in present-day Kerala3 Much less obvioushowever is the awareness of the importance of Barygazarsquos inland connections with Dachinabades a choronym which perhaps not insigni cantly is not found in anyother Greek or Latin work
Dachinabades is the Greek transcript of a Prakrit form of the Sanskrit Dakṣiṇāpatha a notion which evokes ndash in different ways as we shall see ndash thesouthern parts of India From it the concept of Deccan arose4 Exceptionally andquite interestingly the author of thePMEoffers an interpretation of the expression
by referencing the local language5
Even more remarkable however is the fact thathis understanding of the term is strictly speaking both imprecise and incomplete Asa matter of fact the compound transcribed in Greek as Dachinabades is composedfrom two words the adverbdakṣiṇā which actually means laquo southwards raquoand the substantive patha laquo road route raquo which the author omitted to translateAgain strictly speaking his explanation of the term is inexact Since it meanslaquo southwards route raquo the expression Dakṣiṇāpatha cannot be logically explainedwith the circumstance that after Barygaza the land (ἤπειρος) extends from north tosouth
However the omission of patha and the imprecise rendition ofdakṣiṇā arenot casual mistakes At the time of thePME Dakṣiṇāpatha was not understoodanymore in its original sense of laquo southwards route raquo In the everyday languageof the locals the hodonym (name of a road) laquo southwards road raquo had become achoronym designating the region south of the Barygaza region As a consequencea translation of the second element of the compound ( patha) and a proper
Ἀρωμάτων ἀφιέντες οἱ μὲν εἰς Λιμυρικὴν πλέοντες ἐπὶ πλεῖον τραχηλίζοντεςοἱ δὲ εἰς Βαρύγαζαν οἵ τε (Muumlller οἱ δὲ) εἰς Σκυθίαν κτλ Consequently lists ofthe exchanged items are given only for Barbarikon (PME 39) Barygaza (PME 49)and the emporia of the Limyrike (PME 56) Ships and merchants sailing back fromeither Barygaza and the Ariake or Barygaza or Limyrike may call respectively at theemporia of the Somali coast (PME14) or Socotra (PME 31) and Moscha (PME 32)Barygaza is a major destination also for merchants from Muza (PME21) and Cane(PME27) its ships sail to Omana (PME36) it receives goods from Minnagara inAriake (PME41) and Ozene (PME48)
4 Cfr rG B handarkar Early History of the Dekkan Bombay 18952 p 1-2 dk C hakraBarti The Archaeology of the Deccan Routes The Ancient Routes fromthe Ganga Plain to the Deccan New Delhi (2005) p 1-21
5 Elsewhere African and Arabian toponyms are etymologically explained byreferencing the Greek cfrPME 16 [hellip]τὰ Ῥάπτα [τὰ] (L Stuck)λεγόμεναταύτην ἔχον τὴν προσωνυμίαν ἀπὸ τῶν προειρημένων ῥαπτῶν πλοιαρίων 26 Εὐδαίμων Ἀραβία [hellip] Εὐδαίμων δrsquo ἐπεκλήθη κτλ
understanding of the rst (dakṣiṇā) must have appeared respectively misleadingand meaningless to the author of thePME and his Indian informers
This evolution from hodonym to choronym is not unparalleled Similar casesare those of the ancient Italian choronyms Aemilia(modern Emilia) andFlaminiawhich come out from the hodonymsvia Aemiliathe former andvia Flaminiathelatter However such transformations take place only when a new trunk road deeplyaffects the material life of the surrounding areas The commercial importanceof what must have been the original Dakṣiṇāpatha 6 ndash the main trunk road thatwent from Barygaza to Paithana Tagara and probably the coastal regions of theBay of Bengal ndash is still clearly recognizable in the text of thePME In chapter 51
the location of the two major emporia of Dachinabades Paithana and Tagara isaccompanied by speci c information concerning the number of travel days fromBarygaza and Paithana respectively laquo Among those in the same Dachinabadestwo are the most conspicuous emporia Paithana which is twenty days travel southof Barygaza and ten days east from it Tagara another very big city raquo7 Extensionsof the laquo southwards route raquo from Tagara to the coastal regions of the Bay of Bengalare not explicitly mentioned Still they are implied by the merchandise laquo of thecoastal parts raquo rst exported to Tagara and then re-exported to Barygaza8
In its original hodonymic sense the expression Dakṣiṇāpatha occurs in apassage of the Arthaśāstra laquo In the case of a land-route lsquoThe route to the Himavatis preferable to the southwards route Commodities such as elephants horsesperfumes ivory skins silver and gold are of very high valuersquo So say the teacherslsquoNorsquo says Kauṭilya lsquoleaving out commodities such as blankets skins and horses(those items) and commodities such as conch shells diamonds rubies pearls andgold are more plentiful on the southwards routersquo raquo9
6 Cfr eghP r ay Monastery and Guild Commerce under the Sātavāhanas Delhi(1986)
7 PME 51 τῶν δὲ ἐν αὐτῇ τῇ Δαχιναβάδει δύο ἐστὶν τὰ διασημότατα ἐμπόριαΠαίθανα μὲν(Schwanbeck ἐπιφαινόμενα) ἀπὸ Βαρυγάζων ἔχουσα ὁδὸν ἡμερῶνεἴκοσι πρὸς νότον ἀπὸ ltδὲgt (edd)ταύτης ὡς ἡμερῶν δέκα πρὸς ἀνατολὴν ἑτέραπόλις μεγίστη Ταγάρα (Fabricius Ταγάρα μεγίστη) Hereἐν αὐτῇ τῇ =ἐν τῇ αὐτῇ cfr h f risk Le peacuteriple de la mer Eacuterythreacutee Suivi drsquoune eacutetude sur la tradition et lalangue Goumlteborg (1927) p 65-66
This passage is noteworthy in many ways Besides showing the term inits original meaning and con rming that it actually referred to a land-route(sthalapatha) and not to as would have been theoretically possible a water-route(vāripatha) it proves that Dakṣiṇāpatha was a commercial road used by merchantswho could choose to carry on business either along the laquo southwards raquo or the Haimavata route10 It makes therefore clear that the laquo southwards route raquo was acontrastive denomination suggested by the opposition (and competition) betweenthe laquo southwards raquo and the Haimavata routes For a long time trade with Himālayahad been considered preferable to trade along the laquo southwards route raquo but whenthose lines of the Arthaśāstra were conceived the latter turned out to be more
lucrative than the former thereby disproving the eldersrsquo opinions It is thereforeself-evident that the expression Dakṣiṇāpatha was forged in an area around thenorthern end of that route by a commercial milieu whose trading activities werepoised between the Himālaya and Deccan This circumstance explains why Dakṣiṇāpatha is so to speak a one-way hodonym oriented in a direction north-south
It has been observed that the contextual reference to Deccan and Himālayaland-routes implies a standpoint laquo of a person in the midlands neither a southernernor a northerner raquo11 The text of thePME may suggest a more precise localizationIn chapter 47 the list of tribes lying behind Barygaza ndash which includes theArachosians the Gandharians and the Bactrians12 ndash re ects Barygazarsquos commercialconnections with the Himālayan region The same conclusion can be drawn fromchapter 48 where import to Barygaza via Ozene (= Ujjayinī) of nard costus andbdellium from the laquo upper places raquo is mentioned13 Just mid-way between the
10 For archaeological evidence pointing to connections between Deccan and Gandharacfr P BranCaCCio laquo Close Encounters Multicultural Systems in Ancient India raquoin d s rinivasan (ed)On the Cusp of an Era Art in the Pre-Kuṣāṇa World Leiden(2007) p 385ndash398 e ad The Buddhist Cave at Aurangabad Transformationin Art and Religion Leiden (2010) p 67-68 69-70 e ad laquo Terracottas fromWestern Deccan An Exploration of Sources and Transmission of Models in theEarly Historic Period raquoin Ph Granoff (ed) A Pantheon Rediscovered ChangingPerceptions of Early Historic India forthcoming
11 rP k anGle TheKauṭilīya Arthaśāstra Part II An English Translation with Criticaland Explanatory Notes Bombay (19722) p 360 nt 22
Himālaya and Deccan regions the Barygaza-Ozene region was perhaps the onlyIndian area where the alternative between a Haimavata and a Dakṣiṇāpatha land-route made perfect sense
Incidentally I would like to point out another example in which a comparisonwith the text of the Arthaśāstra may improve our understanding of the text ofthe PME Again it relates to the Dachinabades land-route In chapter 51 itemsfrom Paithana and Tagara are brought to Barygazaπορείαις ἁμαξῶν καὶ ἀνοδίαιςμεγίσταις which Casson translates as laquo by conveyance in wagons over very greatroadless stretches raquo14 Muumlllerrsquos Latin rendition is similar laquo per loca maximeinvia plaustris merces devehuntur raquo15 Both these translations assume that such
long journey was made on stretches of land that were simultaneously roadless andaccessible by wagons which is contradictoryEvoked by the author of the Arthaśāstra in the very same chapter containing
the comparison between Dakṣiṇāpatha and Haimavata land-route is thedistinction betweencakrapatha laquo the wheel-track raquo pādapatha laquo the foot-path raquoand kharoṣṭrapatha laquo the ass- or camel-road raquo laquo [hellip] as between a wheel-trackand a foot-path the wheel-track is preferable as it enables undertakings on a largescale Alternatively path for donkeys and camels in conformity with place and time(is to be preferred) raquo16 This line suggests that a long land-route may not be entirelysuitable to wheeled vehicles and may therefore require changes in the mode oftransport such as shifting between wagons and donkeys It is consequently probablethat atPME51 theκαί which connectsπορείαις ἁμαξῶν andἀνοδίαις μεγίσταις is disjunctive from Paithana and Tagara the items are brought laquo with journeysof wagons and long roadless tracts raquo In other words laquo journeys of wagons raquo(πορεῖαι ἁμαξῶν) and laquo roadless tracts raquo (ἀνοδίαι μέγισται) refer to two distinctmodalities of transport each adopted as needed along different legs of the voyageThe long route from Tagara to Paithana and Barygaza was to be travelled withwagons on laquo wheel-track raquo (cakrapatha) only in part for long stretches lsquothe ass- orcamelroadsrsquo (kharoṣṭrapatha) were the only option available
While in the Arthaśāstra the term Dakṣiṇāpatha still retains its originalmeaning of laquo southwards route raquo its earliest inscriptional occurrences already show
Σκυθίας ὅ τε κόστος καὶ ἡ βδέλλα The same items are available at Barbarikon onthe mouth of the Indus PME39
14 l C asson ThePeriplus Maris Erythraei Text with Introduction Translation andCommentary Princeton (1989) p 83
15 C m uumlller Geographi Graeci Minores I Paris (1855) p 294 A more t translationis given byWh s Choff The Periplus of Erythraean Sea New York LondonBombay and Calcutta (1912) p 43 laquo [hellip] brought down to Barygaza from theseplaces by wagons and through great tracts without roads raquo
16 Arthaśāstra VII 12 27 tatrāpi cakrapādapathayoścakrapatho vipulārambhatvātcchreyāndeśakālasambhāvano vā kharoṣṭrapathaḥ
its evolution into a choronym Dakṣiṇāpatha was claimed by Sātavāhana rulers as thegeographical frame of their sovereignty Dakhiṇāpathapati and Dakhiṇāpathesara are the synonymous formulas (both meaning laquo Lord of Dakhiṇāpatha raquo) whichoccur in inscriptions from Nāṇeghāṭ Nāsik and Girnār17 The inscriptions of Nāsikand Girnār refer respectively to Vāsiṣṭhīputra Śrī Puḷumāvi and to his immediatesuccessor Vāsiṣṭhīputra Śrī Śātakarṇi The inscription from Nāṇeghāṭ probablyrefers to Simuka the founder of the dynasty
The lordship over Dakhiṇāpatha claimed by the Sātavāhanas explains howone of them called laquo the elder Saraganos raquo (= Sātakarṇi) by the author of thePME made Kalliena not far from Nāṇeghāṭ a laquo lawful emporion raquo18 It explains also
why Ptolemy locates at Baithana (= Paithana one of the two most conspicuousemporia ofPME rsquos Dachinabades) the royal residence of Sirip(t)olemaios (= SiriPuḷumāvi)19 Nonetheless however powerful (or charlatan) were the Sātavāhanakings20 their domain could have hardly coincided with the Dakṣiṇāpatha of somepassages of thePurāṇas where it includes Pāṇḍya Kerala and Cola in the southernend of the subcontinent21 Disconnected from any political or economical realitythis expanded Dakṣiṇāpatha is the result of a generalization elaborated by the
17 vv m irashi The History and Inscriptions of the Sātavāhanas and the WesternKshatrapas Bombay (1981) n 3 l 2 p 11 (dakhiṇāpathapatiin the Nāṇeghāṭinscription of Nāganikā) ibid n 18 l 11 p 46 (dakhiṇāpathesarain the Nāsikinscription of Vāsiṣṭhīputra Śrī Puḷumāvi) dC s irCar Select Inscriptions Bearingon Indian History and Civilization Volume I From the Sixth Century BC to theSixth Century AD Calcutta (19652) n 67 l 12 p 178 (Vāsiṣṭhīputra Śrī Sātakarṇistyled as dakhiṇāpathapatiin the Rudradāman inscription of Girnār)
18 PME 52 τοπικὰ δὲ ἐμπόρια κατὰ τὸ ἑξῆς κείμενα Ἀκαβαρου Σούππαρα (edd ἀπὸ Βαρυltγάζωνgt Σούππαρα Muumlller in comm Ἀκαβαρους οὔππαρα) Καλλίενα (καὶ ΚαλλίεναFrisk) πόλις ἡ ἐπὶ τῶν Σαραγάνου τοῦ πρεσβυτέρου χρόνωνἐμπόριον (Gelenius ἐμποριω) ἔνθεσμον γενομένη (Fabricius γενόμενον) Forthe identi cation of the laquo elder Saraganos raquo with the husband of queen Nāganikācfr am s hastri The Sātavāhanas and the Western Kshatrapas a historical framework Nagpur (1998) p 55-60
19 Ptol Geogr VII 1 82 Βαίθανα βασίλειον Σιριπτολεμαίου (UKX Σίριοςπολεμίου V Σίριος πτολεμαίου R Σιροπολεμίου A Σιριπολεμαίου Z) ριζ ιη LʹIt is to be noticed that in Bhāgavatapurāṇa IX 1 41-42 the title pratiṣṭhānapati (skt Pratiṣṭhāna = Gk Paithana) is given to a king whose three sons weredakṣiṇāpatharājānaḥ laquo kings ofdakṣiṇāpatha raquo
20 s hastri op cit p 7-12 C s inoPoli laquo On the Edge of the Empire Formand Substance in the Sātavāhana Dynasty raquoin se a lCoCk tn drsquoa ltroy kd m orrison and Cm s inoPoli (eds) Empires Perspectives from Archaeologyand History Cambridge (2001) p 162-178
21 Mātsyapurāṇa 114 46-49 Vāyupurāṇa45 124-128 Mārkandeyapurāṇa57 45-49 where the names of pāṇḍyāḥ keralāḥ andcolāḥ are corrupt in puṇḍrāḥ kevalāḥ andgolāṅgūlāḥ
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
understanding of the rst (dakṣiṇā) must have appeared respectively misleadingand meaningless to the author of thePME and his Indian informers
This evolution from hodonym to choronym is not unparalleled Similar casesare those of the ancient Italian choronyms Aemilia(modern Emilia) andFlaminiawhich come out from the hodonymsvia Aemiliathe former andvia Flaminiathelatter However such transformations take place only when a new trunk road deeplyaffects the material life of the surrounding areas The commercial importanceof what must have been the original Dakṣiṇāpatha 6 ndash the main trunk road thatwent from Barygaza to Paithana Tagara and probably the coastal regions of theBay of Bengal ndash is still clearly recognizable in the text of thePME In chapter 51
the location of the two major emporia of Dachinabades Paithana and Tagara isaccompanied by speci c information concerning the number of travel days fromBarygaza and Paithana respectively laquo Among those in the same Dachinabadestwo are the most conspicuous emporia Paithana which is twenty days travel southof Barygaza and ten days east from it Tagara another very big city raquo7 Extensionsof the laquo southwards route raquo from Tagara to the coastal regions of the Bay of Bengalare not explicitly mentioned Still they are implied by the merchandise laquo of thecoastal parts raquo rst exported to Tagara and then re-exported to Barygaza8
In its original hodonymic sense the expression Dakṣiṇāpatha occurs in apassage of the Arthaśāstra laquo In the case of a land-route lsquoThe route to the Himavatis preferable to the southwards route Commodities such as elephants horsesperfumes ivory skins silver and gold are of very high valuersquo So say the teacherslsquoNorsquo says Kauṭilya lsquoleaving out commodities such as blankets skins and horses(those items) and commodities such as conch shells diamonds rubies pearls andgold are more plentiful on the southwards routersquo raquo9
6 Cfr eghP r ay Monastery and Guild Commerce under the Sātavāhanas Delhi(1986)
7 PME 51 τῶν δὲ ἐν αὐτῇ τῇ Δαχιναβάδει δύο ἐστὶν τὰ διασημότατα ἐμπόριαΠαίθανα μὲν(Schwanbeck ἐπιφαινόμενα) ἀπὸ Βαρυγάζων ἔχουσα ὁδὸν ἡμερῶνεἴκοσι πρὸς νότον ἀπὸ ltδὲgt (edd)ταύτης ὡς ἡμερῶν δέκα πρὸς ἀνατολὴν ἑτέραπόλις μεγίστη Ταγάρα (Fabricius Ταγάρα μεγίστη) Hereἐν αὐτῇ τῇ =ἐν τῇ αὐτῇ cfr h f risk Le peacuteriple de la mer Eacuterythreacutee Suivi drsquoune eacutetude sur la tradition et lalangue Goumlteborg (1927) p 65-66
This passage is noteworthy in many ways Besides showing the term inits original meaning and con rming that it actually referred to a land-route(sthalapatha) and not to as would have been theoretically possible a water-route(vāripatha) it proves that Dakṣiṇāpatha was a commercial road used by merchantswho could choose to carry on business either along the laquo southwards raquo or the Haimavata route10 It makes therefore clear that the laquo southwards route raquo was acontrastive denomination suggested by the opposition (and competition) betweenthe laquo southwards raquo and the Haimavata routes For a long time trade with Himālayahad been considered preferable to trade along the laquo southwards route raquo but whenthose lines of the Arthaśāstra were conceived the latter turned out to be more
lucrative than the former thereby disproving the eldersrsquo opinions It is thereforeself-evident that the expression Dakṣiṇāpatha was forged in an area around thenorthern end of that route by a commercial milieu whose trading activities werepoised between the Himālaya and Deccan This circumstance explains why Dakṣiṇāpatha is so to speak a one-way hodonym oriented in a direction north-south
It has been observed that the contextual reference to Deccan and Himālayaland-routes implies a standpoint laquo of a person in the midlands neither a southernernor a northerner raquo11 The text of thePME may suggest a more precise localizationIn chapter 47 the list of tribes lying behind Barygaza ndash which includes theArachosians the Gandharians and the Bactrians12 ndash re ects Barygazarsquos commercialconnections with the Himālayan region The same conclusion can be drawn fromchapter 48 where import to Barygaza via Ozene (= Ujjayinī) of nard costus andbdellium from the laquo upper places raquo is mentioned13 Just mid-way between the
10 For archaeological evidence pointing to connections between Deccan and Gandharacfr P BranCaCCio laquo Close Encounters Multicultural Systems in Ancient India raquoin d s rinivasan (ed)On the Cusp of an Era Art in the Pre-Kuṣāṇa World Leiden(2007) p 385ndash398 e ad The Buddhist Cave at Aurangabad Transformationin Art and Religion Leiden (2010) p 67-68 69-70 e ad laquo Terracottas fromWestern Deccan An Exploration of Sources and Transmission of Models in theEarly Historic Period raquoin Ph Granoff (ed) A Pantheon Rediscovered ChangingPerceptions of Early Historic India forthcoming
11 rP k anGle TheKauṭilīya Arthaśāstra Part II An English Translation with Criticaland Explanatory Notes Bombay (19722) p 360 nt 22
Himālaya and Deccan regions the Barygaza-Ozene region was perhaps the onlyIndian area where the alternative between a Haimavata and a Dakṣiṇāpatha land-route made perfect sense
Incidentally I would like to point out another example in which a comparisonwith the text of the Arthaśāstra may improve our understanding of the text ofthe PME Again it relates to the Dachinabades land-route In chapter 51 itemsfrom Paithana and Tagara are brought to Barygazaπορείαις ἁμαξῶν καὶ ἀνοδίαιςμεγίσταις which Casson translates as laquo by conveyance in wagons over very greatroadless stretches raquo14 Muumlllerrsquos Latin rendition is similar laquo per loca maximeinvia plaustris merces devehuntur raquo15 Both these translations assume that such
long journey was made on stretches of land that were simultaneously roadless andaccessible by wagons which is contradictoryEvoked by the author of the Arthaśāstra in the very same chapter containing
the comparison between Dakṣiṇāpatha and Haimavata land-route is thedistinction betweencakrapatha laquo the wheel-track raquo pādapatha laquo the foot-path raquoand kharoṣṭrapatha laquo the ass- or camel-road raquo laquo [hellip] as between a wheel-trackand a foot-path the wheel-track is preferable as it enables undertakings on a largescale Alternatively path for donkeys and camels in conformity with place and time(is to be preferred) raquo16 This line suggests that a long land-route may not be entirelysuitable to wheeled vehicles and may therefore require changes in the mode oftransport such as shifting between wagons and donkeys It is consequently probablethat atPME51 theκαί which connectsπορείαις ἁμαξῶν andἀνοδίαις μεγίσταις is disjunctive from Paithana and Tagara the items are brought laquo with journeysof wagons and long roadless tracts raquo In other words laquo journeys of wagons raquo(πορεῖαι ἁμαξῶν) and laquo roadless tracts raquo (ἀνοδίαι μέγισται) refer to two distinctmodalities of transport each adopted as needed along different legs of the voyageThe long route from Tagara to Paithana and Barygaza was to be travelled withwagons on laquo wheel-track raquo (cakrapatha) only in part for long stretches lsquothe ass- orcamelroadsrsquo (kharoṣṭrapatha) were the only option available
While in the Arthaśāstra the term Dakṣiṇāpatha still retains its originalmeaning of laquo southwards route raquo its earliest inscriptional occurrences already show
Σκυθίας ὅ τε κόστος καὶ ἡ βδέλλα The same items are available at Barbarikon onthe mouth of the Indus PME39
14 l C asson ThePeriplus Maris Erythraei Text with Introduction Translation andCommentary Princeton (1989) p 83
15 C m uumlller Geographi Graeci Minores I Paris (1855) p 294 A more t translationis given byWh s Choff The Periplus of Erythraean Sea New York LondonBombay and Calcutta (1912) p 43 laquo [hellip] brought down to Barygaza from theseplaces by wagons and through great tracts without roads raquo
16 Arthaśāstra VII 12 27 tatrāpi cakrapādapathayoścakrapatho vipulārambhatvātcchreyāndeśakālasambhāvano vā kharoṣṭrapathaḥ
its evolution into a choronym Dakṣiṇāpatha was claimed by Sātavāhana rulers as thegeographical frame of their sovereignty Dakhiṇāpathapati and Dakhiṇāpathesara are the synonymous formulas (both meaning laquo Lord of Dakhiṇāpatha raquo) whichoccur in inscriptions from Nāṇeghāṭ Nāsik and Girnār17 The inscriptions of Nāsikand Girnār refer respectively to Vāsiṣṭhīputra Śrī Puḷumāvi and to his immediatesuccessor Vāsiṣṭhīputra Śrī Śātakarṇi The inscription from Nāṇeghāṭ probablyrefers to Simuka the founder of the dynasty
The lordship over Dakhiṇāpatha claimed by the Sātavāhanas explains howone of them called laquo the elder Saraganos raquo (= Sātakarṇi) by the author of thePME made Kalliena not far from Nāṇeghāṭ a laquo lawful emporion raquo18 It explains also
why Ptolemy locates at Baithana (= Paithana one of the two most conspicuousemporia ofPME rsquos Dachinabades) the royal residence of Sirip(t)olemaios (= SiriPuḷumāvi)19 Nonetheless however powerful (or charlatan) were the Sātavāhanakings20 their domain could have hardly coincided with the Dakṣiṇāpatha of somepassages of thePurāṇas where it includes Pāṇḍya Kerala and Cola in the southernend of the subcontinent21 Disconnected from any political or economical realitythis expanded Dakṣiṇāpatha is the result of a generalization elaborated by the
17 vv m irashi The History and Inscriptions of the Sātavāhanas and the WesternKshatrapas Bombay (1981) n 3 l 2 p 11 (dakhiṇāpathapatiin the Nāṇeghāṭinscription of Nāganikā) ibid n 18 l 11 p 46 (dakhiṇāpathesarain the Nāsikinscription of Vāsiṣṭhīputra Śrī Puḷumāvi) dC s irCar Select Inscriptions Bearingon Indian History and Civilization Volume I From the Sixth Century BC to theSixth Century AD Calcutta (19652) n 67 l 12 p 178 (Vāsiṣṭhīputra Śrī Sātakarṇistyled as dakhiṇāpathapatiin the Rudradāman inscription of Girnār)
18 PME 52 τοπικὰ δὲ ἐμπόρια κατὰ τὸ ἑξῆς κείμενα Ἀκαβαρου Σούππαρα (edd ἀπὸ Βαρυltγάζωνgt Σούππαρα Muumlller in comm Ἀκαβαρους οὔππαρα) Καλλίενα (καὶ ΚαλλίεναFrisk) πόλις ἡ ἐπὶ τῶν Σαραγάνου τοῦ πρεσβυτέρου χρόνωνἐμπόριον (Gelenius ἐμποριω) ἔνθεσμον γενομένη (Fabricius γενόμενον) Forthe identi cation of the laquo elder Saraganos raquo with the husband of queen Nāganikācfr am s hastri The Sātavāhanas and the Western Kshatrapas a historical framework Nagpur (1998) p 55-60
19 Ptol Geogr VII 1 82 Βαίθανα βασίλειον Σιριπτολεμαίου (UKX Σίριοςπολεμίου V Σίριος πτολεμαίου R Σιροπολεμίου A Σιριπολεμαίου Z) ριζ ιη LʹIt is to be noticed that in Bhāgavatapurāṇa IX 1 41-42 the title pratiṣṭhānapati (skt Pratiṣṭhāna = Gk Paithana) is given to a king whose three sons weredakṣiṇāpatharājānaḥ laquo kings ofdakṣiṇāpatha raquo
20 s hastri op cit p 7-12 C s inoPoli laquo On the Edge of the Empire Formand Substance in the Sātavāhana Dynasty raquoin se a lCoCk tn drsquoa ltroy kd m orrison and Cm s inoPoli (eds) Empires Perspectives from Archaeologyand History Cambridge (2001) p 162-178
21 Mātsyapurāṇa 114 46-49 Vāyupurāṇa45 124-128 Mārkandeyapurāṇa57 45-49 where the names of pāṇḍyāḥ keralāḥ andcolāḥ are corrupt in puṇḍrāḥ kevalāḥ andgolāṅgūlāḥ
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
This passage is noteworthy in many ways Besides showing the term inits original meaning and con rming that it actually referred to a land-route(sthalapatha) and not to as would have been theoretically possible a water-route(vāripatha) it proves that Dakṣiṇāpatha was a commercial road used by merchantswho could choose to carry on business either along the laquo southwards raquo or the Haimavata route10 It makes therefore clear that the laquo southwards route raquo was acontrastive denomination suggested by the opposition (and competition) betweenthe laquo southwards raquo and the Haimavata routes For a long time trade with Himālayahad been considered preferable to trade along the laquo southwards route raquo but whenthose lines of the Arthaśāstra were conceived the latter turned out to be more
lucrative than the former thereby disproving the eldersrsquo opinions It is thereforeself-evident that the expression Dakṣiṇāpatha was forged in an area around thenorthern end of that route by a commercial milieu whose trading activities werepoised between the Himālaya and Deccan This circumstance explains why Dakṣiṇāpatha is so to speak a one-way hodonym oriented in a direction north-south
It has been observed that the contextual reference to Deccan and Himālayaland-routes implies a standpoint laquo of a person in the midlands neither a southernernor a northerner raquo11 The text of thePME may suggest a more precise localizationIn chapter 47 the list of tribes lying behind Barygaza ndash which includes theArachosians the Gandharians and the Bactrians12 ndash re ects Barygazarsquos commercialconnections with the Himālayan region The same conclusion can be drawn fromchapter 48 where import to Barygaza via Ozene (= Ujjayinī) of nard costus andbdellium from the laquo upper places raquo is mentioned13 Just mid-way between the
10 For archaeological evidence pointing to connections between Deccan and Gandharacfr P BranCaCCio laquo Close Encounters Multicultural Systems in Ancient India raquoin d s rinivasan (ed)On the Cusp of an Era Art in the Pre-Kuṣāṇa World Leiden(2007) p 385ndash398 e ad The Buddhist Cave at Aurangabad Transformationin Art and Religion Leiden (2010) p 67-68 69-70 e ad laquo Terracottas fromWestern Deccan An Exploration of Sources and Transmission of Models in theEarly Historic Period raquoin Ph Granoff (ed) A Pantheon Rediscovered ChangingPerceptions of Early Historic India forthcoming
11 rP k anGle TheKauṭilīya Arthaśāstra Part II An English Translation with Criticaland Explanatory Notes Bombay (19722) p 360 nt 22
Himālaya and Deccan regions the Barygaza-Ozene region was perhaps the onlyIndian area where the alternative between a Haimavata and a Dakṣiṇāpatha land-route made perfect sense
Incidentally I would like to point out another example in which a comparisonwith the text of the Arthaśāstra may improve our understanding of the text ofthe PME Again it relates to the Dachinabades land-route In chapter 51 itemsfrom Paithana and Tagara are brought to Barygazaπορείαις ἁμαξῶν καὶ ἀνοδίαιςμεγίσταις which Casson translates as laquo by conveyance in wagons over very greatroadless stretches raquo14 Muumlllerrsquos Latin rendition is similar laquo per loca maximeinvia plaustris merces devehuntur raquo15 Both these translations assume that such
long journey was made on stretches of land that were simultaneously roadless andaccessible by wagons which is contradictoryEvoked by the author of the Arthaśāstra in the very same chapter containing
the comparison between Dakṣiṇāpatha and Haimavata land-route is thedistinction betweencakrapatha laquo the wheel-track raquo pādapatha laquo the foot-path raquoand kharoṣṭrapatha laquo the ass- or camel-road raquo laquo [hellip] as between a wheel-trackand a foot-path the wheel-track is preferable as it enables undertakings on a largescale Alternatively path for donkeys and camels in conformity with place and time(is to be preferred) raquo16 This line suggests that a long land-route may not be entirelysuitable to wheeled vehicles and may therefore require changes in the mode oftransport such as shifting between wagons and donkeys It is consequently probablethat atPME51 theκαί which connectsπορείαις ἁμαξῶν andἀνοδίαις μεγίσταις is disjunctive from Paithana and Tagara the items are brought laquo with journeysof wagons and long roadless tracts raquo In other words laquo journeys of wagons raquo(πορεῖαι ἁμαξῶν) and laquo roadless tracts raquo (ἀνοδίαι μέγισται) refer to two distinctmodalities of transport each adopted as needed along different legs of the voyageThe long route from Tagara to Paithana and Barygaza was to be travelled withwagons on laquo wheel-track raquo (cakrapatha) only in part for long stretches lsquothe ass- orcamelroadsrsquo (kharoṣṭrapatha) were the only option available
While in the Arthaśāstra the term Dakṣiṇāpatha still retains its originalmeaning of laquo southwards route raquo its earliest inscriptional occurrences already show
Σκυθίας ὅ τε κόστος καὶ ἡ βδέλλα The same items are available at Barbarikon onthe mouth of the Indus PME39
14 l C asson ThePeriplus Maris Erythraei Text with Introduction Translation andCommentary Princeton (1989) p 83
15 C m uumlller Geographi Graeci Minores I Paris (1855) p 294 A more t translationis given byWh s Choff The Periplus of Erythraean Sea New York LondonBombay and Calcutta (1912) p 43 laquo [hellip] brought down to Barygaza from theseplaces by wagons and through great tracts without roads raquo
16 Arthaśāstra VII 12 27 tatrāpi cakrapādapathayoścakrapatho vipulārambhatvātcchreyāndeśakālasambhāvano vā kharoṣṭrapathaḥ
its evolution into a choronym Dakṣiṇāpatha was claimed by Sātavāhana rulers as thegeographical frame of their sovereignty Dakhiṇāpathapati and Dakhiṇāpathesara are the synonymous formulas (both meaning laquo Lord of Dakhiṇāpatha raquo) whichoccur in inscriptions from Nāṇeghāṭ Nāsik and Girnār17 The inscriptions of Nāsikand Girnār refer respectively to Vāsiṣṭhīputra Śrī Puḷumāvi and to his immediatesuccessor Vāsiṣṭhīputra Śrī Śātakarṇi The inscription from Nāṇeghāṭ probablyrefers to Simuka the founder of the dynasty
The lordship over Dakhiṇāpatha claimed by the Sātavāhanas explains howone of them called laquo the elder Saraganos raquo (= Sātakarṇi) by the author of thePME made Kalliena not far from Nāṇeghāṭ a laquo lawful emporion raquo18 It explains also
why Ptolemy locates at Baithana (= Paithana one of the two most conspicuousemporia ofPME rsquos Dachinabades) the royal residence of Sirip(t)olemaios (= SiriPuḷumāvi)19 Nonetheless however powerful (or charlatan) were the Sātavāhanakings20 their domain could have hardly coincided with the Dakṣiṇāpatha of somepassages of thePurāṇas where it includes Pāṇḍya Kerala and Cola in the southernend of the subcontinent21 Disconnected from any political or economical realitythis expanded Dakṣiṇāpatha is the result of a generalization elaborated by the
17 vv m irashi The History and Inscriptions of the Sātavāhanas and the WesternKshatrapas Bombay (1981) n 3 l 2 p 11 (dakhiṇāpathapatiin the Nāṇeghāṭinscription of Nāganikā) ibid n 18 l 11 p 46 (dakhiṇāpathesarain the Nāsikinscription of Vāsiṣṭhīputra Śrī Puḷumāvi) dC s irCar Select Inscriptions Bearingon Indian History and Civilization Volume I From the Sixth Century BC to theSixth Century AD Calcutta (19652) n 67 l 12 p 178 (Vāsiṣṭhīputra Śrī Sātakarṇistyled as dakhiṇāpathapatiin the Rudradāman inscription of Girnār)
18 PME 52 τοπικὰ δὲ ἐμπόρια κατὰ τὸ ἑξῆς κείμενα Ἀκαβαρου Σούππαρα (edd ἀπὸ Βαρυltγάζωνgt Σούππαρα Muumlller in comm Ἀκαβαρους οὔππαρα) Καλλίενα (καὶ ΚαλλίεναFrisk) πόλις ἡ ἐπὶ τῶν Σαραγάνου τοῦ πρεσβυτέρου χρόνωνἐμπόριον (Gelenius ἐμποριω) ἔνθεσμον γενομένη (Fabricius γενόμενον) Forthe identi cation of the laquo elder Saraganos raquo with the husband of queen Nāganikācfr am s hastri The Sātavāhanas and the Western Kshatrapas a historical framework Nagpur (1998) p 55-60
19 Ptol Geogr VII 1 82 Βαίθανα βασίλειον Σιριπτολεμαίου (UKX Σίριοςπολεμίου V Σίριος πτολεμαίου R Σιροπολεμίου A Σιριπολεμαίου Z) ριζ ιη LʹIt is to be noticed that in Bhāgavatapurāṇa IX 1 41-42 the title pratiṣṭhānapati (skt Pratiṣṭhāna = Gk Paithana) is given to a king whose three sons weredakṣiṇāpatharājānaḥ laquo kings ofdakṣiṇāpatha raquo
20 s hastri op cit p 7-12 C s inoPoli laquo On the Edge of the Empire Formand Substance in the Sātavāhana Dynasty raquoin se a lCoCk tn drsquoa ltroy kd m orrison and Cm s inoPoli (eds) Empires Perspectives from Archaeologyand History Cambridge (2001) p 162-178
21 Mātsyapurāṇa 114 46-49 Vāyupurāṇa45 124-128 Mārkandeyapurāṇa57 45-49 where the names of pāṇḍyāḥ keralāḥ andcolāḥ are corrupt in puṇḍrāḥ kevalāḥ andgolāṅgūlāḥ
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
Himālaya and Deccan regions the Barygaza-Ozene region was perhaps the onlyIndian area where the alternative between a Haimavata and a Dakṣiṇāpatha land-route made perfect sense
Incidentally I would like to point out another example in which a comparisonwith the text of the Arthaśāstra may improve our understanding of the text ofthe PME Again it relates to the Dachinabades land-route In chapter 51 itemsfrom Paithana and Tagara are brought to Barygazaπορείαις ἁμαξῶν καὶ ἀνοδίαιςμεγίσταις which Casson translates as laquo by conveyance in wagons over very greatroadless stretches raquo14 Muumlllerrsquos Latin rendition is similar laquo per loca maximeinvia plaustris merces devehuntur raquo15 Both these translations assume that such
long journey was made on stretches of land that were simultaneously roadless andaccessible by wagons which is contradictoryEvoked by the author of the Arthaśāstra in the very same chapter containing
the comparison between Dakṣiṇāpatha and Haimavata land-route is thedistinction betweencakrapatha laquo the wheel-track raquo pādapatha laquo the foot-path raquoand kharoṣṭrapatha laquo the ass- or camel-road raquo laquo [hellip] as between a wheel-trackand a foot-path the wheel-track is preferable as it enables undertakings on a largescale Alternatively path for donkeys and camels in conformity with place and time(is to be preferred) raquo16 This line suggests that a long land-route may not be entirelysuitable to wheeled vehicles and may therefore require changes in the mode oftransport such as shifting between wagons and donkeys It is consequently probablethat atPME51 theκαί which connectsπορείαις ἁμαξῶν andἀνοδίαις μεγίσταις is disjunctive from Paithana and Tagara the items are brought laquo with journeysof wagons and long roadless tracts raquo In other words laquo journeys of wagons raquo(πορεῖαι ἁμαξῶν) and laquo roadless tracts raquo (ἀνοδίαι μέγισται) refer to two distinctmodalities of transport each adopted as needed along different legs of the voyageThe long route from Tagara to Paithana and Barygaza was to be travelled withwagons on laquo wheel-track raquo (cakrapatha) only in part for long stretches lsquothe ass- orcamelroadsrsquo (kharoṣṭrapatha) were the only option available
While in the Arthaśāstra the term Dakṣiṇāpatha still retains its originalmeaning of laquo southwards route raquo its earliest inscriptional occurrences already show
Σκυθίας ὅ τε κόστος καὶ ἡ βδέλλα The same items are available at Barbarikon onthe mouth of the Indus PME39
14 l C asson ThePeriplus Maris Erythraei Text with Introduction Translation andCommentary Princeton (1989) p 83
15 C m uumlller Geographi Graeci Minores I Paris (1855) p 294 A more t translationis given byWh s Choff The Periplus of Erythraean Sea New York LondonBombay and Calcutta (1912) p 43 laquo [hellip] brought down to Barygaza from theseplaces by wagons and through great tracts without roads raquo
16 Arthaśāstra VII 12 27 tatrāpi cakrapādapathayoścakrapatho vipulārambhatvātcchreyāndeśakālasambhāvano vā kharoṣṭrapathaḥ
its evolution into a choronym Dakṣiṇāpatha was claimed by Sātavāhana rulers as thegeographical frame of their sovereignty Dakhiṇāpathapati and Dakhiṇāpathesara are the synonymous formulas (both meaning laquo Lord of Dakhiṇāpatha raquo) whichoccur in inscriptions from Nāṇeghāṭ Nāsik and Girnār17 The inscriptions of Nāsikand Girnār refer respectively to Vāsiṣṭhīputra Śrī Puḷumāvi and to his immediatesuccessor Vāsiṣṭhīputra Śrī Śātakarṇi The inscription from Nāṇeghāṭ probablyrefers to Simuka the founder of the dynasty
The lordship over Dakhiṇāpatha claimed by the Sātavāhanas explains howone of them called laquo the elder Saraganos raquo (= Sātakarṇi) by the author of thePME made Kalliena not far from Nāṇeghāṭ a laquo lawful emporion raquo18 It explains also
why Ptolemy locates at Baithana (= Paithana one of the two most conspicuousemporia ofPME rsquos Dachinabades) the royal residence of Sirip(t)olemaios (= SiriPuḷumāvi)19 Nonetheless however powerful (or charlatan) were the Sātavāhanakings20 their domain could have hardly coincided with the Dakṣiṇāpatha of somepassages of thePurāṇas where it includes Pāṇḍya Kerala and Cola in the southernend of the subcontinent21 Disconnected from any political or economical realitythis expanded Dakṣiṇāpatha is the result of a generalization elaborated by the
17 vv m irashi The History and Inscriptions of the Sātavāhanas and the WesternKshatrapas Bombay (1981) n 3 l 2 p 11 (dakhiṇāpathapatiin the Nāṇeghāṭinscription of Nāganikā) ibid n 18 l 11 p 46 (dakhiṇāpathesarain the Nāsikinscription of Vāsiṣṭhīputra Śrī Puḷumāvi) dC s irCar Select Inscriptions Bearingon Indian History and Civilization Volume I From the Sixth Century BC to theSixth Century AD Calcutta (19652) n 67 l 12 p 178 (Vāsiṣṭhīputra Śrī Sātakarṇistyled as dakhiṇāpathapatiin the Rudradāman inscription of Girnār)
18 PME 52 τοπικὰ δὲ ἐμπόρια κατὰ τὸ ἑξῆς κείμενα Ἀκαβαρου Σούππαρα (edd ἀπὸ Βαρυltγάζωνgt Σούππαρα Muumlller in comm Ἀκαβαρους οὔππαρα) Καλλίενα (καὶ ΚαλλίεναFrisk) πόλις ἡ ἐπὶ τῶν Σαραγάνου τοῦ πρεσβυτέρου χρόνωνἐμπόριον (Gelenius ἐμποριω) ἔνθεσμον γενομένη (Fabricius γενόμενον) Forthe identi cation of the laquo elder Saraganos raquo with the husband of queen Nāganikācfr am s hastri The Sātavāhanas and the Western Kshatrapas a historical framework Nagpur (1998) p 55-60
19 Ptol Geogr VII 1 82 Βαίθανα βασίλειον Σιριπτολεμαίου (UKX Σίριοςπολεμίου V Σίριος πτολεμαίου R Σιροπολεμίου A Σιριπολεμαίου Z) ριζ ιη LʹIt is to be noticed that in Bhāgavatapurāṇa IX 1 41-42 the title pratiṣṭhānapati (skt Pratiṣṭhāna = Gk Paithana) is given to a king whose three sons weredakṣiṇāpatharājānaḥ laquo kings ofdakṣiṇāpatha raquo
20 s hastri op cit p 7-12 C s inoPoli laquo On the Edge of the Empire Formand Substance in the Sātavāhana Dynasty raquoin se a lCoCk tn drsquoa ltroy kd m orrison and Cm s inoPoli (eds) Empires Perspectives from Archaeologyand History Cambridge (2001) p 162-178
21 Mātsyapurāṇa 114 46-49 Vāyupurāṇa45 124-128 Mārkandeyapurāṇa57 45-49 where the names of pāṇḍyāḥ keralāḥ andcolāḥ are corrupt in puṇḍrāḥ kevalāḥ andgolāṅgūlāḥ
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
its evolution into a choronym Dakṣiṇāpatha was claimed by Sātavāhana rulers as thegeographical frame of their sovereignty Dakhiṇāpathapati and Dakhiṇāpathesara are the synonymous formulas (both meaning laquo Lord of Dakhiṇāpatha raquo) whichoccur in inscriptions from Nāṇeghāṭ Nāsik and Girnār17 The inscriptions of Nāsikand Girnār refer respectively to Vāsiṣṭhīputra Śrī Puḷumāvi and to his immediatesuccessor Vāsiṣṭhīputra Śrī Śātakarṇi The inscription from Nāṇeghāṭ probablyrefers to Simuka the founder of the dynasty
The lordship over Dakhiṇāpatha claimed by the Sātavāhanas explains howone of them called laquo the elder Saraganos raquo (= Sātakarṇi) by the author of thePME made Kalliena not far from Nāṇeghāṭ a laquo lawful emporion raquo18 It explains also
why Ptolemy locates at Baithana (= Paithana one of the two most conspicuousemporia ofPME rsquos Dachinabades) the royal residence of Sirip(t)olemaios (= SiriPuḷumāvi)19 Nonetheless however powerful (or charlatan) were the Sātavāhanakings20 their domain could have hardly coincided with the Dakṣiṇāpatha of somepassages of thePurāṇas where it includes Pāṇḍya Kerala and Cola in the southernend of the subcontinent21 Disconnected from any political or economical realitythis expanded Dakṣiṇāpatha is the result of a generalization elaborated by the
17 vv m irashi The History and Inscriptions of the Sātavāhanas and the WesternKshatrapas Bombay (1981) n 3 l 2 p 11 (dakhiṇāpathapatiin the Nāṇeghāṭinscription of Nāganikā) ibid n 18 l 11 p 46 (dakhiṇāpathesarain the Nāsikinscription of Vāsiṣṭhīputra Śrī Puḷumāvi) dC s irCar Select Inscriptions Bearingon Indian History and Civilization Volume I From the Sixth Century BC to theSixth Century AD Calcutta (19652) n 67 l 12 p 178 (Vāsiṣṭhīputra Śrī Sātakarṇistyled as dakhiṇāpathapatiin the Rudradāman inscription of Girnār)
18 PME 52 τοπικὰ δὲ ἐμπόρια κατὰ τὸ ἑξῆς κείμενα Ἀκαβαρου Σούππαρα (edd ἀπὸ Βαρυltγάζωνgt Σούππαρα Muumlller in comm Ἀκαβαρους οὔππαρα) Καλλίενα (καὶ ΚαλλίεναFrisk) πόλις ἡ ἐπὶ τῶν Σαραγάνου τοῦ πρεσβυτέρου χρόνωνἐμπόριον (Gelenius ἐμποριω) ἔνθεσμον γενομένη (Fabricius γενόμενον) Forthe identi cation of the laquo elder Saraganos raquo with the husband of queen Nāganikācfr am s hastri The Sātavāhanas and the Western Kshatrapas a historical framework Nagpur (1998) p 55-60
19 Ptol Geogr VII 1 82 Βαίθανα βασίλειον Σιριπτολεμαίου (UKX Σίριοςπολεμίου V Σίριος πτολεμαίου R Σιροπολεμίου A Σιριπολεμαίου Z) ριζ ιη LʹIt is to be noticed that in Bhāgavatapurāṇa IX 1 41-42 the title pratiṣṭhānapati (skt Pratiṣṭhāna = Gk Paithana) is given to a king whose three sons weredakṣiṇāpatharājānaḥ laquo kings ofdakṣiṇāpatha raquo
20 s hastri op cit p 7-12 C s inoPoli laquo On the Edge of the Empire Formand Substance in the Sātavāhana Dynasty raquoin se a lCoCk tn drsquoa ltroy kd m orrison and Cm s inoPoli (eds) Empires Perspectives from Archaeologyand History Cambridge (2001) p 162-178
21 Mātsyapurāṇa 114 46-49 Vāyupurāṇa45 124-128 Mārkandeyapurāṇa57 45-49 where the names of pāṇḍyāḥ keralāḥ andcolāḥ are corrupt in puṇḍrāḥ kevalāḥ andgolāṅgūlāḥ
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
following Castoriuscosmographus includes Maziris namely Muziris in the IndiaDimirica27
Both the forms Limyrike and DamiricaDymirica etc must stem from theethnonymTamil and clearly echo the sense of linguistic otherness of DravidianIndia TheTolkāppiyam the earliest extant literary document in Tamil projects thelinguistic identity of the Tamils in a space de nedTamiḻkūṟu whose landmarksare the Vēṅkaṭa mountain to the north Cape Comorin to the south and the seaboth to east and west28 The choice of the Vēṅkaṭa hill as its northern landmarkshows ndash quite unsurprisingly ndash that the Tamils themselves laquo felt raquo the axis of theDravidian India east of the Western Ghats a perception which is consistent with
the fact that the residences of the Chera Pandya and Chola kings too are east ofthe Western Ghats and also that most of the Tamil-Brahmi inscriptions come fromwhat is now Tamil Nadu
Land connections across the Western Ghats between the centers of Tamilpolitical power and the maritime emporia of the Malabar Coast were rather poorat the age of thePME Its author so well informed about the distance betweenBarygaza Paithana and Tagara knows only that both the Chera and the Pandyakings dwell somewhere in the interior29 Merchants from Egypt are not said tobring any special items for those rulers30 Unsurprisingly the perception of the Limyrike by the author of thePME is quite different from that of theTamiḻkūṟu bythe author of theTolkāppiyam it begins with Naoura and Tyndis the northernmostemporia of the Kerala coast31 and does not include the emporia Kamara Podoukeand Sopatma on the Coromandel coast32 We are not however informed exactly
27 Geogr Rav p 15 64 Schn
28 Tolkāppiyam Pāyiram vaṭa vēṅkaṭa teṉ kumari āyiṭait tamiḻkūṟum nal ulakattuThe same idea occurs also in several other later texts cfr kk P illai A Social History of the Tamils Madras (1975) p 14-15 nt 3
30 The list of the exports atPME 56 includes no special items for the Chera or Pandyakings whilePME 6 mentions special items for the king Zoskales PME 24 specialitems for the king Charibael and the tyrannos Cholaibos PME 28 special items forthe king Eleazos PME 49 special items for the king Manbanes With the possibleexception of Zoskales none of these rulers dwells in a coastal emporium Plinyknows that the Pandya king dwellslonge ab emporio in mediterraneo distanteoppido quod vocatur Modura (Plin NH VI 103) We have to wait until Ptolemy toget a vision of the inland regions of the southern end of India
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
how far the Limyrike stretched south of Naoura or how deeply inland from thecoastal line it penetrated
As for the Limyrikersquos north-south extension some information might havebeen provided by chapter 58 of thePME if the text didnrsquot come to us with a lacuna ἀπὸ δὲ Βακαρὴ (Stuck ἀπrsquoἐλαβακαρὴ) τὸ λεγόμενον Πυρρὸν ὄρος ltκαὶgt (Muumlllerin comm)ἄλλη παρήκεltιgt (edd)χώρα τη κης ἡ Παραλία λεγομένη πρὸς αὐτὸντὸν νότον (laquo After Bakare there is the mountain called Pyrrhon (= lsquoRedrsquo) andanother region of the extends calledParalia (= lsquoSeaboardrsquo) directly to theSouth raquo) ApparentlyParalia is a district of a bigger region whose name endingwith the suf x -ικη became almost completely illegible in the exemplar In the
Heidelberg manuscript between aτη which must be the beginning of the genitivefeminine form of the article (τῆς) and theκης which must be the ending of a majorchoronym includingParalia there is a blank space long enough for approximatelythirteen letters The adjectiveἄλλη shows that the missing choronym was alreadymentioned in the text
Restorations of the defective text have been proposed by Muumlller33 whosuggested eitherτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς or τῆltς καλουμένης Κοττοναριgtκῆς and by Frisk34 who suggested eitherτῆltς πρὸς νότον Ἰνδιgtκῆς or τῆltς Ἰνδικῆςἐπιμήgtκης All these restitutions entail the idea that the coast immediately south ofBakare was already outside the Limyrike an idea which agrees exactly with whatPtolemy implies35 Nonetheless are we justi ed in assuming thatPME rsquos Limyrike coincided with Ptolemyrsquos Limyrike
Passages such asPME 47 where the Limyrike is juxtaposed with laquo thesouthern parts of India raquo (τὰ νότια τῆς Ἰνδικῆς = Dakṣiṇāpatha)36 or PME 56 where the Lakshadweep islands are identi ed as those laquo lying off the same Limyrike raquo37 strongly suggest that in the perception of the author ofPME Limyrike was a choronym of considerable extent Moreover a reference to Ἰνδική withoutany further speci cation (ἄλλη παρήκεltιgt χώρα τῆltς Ἰνδικῆς ἐπιμήgtκης) herewould be too generic The context I believe requires a more speci c and limited
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
choronym On the other hand periphrases such asτῆltς ὑπὸ Πανδίονι Ἰνδιgtκῆς orτῆltς πρὸς νότον Ἰνδιgtκῆς seem either alien from theusus scribendiof the author38 or awkward39 Finally if the missing choronym wereKottanarike 40 andParalia was laquo another region raquo of theKottanarike we should conclude that in the mind ofthe author bothParalia and Limyrike right north of the former were just districts oftheKottanarike Such a view however is hardly compatible withPME 47 where Limyrike ndash notKottanarike ndash is contrasted with laquo the southern parts of India raquo Insum it seems to me much more probable that in thePME rsquos mind Limyrike wasa hierarchically superior choronym and thatKottanarike and Paralia were justdistricts of it
If this conclusion is right and the blank space of the Heidelberg manuscriptreplicates with some precision the length of the lacuna we can propose the followingrestoration [hellip]καὶ ἄλλη παρήκεltιgt (edd) χώρα τῆltς αὐτῆς 41 Λιμυριgtκῆς ἡΠαραλία λεγομένη πρὸς αὐτὸν τὸν νότον laquo [hellip] and another region of the same Limyrike extends calledParalia (= lsquoSeaboardrsquo) directly to the South raquo
Admittedly this restoration of the corrupt text compels us to extendPME rsquos Limyrike beyond the southern limits given to it by Ptolemy perhaps up to Kolchoiafter which is Aigialos (= Strand) laquo having an inland region named Argalou raquo42However it must be emphasized that from the time of thePMEto that of Ptolemythe political geography of the Dravidian India underwent considerable changes
38 If he really wanted to refer to another land of the Pandya kingdom he would haveprobably writtenἄλλη ὑπὸ τὸν Πανδίονα παρήκει χώρα or maybe betterἄλληπαρήκει χώρα τῆς Πανδίονος βασιλείας cfr PME 54 βασιλείας δέ ἐστιν ἑτέραςτῆς Πανδίονος In thePME land controlled by the kings is quite often referredto as theirβασιλεία βασιλεία Χαριβαήλ (PME26)βασιλεία Ἐλεάζου (PME27)Μανβάνου βασιλεία (PME41)βασιλεία Κηπροβότρου (PME54)
39 The result would beἄλλη παρήκεltιgt χώρα τῆltς πρὸς νότον Ἰνδιgtκῆς ἡ Παραλίαλεγομένη πρὸς αὐτὸν τὸν νότον
40 Where the black pepper sold at Muziris and Nelkynda grows or comes from cfrPME 56 φέρεται δὲ πέπερι μονογενῶς ἐν ἑνὶ τόπῳ τούτων τῶν ἐμπορίων (Muumlller τούτῳ τῷ ἐμπορίῳ) γεννώμενον πολύ λεγομένῃ Κοττοναρικῇ (Muumlller λεγομένηΚοττοναρική) Plin NH VI 105 regio autem ex qua piper monoxylis lintribus Becaren convehunt vocatur Cottonara Ptolemy (Geogr VII 1 9) mentions aΚοττιάρα μητρόπολις in the land of the Aioi
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
in the north Naoura had fallen in the hands of the pirates43 and in the centerBakare was no longer controlled by the Pandya king (whose domain even in theinterior was not conterminous with the Limyrike 44) The territorial losses of thePandya kings may have affected the commercial practice In the mid I century ADthe author of thePME mentions generically the Limyrike as the destination of thewestern merchants45 and refers to both Muziris and Nelkynda as the then (most)dynamic emporia46 Around the same years Pliny even suggests to go rather thanto Muziris to the laquo more serviceable port belonging to the Neacyndes tribe calledBecare raquo47 On the contrary in the II cent AD lsquoMuziris papyrusrsquo the standardloan contracts for voyages to south India are referred to as laquo the loan contracts for
Muziris raquo48
This concentration of the Roman trade in Muziris in the II cent ADmay have at that time suggested a narrower notion of Limyrike basically restrictedto the portion of the Malabar coast from Tyndis to Bakare and the inland regionsof the Chera domain The notion of a geographically wider Limyrike howeversurvived in the western geography surfacing again in the DamiricaDymirica etcof theGeographus Ravennas and the Peutinger Table
The linguistic distinctiveness of southern India was apparent to bothDravidian and non-Dravidian speakers in ancient India In non-Dravidian India are ection of this perception is found in the Hāthīgūmphā inscription of Khāravelawhere atramira-deṣa-saṁghātam(or tamara-deha-samghātam or tamira-daha-saṁghāta 49) a laquo confederacy of the T(r)amira countries raquo appears The choronym Limyrike was inspired by the same perception and is likely to have penetrated intothe Greek geographic literature through the mediation of the very same merchantsfrom Barygaza who had popularized the choronym Dachinabades As a matterof fact besides the Himālaya and Deccan regions Barygaza merchants also had
43 PtolGeogr VII 1 7 But it is still anἐμπόριον
44 Along the coast the land of Pandion begins only with the gulf Argarikos after CapeComorin and Kolchoi (PtolGeogr VII 1 11) In the interior between the Limyrikewhich includes the residence of the Chera king and the laquo land of Pandion raquo areagain the Aioi (PtolGeogr VII 1 86-89)
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
commercial relationships with Muziris50 the wording ofPME 47καταλιπὼν τήντε Λιμυρικὴν καὶ τὰ νότια τῆς Ἰνδικῆς laquo leaving aside both the Limyrike and thesouthern parts of India raquo re ects their vision of South India characterized by theduality between an inland Dakṣiṇāpatha and a maritime Limyrike
Federicod e r omanis
Universitagrave di Roma lsquoTor Vergatarsquo
50 PME 54 ἡ δὲ Μούζιρις βασιλείας μὲν τῆς αὐτῆς ἀκμάζουσα δὲ τοῖς ἀπὸ τῆςἈριακῆς εἰς αὐτὴν ἐρχομένοις πλοίοις καὶ τοῖς Ἑλληνικοῖς Cfr PME 51 wherethe all coasting navigation as far as the Limyrike (probably from the promontoryAstakapra just opposite Barygaza) is estimated 7000stadioi ὁ δrsquoὅλος παράπλουςμέχρι τῆς Λιμυρικῆς ἐστὶν σταδίων ἑπτακισχιλίων For the implications of suchdistance estimates cfrP a rnaud laquo De la dureacutee agrave la distance lrsquoeacutevaluation desdistances maritimes dans le monde greacuteco-romainraquo Histoire amp Mesure 8 (1993)p 225-247
