1

Mīmāṃsābetweenepistemologyandhermeneutics:Thehistoryofarthāpatti1

elisafreschi

1 Introduction

Arthāpattiis,accordingtoMīmāṃsāauthorsstartingwithŚabara,adistinctinstrumentofknowledge,leadingonetoknowoutofagivensetoffactsthatsomethingelseneedstobethecase.Thestandardexamplesof itare“Caitra,who isknowntobealive, isnothome”,leading to “Caitra isoutside”and“The fatDevadattadoesnoteatatday time”, leading to“Devadatta eats at night”. Mīmāṃsā authors contend, against most Naiyāyika ones, thatarthāpattiisdistinctfrominference.2Theyalsoclaimthatitdeliverscertainknowledge,i.e.,that it is not just what contemporary philosophers might call “inference to the bestexplanation”or“abduction”.ThisisconfirmedbythefactthatsincethetimeoftheVṛttikārathe standardexampleof arthāpatti startswith “Caitrawho is alive isnothome”,whereas“Caitraisnothome”wouldbeasuitablestartforaninferencetothebestexplanation.3

1Research for this article has been funded by the WWTF project M16_028. I read anddiscussed all Sanskrit texts used within this article together with Andrew Ollett (ourtranslationswillbepublishedasFreschiandOllett forthcoming(a)andFreschiandOllettforthcoming(b)) and I am deeply grateful for many engaging discussions and insights Iborrowed from him, both in Vienna (2016 and 2017) and in Tsukuba (2018). I am alsograteful toKiyotakaYoshimizu fordiscussing the topicwithmeper email and inperson,againinTsukuba,andtoKeesvanBerkelforpatientlyexplainingmeseveraltopicsoflogic.Last, I started reading Vātsyāyana’s Nyāyabhāṣya on 2.2.1–7, and Gaṅgeśa’s andRaghunātha’s texts on arthāpatti during aworkshop on arthāpatti organised byMalcolmKeatinginSingapore(2018).LetmethereforealsoexpressmythankstoMalcolm.IamalsothankfultoMarcoLauriforgoingthroughtheEnglishformofthearticleandtoDaniaHuberforherhelpwiththefinalformatting.

Last,IwouldliketoaddherealsomymostheartfeltthankstoJohnTaberforhavinghelpedme at the very beginning of myMīmāṃsā journey, when I was full of curiosity but hadhardlyanygraspofthisschool.2Inthefollowing,Iwilluseanumānaand‘inference’interchangeablyandthesameappliestotheothernamesoftheinstrumentsofknowledge.3IderivethispointfromYoshimizuforthcoming.

2

Thepresent article takes the above facts as given andattempts to reconstruct theearlyhistoryofthetheoryofarthāpattianditsdevelopmentwithinMīmāṃsā.ThisallowsonetoshedlightontheearlyhistoryofMīmāṃsāanditsinnercontroversies,showingthatitwaswaymoremanifoldthanonemighthavethought.

1.1 Mīmāṃsāfromhermeneuticstoepistemologyandback

TheMīmāṃsā is a school of Indian philosophywhich developed since the last centuriesBCE. Its foundational text, theMīmāṃsāsūtra (orPūrvamīmāṃsāsūtra,henceforthPMS) istraditionallyattributedtoJaimini(2ndc.BC?)andisprobablythemostancientamongthefoundational texts of the various philosophical schools in India.4 The PMS has beencommenteduponbyoneormoreauthorswhosework is lostandwhoarementionedbylater ones. Notable among them is especially the so-called vṛttikāra ‘author of thecommentary’, a termwhich ismentionedbyMīmāṃsāauthorsandbyauthorsofvariousVedānta schools (although it is not completely surewhetherone and the sameperson ismeant by all). The first extant commentary is the one which became the standard one,namely the one authored by Śabara (5th c. CE?).5 Śabara’s commentary (or Bhāṣya,henceforthŚBh)hasbeeninturncommenteduponbyKumārilaBhaṭṭa(6–7thc?)andbyhisyoungercontemporaryPrabhākaraMiśra.6Tothesetwoisattributedthefoundationofthe two sub-schools of Mīmāṃsā, named after them Bhāṭṭa and Prābhākara Mīmāṃsā.JaiminiandŚabaradonotappeartobeinterestedinepistemologyasdistinctfromthecoremission of Mīmāṃsā, namely the hermeneutics of Vedic texts; Prabhākara keeps theprimaryfocusonVedichermeneutics.Bycontrast,Kumārila’stargetaudienceisclearlythatof thebroaderarenaof Indianphilosophers.7For instance,Kumārila’sworkwouldnotbeconceivable without taking into account that of the Buddhist philosopher Diṅnāga(approximately 480–540 CE) and he is often engaged in countering his and others’

4Iuse theabbreviationPMS forpractical reasonsandnotasahint to theallegedoriginalunityoftheMīmāṃsāsūtraandtheBrahmasūtra.5ForthedatesofJaiminiandŚabaraseeFreschiandPontillo2013.6On thedateofDiṅnāga,Bhāviveka andDharmakīrti, and, consequently, ofKumārila andPrabhākara, see Krasser 2012, which revises Frauwallner’s traditional dates and movesDharmakīrti (and consequently Kumārila and Prabhākara) back to ca. 550. Steinkellner2013 (with some caution) and Eltschinger 2014 follow Krasser’s suggestion, whereasFranco forthcoming opposes it. Crucial in this connection are the relationship betweenBhāvivekaandDharmakīrti,thestatementsofsomeChinesepilgrimstravellingtoIndiaandCandrakīrti’s lack ofmention ofDharmakīrti. Directly on the chronology of Kumārila seeYoshimizu 2015, fn. 1, where the dates suggested are “ca 560–620”, on the basis ofBhāvivekaasterminuspostquemandthepoetSubandhuasterminusantequem.7TracesofthisinclinationcanbefoundalsointhesectionontheinstrumentsofknowledgebytheVṛttikāraquotedbyŚabarainhiscommentaryonPMS1.1.5.

3

objections to the Mīmāṃsā epistemological tenets. This inclination will not change inKumārila’scommentators.8

1.2 Etymologyandidentityofarthāpatti

Howshouldoneinterpretthecompoundarthāpatti?Anditssingleterms?

As for the interpretation of the compound, several early śāstricworks (Patañjali’sMahābhāṣya, the Caraka Saṃhitā and the Suśruta Saṃhitā) use expressions like arthādāpanna‘obtainedbecauseoftheartha’.9Lookingattherespectivecontexts,thisexpressionmightmean something like arthāt, i.e., ‘(obtained) implicitly’, ‘on the strength of things’.This interpretation of the compound is explicitly attributed to arthāpatti already inVātsyāyana’sNyāyabhāṣyaad2.2.1.

As for its constituent terms, artha is a complex term, since it has at once anontological,epistemological, linguisticanddeonticmeaning.Inthefirstcase,itdenotesanobject,afactorastateofaffairsandcanalsobeusedindefinitelyastheEnglish‘thing’.Inthe second it denotes the content of a cognition, in the third themeaning of a linguisticexpression and in the last case it denotes a purpose.WithinMīmāṃsā, all these aspectsintersect, since theMīmāṃsā’smain focus lies in thedeontic sectionsof theVedaandanarthaisthereforeatonceapurposetoberealisedandthemeaningofaVedicsentence.Atthesametime,BhāṭṭaMīmāṃsakasweredirectrealistsinregardtoworldlyexperience,sothatanartha intheworldlysensewasforthemameaninginFrege’ssenseofBedeutung,thus linking together linguistic and ontology through an epistemologically soundcognition.10

āpatti isanomenactionis fromtherootāpad-andit indicatesthefactofobtainingautomaticallysomething,likeprasañj-.InclassicalSanskrit,likeprasaṅga,itindicatesalsoanegativeconsequence,evenacalamity.By lookingatŚālikanātha(seesection5)oneseesthathewasunderstandingarthāpattiinthisway,namelyasthegettingintroubleofathingby means of a (possibly different) thing. However, looking back at earlier Mīmāṃsā,arthāpattidoesnothavethismeaning.Thus,thetermmighthavewitnessedadevelopmentleading from ‘obtainment’ (Mahābhāṣya, Caraka Saṃhitā, Suśruta Saṃhitā) through thepossible intermediate step of ‘automatic obtainment (due to the logical or factualinconsistencyof thesituationpriorto itsobtainment)’(Śabara,Vṛttikāra,Yuktidīpikā (see

8ReasonsofspaceforcedmetoleaveoutherethediscussionoftheviewsonarthāpattiofKumārila’scommentators.ThiscanbereadinFreschiforthcoming.9I could locate these occurrences since they are listed and reproduced in Oberhammer,Prets,andPrandstetter1991,s.v.10OnarthainaNyāyaauthorwhoknewalotofMīmāṃsā,JayantaBhaṭṭa,seeFreschiandKeidan 2017,which includes also a discussion of possible translations inspired by Fregeandotherphilosophersoflanguage.

4

section 2.2)) and until ‘endangerment’ (Śālikanātha). In all these cases, artha could beunderstood as the reason for the āpatti (i.e., as arthād āpatti, like in the pre-MīmāṃsāoccurrencesandintheYuktidīpikā).Inthepost-JaiminiMīmāṃsāoccurrences,āpattilinkshowever two arthas, one out of which the āpatti takes place and onewhich is obtainedthroughit.

Thedifferenceisnotonlyterminological,sincetheissuehastodowiththedefinitionofarthāpattiandthewayitisdifferentiatedfrominference.ThiswaspossiblynotanissuefortheearlyMīmāṃsā(Jaiminiisnotpartoftheepistemologicaldebate),sothatitisonlyafter the Vṛttikāra —whose quotations are embedded in Śabara’s commentary— thatMīmāṃsāauthorsbecameawareoftheneedtodistinguishitfrominference.Itispossiblybecauseofthat,thattheVṛttikāraidentifiedthe“impossibilityotherwise”asthedistinctivemarkofarthāpatti.

However (as discussed in the Prakaraṇapañcikā, see section 5 and 4), thecharacteristicofbeingotherwiseimpossiblecouldbesharedalsobyinference(ifwetakeasinstance the caseof a sproutwhich leadsone to infera seed,one couldalso say that thesproutisimpossibleifnotthroughaseed).Hence,someMīmāṃsāauthorsfelttheneedtofurther specify this impossibilityotherwise inadistinctiveway.Thisprocedure took twodistinctiveways:

1. explainingthatthe impossibilityotherwise leadstoarealepistemologicaldoubt inthe knower (this path was taken by the unnamedMīmāṃsā author embedded inDiṅnāga’sPramāṇasamuccaya211andbyŚālikanāthaanditcouldbethereasonforthe critique of arthāpatti found in the *Tarkaśāstra12 and in theNyāyabhāṣya, forwhichseesection2.2)

2. noting the specific distinctive elements of arthāpatti (this path was taken byPrabhākaraandKumārila)

The unnamed Mīmāṃsā opponent quoted within Diṅnāga’s Pramāṇasamuccaya evokeddoubt(whichwillplayadecisiveroleinŚālikanātha’sreconstructionofthewayarthāpattiworks),possiblywithnotmuchemphasis.Diṅnāgaruledouttheproposalimmediatelybysaying that if doubt played a role within it, arthāpatti would not be an instrument ofknowledge.ThispowerfulargumentmighthavebeenthereasonwhyKumārilapreferredacompletely different strategy, remained on the safe side of undoubted cognitions and

11As reconstructed on the basis of the quotes within Jinendrabuddhi’s commentary inSteinkellner,Krasser,andLasic2005.

12*TarkaśāstraisthereconstructedtitleofalogictextofwhichonlyaChinesetranslationisextant.ThetranslationdatestotheLiangDynasty(552–557CE).GiuseppeTucciattemptedto reconstruct the Sanskrit text on its basis (seeTucci 1929). If Tucci’s reconstruction isreliable,the*Tarkaśāstraisakeytextforthetheoryofanumāna‘inference’.IamgratefultoMalcolmKeatingforhavingmentionedthepassageonarthāpattiwithinthe*Tarkaśāstra.

5

stressed other dissimilarities between arthāpatti and inference. By contrast, Śālikanāthamight have decided to grasp back to thismore ancient proposal because of his desire todifferentiatethemmorestrongly,evenattheriskofhavingtofaceobjectorsclaimingthatinthiswayoneloosespartoftheepistemicsecurityofarthāpatti.

1.2.1 Etymologyandinterpretationofdṛṣṭārthāpattiandśrutārthāpatti

TheVṛttikāramentionsaśrutaandadṛṣṭaarthainconnectionwitharthāpatti(seesection2.1).Thisledtolongdiscussionsaboutdṛṣṭa-andśrutārthāpattiinKumārilaandamongthePrābhākaras.ButhowarethesetermstobeinterpretedinearlyMīmāṃsāandamonglaterauthors?

Out of symmetry with dṛṣtārthāpatti, possibly meaning ‘reaching an artha out ofsomething experienced’,we can assume that the compound śrutārthāpattiwas originallymeant to be interpreted as ‘reaching anartha out of something heard’. Out of the belowdiscussion(seesection3.3),however,itwillbeclearthatinlaterMīmāṃsā,itsdistinctiveelementwasnolongerthefactofbeingoriginatedoutofanauditorycognition,butratherthe fact of resulting in one. In this sense, a śrutārthāpatti became rather an arthāpattiresulting in the postulation of something hearable. In Uṃveka’s words: “Since theśrutārthāpattihasasitscontentaninstrumentofknowledge(i.e.,asentence),itisdifferentfromthedṛṣṭārthāpatti”(adŚV,arthāpatti76).

Whythisdissymmetry?Thequestionhastodowiththedistinctionoftwotypesofarthāpatti,possiblyalreadyintheVṛttikāra’squotewithinŚabara,forwhichseesection2.1.Itis,aswillbeseen,possiblethattheVṛttikāraconceivedofdṛṣṭaandśrutaarthasinawaydifferentthanwhatbecameofthetwoarthāpattisatalaterstage.Thelaterelaborationofwhatwasthedistinctiveelementinthetwoarthāpattisledthetwofurtherfromeachother(seesection8inFreschiforthcoming).

2 arthāpattiinearlyMīmāṃsā

2.1 śruta and dṛṣṭa objects in Jaimini and the Vṛttikāra:epistemologicalandhermeneuticconcernsatstake

Jaimini does not define arthāpatti, whereas the Vṛttikāra defines it in just the followingshortpassage:

arthāpattir api dṛṣṭaḥ śruto vārtho ’nyathā nopapadyata ity arthakalpanā, yathā jīvato devadattasyagṛhābhāvadarśanenabahirbhāvasyaadṛṣṭasyakalpanā.

Andthearthāpatti isthepostulationofsomethingwhenaseenorheardthingwouldnototherwisemakesense.Forexample:OnehasexperiencedthatDevadatta,whoisalive,isnotathome.Throughthatonepostulatesthatheisoutside,whichhadnotbeenexperienced.

It is noteworthy, given the shortness of the definition, that the Vṛttikāra feels theneedtodistinguishbetweenadṛṣṭa ‘seen’andaśruta ‘heard’arthāpatti. In fact,Kumārilaand his subcommentators all discussed the issue at length and offered acuteepistemological explanations (see section 8 in Freschi forthcoming),which are, however,unlikelytobetheonestheVṛttikārahadinmind.

6

ThetermarthāpattiisusedseveraltimesasdesignatingahermeneuticaldevicealreadyinthePMS(PMS7.4.16;7.4.18;10.1.30;10.3.35;10.4.35),withnofurtherattribute.Thereitis(asmostlythecasewithinthePMS)appliedtoVedicsentences.ThismakesonethinkthatarthāpattidevelopedwithinMīmāṃsāasahermeneuticdevice tobeapplied to theVeda.Furthermore,thePMSingeneralandthespecificcontextofthesequotesinparticularpointinthedirectionofarthainanon-descriptivesenseas‘effect’,witharthāpattias‘obtainmentofanartha(i.e.,obtainmentofaneffect)’,againsttheinterpretationofarthāpattiasderivingfromarthādāpattiḥinthe*TarkaśāstraandintheNyāyabhāṣya(seesection1.2).13

GiventheVediccontextofthefirstusagesofarthāpattiinMīmāṃsā,theworldlyuseofarthāpatticouldbeasecondarydevelopment,oratleasttheconnectionofbothunderasinglehatcouldbe.This innovationwouldhavecalledforadistinguishing label.Hence, itmightbe imagined, theVṛttikāra’sdecision todistinguisha śruta andadṛṣṭa case,whichwouldreferrespectivelytoanarthāpattiregardingtheVeda(generallycalledśruti)andoneregardingcommonexperience(generallycalleddarśana).ThiswouldalsoexplainwhytheVṛttikāra felt the need to add an example only of the second type: the first one wasassumed to be clear to the Vṛttikāra’s readers, who were Mīmāṃsakas and thereforeconversantwith Vedic exegesis. Last, the criticism of arthāpatti as not leading to certainknowledge as attested in Vātsyāyana and in the *Tarkaśāstra might have been currentalreadybeforethetimeoftheVṛttikāraandhavepromptedhimtoinsertthequalificationjīvan ‘alive’totheexampleofthepersonabsentfromhome.Inthisway,theVṛttikārawassure to link arthāpatti to certainty. All three moves will be reinforced by Kumārila’ssystematization.

Whytaking thesedecisions?TheVṛttikārawasprobably the firstMīmāṃsāauthorwhoattemptedtoemphasisethegeneralepistemologicalsideofMīmāṃsā,asattestedbythe fact that he was the first one discussing pramāṇas ‘instruments of knowledge’ asworldlyepistemicinstrumentsandnolongeronlyasVedicexegeticaldevices.Inthissense,theVṛttikārapossiblyanticipatedKumārila’sagendaofplacingMīmāṃsāinthemiddleofthe philosophical arena. This attitude was possibly at play also in the treatment ofarthāpatti.

The early history of arthāpatti as exegetical device was possibly lost in thebackground after Kumārila’s reconfiguration of the topic as a chiefly epistemologicalproblem,sothathiscommentatorshadtoargueat lengthabouttheneedandthecorrectway of distinguishing a śruta and a dṛṣṭa type of arthāpatti (see section 8 in Freschiforthcoming).

13SeethediscussionofthesepassagesinYoshimizu2000bandanexcerptofitbelow,p.9.

7

2.2 arthāpatti in pre-Kumārila philosophy: epistemological andhermeneutictrends

Traces of a competing discourse on arthāpatti can be detected in the *Tarkaśāstra, theNyāyabhāṣya, Diṅnāga’s refutation of a Mīmāṃsā Vṛttikāra and also in the Yuktidīpikā(henceforthYD)(WezlerandMotegi1998,p.73).14Diṅnāga’stext(availableinTibetanandin a German translation in Frauwallner 1968, p. 91 and then within Jinendrabuddhi’scommentarythereoninSteinkellner,Krasser,andLasic2005)mentionsthesameexampleofDevadatta’sbeingoutofhome(incorrectlyoralternativelyconstruedasanexampleofabhāva ‘absence as an instrument of knowledge’) but before that divides arthāpatti inconclusive and false. The same partition can be found in the YD which calls themvyabhicārin ‘deviating’ (from truth) and avyabhicārin ‘not deviating’. By that, authorsopposingtheideaofarthāpattiasadistinctinstrumentofknowledgerefertothecasesofarthāpattiwhicharejustmisleadingandthecasesofarthāpattiwhichyieldvalidcognitions,but shouldbe considered inferences.Both texts then speakofworldlyexamples.More indetail,Diṅnāga’stextmentionstheexampleofknowingoutofseeingpreparedfoodthatitmust have been cooked. It alsomentions the snake-mongoose example (you see a deadsnakeandconcludethatitmusthavebeenkilledbyamongoose),butasacaseofinference,possiblybecauseitwasconsideredassuchbySāṅkhyaauthorsandthesewereforDiṅnāgamoreinfluentialasopponentsthanMīmāṃsāauthors(seeSteinkellner,Krasser,andLasic2005). TheYDmentions the same example in connectionwith arthāpatti, but usingboarand lion insteadof snakeandmongoose. It alsomentions thecaseofknowingsweetnessuponseeing jaggeryorhearing theword jaggery.TheYDaccount, thoughshort, containsfurther elements which will be found again in Mīmāṃsā discussions, so that it seemsplausiblethatitisreactingtoapositionheldbyaMīmāṃsāauthor.Thesearethementionof the relation (sambandha) between trigger and thing known (which are used by thesiddhāntintoshowthattheallegedarthāpattiisnothingbutaninference)andthementionofdarśana‘seeing’andśravaṇa‘hearing’astwoalternativesourcesforarthāpatti,asintheVṛttikāra.

Within Nyāya, theNyāyasūtras mention arthāpatti as not being an instrument ofknowledge, insofar as it is not conclusive (NS 2.2.3). Vātsyāyana (late fifth century?)mentionsarthāpattiwhilecommentingontheNyāyasūtras2.2.1–6.Hisshortcommentsare

14HereandinthefollowingIamassumingthattheMīmāṃsāpositionembeddedintheYDisolderthanKumārila’s.WezlerandMotegidatetheYDto680–720duetoanallegedquotefromtheKāśikāvṛtti,whichishowevernotnecessarilyaquoteandcouldbeonlyasimilargrammaticalformulation(IowethisargumenttoKarinPreisendanz,whodiscusseditinaclasson theYD in2008).Even if theYDhadbeenwrittenafterKumārila, itspositiononarthāpattiattestsofapre-Kumārilastageofthereflectiononthisinstrumentofknowledge,sinceitelaboratesonexamples,suchastheoneaboutnaturalenemieswhichwererejectedbyKumārila.

8

interesting for several reasons. First, Vātsyāyana analyses arthāpatti as arthād āpattiḥ.Second,hecriticisesarthāpattiasnotaconclusiveinstrumentofknowledgeonthebasisofthefollowingexample:raindoesnotoccurwhentherearenoclouds,hence,whentherearecloudsthereisrain.ThismightbetheremotecauseofanexamplementionedbyGaṅgeśa(see section 6). According to Giuseppe Tucci’s reconstruction, the same etymology forarthāpattiandthesameexamplearefoundalreadyinthe*Tarkaśāstra.15

Summing up, arthāpatti must have been already discussed as an instrument ofknowledge different from inference, possibly at the same time of the Vṛttikāra, andwaspossiblycriticisedbecauseitdidnotattaincertainty.

Onewonders, therefore, if therewerenot twocompeting trendsonarthāpatti, thehermeneutical one found in Jaimini and the epistemological-worldly one reflected in theNyāyasūtraanditsBhāṣya,the*Tarkaśāstra,DiṅnāgaandintheYDalreadybeforethetimeof Kumārila. On a more speculative note, one might even suggest that the mention ofdarśanaandśravaṇaintheYDmightbeahintofthefactthatsomethinkerswerealreadytryingtobringthetwotogetherunderthesamehat.

TheMīmāṃsakavṛttikāra‘authorofthecomment’quotedbyDiṅnāgawouldthusbe,like theonequotedbyŚabara,16 an engagedepistemologist, opposing the (possiblymoretraditional)trendofpurelyMīmāṃsāhermeneutics.

Śabara falls somehow in between, as in many other cases. Among his usages ofarthāpatti,severalonesarehermeneutically-oriented(e.g.,whenheusessomeformofwhatKumārilawouldcallśrutārthāpatti inhiscommentaryonPMS3.1.10,3.3.14,3.6.37,6.1.1,6.1.3 in order to justify a given Vedic interpretation), but arthāpatti is also mentioned(interestingly, by opponents)within epistemological issues such as the postulation of anauthoroftherelationbetweenlinguisticexpressionsandmeanings(ŚBhonPMS1.1.5),ofacapacityoflanguagetoconveythesentence-meaningindependentlyoftheword-meanings(ŚBhonPMS1.1.25)andofanauthoroftheVeda(ŚBhonPMS1.1.30).However,thefactthatarthāpattiismentionedinthesecasesbyopponentsembeddedintheŚBhcouldalsobereadas confirming that this trendwaspresent inMīmāṃsā,but remainedmarginaluntilKumārila.

ThesetwotrendsideallycontinuedintheworkofPrabhākara(morehermeneuticalandVeda-centered,forwhomarthāpatti isessentiallyahermeneuticaldeviceandthereis

15See Tucci 1929, p. 25 of the *Tarkaśāstra Sanskrit text: ko ’sau nyāyo yenaitad arthādāpattir bhavet. yad anabhivyaktaṃ tad atyantam asad iti naitad arthād āpadyate.abhivyaktaṃ dvividham anarthāpattir arthāpattiś ca. yadi vṛṣṭir bhavati tadāmeghenāpibhavitavyam.meghesatyapitukadācidvṛṣṭirbhavatikadācinnabhavatītyanaikāntikatā.dhūmenāgneranumānam.nārthādāpattiḥ.[…]kasmāditicet.taptāyaḥpiṇḍelohitāṅgārecadhūmābhāve’pyagneḥsadbhāvaḥ.tasmādabhivyakteṣvarthāpattikhaṇḍanamabhūtam.16SeeFrauwallner1968,p.98forarefutationoftheiridentification.

9

no need to separate between a dṛṣṭa- and a śruta- type) and Kumārila (moreepistemological and interested in wider philosophical debates). It is, accordingly, notsurprisingthatKumārila’scommentatorsstruggledwithfindingapossibleexplanationfortheoppositionbetween śruta anddṛṣṭa in the ŚBh.Theoppositionhad in fact originatedwithin a different context (the conflation of a primarily epistemological and primarilyhermeneuticarthāpatti)thantheoneofthesecommentators’philosophicalengagement.

A further element which is typical of early Mīmāṃsā is that it is much morevariegatedthanitbecamelater.TheopponentsembeddedinDiṅnāgaandintheYD,aswellasthevoicesfoundinthe*TarkaśāstraandintheNyāyabhāṣyausedifferentexamplesandthe YD even avoids mentioning the one which later became the standard one, namely“Caitra,whoisalive,isnothome,thereforehemustbesomewhereoutside”.

A last question needs some closer examination, namely whether the twounderstandingsofarthāpattiareoriginallyfullyunrelated.Inordertoanswer,letmestartby summing up the difference between arthāpatti as used by Vedic exegetes and byepistemologists:

exegetesepistemologists

field of application examples

artha

Vedic sentences Vedic sentences

dynamic

worldly experience absence from home, eating,

snake… descriptive

Onecouldnowimaginethatonehastodowithtworadicallydifferentconcepts,oneof which developed within Mīmāṃsā whereas the other came to it from outside. Thishypothesis is theoneendorsed in theonly rigorous studyof theprehistoryofarthāpatti,namelyYoshimizu2000b:

IntheJS[=PMS],theterm“arthāpatti”alwaysmeans“takeeffect”(*arthamāpadyate).Therefore,wemaysaythatMīmāṃsaka’sconventionofcallingakindofpramāṇa‘instrumentofknowledge’withtheterm“arthāpatti”cannotbe tracedback to the timeof the compilationof the JS, but rather itwas introduced fromoutside, such as thetraditionoftantrayukti.Oneandthesamewordwasestablishedinatotallydifferentmeaning.17

17Since I cannot read Japanese, I completely depend on the explanation of the articleKiyotakaYoshimizuwaskindenough toofferme.Thepassageabove is theconclusionofthefirstsectionofYoshimizu2000b,p.1113.

10

Abenefitofthisviewisthatitaccountsforthecompetingetymologiesofarthāpattiandforthe presence of the term arthāpatti in different contexts, i.e., as a pseudo logical reason(called arthāpattisama hetu or arthāpattyābhāsa). The problem with this view is that itpresupposes the existence of a group of people (let me call them again “theepistemologists”)whowere outsideMīmāṃsā but did not influenceany other school butMīmāṃsā. Therefore, I would like to fine-tune Yoshimizu’s theory by rather speaking ofdifferenttrendsactivenotonlyoutsideofMīmāṃsā,butalsoalreadywithinearlyMīmāṃsā,someofwhichwererepresentedinthePMSandsomenot.ThePMSwouldbethechieftextfor(partof)theexegeticalcomponentwithinMīmāṃsā,butnotnecessarilyforthewholeofMīmāṃsā. In this view, the first two differences mentioned in the table above could beexplained as part of a difference in themain focus. The different understanding ofarthacouldalsobeexplainedonthebasisoftheVedicvs.worldlyfocus,insofarashavingaVedicfocusmeansfocusingonthedeonticdimensionandthereforeonarthaassomethingtobedoneandnotasthedescriptionofastateofaffairs.

Further, the very inner-Mīmāṃsā history of the term arthāpatti discussed byYoshimizuwould have been problematic from the viewpoint of the “epistemologists”, sothatitisdifficulttoimaginethattheydevisedsuchalabelfortheinstrumentofknowledgethey had in mind (and which they could have more easily called anyathānupapatti‘impossibility otherwise’ or kalpanā ‘postuation’). The re-semantisation of a term whichwasalready inuse in theschoolappears,bycontrast, tobemoreconvincing.LarsGöhlerhaspointedoutsimilarcasesofanepistemologicalre-semantisationwithinMīmāṃsā,e.g.,inthecaseofupamāna‘analogy’(Göhler2011).

Last,thearthāpattiexamplesmentionedinthe*Tarkaśāstra(andbyVātysāyana),byDiṅnāgaandintheYD(andbyDiṅnāga’scommentatorJinendrabuddhi)willbeignoredorexplicitlyrefuted(inthecaseofthesnake-mangooseone)byKumārilaandlaterMīmāṃsāauthors.AtleastoneoftheexampleswhichbecamestandardforarthāpattiisdiscussedbyDiṅnāga in the topic ofabhāva,whereas further exampleswill be recognised as cases ofinference. It appears, in sum, that the pre-Diṅnāga or pre-Kumārila philosophywas stillstrugglingtofinduniformdefinitionsforeachpramāṇa.

3 Kumārila:arthāpattiwithintheepistemologicallandscape

AllIndianphilosophicalschoolsbuttheCārvākasacceptthatsenseperception(pratyakṣa)andinferenceareinstrumentsofknowledge.AlltheremainingonesbuttheVaiśeṣikaandthe Diṅnāga-Dharmakīrti school accept also linguistic communication as a distinctinstrumentofknowledge(śabda).Amongtheremainingones,NyāyaandMīmāṃsāacceptalso upamāna ‘analogy’. Mīmāṃsakas remain alone in accepting also arthāpatti and thePrābhākara subschool of Mīmāṃsā parts company from the Bhāṭṭa subschool when itcomestoabhāva ‘absenceasaninstrumentofknowledge’,sincetheBhāṭṭasaccept itandthePrābhākarasdonot.Theschoolswhichdonotacceptlinguisticcommunication,analogy,

11

arthāpattiandabsenceas independent instrumentsofknowledge,generally try toreducethem to a form of inference. This choice is due to the basic distinction between senseperception and inference and to the clear precinct of application of the former, so thatreductionismautomaticallytranslatesintoreductiontoinference.18

It is therefore important to highlight the basic features of inference in Indianthought. In its basic form accepted by all schools, this links a known property which ispresent in a given locus to an inferable property present in the same locus. The non-accidentalnatureoftherelationbetweentheknownpropertyandtheinferableoneneedsto be confirmed with an example of the same concomitance. The known property (orprobans)iscalledliṅga,dharmaorhetu(H),thelocusiscalledpakṣa(P)ordharminandtheinferableproperty(orprobandum)iscalledsādhya(S).UnlikeinAncientGreekphilosophy,thepossibilityofsyllogismswithemptytermsisruledoutbythenecessarypresenceofanexampleinwhichSandHco-occur.Theexampleiscalleddṛṣṭānta(D)orudāharaṇa.Thus:

That is: the presence of the probans (H) in the locus (P) leads one to infer that theprobandum (S) is also found in the locus (P). In the standard example used by Indianauthors:

Onthemountain (P) there is fire (S),because there is smoke (H),as in thekitchen (wheresmokeand fireoccurtogether)(D).

AtleastonegenerationbeforeKumārila,Diṅnāgaformulatedthethreerequirementsforavalidinference(trairūpya),namely:

1. Presenceoftheprobansinthelocus(e.g.,presenceofsmokeinthemountain)

2. Presence of the probans in locations similar to the locus (called sapakṣa) (e.g.,presenceofsmokeinthekitchen)

3. Absenceof theprobans in locationsdissimilar fromthe locus(calledvipakṣa)(e.g.,absenceofsmokefromthelake)

Thisdefinedthekeyelementsofavalidinferenceasfollows:

fire:sādhyaorthingtobeinferred

smoke:hetuorinferentialreason

mountain:pakṣaorlocus

kitchenandotherfieryplaces:sapakṣaorsimilarinstances

pondandotherplacesdevoidoffire:vipakṣaordissimilarinstances

18ApartialexceptionistheNyāyaattempttoreadabhāvaasperceptionofsomethingelseandUdayana’sisolatedattempttoreduceupamānatośabdapramāṇa.

12

Muchoftheanti-reductionistdiscussioninthechaptersofKumārila’sŚlokavārttika(henceforth ŚV, the commentary dedicated byKumārila to the epistemological section ofthe PMS and the ŚBh) dealing with linguistic communication, analogy, arthāpatti andabsence focuses onwhether or not a correct inference can be built in these cases.19 Forinstance, the chapter on language refutes the idea that linguistic communication can bereducedto inferencebyexplaininghowtherelationbetweenśabda ‘linguisticexpression’and artha ‘meaning’ cannot be formalised as above, since linguistic expressions wouldfigure as both the locus and the probans. A similar strategy will be implemented byKumārilaalsointhechapteronarthāpatti(seebelow,section3.2).

A further building block of Kumārila’s refutation of the reduction of arthāpatti toinference presupposes his main epistemological tenet. In fact, as for epistemology ingeneral, Kumārila’s school upholds the svataḥ prāmāṇya, that is the ‘intrinsic validity’ ofcognitions, according to which each cognition should be credited with validity withoutfurtherverificationsimmediatelyanduntilitisfalsifiedbyabādha‘subsequentinvalidatingcognition’.20Thetopicofintrinsicvalidityisoftenevokedinthecontextofarthāpatti,sincethereductionistsidecontendsthatifarthāpattiisnotreducedtoinference,andisthereforenotseenasgroundedonaninvariableconcomitance,thereisnostrongbasisforitsvalidity.Kumārila and his commentators answer that the relation of invariable concomitance isindeedthecausefortheproductionofinferentialcognitions,butitisnotthecauseforthevalidityofneitherthesenoranyothercognition.Bycontrast,theyclaimthatacognitionisvaliduntilaninvalidatingcognitionarises(seevv.79–86andSucarita’scommentaryonv.29).

3.1 StructureoftheargumentintheŚlokavārttika

ThearthāpattisectionisacomparativelyshortoneanditislocatedinthesectionoftheŚVdedicatedtoananalysisofallinstrumentsofknowledge.Kumārila’smainpurposesinthissectionare:

1. Highlightingthemainfeaturesofarthāpatti

2. Provingthatitisdifferentfrominference

3. Distinguishingbetweenadṛṣṭaandaśrutatypeofarthāpatti

The first and the secondobjectives are deeply connected, since showing the specificcharactersofarthāpattiamountstoprovingthatitisnotjustaninference.

19Several studies have been dedicated to the various articulations of the Mīmāṃsāantireductionism. On abhāva in Mīmāṃsā see Taber 2001, on śabda and inference, seeTaber1996andTaber2002.20Formoreonthetopic,seethegroundbreakingandinsightfulTaber1992.

13

3.2 Kumārila’sstrategy

Just like throughout the Ślokavārttika, Kumārila’s strategy in the arthāpatti chapter ischaracterisedbysomeleitmotivs:

1. Defendingarthāpattiaspartofanepistemologicaldebate.TheimpactonMīmāṃsāoftheadmissionofarthāpattiisdealtwithonlylaterally,andKumārilaclearlywantstoground thevalidityof arthāpatti independentlyofVedichermeneutical reasons.Therefore,hepostponesthediscussiononthevarietyofarthāpattiwhichwouldbemore relevant for Vedic exegesis, namely the śrutārthāpatti after a first, purelyepistemological discussion on dṛṣṭārthāpatti. Similarly, the discussion of its Vedicapplications is only dealt with at the end of the epistemological analysis ofśrutārthāpatti. Last, one of the fields of application of arthāpatti, namely theexistenceofapūrva, isnotdiscussedbyKumārila(itwillbe,bycontrast,discussedbyGaṅgeśa,seesection6).

2. Elasticityintheapproach:Kumārilaisradicallynon-dogmaticandisalwayswillingto adopt one or the other solution, if only the main tenets of the system arepreserved(seehiswillingnesstoconsiderthedeductionofpastrainonthebasisofafloodedriverasaninferenceoranarthāpatti,v.18;orhisopennesstowardscallingarthāpattianumāna,ifonewants,v.88).21

3. Antireductionism: Kumārila is against the attempt to reduce the multifacetedprocessesofcognition-acquisitiontotherigidschemeofinference.

The above points converge in Kumārila’s epistemological agenda, which mainly aims atjustifying as much as possible of common cognitions (as explained in Taber 2001 withregardtoabsenceasaninstrumentofknowledge).Itprovidesthereforeaframeworkwhichisflexibleenoughtoaccommodateasmuchaspossibleofcommoncognitionswhichcanbeinterpreted as instances of arthāpatti and therefore deliberately avoids proposing an all-too-rigidschemeforthefunctioningofarthāpatti.

Concerningtheantireductionism,Kumārilahasthreemainarguments:

1. i arthāpatti lacks the formal structureof inference (seeŚVarthāpatti vv.11–24)

2. ii Inarthāpattionedoesnotneedtobeawareofaninvariableconcomitance(v.30andvv.79–86)

3. iii Inarthāpatti, theconclusion is implied in thepremises(seeŚVarthāpattiv.29).

21For the sameattitude in thechapteronsenseperception, see the introductorystudy inTaber2005.

14

Theconfutationoftheidentificationofarthāpattiwithasortofinferencestarts,justliketheconfutationoftheidentificationoflinguisticcommunicationwithasortofanumānain the relevant chapters of the ŚV (śabdapariccheda and vākyādhikaraṇa),with a formalanalysis.Thisshowsthatitisimpossibletodetectthefixedconstituentsofavalidanumānain the case of arthāpatti. That is, and using the standard examples of arthāpatti (“Caitra,whoisalive,isnotathome,thereforeheisoutside“)andinference(“Onthemountainthereis fire because there is smoke”), the former lacks the constituent elements of the latter,namelyaprobans(smoke),alocus(themountain)andaprobandum(fire).22

Thesecondpointmightlookcontroversial,sinceonecouldobjectthatonedoesneedtoknowsomethingaboutCaitra’sbeingeitherhomeoroutsideinordertomovefromtheknowledgeofhisabsence fromhometohispresenceoutside.However,Kumārila isquitestraightforward in denying the pre-existing knowledge of a relation. Rather, he explains,one becomes aware of the relation only at the end, through arthāpatti, so that,paradoxically, at the end of the whole process, one knows the relation of invariableconcomitanceonthebasisofwhichonecouldbeabletostartananumāna(whichwouldbeuselessforthecurrentcase).Iwillcomebacktothistopicinafewlines(section3.2.1).

The third point is discussed also as point (ii) in Yoshimizu forthcoming and inYoshimizu2007anditistheonewhichmoredirectlysuggeststhatarthāpattiisdeductiveinnature,sincetheconclusionsarecontainedinthepremises.23

Kiyotaka Yoshimizu (Yoshimizu 2007, pp. 321–324 and Yoshimizu forthcoming)elaboratesonafurthermovebyKumārila,whichislargelyimplicitinKumārila;YoshimizuneededtodigitoutthroughaformalisationofKumārila’sarguments:24

4. ivIn the view of Kumārila, the basic operative unit for arthāpatti is a sentence(proposition),notaterm,unlikeinthecaseofDiṅnāga’slogic,whichusedsingletermsfor expressing the universal relationship between dharma and dharmin in a certaindomainofdiscourse(seeŚVarthāpattiv.64).

22SeeŚVarthāpatti,vv.11–15,translatedandanalysedinFreschiandOllettforthcoming.

23Unlike that, in inference based on causal connections, there is nothing intrinsic in theeffect itself (e.g., smoke)whichwould necessarily lead to its cause (e.g., fire), unless oneknewabout theirconnection.Theconclusion is,bycontrast, contained in thepremises intheinferencebasedontādātmya‘identity’,suchastheonefromśiṃśapā‘Aśokatree’toitsbeingatree.OncausalityandidentityasthetwokindsofrelationacceptedinDharmakīrti’stheory of syllogism and on their application to the problem of reducing linguisticcommunicationtoacaseofinference,seeTaber,Krasser,andEltschinger2012,fn.113.24I gratefully acknowledge Kiyotaka Yoshimizu’s help (per email and in person) inunderstandingthisaspectofKumārila’sstrategy.

15

In other words, an inference operates on the basis of general valid connections amongproperties and property-bearers. By contrast, an arthāpatti works on specific situations,which —insofar as they are specific— can only be represented through sentences, notnames (as explained in ŚV vākyādhikaraṇa). It does not work, e.g., on the basis of thegeneral connectionbetween theproperty of being somewhere and theproperty of beingabsentanywhereelse.Rather,itworksonthebasisofCaitra’sconcreteabsencefromhome.AlthoughKumāriladoesnotspelloutthisdifference,itisinthissensetellingthathealwaysspeaksofCaitraandnotof‘someone’ingeneral.Theonlyexceptionisv.46,speakingofaperson (puruṣa) but this verse is part of the reasoning explaining that arthāpatti ispreliminary to inference. In fact, Kumārila explains that one cannot possibly grasp theinvariableconcomitancebetweenbeinginoneplaceandnotbeinganywhereelse(orviceversa)aslonganduntilonedoesnotknowboththeconnectedelements(sambandhin)andonedoesnotknowthembeforethearthāpattihastakenplacebecauseoneofthetwo(forinstance,Caitra’sbeingsomewhereelse)isjustnotavailabletoone’sawareness.Kumārilaexplains that once one has known both connected elements, one can implement aninference.

3.2.1 Onarthāpattiandtheestablishmentoftheinvariableconcomitance

Itappears that (seev.32)one firstneeds tousearthāpattiand thencanuse inference infutureoccurrencesofasimilarcase.25Whatdoesthisexactlymean?Theparsdestruensofthe argument is clear, namely that one cannot perform an inference since one does notknow the invariable concomitance.But inwhat sense canone thenperformananumānaafter having performed the arthāpatti? Which invariable concomitance did one becomeawareof?TheinvariableconcomitancebetweenCaitra’sbeingaliveandnothomeandhisbeingoutside?Ifso,thenthismeansthatonewillonlybeabletoinferthatCaitraisoutsidethenexttimeoneknowsthatheisaliveandseesthatheisnothome.Thisseemslimited.Alternatively,onecouldimaginethat,afterhavingknownthrougharthāpattithatCaitraisoutside, onewere able to infer thatDevadatta is outsidebyknowing thathe is alive andseeing thathe isnothome. If itwere so,however, arthāpattiwouldworkon thebasisofterms which can be generalised, and not just particular sentences. Could it be that onecannotinferthatCaitraisoutsidebecauseoneisnotawareoftheinvariableconcomitancebetweenbeinghomeandaliveandbeingoutside,andthatarthāpattiisneededinordertobecome aware of something which was implicit in the premises, but not cognitivelyavailable to the epistemic subject? In this sense, arthāpatti would be needed to becomeawareoftheinvariableconcomitance.ThisinterpretationcomesclosetoPrabhākara’sideathatinthecaseofarthāpattioneonlyknowstheconnectioningeneralbetweentwothingsandnottheparticularisedone,seesection4.Kumāriladoesnotdiscussthisinterpretation,possiblybecausehewasnotawareofPrabhākara’sview.

25ThismoveisparticularlyclearinSucarita’scommentary.

16

3.2.2 WhoareKumārila’smainopponentsinthischapter?

In the first part, dedicated to the epistemological foundation of arthāpatti, the mainopponentisanepistemologist,whowantstoreducearthāpattitoanumāna.AmongIndianphilosophicalschoolsthetwowhichweremorecommittedtoreductionismtoinferencearethe Nyāya (later enlarged to embrace Vaiśeṣika) and the Diṅnāga-Dharmakīrti school.AgainsttheidentificationofKumārila’smainopponentswithphilosophersoftheDiṅnāga-Dharmakīrtischoolrunsthefactthattheauthorsofthisschooldedicatemuchenergytothereduction of linguistic communication to inference, whereas arthāpatti is not a relevanttopic for them.Bycontrast,Nyāyaauthorshaveextensivesectionson it andarthāpatti iscriticised already in the Nyāyasūtra. However, the Nyāya school is never mentioned inKumārila’sarthāpattisection,whereasanexplicitmentionofbauddha ‘Buddhists’isfoundinSucarita’scommentaryonv.40andmoresignificantlyinPārthasārathi’scommentaryonv.83.Last,Kumārilahimselfpointsoutagainandagainthatthesamediscussionhasbeenraisedinthecontextoftheśabdādhikaraṇa,inthesectionwherethe(Vaiśeṣika)opponenttriedtoconstrueaninferenceonthebasisoftherelationbetweenlinguisticexpressionandmeaning. That the first section of the arthāpatti chapter is devised against this type ofopponent isconfirmedbythefactthatnowheredoesKumārilahintatthepossibilitythatCaitra’s absence from home could be known through abhāvapramāṇa ‘absence’, aninstrumentofknowledgerejectedbyall schoolsapart fromBhāṭṭaMīmāṃsā,whereashediscusses anupalabdhi ‘non-apprehension’, which is a valid probans in an inferenceaccordingto theDiṅnāga-Dharmakīrtischool.Summingup, the firstpartof thearthāpattichapter is targeted to non-Mīmāṃsakas, possibly to Nyāya-Vaiśeṣikas or Buddhistepistemologists. It ispossible that scholarsknowonly little about the latter’spositiononarthāpatti because the texts of their own school had no direct interest in recording thepolemicswithMīmāṃsāonarthāpatti.

The situation changes in the second part of ŚV arthāpatti, dedicated tośrutārthāpatti,wheretheinterlocutorsaremainlyotherMīmāṃsakasandwhatisatplayischieflythecorrectinterpretationofVedichermeneuticaldevicessuchasūha(aboutwhichsee fn. 29). Interestingly, Kumārila does not seem to be aware of Prabhākara’s position,whichrefusesthedistinctionbetweentwotypesofarthāpattiandwhichbecamethetopicof a hot debate in later texts. Kumārila does not address at all the question of whetherśrutārthāpattishouldbedeniedaseparatepositionandseemstostartwiththeassumptionthatoneonlyneedstodefinewhatisśrutārthāpatti.26

26ThiscouldbeafurtherhintattheideathatKumārila,thoughawareofsomePrābhākara-like ideas, was not aware of Prabhākara himself and was thereforemost likely a seniorcontemporaryofhim.Onthemajorargumentsforthisdate,seeYoshimizu1997.

17

3.3 śrutārthāpatti

IthasalreadybeenstatedthatwhatwillbelatercalledśrutārthāpattiwasprobablythefirstkindofarthāpattidevisedbyearlyMīmāṃsāauthors,whowerebyfarmoreinterestedinVedicexegesisthaninfindingoutwhetherDevadatta’sbeingoutsideoncehehasnotbeenseenathomeisepistemologicallygrounded.

Thehermeneutical backgroundof the śrutārthāpatti is acknowledgedbyKumārilatooinŚVarthāpattivv.87–88,wherehestatesthatmanyMīmāṃsāhermeneuticstrategieswouldbecomeinvalidifśrutārthāpattiwerenotaccepted.27

Still, in harmony with his general agenda (see section 3.2), Kumārila tries toestablishhisepistemologyindependentlyoftheneedsofVedicexegesis,sothatMīmāṃsāismade fit to fight a purely philosophical battle, without having to recur to hermeneuticalapplicationsofit.

Accordingly,Kumārila(orperhapsoneortheothervṛttikārabeforehim,seesection2.2andfn.6)proposedaworldlyexampleforśrutārthāpatti,namely:

"ThefatDevadattadoesnoteatatdaytime"

The linguistic incongruity raised by this sentence, according to Kumārila, expects alinguisticcompletion,namelythesentence:

"Heeatsatnight"

The latter sentence is not actually heard and needs therefore to be postulated throughśrutārthāpatti.Anopponent in v. 77askswhether the linguistic incongruity couldnotbeappeasedbyafact,ratherthanasentence,suchasseeingthesamepersoneatingatnight.Thecommentatorshavetroublesansweringandfinddifferentwaystodealwithit.Actually,theexamplementionedbyKumārilatoprovetheneedofaseparateśrutārthāpatti isonlypartly adequate, since the incongruity in the sentence “The fatDevadattadoesnot eat atdaytime”seemstobemoreafactualthanalinguisticone,moresimilartotheincongruityofCaitra’s28absencefromhomethantoapurelylinguisticincongruity(suchas,perhaps,thelack of a verb in a sentence). The examples of śrutārthāpatti derived from Vedichermeneuticsappeartobemuchmoreconvincing,sincewithinVedicexegesisitisoftenthecase that one needs to supply linguistic expressions, not just concepts. For instance,

27Further notes on Kumārila’s introduction of śrutārthāpatti can be read in Freschiforthcoming.

28Kumārila appears to be the first one who uses Devadatta as the name used for theexample about the fat person and Caitra as the name used for the example about beingoutside. I will stick with this use in order to keep them apart, even while discussingPrabhākara,whogoesback to theVṛttikāra anduses thenameDevadatta for thepersonabsentfromhome.

18

mantrasneedtobeadaptedbymeansofspecificwords,whentheyneedtobeutteredinawaywhichfitsthechangedcircumstances(e.g.,amantraforagivendeitymayneedtobeuttered in a different ritual andbe dedicated to a different one, so that the deity’s nameneedstobeupdated).29

A further role for śrutārthāpatti could be the establishment of apūrva.30 This is, inKumārila’sreinterpretation, theunseen forcebridgingthegapbetweenasacrificeand itsresult.31Thisneedstobepostulated,accordingtoKumārila, inordertosolvetheseemingparadoxbetweentwofacts:

• a)TheVedamustbetruewhenitsaysthatasacrificeleadsonetoitscorrespondingresult.

• b)Thesacrifice is completedsincea long timeby the time theresult shouldarise,henceitcannotbeitscause.

Thesolutionisthatthesacrificeproducesanew(apūrva)forceleadingtotheresult.Thisargumentation seems to suggest that the instrument to know about apūrva must bearthāpatti.Moreindetail,intheTantravārttika(henceforthTV)commentaryadPMS2.1.5(BSSp.361),Kumārilastatesthattheapūrvacanbeknownoutofśrutārthāpatti.However,asnotedbyYoshimizu(1999),apūrva isacontent,nota linguisticexpressionand in thissenseitshouldbeobtainedthroughdṛṣṭa-,notśrutārthāpatti,notwithstandingKumārila’sassertion. Yoshimizu therefore concludes that there are indeed some discrepancies inKumārila’sattempttoembedapūrvainhistheory,duetothefactthatheneededtoembedtwopre-existingtheories(arthāpattiandapūrva)intohisnewinterpretation.Iwouldaddthat the ideaof linkingapūrvawithśrutārthāpattimakessense ifoneconsiders theearlyhistory of śrutārthāpatti as described above (section 2.1). Kumārila’s apūrva could beobtainedthroughJaiminiandŚabara’sśrutārthāpattibecauseitisaVedicconcept,althoughonecannotuseKumārila’sśrutārthāpattiforit,sinceitisaconcept,notasentence.TheTVpassage could be interpreted in this light as a slip into the preceding terminology aboutśrutārthāpatti. Alternatively, it could be interpreted as just the abbreviated form of thewhole argument: Through śrutārthāpatti one postulates a further Vedic sentence aboutapūrva, inorder tomakesenseof thecontrastbetweenVedicsentencesabouta)andthefactb)(above).Outof thispostulatedsentence, throughśabdapramāṇa,oneknowsaboutapūrva as concept.The latter interpretation is theonepresentedbyGaṅgeśa (seebelow.section6).

29Thisprocedureiscalledūha.30This issuggested inYoshimizu1999and isexplicitlypresent inGaṅgeśa’sdiscussionofarthāpatti in his Tattvacintāmaṇi. I discussed the topic with Stephen Phillips during theworkshoponarthāpattiorganisedinSingaporebyMalcolmKeating(August2018).31Yoshimizu2000aexplainshowKumārilainnovatedwithregardtoŚabara’sandJaimini’sconceptofapūrva.ANDCLOONEY?CHECK

19

3.4 Distinctterminologyforarthāpatti

Since Kumārila is adamant in distinguishing arthāpatti and inference, he and hiscommentators need to avoid the terminology employed in the case of the latter and todevise new ways to refer to the constituent elements of arthāpatti. In most cases, thisamountstothefactthatKumārilaandhiscommentatorsusemoregenerictermsthantheonesusedinthecontextofinference.Forinstance,theyspeakoftheabsencefromhomeasagamaka‘elementcausingonetounderstand’(inthefollowing‘trigger’),insteadofcallingitahetu‘inferentialreason’,ofthebeingoutsideasagamya‘thingtobeunderstood’insteadof a sādhya, and of the concomitance of absence from home and presence outside as asāhitya ‘co-occurrence’ instead of avinābhāva ‘invariable concomitance’. The termsambandha‘connection’appearsalsotobeusedcoveringbothcases.

By contrast, some terms acquire a technicalmeaning in the context of arthāpattiandaresostronglycharacterisedasconnectedtoitthattheycanbeusedassynonymsofthe very term arthāpatti. The foremost among them is anyathānupapatti ‘impossibilityotherwise’,oftwocognitions,suchasCaitra’sbeingaliveandhisbeingoutofhome,whichistherealtriggerofarthāpatti.Similarly,kalpanā‘postulation’definesthecognitiveprocesswhich cannotbe called inferential andoutofwhichone reaches the conclusion that, e.g.,“Caitra is outside”. Althoughkalpanāmay have a negative connotation in other contexts,here it refers to theproductiveaspectofarthāpattiwhich leadsone toanewconclusion,whileatthesametimepreservingthesamelevelofcertainty.

4 arthāpattiinPrabhākara

Prabhākara’scommentary(calledBṛhatī,henceforthBṛ)ontherelevantpassageoftheŚBhisrelativelyshortandfocusesonthreeelements:

1. distinctfeaturesofarthāpatti

2. arthāpattiasdistinctfrominference

3. denialofadistinctśrutārthāpatti

AswithKumārila (see above, section 3.2), the points 1 and 2 are deeply connected.Prabhākara is also aware of the objection from the side ofNyāya-Vaiśeṣika (or from theDiṅnāga-Dharmakīrtischool)concerningtheidentityofinferenceandarthāpatti.

On top of that, further (ekadeśin-)objectors, possibly from a Bhāṭṭa or proto-Bhāṭṭastandpoint,proposetwoothertheoriesofarthāpatti,whicharebothrefuted:32

32Onecouldalsoconsiderthesecondtheoryasjustaclarificationofthefirstone.

20

1. Arthāpatti isduetotheimpossibilityotherwise(anyathānupapatti),whichconsistsinthefactthatathingXisnotpossiblewithoutanotherthingY.

2. Arthāpatti is as defined above, but it is distinct from inference because the latterpresupposesthatoneknowstheconnection(sambandha)betweentriggerandthingtobeknown.

Thefirsttheoryisjustsketched,butseemstobetantamounttoKumārila’soneinsofarasitisbasedontheclashofcognitionsduetoanyathānupapatti.Itisrefuted,sinceitissaidtodefinenothingbutaninferenceofthecauseonthebasisofitseffect.Infact,asŚālikanāthaexplains inhis commentary, also smoke is impossiblewithout fire, so that this definitiondoesnotuniquelydistinguisharthāpatti.

Thesecondposition isalsoakintoKumārila’sandstates that thedifference lies inthefactthatinthecaseofinferenceoneneedstoknowtherelationbeforehand(cf.YD,p.73,andseeabove,section2.2).Thispositionwillberefutedinsofarasifonedidnotknowofanyconnection,noimpossibility(anyathānupapatti)wouldbegrasped.

Prabhākara’sfinalpositionwillbeconnectedtothisrefutation,butwiththenuancethat in the caseof arthāpatti oneknowsa connection ingeneral andnot the specificoneneeded. He calls the connection one needs for arthāpatti sambandhamātrajñāna ‘thecognitionofaconnectioningeneral’andtheoneonedoesnotneedjñātasambandhitā‘thefactofhavingaconnectedelementalreadyknown’.

AccordingtoŚālikanātha’scommentaryontheBṛ,calledṚjuvimalā,thismeansthatoneonlyneedstoknowaboutthecorrelationbetweenbeingaliveandbeinginoneplace,not the one between not being home and being outside, so that the connectionwith thespecific trigger of the arthāpatti (the absence from home) is not known. Why wouldPrabhākara not say, like Kumārila, that a relation is not needed at all in the case ofarthāpatti?Becauseheisconvincedthatoneneedssomebackgroundinformationinorderfortheimpossibilitytobeperceived.Onemightaddthataverysimpleperson,achildoranalien,wouldnotperceiveanyinconsistencyatseeingsomeone’sabsencefromhome(afterall, themissing person could have become invisible or disappeared from this dimensionetc.). One needs to be aware of the general connection between being alive and beingsomewhereintheworldinordertograspthatthereissomethingoutofplaceifsomeonewhoisaliveisnotinherusualplaceorevenjustthatsomethingcannottakeplacewithoutsomethingelse(asinPrabhākara’stheoryofanyathānupapatti,seebelow).

Having ruled out the two proposals by the ekadeśins, what can be the distinctspecificityofarthāpatti?Prabhākara’sformalisationofinferenceispossiblylessstrictthanKumārila’s, insofar as Kumārila has been in this respect deeply influenced by Diṅnāga’stheoryofthetrairūpya(seesection3).SincePrabhākara’sformulationof inferenceis lessstrictly formalised, themain reason for the distinction between arthāpatti and inferencecannot rely —like in Kumārila— on technical aspects about the formalisation of pakṣa‘locus’andhetu‘inferentialreason’.Moreover,Śālikanāthaevenexplicitlystatesthatalsointhecaseofinferencethereisanimpossibilityotherwise(anyathānupapatti)atstake,sincethesmokewouldnotbepossiblewithoutthefire.

21

Incontrast,Prabhākaradistinguishesarthāpatti frominferencesince inthe formerthe thing which would not be possible is exactly the thing to be known. In the case ofinference, it is the hetu (smoke) which would be impossible without the sādhya (fire),whereas in the caseof arthāpatti, by contrast, it is the thing tobeknown(gamya)whichwouldnotbepossiblewithoutitstrigger.

Prabhākara adds in this connection an etymological explanation of arthāpatti,highlightingtheidentityofgamyaandanupapannaelements:

anyathānupapadyamānatāmāpādayannarthāntaraṃgamayati.

It causes one to know a different thing by causing [it] to attain the condition of not being possible otherwise.(RāmanāthaŚāstrīandSubrahmanyaSastri1934–1967,p.113)

Thispossiblymeansthatheanalysedthecompoundarthāpattias‘thefactofcausingtoattainathing’.Śālikanāthawillthenelaboratefurtherandaddadifferentunderstandingofarthainthecompound(seebelow,section5.3.1).

4.1 Identificationofthegamyaofarthāpatti

HavingsaidthataccordingtoPrabhākarathedifferencebetweenarthāpattiandinferencelies in the fact that in the latter it is thegamyawhichwouldnotmake sensewithout itstrigger,whatexactly is thisgamya?Thefirstsolutionexaminedis that itmustbethe ‘notbeinginanotherplace’(anyatrābhāva),whichcannotmakesenseonceonehasseenCaitra’sabsencefromhome.Thenotbeingelsewherewouldinfactclashwiththefactofnotbeingpresentathome.Notbeingoutsidedoesindeedclashwithnotbeinghome,sothatitcouldmakesensetocallitananupapatti‘logicalimpossibility’,sinceitcannotbethecasethatoneisnotathomewhilenotbeingelsewhere.

However, the non presence elsewhere would not go, because Prabhākara has alreadypostulated that the thingwhichdoesnotmake senseneeds to be thegamya, in order todistinguisharthāpattifrominference,andwhatoneknowsoutofthementionedexampleisthe fact that Caitra is outside, and not that he is not outside. The argumentationmovesthereforeto(a)furthercandidate(s)intheshortandsomehowenigmaticconclusionofthediscussion:

kasyatarhi.bhāvasya,nacāsaugṛhābhāvadarśanenopapadyate.bāḍhaṃnopapadyate.nahigṛhābhāvadarśanenavinābahiḥbhāvaupapadyate.

Whatisthen[thegamya]?Theexistence.Andthisdoesnotmakesensesinceonehasseentheabsence[ofCaitra]fromhome.

Surelyitdoesnotmakesense!For,itisnotthecasethattheexistenceoutsidemakessensewithouttheexperienceoftheabsencefromhome.

Thefirstlineseemstosaythatexistence(bhāva)doesnotmakesenseonceonehasseen the absence from home. And the context suggests that after the proposal ofanyatrābhāva,nowadifferentproposal for somethinganupapanna ismade.Theproblemhere is that the next line seems to state the opposite, namely that bhāva, now betterspecifiedasbahirbhāvadoesnotmakesensewithouttheabsencefromhome.Whatshouldone make of this? I am inclined to think that the latter suggestion is to be taken asPrabhākara’s conclusiveview, since it closes theparagraph.Prabhākarawould thenhave

22

movedfromtheideaofanyatrābhāvatothatofbhāvaingeneralandfinallytobahirbhāvaasthegamyawhichisanupapanna.Thepositionaboutthebhāvaingeneral(possiblytobeidentified,asinŚālikanātha,withthe‘beingalive’)beinganupapannacouldbeattributedtoanopponent.33

If this interpretation is correct, Prabhākara identifies thegamya to be, rather, the‘presence outside’ (bahirbhāva). Aswill be shown below, Śālikanāthawill identify a flawwithinthisargument.

In which sense can one now say that there is an anyathānupapatti? Prabhākaraexplainsanyathānupapattiasanenavinānaupapadyate“itdoesnottakeplacewithoutit”andgoesonspecifyingthatwithouthavingseenCaitra’sabsencefromhome,onewouldnotcometotheconclusionthatheisoutside.34

GiventhatPrabhākarahasgivenuptherequirementofalogicalinconsistency,it isverymuchlikelythathisarthāpattiwillnotbedeductive,likeKumārila’s.Butwhatcanleadfromonestateofaffairstoanother,onlylooselyconnectedtothefirst,ifnotadeduction?What happens, according to Prabhākara, is that one rephrases in a new way the looseconnectionsoastobecomeawareofanewconclusion.Inotherwords,alightformofbeliefrevision isat stake inhisarthāpatti.Belief revisionsneverstartwithone’smost stronglyheld beliefs, but rather with one’s weaker ones. In this case, there is no negation of aprevious belief, but rather its update in reference to the case at stake. Why would oneupdatethelooseconnectioninsteadofgivingupotherbeliefs(e.g.,distrustingone’ssenseperception of the empty house or one’s knowledge of Caitra’s being alive)? Because theupdatehastheleastepistemicconsequences,insofarasitdoesnotforceonetogiveupanyotherbelief.35

33Alternatively,onemightsuggestthatthefirstlineofPrabhākara’spassagesimplymissesalongaandshouldbereadasgṛhābhāvādarśanena.

34Thewholepassagereads:asyānenavināsadbhāvonopapadyate“Therealexistenceofthisdoes not take place without that”. The passage is attributed to an objector, but theinterpretation of anyathānupapatti is not refuted (rather, the missing distinction fromanumāna is). Śālikanātha glossing on the same passagewrites:nopapadyate ity anenāpināstitvam“Eventhroughtheclause‘itdoesnotmakesense’[Prabhākaraindicates]thatitdoesnotexist”.35“Evenifallsentencesinabeliefsetareacceptedorconsideredasfacts(sothattheyareassignedmaximalprobability),thisdoesnotmeanthatallsentencesareofequalvalueforplanningorproblem-solvingpurposes.Certainpiecesofourknowledgeandbeliefsaboutthe world are more important than others when planning future actions, conductingscientificinvestigations,orreasoningingeneral.Wewillsaythatsomesentencesinabeliefsystem have a higher degree of epistemic entrenchment than others. This degree ofentrenchmentwill, intuitively,haveabearingonwhatisabandonedfromabeliefset,and

23

In formal terms, and assuming that beinghome isA, beingoutside isB andbeingaliveisC:

Step1.()(sambandhamātrajñāna:beingaliveisbeingsomewhere,eitherhomeornot)

Step2.[C]

Step3.(beliefrevision:inStep1.istantamountto)

NotethatPrabhākaradoesnotspelloutthatoneneedstobeawareofthefactthatCaitraisaliveinStep2(hencemyuseofsquarebrackets),heappearstoimagineitasselfassumed.That this is thecase isconfirmedbythe fact thatotherwisetheknowledgethatCaitraisoutsidecouldnotbereached.Infact,ifoneputs()(asintherevisionoftheStep1belief in Step3), in a truth table, the sheer fact of notbeinghome could lead to eitherbeing outside as true or being outside as false. In the latter case, the falsity of not-beingoutsideandthetruthofnot-beingathomewould imply the falsityof theantecedent(i.e.,beingalive) (whichwouldmake the implication true).Thus, theonlyway to come to theconclusion Prabhākara asserts is to assume C in Step 2. As will be discussed below,Śālikanāthawillexaminetheriskoffalsifyingtheantecedent.Prabhākara,instead,doesnoteven hint at this possibility, so that I thinkwe can safely assume that he, like Kumārila,consideredtheabsencefromhometogotogetherwiththefactofbeingalive.Prabhākaramost likely does not spell this out for the same reasons why Kumārila did not feel heneeded to explain the reasons forourknowledge thatC is the case.Thedoubt about thepremise,thoughpresentintheearlyhistoryofMīmāṃsā(seeabove,section1.2)wasmostprobablyaminoritypositionby the timeof Śabara,Kumārila andPrabhākara, so thatnomuch energy was needed to rule out this option. By contrast, after Śālikanātha’sreaffirmationofthisposition,laterMīmāṃsakaswillfeeltheneedtospelloutwhytheyareso sure about Caitra’s being alive (their typical explanation, as found, e.g., in Gaṅgeśa’sTattvacintāmaṇi,willbethatonehascheckedCaitra’sastrologicaltableandknowsthathewilllivealonglife).

The above reconstruction implies also thatanyathānupapatti has in Prabhākara awelldifferentmeaningthaninKumārila.ForKumārila,itrepresentedaclashbetweentwocognitive data, possibly to be understood as a logical inconsistency (see Yoshimizuforthcoming). For Prabhākara, it seems to represent just the impossibility of something

whatisretainedwhenacontractionorarevisioniscarriedout.[…][W]henabeliefsetKisrevised or contracted, the sentences in K that are given up are those having the lowestdegreeof epistemic entrenchment. Fagin,UllmanandVardi (1983), pp.358 ff., introducethenotionof“databasepriorities””(Gärdenfors1992,p.17,emphasisoriginal).

IcametotheideaofbeliefrevisioninthecaseofPrabhākara’stheoryofarthāpattiduring a long discussion on the topic with Malcolm C. Keating and Kiyotaka Yoshimizu(Tsukuba,March2018).IamgratefultoKeesvanBerkelforhelpingmewithmyfirststepsinthislogicandtoMalcolmKeatingforpointingoutGärdenfors1992.

24

without somethingelse, i.e., aprecondition. In this sense,naupapadyate endsupbeingasynonymofnasambhavati,both in thesenseof ‘isnotpossible’aswellas ‘doesnot takeplace’.Accordingly,heconstantlyglossesanyathāasvinā.

The next two tables sum up (in Sanskrit and English respectively) the maindifferencesbetweenKumārila’sandPrabhākara’sunderstandingofarthāpatti.MoredetailsonthePrābhākarapositioncanbereadinsection5.3.

gamaka gamya anupapanna

K jīvatogṛhābhāvadarśana bahirbhāva jīvatogṛhābhāva

P (jīvato)gṛhābhāvadarśana bahirbhāva gamya

trigger thingtobeknown non-sensicalelement

K not-being-home&alive beingout not-being-home&alive

P not-being-home(&alive) beingout thingtobeknown

Thus, theanupapanna element is the existence outside, not in the sense that it islogicallyinconsistentwiththeexperienceofCaitra’snotbeinghome,butratherinsofarasthecognitionofbeingoutsidewouldnothavetakenplacewithouttheexperienceofCaitra’snotbeinghome.

The identification of the otherwise impossible element with the gamya, however,leadstoahermeneuticalproblem,sinceŚabarahadstatedthefollowing:

arthāpattirapidṛṣṭaḥśrutovārtho’nyathānopapadyateityarthakalpanā.

Whichcouldbestraightforwardlyinterpretedas:

Asforarthāpatti,aseenorheardthingdoesnotmakesenseotherwise,hencethereisthepostulationofathing.36

36Since, as discussed above, arthāpatti always links two things, arthakalpanā could heremean‘postulationof[another]thing’(postulationofthegamya)or‘postulationonthebasisofthe[inconsistent]thing’(postulationonthebasisofthegamaka).

25

Thisseems to imply thatoncesomethingdoesnotmakesense,onepostulatessomethingelse.Hence,itseemstosupportKumārila’sidentificationoftheimpossibleelementwiththegamakaratherthanwiththegamya.Prabhākaramustfindasolutionandreads,therefore,thepassagebyŚabaraasfollows:

Asforarthāpatti,aseenorheardthing[when]theensuingcognitiondoesnotmakesense,isthepostulationofthething.

That is,heconnectsdṛṣṭaḥ śrutovādirectlywith ityarthakalpanā andaddsanunspokenpramiti to be connected with anyathā nopapadyate. The proposal is inventive, since itmanagestoforceŚabara’stextinadirectionquitefarfromitsmostnaturalinterpretation,anditcanhardlybethoughttobeŚabara’soriginalintention.

4.2 Rejectionofadistinctśrutārthāpatti

Last,PrabhākaraconcludesthediscussionwiththeexplanationthatŚabara’swordingdoesnot mean that there is a distinct śrutārthāpatti, because the expectation of incompletesentences is appeased by meanings, not words. An objector then asks why would haveŚabarausedbothwords,giventhatdṛṣṭa‘experienced’wouldhavebeenenough.Thereplyisthatitisjustanotherexpression.Śālikanāthawillexplaininhiscommentarythereonthat“we are worldly people” and that therefore it does not make sense to discuss worldlylinguisticusages.

The striking element here is that Prabhākara’s discussion is short and essential.Prabhākara’smainconcernistoexplainwhyŚabaramentionedaśrutaḥarthaḥbuthedoesnot attack Kumārila’s arguments in favour of a śrutārthāpatti.Why so? Possibly becausePrabhākara thought that he was just mentioning the standard traditional approach toarthāpatti,sothathedidnotneedtoexplainmuch.Thiscouldbeconfirmedbythefactthat,bycontrast,KumārilaaddressedaPrābhākara-likeobjection,sothatonecanassumethatitwasKumārilawhowasinnovatinghere.

5 Śālikanātha:distinctionfrominferencebecauseofdoubt

The following sections deal with the part on arthāpatti within Śālikanātha’s Ṛjuvimalācommentary on Prabhākara’s Bṛhatī (in turn commenting on the short quote by theVṛttikāra on arthāpatti embedded in the ŚBh), and with the chapter on arthāpatti inŚālikanātha’sPramāṇaparāyaṇa of the so-calledPrakaraṇapañcikā (henceforth PrP). ThePrPisacollectionofessaysonvarioustopicswhichalreadysoonafterPrabhākarabecamethe standard reference work for the Prābhākara philosophy.37 Śālikanātha is in fact a

37ThePrPwasmostprobablyput togetherafter thecompositionof itsconstituentbooks,butthisdoesnotchangeanythinginthecaseofthebookonepistemology.

26

clearer and more systematic writer than Prabhākara, so that most authors attackingPrābhākarapositionsindeedattackedŚālikanātharatherthanPrabhākaradirectly.

As will be explained, both texts by Śālikanātha discuss all the main issues aboutarthāpatti,namely:

1. Howisarthāpattidistinguishedfrominference?

2. Howdoesarthāpattiwork?

3. Isthereaśrutārthāpattidistinguishedfromthenormalarthāpatti?

LikeinthecaseofKumārilaandPrabhākara,thefirsttwoissueswillbedealtwithjointly,since they are deeply intertwined. I will then discuss in detail the third one. SinceŚālikanāthaisoneoftheearliestauthorsdealingwiththetopicofarthāpattiafterKumārila,Iwillpayspecialattentiontoacomparisonbetweenthetwo.

5.1 Śālikanātha’sindebtednesstoKumārila

Śālikanātha appears to have been heavily influenced by Kumārila approach to the topic.Firstofall,hestartsthediscussionbystatinghisownposition,whichovertlydiffersfromKumārila’s.However,asasecond step,he introducesaNaiyāyikaobjectorwhoseems todirectlyreacttoKumārila’sattacks,insofarashetriestoconstruethearthāpattiasavalidinferenceandespeciallyfocusesonthepossibilityofidentifyingalocus.Now,thisseemsatfirst sight difficult, because inferences depend on the fact that the probans and theprobandumco-occurinthesamelocus,butthebeingoutsideofCaitraandhisbeinghomehavenecessarily two spatially distinct substrates. Therefore, theNaiyāyika proposes twoalternativeloci,namelythesametimeinwhichbothoccurandthesamepersoninrelationtowhichtheyoccur.Asanexampleofthefirstkindofco-occurrencehecanquotethecaseofhightideandtheraiseof themoon,whichareclearlycausally linkednotbecausetheyoccur in the sameplace, but because they occur at the same time. The third and fourthstepsconsist,aswithKumārila,inŚālikanātha’srefutationofthesesuggestions.Apartfromaddinganewreason,namelythattheprobanswouldbedoubtfulandthisisimpossibleinan inference, Śālikanātha also repeats an argument found in Kumārila, namely that apossible probans, the sheer absence from home, would lead to excessive consequences(namely,thebeingoutsidealsoofpeoplewhodonotexistatall),whereasanotherone(theabsencefromhomeofsomeonewhoisalive)wouldonlyworkif itweremaderedundant(ŚVarthāpattiv.25).

Toelaborate:theNaiyāyikaobjectorsuggeststousetheabsencefromhomeofCaitraasprobans as soonas themomentofdoubthasbeenovercome.But theupholderof theconcluding view explains that the doubt is overcome exactly through arthāpatti, so thatafter that moment nothing is left to be inferred. In this sense, the argument is akin toKumārila’soneabout the fact thatavalid inferencecould in fact takeplaceonlyafter theseeming impossibility has been cleared out through arthāpatti—andwould therefore beuseless(seeŚVarthāpattiv.24,andespeciallyPārthasārathi’scommentarythereon).

Then, as a fifth step, Śālikanātha, just like Kumārila before him, examines thedifferent elements needed for inference and arthāpatti. The distinctions between the

27

detailed analysis of these elements in Kumārila and its much less deep discussion inŚālikanāthaisdiscussedinsection5.3.

5.2 Śālikanāthaonanyathānupapatti

In arthāpatti an element causes another element (see section 4.1 on how these aredifferently identified) to be impossible unless one postulates a third element, namelyCaitra’s being out. Kumārila’s and Prabhākara’s schools discuss at length about theidentificationofthefirsttwoelements,whereastheyremainrelativelysilentconcerningthethirdoneandthenatureoftheimpossibility(seealsoabove,section4.1).38

The impossibilityappears tobe twofold, in thesense that it isbasedonamaterialimpossibility(howcomesthatouroldfriendCaitra,whoisalwayshomeanddidnotdie,isnotthere?)butisexplicitlysaidbyPrabhākaraandbyŚālikanātha(notinthePrPbutintheṚjuvimalā)tobesomethingthesubjectneedstobeawareof.Inotherwords,anontologicalimpossibilityitselfisnotenough,oneratherneedstobeintheepistemologicalconditionofbeing aware ofwhat looks like an impossibility. Thiswill then be appeased through thepostulationof,forinstance,Caitra’sbeingoutofhome.

Summingup,forKumārilaanyathānupapattiisalogicalinconsistency.39Prabhākaradiffersslightly,sinceheinterpretstheanyathānupapattiastheimpossibilityofsomethingwithout something else, i.e., as expressing merely a precondition, with no need of acognitive or logical clash (see 4.1). By contrast, for Śālikanātha the anyathānupapattiappears to be a factual impossibility of which one needs to be aware. In short:

Kumārila Prabhākara Śālikanātha

logicalinconsistency impossibilitywithout factualimpossibilityoneisawareabout

38I also discussed the topic from the point of view of Śālikanātha here:http://elisafreschi.com/2019/03/05/against-arthapatti-as-only-technically-distinguished-from-inference-in-salikanatha/39ThelogicalnatureoftheimpossibilityhasbeenshowninYoshimizuforthcoming.

28

5.3 Howisarthāpatticonceptualised?

Śālikanātha’sunderstandingofarthāpatti sharessomeelementswithKumārila’s,butalsodivergessignificantlyfromit.Whattheyshareis:

• arthāpattiisaseparateinstrumentofknowledge,differentthaninference

• Inarthāpattithereisaclashbetweenabackgroundknowledgeandanewlyacquiredone(inthestandardexample:betweenone’sknowingthatCaitraisaliveandone’sseeingthatheisnothome)

The firstpoint is a concern sharedbyallMīmāṃsakas andhotlydebated,mainly againstNaiyāyikaopponents.Bycontrast,Mīmāṃsakasdivergeasfor:

1. whatisthetrigger(andthethingknown)intheprocessofarthāpatti

2. theinterpretationofanyathānupapatti

3. theroleofdoubt

4. thereasonswhyarthāpattiisnotaninference

Concerningpoint1(seealsothetableinsection4.1),thetriggerisforKumārilatheabsencefrom home of Caitra specified by the fact that he is alive. By contrast, the trigger forŚālikanātha (see below, and Subrahmanya Sastri 1961, pp. 274t he) is the sheer absencefromhome.Howcanthisbe?AnopponentembeddedinŚālikanātha’stextcaneasilyobjectthat if thesheerabsence fromhomewere the triggerof thecognitionofaperson’sbeingoutside,thenitwouldleadonetoconcludethatalsoadeadpersonXoranunbornpersonY,whoarealsoabsentfromhome,aresomewhereoutside.ŚālikanāthacanexplainawaythisobjectionthroughpointNo.2.Infact,forKumārilatheclashbetweencognitionsisgivenintheverymomentyouenterCaitra’shomeand seehe isnot there. Since these cognitionscould not be possible otherwise, one postulates that Caitra is (alive and) outside. ForŚālikanātha, by contrast, one reaches the awareness of an impossibility through theexperiencedtriggeronly.Inotherwords,Caitra’sabsencefromhomemakeshisbeingalivelook impossible and the trigger is such exactly insofar as it leads one to think thatsomethingonethoughttoknowseemsimpossible.Accordingly,Caitra’sabsencefromhomeisatriggerofthecognitionofhisbeingoutsidebecauseitleadsonetothinkthatwhatoneknew about Caitra, namely his being alive, is impossible. By contrast, the absence fromhomeofX(adeadperson)orY(anunbornone)doesnothavethesameeffect,sinceitdoesnotleadtoanyimpossibility.AccordingtoŚālikanātha,thisimpossibilitymakesonedoubtforamomentaboutCaitra’sbeingalive(“Howcanhebealive,giventhatheisnothome?",SubrahmanyaSastri1961,p.273).Next,onepostulatessomething(suchasCaitra’sbeingoutside)whichappeasesthedoubtandtheseemingimpossibilityandharmonisesagainallone’scognitions.

Thus, Śālikanātha deeply innovates on Prabhākara not just concerning the role ofdoubt,butalsoconcerningtheidentificationofwhatdoesnotmakesenseandwhatisthegamya, which is no longer the being outside, but the being (alive and outside). Why?Possibly because he was convinced of a logical flaw of Prabhākara’s argument. In theṚjuvimalā he explains, in fact, that the existence outside in itself is not impossible. This

29

means that he is again reading anyathānupapatti as involving a (factual or logical)impossibility and not just as the fact that a given cognition does not take place (as inPrabhākara). In contrast to Prabhākara’s proposal, having seen a person’s absence fromhomeitisverymuch(logicallyorfactually)possiblethatsheisoutside.Norcanthebeingoutsideinitself, i.e. independentlyoftheexperienceofherabsencefromhome,besaidtobe(logicallyorfactually)impossible.Bycontrast,theonlythingwhichmaylooklogicallyorfactuallyimpossibleonceonehasexperiencedsomeone’sabsencefromhomeisherbeingstillalive.Śālikanāthacannonethelesskeep the identitybetweengamya andanupapannaelement, thus remaining faithful to a crucial point in Prabhākara’s account of arthāpatti.Śālikanāthacankeepthisidentityinsofarasboththegamyaandtheanupapannaelement(see the table below) aredefined as justbhāva ‘existence’. In the case of thegamya, thisexistenceisfurthertobeunderstoodastheexistenceconnectedwithanexternalplace(i.e.,bahirdeśasambaddhabhāva). By contrast, in the case of the definition of the anupapannaelement,‘existence’shouldbeunderstoodasasynonymof‘beingalive’.

In this connection it is also worth mentioning that, although Śālikanātha clearlyknew Kumārila, unlike Kumārila he did not enter into the detailed discussion of theelements of the inference and how they are missing in the case of arthāpatti. In fact,whereasKumārilaexplainedatlengthhowonecannotconstrueavalidlocusandprobanson the basis of the elements at hand in the standard example of arthāpatti, Śālikanāthachiefly deals with the identification of the trigger alone. Why so? Possibly for reasonssimilartotheonesdiscussedinthecaseofPrabhākara.KumārilaacceptedDiṅnāga’sformaldefinitionofinferenceasentailingalocus,aprobansandaprobandum,aswellasasimilarand dissimilar instances (see section 3), whereas Śālikanātha defines (perhaps lesstechnically) inference as a cognition of the other correlate due to the experience of onecorrelateforonewhoknowstheirinvariableconcomitance.40

Nonetheless, Śālikanātha probably considered that Prabhākara’s defence ofarthāpatti as a distinct instrument of knowledge was weakened by the flaw discussedabove,butalsobyhisloosetheoryofinference.Infact,Prabhākara’stheoryofinferenceisstillpre-Diṅnāga,whereasKumārilaandthenŚālikanāthacouldnot ignorethenewbasesforinferencelaiddownbyDiṅnāgaand,atleastinthecaseofŚālikanātha,Dharmakīrti(hisknowledge of the trairūpya distinction is evident in the Ṛjuvimalā, where the summaryversesrefertotheconceptofvipakṣa‘dissimilarinstance’).

Consequently, the distinction of a sambandhamātrajñāna vs. jñātasambandhitā(aboutwhich see section 4 above)would no longer be enough as the basis of a distinctinstrumentofknowledge,sincethedistinctiondependedonanimprecisedefinitionoftheconnectionatstake.Kumārila’ssolutionwouldnotbeviableeithersinceŚālikanāthasharesPrabhākara’s idea that in order for one toperceive an impossibility shemust have some

40Cf. jñātasambandhaniyamasyaikadeśasya darśanāt | ekadeśāntare buddhir anumānamabādhite||,PrP,pramāṇaparāyaṇa,anumānav.1(1961p.196)

30

expectation of what should have been the normal case, so that some (though vague)connectionmustbeknownbeforehand.

Therefore,unlikeKumārila(whomostlyfocusedontechnicalreasonsdistinguishingarthāpatti and inference and on the absence of prior knowledge of the relation) andPrabhākara (who focused only on the absence of a specific prior knowledge and on theimpossibilitypertainingtothegamyanottothegamaka),Śālikanāthachosetoadddoubtasthe distinct element of arthāpatti. He thus had a direct and easy way to distinguisharthāpattiandinference,butonethatriskedtojeopardisetheepistemicstatusofarthāpattiasaninstrumentofnecessarilyvalidcognition.

Inotherwords,onecomestovisitCaitraexpectinghimtobehomeandseesthatheisnotthere.Atthispoint,one’sbeliefabouthisbeingaliveisendangered,butonedoesnotwanttogiveitup.Immediately,onecomestotherightsolution:Heisalive,butelsewhere.Inlogicalterms,onecouldsaythatstartingwiththesituation

(A)*beingalive(beinghomebeingoutside)

onerisksabeliefrevision(asexplainedinsection4.1,“beingalive”couldbefalse),butdoesnotreallyperform it, since the twoconflictingbeliefsarenoton thesame level,and there is a metarule stating that the first one should not be given up, so that oneautomaticallyoptsforthesecondsolution(Caitraisoutside).However,thisreconstructionhas the disadvantage that it transgresses the requirement of Prabhākara that thegamyaand theanupapanna element are the same thing.Moreprecisely, theunityofgamya andanupapanna can be maintained only by cheating a little bit and making them the samealthoughoneofthetwo(theanupapanna)istheexistencealiveandtheother(thegamya)istheexistenceoutside.

In order to keep a real unity, one can, by contrast, suggest the followingreconstruction.One startswith the sambandhamātrajñāna thatwhoever is alivemust besomewhere,butmisconstructsitasbeingabidirectionalrelationlinkingbeingalive(C)andbeinghome(A)(wheneverthereisAthereisCandviceversa).OnethennoticesthatAisnot the case, something which would necessarily lead to the negation of C, if thebidirectional link were the case. However, one knows C to be the case, therefore onereconsiders the other seemingly conflicting belief, namely one’s initial belief, thesambandhamātrajñāna.41 Inthisway,oneunderstandsitasentailingnotthebidirectionallink, but a simple material implication, linking A and C. In a material implication, theantecedent is not negated if the consequent is negated (affirming that by negating theantecedentyounegatetheconsequent is the fallacyofdenyingtheantecedent).Similarly,smokeandfireare linkedbyamaterial implicationandinfact if thereisnosmoke,therecanstillbefire(e.g.,inthecaseofmeltinggold).Onceonehasrephrasedone’sinitialbelief

41Readers will notice that also in this case, the belief revision proceeds according to aparsimonyprinciple,seefn.35.

31

inthisway,onecandiscoverthatgiventheabsenceofA,thepersistenceofCneedstobethepersistenceofCunderstood ina slightlydifferentway, i.e., asCdevoidofA. It is stillbhāva‘existence’,butitisexistenceoutsideofone’shome,i.e.,itisbahirbhāva.

(B)Step1:beingalivebeingsomewhere(sambandhajñānamātra) Step2:(misattribution) Step3: Step4:((()))(anupapatti) Step5:(beliefrevision:newconstructionofsambandhajñānamātra) Step6:

This reconstruction makes clear also why the risk of doubt is not enough, asŚālikanāthaexplains,tomakethearthāpatti-deliveredcognitionunsure.Infact,Conlyrisksto beput intodoubt, but in fact one knows it surely to be the case and thereforemovesimmediatelytotherevisionofanother,weakerbelief.

This partial innovation concerning thegamya and theanupapanna elements leadsŚālikanātha to force the interpretation of a terse statement by Prabhākara. He reads astatementaspartofanobjectorandre-readsanotheronewithasomehowforcedsyntax(just like Prabhākara had done with Śabara). In the following quote, first comesPrabhākara’s textand its straightforward translation, thenŚālikanātha’s interpretationasexplainedintheṚjuvimalā:

kasyatarhi.bhāvasya,nacāsaugṛhābhāvadarśanenopapadyate.bāḍhaṃnopapadyate.nahigṛhābhāvadarśanenavinābahiḥbhāvaupapadyate.

P:Whatisthen[thegamya]?Theexistence.Infact,thisdoesnotmakesenseonceonehasseentheabsencefromhome.Itsurelydoesnotmakesense,for,withouttheexperienceoftheabsencefromhometheexistenceoutsidedoesnotmakesense.42

Ś:What is then[thegamya]? [Obj.:] It is theexistence[outside]. In fact, thisdoesnotmakesenseonceonehasseen the absence from home (so, thegamya would be the existence outside). [(Alleged) reply by Prabhākara:]Surelynot.This(existenceoutside)doesmakesense.For,onceonehasexperiencedtheabsencefromhome,itistheexistence,withouttheoutside,whichdoesnotmakesense(thus,thegamyaisthesheerexistence).

The following two tables sum up the key elements of arthāpatti in the authorsexamined:

gamaka gamya anupapannaelement

42Note that Prabhākara reads anyathā in anyathānupapatti as vinā ‘without’. AlsoŚālikanātha comments anyathāśabdena vināśabdārtha evocyate “With the word anyathā‘without’ only is meant” (Subrahmaniya Sastri 1967, p. 112). For various proposalsconcerningthemeaningofthispassageinPrabhākara,seesection4.1.

32

K jīvatogṛhābhāvadarśana bahirbhāva jīvatogṛhābhāva

P gṛhābhāvadarśana bahirbhāva gamya

Ś gṛhābhāvadarśana (bahirdeśe)bhāva (jīvana=)bhāva

sambandha anupapatti lakṣaṇaofarthāpatti

K existent,butnotknown logicalinconsistency nosambandhanopakṣadharmatā

nouniv.quantification

P sambandhamātrajñāna nottakingplace gamyasyānupapattisambandhamātrajñāna

Ś sambandhamātrajñāna factualimpossibilityofwhichoneisaware

gamyasyānupapattisambandhamātrajñāna

sandeha

ThefollowingtablesummarisestheinnovationsbyŚālikanātha:

distinctionfrominference gamya anupapannaelement

K ajñātasambandha bahirbhāva jīvatogṛhābhāva

P gamyasyānupapatti bahirbhāva bahirbhāva

Ś sandeha (bahirdeśe)bhāva (jīvana=)bhāva

5.3.1 AsecondPrābhākarawayofunderstandingarthāpatti

TheaboveinterpretationisadversedbysomeunidentifiedopponentswithinthePrP,whoclaimthattherealtriggeroftheprocessisnottheabsencefromhome,butthewell-known

33

fact of Caitra’s being alive. Caitra’s being alive itselfmakes one doubt about it and thenpostulate something else (namely, its being connected with a new location, outside ofhome),inordertoappeasethedoubt.Whythisdifferentinterpretation?Thereasonisnotspelt out in full, but Śālikanātha does say that this position depends on perception(darśana).Thismightmeanthatitdoesnotdependonsomethingunseen,likeanabsence.Inthissense,theseopponentsmightwanttoavoidtheideaofhavingabsencefromhomeasthe triggerof theprocess,because theywant toavoidabsences ingeneralasplayinganyrole, since, according to the Prābhākara epistemology, absences are nothing but theexistenceofsomethingelse.Accordingly,theseopponentstrytosaythatthebeingaliveisthenthrownintodoubtbythefactthatoneexperiencesitinanewform(rūpa),namely,inconnectionwithadifferentlocation.Thetextsummarisesthisandthepreviouspositionintwo verses (Subrahmanya Sastri 1961, p. 275), which also attempt an etymologicalexplanationofarthāpatti,analysedrespectivelyasthe‘falling[intodoubt]becauseofastateofaffairs’orasthe‘falling[intodoubt]ofastateofaffairs’.Thetwopositionsarethenalsoevokedsuccinctlyattheendofthegeneraldiscussion(SubrahmanyaSastri1961,p.278),wheretheactiveandpassiveunderstandingof‘falling[intodoubt]’arealsohighlighted:inonecase,thestressisonthetrigger,insofarasitcausessomethingelsetofallintodoubt,intheotheritisonthetrigger,insofarasitisitselfthethingwhichisfallenintodoubt.Thus,the thing which falls into doubt (the being alive) is the same in both cases, and thedifferencesarepartlylexicalandpartlymorefundamental:

1. Thething fallen intodoubt,e.g.,Caitra’sbeingalive, is fallen intodoubtbecauseofhisabsencefromhome(Śālikanātha)orbecauseofitsbeinginadifferentmodality(opponentswithinthePrābhākaragroup).

2. Thetriggerofthepostulation,whichthenappeasestheseeminginconsistency,isinonecasethethingwhichcausesthedoubt(Śālikanātha)andintheotherthethingfallenintodoubt(Prābhākaraopponents).

Therefore,therearemanysimilaritiesbetweenthetwoexplanationsandŚālikanāthaseemsmorethankeenonplayingdownthedifferences,byneglectingthefirstoneandexplainingthat thesecondone isnot important. Inhis summary(as foundat theendof thegeneraldiscussion, Subrahmanya Sastri 1961, p. 278), both interpretations see a same thing,namelytheabsencefromhome,puttingindoubtthesameotherthing,namelythefactthatCaitraisalive.Acontemporaryreadermay,however,notethattheseobjectorsnegatethemain characteristic of arthāpatti according to Prabhākara, namely the fact that it is thegamyaandnotthegamakawhichisanupapanna.

5.3.2 Śālikanātha’ssummaryverses

Like Kumārila in the TV and Jayanta in theNyāyamañjarī, Śālikanātha writes mostly inprose,butthenaddssomeversessummarisinghisposition.Theseversesarenotexplicitlyattributed and summarise distinctly Śālikanātha’s original contribution, so that one cansafelyattributethemtoŚālikanāthahimself.

TheinitialverseinŚālikanātha’spresentationinthePrPdescribesthearthāpattiasapostulation triggered by a seen thing (e.g., Caitra’s absence from home) which leads anunseenthing(Caitra’sbeingalive)toastateof logical impossibilityunlesssomethingelse(Caitra’sbeingoutside)ispostulated.

34

Furthertwoversesindicatetheendofasectionandthebeginningofanewoneinthe arthāpatti section of the PrP. Within the Ṛjuvimalā commentary on arthāpatti,Śālikanātha mentions seven verses which condense his contribution, at the end of thegeneraldiscussionandbeforeenteringintotherefutationofadistinctśrutārthāpatti.

The summary verses in theṚjuvimalā are also away for Śālikanātha to condensewhatheconsideredtobehiscontributiontotheteachingonarthāpatti.Thisisdescribedasbeingthreefold:1.Arthāpattiisestablishedtobeadistinctinstrumentofknowledge,2.Itisdistinctduetotheroleofdoubt,3.Botharthāpattiandinferencecanbeconstruedasalinkbetweenagamaka‘trigger’andagamya‘triggered’(thethingtobeknown),butinthecaseofinferenceitisthegamaka(forinstance,smoke)whichdoesnotmakesensewithoutthegamya,whereasinthecaseofarthāpattiitisthegamyaitselfwhichwouldnotmakesenseotherwise.

6 Howdoesthestorycontinue?

Thediscussionsonarthāpattishowhowdifferentauthorssupportingitsharehardlymorethan the label and the fact that arthāpatti is amethod of reasoning connecting premiseswithconclusionsbutdifferentthaninference.

Thisstateofaffairs isverymuchpresent intheearlyhistoryofarthāpatti,rangingfromtheTarkaśāstra(pre-Dignāga)asreconstructedbyGiuseppeTuccitoVātsyāyana(seeOberhammer, Prets, and Prandstetter 1991 and section 1.2). This article cannotaccommodatethesuccessivedevelopmentsofthedebate,butIwouldliketoanticipatethatlater authors continue to diverge in their approaches to arthāpatti. As a rapid survey,Vācaspati(10thc.)inhisNyāyawork(Nyāyavārttikatātparyaṭīkā)seemstohavebeenthefirst one to introduce the topic of virodha ‘conflict’ among cognitions as the standardinterpretationofanyathānupapatti.VācaspatiisalsoapparentlythefirstauthorreferringtoŚālikanātha’s position as just implying doubt, although he does not discuss it separatelyfrom the Bhāṭṭa version. This then became the standard way Nyāya and Bhāṭṭa authorsattacked the Prābhākara position.Cidānanda, a BhāṭṭaMīmāṃsāwho lived in the early14thc.,rephrasesthevirodhaasaconflictbetweenageneralandaspecificcognition.Thisidea has been thenpopularisedbyNārāyaṇaBhāṭṭa in his 17th c.Mānameyodaya,whichbecameastandardtextbookofBhāṭṭaMīmāṃsā.

Śāntarakṣita(8thc.),inhisTattvasaṅgraharepresentationofKumārila’sview,addsthatoneknowsthatCaitraisalivebecauseoflinguisticcommunication(śabda)(TS1644).Thismightmean that themention of śabdawas alreadypresent inKumārila’sBṛhaṭṭīkā.Gaṅgeśa (14th c.), possibly following some previous thinker(s), seems to have been thefirstonetodeveloprationalisationsforthecogniser’sreasonstobelievethatCaitrashouldbe home and that he is alive, namely the general habit of old people to stay home andastrologicalchartsaboutthelengthofhislife.Gaṅgeśaalsoattackedtheideaofdoubtasafoundationofarthāpatti,andmadeclearthatapurposeofarthāpattiistheestablishmentofapūrva.Kumāriladoes indeedusearthāpatti toestablishapūrva,buthedoesnotaddresstheproblemwithinthearthāpattisection.

35

RaghunāthaŚiromaṇi(16thc.),agenialcommentatoronGaṅgeśa,explainedhowarthāpatti, according to the own standards ofNyāya, should be a separate instrument ofknowledge(seeNilanjanDas(forthcoming)inKeating).

As forsupplyingpartsofamissingstatement, calledbyKumārilaśrutārthāpatti, italsotookapathwhichledit furtherawayfromdṛṣṭārthāpattiafterKumārila’sattempttobring them together as two cases of arthāpatti. After Jayanta’s mention of expectancy(ākāṅkṣā)inconnectionwithśrutārthāpatti,whichisfollowedbySucarita’semphasisonit(aboutwhich,seeFreschiforthcoming),śrutārthāpattibecomespartofadiscussionhavingimportant bearings on philosophy of language, and focusing especially on the case ofellipsis.

7 Conclusions

Arthāpatti is an interesting case study for the intersections of distinct currents alreadywithin early Mīmāṃsā. On the one hand, one notices a hermeneutic concern, whichprobably constitutes thehistorical reason for thedevelopment of arthāpatti according tothehypothesisbyLarsGöhler(Göhler2011).Ontheother,theepistemologicalconcerniswelldelineatedatanearlystageandbecomespre-eminentintheworkofKumārila.

The case of arthāpatti also shows how the two thinkers later credited with thecreation of the two subschools of Mīmāṃsā, namely Kumārila and Prabhākara, at timesignoredeachothercompletely.InthecaseofKumārila,hisneglectofPrabhākara’spositionmight be a sign of the fact that there were no important predecessors for the latter’sdistinctive positions on arthāpatti (especially for the identification of the nonsensicalelement with the thing to be known). In the case of Prabhākara, he attacks Kumārila’sposition in the case of the definition of arthāpatti in general, which could be a furtherevidenceofthefactthatPrabhākarawasinnovatingandneededtopersuadehisaudienceofhisnewinterpretations.Bycontrast,inthecaseofśrutārthāpattiPrabhākaradoesnotreferatall toKumārila’spositions,perhapsbecauseherePrabhākarawasjustgraspingbacktowhatheconsideredtobethemainstreampositionofMīmāṃsā.

Besides, the discussions on arthāpatti show how different Mīmāṃsā authorspresupposedandimplementeddifferentapproaches,correspondingtowhatcontemporaryreaderscouldidentifyasdifferentlogicaltheories,rangingfromnaturaldeductionandfirstorder logic (Kumārila)43 to belief revision logic (Prabhākara and Śālikanātha). The exactformalisation behind each of such approaches as conceived by Mīmāṃsā authorsthemselves will probably never be settled. Nonetheless, attempting to reconstruct andformalise them can help us inmaking sense of the ownwords of the various authors at

43For a fuller discussion of natural deduction and first order logic in Kumārila, seeYoshimizuforthcoming.

36

stakeandoftheseemingparadoxestheyentail(e.g.,asfortheidentityofthegamyaandtheanupapannaelementinŚālikanātha).

Abbreviations

BṛBṛhatībyPrabhākara,seeRāmanāthaŚāstrīandSubrahmanyaSastri1934–1967 PMSMīmāṃsāSūtrabyJaimini,seeApte(Āpaṭe)1929PrPPrakaraṇapañcikābyŚālikanāthaMiśra,seeSubrahmanyaSastri1961SBh Sabarabhasya by Sabara, see Apte (Apate) 1929 TV Tantravarttika by Kumarila, see Apte (Apate) 1929 YDYuktidīpikā,seeWezlerandMotegi1998

References Apte (Āpaṭe), Vināyaka Gaṇeśa, ed. (1929). Śrīmajjaiminipraṇīte Mīmāṃsādarśane [...]

prathamas tarkapādaḥ [...]śrīvaidyanāthaśāstripraṇītaprabhābhidhavyākhyāsametaśābarabhāṣyopetaḥ.Ānandāśramamudrāṇālaya.

Das, Nilanjan (forthcoming). “Raghunātha on Postulation”.In: Controversial Reasoning inIndian Philosophy: Major Texts and Arguments on Arthâpatti. Ed. by Malcolm C. Keating.Bloomsbury.Eltschinger,Vincent(2014).Buddhistepistemologyasapologetics:studiesonthehistory,self-

understandinganddogmaticfoundationsoflateIndianBuddhistphilosophy.Beiträgezur Kultur- und Geistesgeschichte Asiens. Wien: Verlag der ÖsterreichischenAkademiederWissenschaften.

Franco, Eli (forthcoming). “Xuanzang’s Silence and Dharmakīrti’s Dates”. In: WienerZeitschriftfürdieKundeSüdasiens.

Frauwallner,Erich,ed.(1968).MaterialienzurältestenErkenntnislehrederKarmamīmāṃsā.Kommission für Sprachen und Kulturen Süd- und Ostasiens; Philosophisch-Historische Klasse; Sitzungsberichte.Wien: Verlag der Österreichischen AkademiederWissenschaften.

Freschi, Elisa (forthcoming). “Reconstructing an episode in the history of philosophy:arthāpatti in Kumārila’s commentators”. In: ed. by Francesco Sferra and VincenzoVergiani.

Freschi, Elisa and Artemij Keidan (2017). “Understanding a philosophical text: a multi-disciplinaryapproachtotheproblemof“meaning”inJayanta’sNyayamañjarī,book5”. In: Reading Bhaṭṭa Jayanta on Buddhist Nominalism. Ed. by Patrick McAllister.Wien: Verlag der Österreichischen Akademie der Wissenschaften, pp. 251–290.Freschi,ElisaandAndrewOllett(forthcoming[a]).“arthāpattiinPrabhākara’sBṛhatīand its commentary”. In:ControversialReasoning in IndianPhilosophy:MajorTextsandArgumentsonArthâpatti.Ed.byMalcolmC.Keating.Bloomsbury.

— (forthcoming[b]). “arthāpatti in Kumārila’s Ślokavārttika and its commentaries”. In:Controversial Reasoning in Indian Philosophy: Major Texts and Arguments onArthâpatti.Ed.byMalcolmC.Keating.Bloomsbury.

37

Freschi, Elisa and Tiziana Pontillo (2013). Rule-extension-strategies in Ancient India:Śrautasūtra,MīmāṃsāandGrammarontantra-andprasaṅga-principles.Wien:PeterLang.

Gärdenfors,Peter(1992).“BeliefRevision:AnIntroduction”.In:BeliefRevision.EdbyPeterGärdenfors. Cambridge University Press, pp. 1—28. Available online at:https://pdfs.semanticscholar.org/2aae/6cdb60a582ec6a38d5d07266cbfa24210b5b.pdfGöhler,Lars(2011).ReflexionundRitualinderPūrvamīmāṃsā.StudiezurfrühenGeschichte

derPhilosophieinIndien.BeiträgezurIndologie44.Wiesbaden:Harrassowitz.

Krasser,Helmut(2012).“Bhāviveka,DharmakīrtiandKumārila”.In:Devadattīyam.JohannesBronkhorstFelicitationVolume.Ed.byFrançoisVoegelietal.WorldsofSouthandInnerAsia5.Bern:PeterLang,pp.535–594.

Oberhammer, Gerhard, Ernst Prets, and Joachim Prandstetter (1991). Terminologie derfrühen philosophischen Scholastik in Indien. Ein Begriffswörterbuch zur altindischenDialektik, Erkenntnislehre und Methodologie. Band 1: A-I. Ed. by GerhardOberhammer.Wien:VerlagderÖsterreichischenAkademiederWissenschaften.

Rāmanātha Śāstrī, S.K. and S. Subrahmanya Sastri, eds. (1934–1967). Bṛhatī [byPrabhākara]adŚābarabhāṣya.Madras:UniversityofMadras.

Steinkellner,Ernst (2013).Dharmakīrtis früheLogik.AnnotierteÜbersetzungder logischenTeilevonPramāṇavārttika1mitderVṛttiII.Introduction,Anmerkungen,Anhängeetc.StudiaPhilologicaBuddhicaMonographSeriesXXIXb.TheInternationalInstituteforBuddhistStudies.

Steinkellner, Ernst, Helmut Krasser, and Horst Lasic, eds. (2005). Jinendrabuddhi’sViśālāmalavatī Pramāṇasamuccayaṭīkā Chapter 2. Sanskrit Texts from the TibetanAutonomous Region. Beijing and Vienna: China Tibetology Publishing House andAustrianAcademyofSciences.

SubrahmaniyaSastri,S.,ed. (1967).BṛhatīofPrabhākaraMiśrawithṚjuvimalāPañcikāofŚālikanātha.MadrasUniversitySanskritSeries26.Madras:UniversityofMadras.

SubrahmanyaSastri,A.,ed.(1961).PrakaraṇaPañcikāofŚālikanāthaMiśrawiththeNyāya-Siddhi of Jaipuri Nārāyaṇa Bhaṭṭa. Darśana Series 4. Benares: Benares HinduUniversity.

Taber, John(1992).“WhatdidKumārilaBhaṭṭameanbysvataḥprāmāṇya?” In: JournaloftheAmericanOrientalSociety112.2,pp.204–221.

— (1996). “Is verbal testimony a form of inference?” In: Epistemology, Meaning andMetaphysicsafterMatilal. Ed.byArindamChakrabarti.Vol.3.2. Simla: Indian InstituteofAdvancedStudy,pp.19–31. — (2001).“Much Ado about Nothing: Kumārila, Śāntarakṣita, and Dharmakīrti on theCognitionofNon-Being”.In:JournaloftheAmericanOrientalSociety121.1,pp.72–88. —(2002).“Mohantyon sabdapramana”. In:JournalofIndianPhilosophy30,pp.161–190.— (2005). A Hindu critique of Buddhist epistemology: Kumarila on perception: the“Determinationofperception”chapterofKumarilaBhatta’sSlokavarttika/ translationandcommentary. RoutledgeCurzon Hindu Studies Series. London and New York: RoutledgeCurzon. Taber,John,HelmutKrasser,andVincentEltschinger(2012).“Translation[ofPVSV164.24–

176.16]”.In:CantheVedaSpeak?DharmakīrtiagainstMīmāṃsāexegeticsandVedicauthority. An annotated translation of PVSV 164,24–176,16. Wien: Verlag derÖsterreichischenAkademiederWissenschaften,pp.31–78.

38

Tucci, Giuseppe (1929). Pre-Diṅnāga Buddhist Texts on Logic from Chinese Sources.Gaekwad’sOrientalSeriesXLIX.Baroda:OrientalInstitute.

Wezler, Albrecht and Shujun Motegi, eds. (1998). Yuktidīpikā: The Most SignificantCommentaryontheSāṃkhyakārikā.Stuttgart:F.Steiner.

Yoshimizu, Kiyotaka (1997). Der “Organismus” des urheberlosen Veda: eine Studie derNiyoga-Lehre Prabhākaras mit ausgewählten Übersetzungen der Bṛhatī. Vol. 25.Wien:DeNobiliResearchLibrary.

— (1999). ““Śrutārthāpatti niyoru ninshiki no taishō ni tsuite” (On the object of thecognitionbyśrutārthāpatti)”.In:Bukkyō-gaku(JournalofBuddhistStudies)40,pp.1–22. —(2000a). “ChangeofViewonApūrva fromŚabarasvāmin toKumārila”. In:TheWay toLiberation.IndologicalStudiesinJapan.Ed.bySengakuMayeda.Vol.1.JapaneseStudiesonSouthAsia.Manohar,pp.149–165. —(2000b).“MīmāṃsāsūtraniokeruArthāpattitoAnumāna(ArthāpattiandAnumānainthe Mīmāṃsāsūtra)”. In: Indogaku Bukkyogaku Kenkyū (Journal of Indian and BuddhistStudies)48.2,pp.1115–1109. —(2007). “Kumārila’sPropositionalDerivation (Arthāpatti)withoutPervasion (Vyāpti)”.In:ExpandingandMergingHorizons.ContributionstoSouthAsianandCross-CulturalStudiesin Commemoration of Wilhelm Halbfass. Ed. by Karin Preisendanz. Wien: Verlag derÖsterreichischenAkademiederWissenschaften,pp.315–335. —(2015).“Kumārila’sCriticismofBuddhismasaReligiousMovementinHisViewsontheSourcesofDharma”. In:ActaAsiatica.Bulletinof the InstituteofEasternCulture.TheTōhōGakkai.SpecialIssue:BuddhismandDebate:TheDevelopmentofMahāyānaBuddhismandItsBackgroundinTermsofReligio-PhilosophicalHistory108,pp.43–62. — (forthcoming). “Physical Existence of a Living Being and Kumarila’s Theory ofarthapatti”.In:ControversialReasoninginIndianPhilosophy:MajorTextsandArgumentsonArthâpatti.Ed.byMalcolmC.Keating.Bloomsbury.