4LASTsummertheIndianpress carriedsensationalstoriesannouncing the final decipherment of the Harappan or IndusValley script. A United Newsof India dispatch on July 11, 1999,picked up throughout South Asia, reported on new research bynoted historian, N.S. Rajaram, who along with palaeographist Dr.Natwar Jha, hasread and deciphered the messageson more than2,000 Harappan seals. Discussion of the messageswaspromisedin Rajaram and Jhasupcoming book, TheDeciphered IndusScript.For nearly a year, the Internet wasabuzz with reportsthat Rajaramand Jha had decoded the full corpusof IndusValley texts.This was not the first claim that the writing of the IndusValley Civilisation (fl. c. 2600-1900 BCE) had been cracked. Ina 1996 book, American archaeologist Gregory Possehl reviewedthirty-five attempted decipherments, perhapsone-third the actu-al number. But the claimsof Rajaram and Jha went far beyondthose of any recent historians. Not only had the principles ofdeciphermentbeendiscovered,buttheentirecorpus oftextscould now be read. Even more remarkable were the historicalconclusionsthat Rajaram and hiscollaborator said were backedby the decoded messages. TheUNIstorywas triggeredbyannouncements thatRajaram and Jha had not only deciphered the IndusValley sealsbut had read pre-Harappan textsdating to the mid-fourth mil-lennium BCE. If confirmed, thismeant that they had decodedmankindsearliest literary message. The texts were a handfulofsymbols scratchedonapotterytabletrecently discovered by Harvard Universityarchaeologist Richard Meadow. The oldestof these, Rajaram told the UNI, wasa textthat could be translated Ila surroundstheblessed land an oblique but unmistakablereferencetotheRigvedas Saraswatiriver.The suggestion wasthat mansearliest mes-sagewas linkedtoIndias oldestreligioustext.1The claim washardly trivial, since thiswasover 2,000 yearsbefore Indologistsdatethe Rigveda and more than 1,000 yearsbeforeHarappancultureitselfreachedmaturity. Raj aram s Worl dAfter months of media hype, Rajaramand JhasTheDeciphered IndusScript2madeit to print in New Delhi early thisyear. By midsummer the bookhad reached the West and wasbeing heatedly discussed via theInternet in Europe, India, and the United States. The book gavecredit for the decipherment method to Jha, a provincial religiousscholar, previously unknown, from Farakka, in West Bengal. Thebookspublicity hailshim asone of the worldsforemost Vedicscholarsand palaeographers. Jha had reportedly worked in iso-lationfortwentyyears,publishingacurious 60-pageEnglishpamphlet on his work in 1996. Jhas study caught the eye ofRajaram,whowas alreadynotorious inIndologicalcircles.Rajaram took credit for writing most of the book, which heavi-ly politicised Jhas largely apolitical message. Rajarams onlinebiography claimsthat their joint effort isthe most importantbreakthrough of our time in the history of Indian history andculture.Boasts likethis donotsurprisebattle-scarredIndologistsfamiliar with Rajaramswork. A U.S. engineering professor inthe 1980s, Rajaram re-invented himself in the 1990sasa fieryHindutva propagandist and revisionist historian. By the mid-1990s, he could claim a following in India and in migrcirclesin the U.S. In manufacturing hispublic image, Rajaram tradedheavily on claims, not justified by his modest research career,that before turning to history he was one of Americas best-known workersin artificial intelligence and robotics. Hyperboleaboundsin hisonline biography, posted at the ironically namedSwordofTruthwebsite.TheHindutvapropaganda site, located in the United States,pictures Rajaramas aworld-renownedexpertonVedicmathematicsandanauthorityonthehistoryofChristianity.The last claim issupported by violently anti-Christianworks carryingtitles likeChristianitys Collapsing Empire andItsDesignsin India. Rajaramspapersinclude hisSearch for the historical Krishna (found inthe IndusValley c. 3100 BCE); attack a longlistofHindutvaenemiesincludingChristianmissionaries,Marxistacademics,leftistpoliticians,IndianMuslims,andWesternIndologists;andglorifythemobdestruction of the Babri Mosque in 1992 asa symbol of Indiasemergence from the gripof alien imperialistic forcesand their surro-HORSEPLAY IN HARAPPA The Indus Valley Decipherment Hoax MICHAEL WITZEL, a Harvard University Indologist, and STEVEFARMER, a comparative historian, report on media hype, faked data, and Hindutva propaganda in recent claims that the Indus Valley script has been decoded.C O V E R S T O R Y1 For theUNI dispatch, seehttp://www.indiaserver.com/thehindu/1999/07/12/stories/0212000l.htm. Typically enough, in light of what weshow below, Rajaram misidentified theearlytextdiscoveredbyMeadow,workingoffa photoofa different potsherdpublishedinerrorbya BBCreporter.Forthe storyofthis Rajaramfiasco,withlinks,seehttp://www.safarmer.com/meadow.html.2 N. Jhaand N.S. Rajaram, TheDeciphered IndusScript: Methodology, readings, interpretations, AdityaPrakashan, New Delhi, 2000; pagesxxvii + 269, Rs. 950.5gates. All Indian history, Rajaram writes, can be pictured asastruggle between nationalistic and imperialistic forces. In Indology, the imperialistic enemy isthe colonial-mission-ary creation known asthe Aryan invasion model, which Rajaramascribesto Indologistslong after crude invasion theorieshave beenreplaced by more sophisticated acculturation modelsby seriousresearchers. Rajaramscartoon image of Indology isto be replacedby a path of study that combinesancient learning and modernscience. What Rajaram meansby science issuggested in one ofhis papers describing the knowledge of the Rigveda poets. TheRigveda rishis, we find, packed their hymnswith occult allusionsto high-energy physics, anti-matter, the inflational theory of theuniverse, calculationsof the speed of light, and gamma-ray burstsstriking the earth three timesa day. The latter isshown in threeRigveda verses(3.56.6, 7.11.3, 9.86.18) addressed to the god Agni.ThesecondRajaramtranslates:OAgni!Weknowyouhavewealth to give three timesa day to mortals.One of Rajaramsearly Hindutva pieceswaswritten in 1995with David Frawley, a Western New Age writer who likestofind allusions to American Indians in the Rigveda. Frawley istransformed via the Sword of Truth into a famousAmericanVedic scholar and historian. The book by Rajaram and Frawleyproposesthe curiousthesisthat the Rigveda wasthe product ofacomplexurbanandmaritimecivilisation,nottheprimitivehorse-and-chariot culture seen in the text. The goal isto link theRigveda to the earlier IndusValley Civilisation, undercutting anypossibilityoflaterAryanmigrations orrelocations oftheRigveda to foreign soil. Ancient India, working through a mas-sive (but lost) Harappan literature, wasa prime source of civili-sation to the West. TheDeciphered IndusScript makessimilar claimswith dif-ferent weapons. The Indus-Saraswati Valley again becomesthehome of the Rigveda and a font of higher civilisation: Babylonianand Greek mathematics, all alphabetical scripts, and even Romannumeralsflow out to the world from the IndusValleysinfinitelyfertile cultural womb. Pressreleasespraise the work for not onlysolvingthemostsignificanttechnicalprobleminhistoricalresearch of our time deciphering the Indusscript but fordemonstrating aswell that if any cradle of civilisation existed,it waslocated not in Mesopotamia but in the Saraswati Valley.The decoded messagesof Harappa thusconfirm the Hindutvapropagandistswildest nationalistic dreams. Raj aram s Pi l t down HorseNot unexpectedly, Indologistsfollowed the pre-presspub-licity for Rajaramsbook with a mix of curiosity and scepticism.Just asthe book hit the West, a lively Internet debate wasunderway over whether anysubstantial textsexisted in Harappa letalone the massive lost literature claimed by Rajaram. IndusValleytextsare cryptic to extremes, and the script showsfew signsofevolutionary change. Most inscriptionsare no more than four orfive characterslong; many contain only two or three characters.Moreover, character shapes in mature Harappan appear to bestrangelyfrozen,unlikeanythingseeninancientMesopotamia,EgyptorChina.This suggests thatexpectedscribal pressures for simplifying the script, arising out of therepeated copying of long texts, waslacking. And if thisistrue,the Indusscript may have never evolved beyond a simple proto-writing system. Once Rajaramsbook could actually be read, the initial scep-ticism of Indologiststurned to howlsof disbelief followed bychargesof fraud. It wasquickly shown that the methodsof JhaHarappa, area of t he paral l elwal l s.Court esy of t he Archaeol ogi calSurvey of Indi a, Punj ab Phot ographi c Vol ume 463/ 86.6and Rajaram were so flexible that virtually any desired messagecould be read into the texts. One Indologist claimed that usingmethodslike these he could show that the inscriptionswere writ-ten in Old Norse or Old English. Otherspointed to the fact thatthedecodedmessages repeatedlyturnedupmissinglinksbetween Harappan and Vedic cultures supporting RajaramsHindutvarevisions ofhistory.ThelanguageofHarappawasdeclared to be late Vedic Sanskrit, some 2,000 yearsbefore thelanguage itself existed. Through the decoded messages, the horse-lessIndusValley Civilisation distinguishing it sharply from theculture of the Rigveda wasawash with horses, horse keepers,and even horse rustlers. To support hisclaims, Rajaram pointedto a blurry image of a horse seal the first pictorial evidenceever claimed of Harappan horses. Chaosfollowed. Within weeks, the two of usdemonstratedthat Rajaramshorse seal wasa fraud, created from a comput-er distortion of a broken unicorn bull seal. Thisled Indologistwagsto dub it the IndusValley Piltdown horse a comic allu-sion to the Piltdown man hoax of the early twentieth century.The comparison was, in fact, apt, since the Piltdown man wascreated to fill the missing link between ape and man just asRajaramshorse seal wasintended to fill a gap between Harappaand Vedic cultures. Once the hoax wasuncovered, $1000 wasoffered to anyonewho could find oneHarappan researcher who endorsed Rajaramshorse seal. The offer found no takers. The Piltdown horse story hasitscomic side, but it touch-eson a central problem in Indian history. Horseswere critical toVedic civilisation, aswe see in Vedic textsdescribing horse sac-rifices, horse raids, and warfare using horse-drawn chariots. IfRigvedic culture (normally dated to the last half of the secondmillennium BCE) isidentified with Harappa, it iscritical to findevidence of extensive use of domesticated horsesin India in thethird millennium BCE. In the case of Hindutva revisionistslike Rajaram, who push the Rigveda to the fourth or even fifthmillennium, the problem isworse. They must find domesticat-ed horsesand chariotsin South Asia thousandsof yearsbeforeeither existed anywhere on the planet. Evidence suggeststhat the horse (Equuscaballus) wasabsentfrom India before around 2000 BCE, or even as late as 1700BCE,whenarchaeologyfirstattests its presenceintheIndusplainsbelow the Bolan pass. The horse, a steppe animal from thesemi-temperate zone, wasnot referred to in the Middle East untiltheendofthethirdmillennium,whenitfirstshows upinSumerian asanshe.kur (mountain ass) or anshe.zi.zi (speedy ass).Before horses, the only equidsin the Near East were the donkeyandthehalf-ass (hemione,onager).Thenearlyuntrainablehemioneslook a bit like horsesand can interbreed with them, ascandonkeys.InIndia,thehemioneorkhor (Equushemionuskhur) wasthe only equid known before the horse; a few speci-mensstill survive in the Rann of Kutch. The appearance of domesticated horses in the Old Worldwasclosely linked to the development of lightweight chariots,which play a central role in the Rigveda. The oldest archaeolog-ical remainsof chariotsare from east and west of the Ural moun-tains, where they appear c. 2000 BCE. In the Near East, theiruse isattested in picturesand writing a little later. A superb fif-teenth-century Egyptian example survives intact (in Florence,Italy); othersshow up in twelfth-century Chinese tombs. Chariotslike these were high-tech creations: the polesof theEgyptian example were made of elm, the wheels felloes(outerrim) of ash, itsaxlesand spokesof evergreen oak, and itsspokelashingsof birch bark. None of these treesare found in the NearEast south of Armenia, implying that these materialswere import-ed from the north. The Egyptian example weighsonly 30 kg orso, a tiny fraction of slow and heavy oxen-drawn wagons, weigh-ing 500 kg or more, which earlier served as the main wheeledtransport. These wagons, known since around 3000 BCE, areFi gure 7.1a: The Horse Seal (Mackay 453) Fi gure 7.1b: The Horse Seal (Art i st s reproduct i on)Rajarams computer enhancement of Mackay 453, transforming it into a horse seal(From the book The Deciphered Indus Script, p. 177)7similar to those still seen in partsof the Indian countryside.The result of all thisisthat the claim that horsesor chariotswere found in the IndusValley of the third millennium BCE isquite a stretch. The problem isimpossible for writerslike Rajaramwho imagine the Rigveda early in the fourth or even fifth mil-lennium, which islong before anywheeled transport let alonechariots existed. Even the late Hungarian palaeontologist S.Bknyi, who thought that he recognised horses bonesat oneIndussite, Surkotada, denied that these were indigenousto SouthAsia. He writesthat horsesreached the Indian subcontinent inanalreadydomesticatedformcomingfromtheInnerAsiatichorse domestication centres. HarvardsRichard Meadow, whodiscovered the earliest known Harappan text (which Rajaramclaims tohavedeciphered),disputes eventheSurkotadaevi-dence. In a paper written with the young Indian scholar, AjitaK. Patel, Meadow argues that not oneclear example of horsebones exists in Indus excavations or elsewhere in North Indiabefore c. 2000 BCE.3All contrary claims arise from evidencefrom ditches, erosional deposits, pitsor horse gravesoriginatinghundredsor even thousandsof yearslater than Harappan civil-isation. Remainsof horses claimed by early Harappan archae-ologistsin the 1930swere not documented well enough to let usdistinguish between horses, hemiones, or asses. All thisexplainsthe need for Rajaramshorse inscriptionsandhorse seal. If this evidence were genuine, it would trigger amajor rethinking of all Old World history. Rajaram writes, inhisaccustomed polemical style:The horse seal goesto show that the oft repeated claimof No horse at Harappa isentirely baseless. Horse boneshave been found at all levelsat Harappan sites. Also... theword asva (horse) isa commonly occuring (sic) wordontheseals.ThesupposedhorselessnessoftheHarappansisa dogma that hasbeen exploded by evidence.But like its cousintheAryaninvasion,itpersists forreasonshaving little to do with evidence or scholarship. Rajaramshorse, which lookssomething like a deer to mostpeople,is abadlydistortedimageprintednexttoanartistsreproduction of a horse, located below a Harappan inscription.4The original source of the image, Mackay 453, isa tiny photoon Plate XCV of Vol. II of Ernest MackaysFurther Excavationsof Mohenjo-Daro(NewDelhi,1937-38).Thephotowas sur-prisingly difficult to track down, since Rajaramsbook doesnottell you in which of Mackaysarchaeological works, which con-tain thousandsof images, the photo islocated. Finding it andothersrelated to it required coordinating resourcesin two of theworldsbest research libraries, located 3,000 milesapart in theUnited States. Once the original wasfound, and compared over the Internetwith hisdistorted image, Rajaram let it slip that the horse sealwasa computer enhancement that he and Jha introduced tofacilitate our reading. Even now, however, he claimsthat theseal depictsa horse. To deny it would be disastrous, since todo so would require rejection of his decipherment of the sealinscription which supposedly includesthe word horse.OnceyouseeMackays originalphoto,itis clearthatRajaramshorse seal is simply a broken unicorn bull seal,the most common seal type found in Mohenjo-daro. In context,itsidentity isobvious, since the same page containsphotosofmore than two dozen unicorn bulls any one of which wouldmake a good horse seal if it were cracked in the right place. What in Rajaramscomputer enhancement lookslike theneck and head of a deer isa Rorschach illusion created bydistortion of the crack and top-right part of the inscription. Anysuggestion that the seal representsa whole animal evaporatesassoon asyou see the original. The fact that the seal isbroken isnot mentioned in Rajaramsbook. You certainly cannot tell it isbroken from the computer enhancement.While Rajaramsbogushorse seal iscrude, because of therelative rarity of the volume containing the original, which isnotproperlyreferencedinRajarams book,onlyahandfulofresearcherslucky enough to have the right sourcesat hand couldtrack it down. Rajaramsevidence could not be checked by histypical reader in Ahmedabad, say or even by Indologistsusingmost university libraries. The character of the original seal becomesclearer when youlook more closely at the evidence. Mackay 453, it turnsout, isnot the photo of a seal at all, asRajaram claims, but of a mod-ern clay impression of a seal (field number DK-6664) dug up inMohenjo-daro during the 1927-31 excavations. We have locat-ed a superb photograph of the original seal that made the impres-sion(identifiedagainbyfieldnumberDK-6664)intheindispensableCorpus ofIndus Seals andInscriptions (Vol.II:Helsinki 1991, p. 63). The work wasproduced by archaeologistsfromIndiaandPakistan,coordinatedbytherenownedIndologist Asko Parpola. According to a personal communica-tion from Dr. Parpola, the original seal was photographed inPakistan by Jyrki Lyytikk specifically for the 1991 publication. Like everyone else looking at the original, Parpola notesthatRajaramshorse seal issimply a broken unicorn bull seal, oneof numerousexamplesfound at Mohenjo-daro. Rajaram hasalsoapparently been told thisby Iravatham Mahadevan, the leadingMackay 453 before i t s comput er enhancement by Raj aram.When you l ook att he ori gi nalpi ct ure, i ti s cl ear t hatt he seali mpressi on i s cracked.4 For theoriginal story of thedebunking of thehorseseal, with linksto other evidence, seehttp://www.safarmer.com/horseseal/update.html.3 Seethecomment by Meadow and Patel on Bknyiswork in South Asian Studies13, 1997, pp. 308-315.9Indian expert on the Indusscript. Mahadevan isquoted, with-out name, in Rajaramsbook asa well known Dravidianistwhopointedouttohimtheobvious.But,Rajaraminsists,acomparison of the two creatures[unicornsand horses], espe-cially in [the] genital area, showsthisto be fallacious. Rajaramhasalso claimed on the Internet that the animalsbushy tailshowsthat it isa horse. Below, on the left, we have reproduced Lyytikkscrisp photooftheoriginalseal,compared(ontheright)withtheseven-decade-oldphoto(Mackay453)oftheimpressionRajaramclaimsisa horse seal. We have flipped the image of the origi-nal horizontally to simplify comparison of the seal and impres-sion. The tail of the animal isthe typical rope tail associatedwith unicorn bull sealsat Mohenjo-daro (seen in more imagesbelow). It isclearly not the bushy tail that Rajaram imagines although Rajaramsstory iscertainly a bushy horse tale. Checking Rajaramsclaimsabout the genital area, we findno genitalsat all in M-772A or Mackay 453 for the simple rea-son that genitalson unicorn bullsare typically located right wherethe seal iscracked! Thisisclear when we look at other unicornsealsor their impressions. One seal impression, Parpola M-1034a(on the right), hasa lot in common with Rajaramshorse seal,including the two characterson the lefthand side of the inscrip-tion. The seal isbroken in a different place, wiping out the right-hand side of the inscription but leaving the genitalsintact. Onthis sealimpressionweseethedistinctiveunicorngenitals,identifiedbythelongtufthangingstraightdown.The genitals arelocatedwherewewouldfindthemonRajaramshorse seal, if the latter were not broken. Otherunicornbullsealimpressions,liketheoneseeninParpola M-595a (see next page), could make terrific horse sealsif cracked in the same place. Unfortunately, Parpola M-595a isnot broken, revealing the fact (true of most Harappan seals) thatit representsnot a real but a mythological animal. (And, of course,neither thisnor any other unicorn hasa bushy tail.) A Russian Indologist, Yaroslav Vassilkov, haspointed to asuspiciousdetail in Rajaramscomputer enhancement that isnot found on any photo of the seal or impression. Just in frontAs shown by t hei r i dent i calarchaeol ogi calfi el d numbers (DK-6664), M-772A (publ i shed i n Vol . II of Corpus of Indus Seal s andInscri pt i ons, 1991) i s t he ori gi nalsealt hatseven decades ago creat ed t he seali mpressi on (Mackay 453) t hatRaj aram cl ai msi s a horse seal .M-772A (fl i pped hori zont al l y)Mackay 453M-1034a10of the animal, we find a small objectthat lookslike a partial image of a com-moniconinanimalseals:afeedingtrough that looksa little like an old-styletelephone.Whoinserteditintothe distorted image of the horse sealis notknown.Rajaramhas notresponded to questionsabout it. Below, we show Rajarams com-puter enhancement next to picturesofMohenjo-daro copper platesthat con-tain several versionsof the object. Lat e Vedi cSanskri t 2000Years Bef ore Schedul e Thehorsesealis onlyonecaseofbogus datainRajarams book.Knowledge of Vedic Sanskrit isneededtouncoverthoseinvolvinghis deci-pherments. That isnot knowledge thatRajaramwouldexpectinhis averagereader, since (despite itspretensions) thebook isnot aimed at scholarsbut at a layIndian audience. The pretence that thebookis addressedtoresearchers (towhom the fraud isobvious) isa smoke-screentoconvincelayreaders thatRajaram isa serioushistorical scholar. Thedeciphermentissueexplainswhy Rajaram continuesto defend hishorseseallongafterhis ownsup-portershave called on him to repudi-ate it. He haslittle choice, since he haspermanentlyweddedhis Piltdownhorsetohis deciphermentmethod.The inscription over the horse, he tellsus,reads (abitungrammatically)arko-hasvaor arko ha asva Sunindeed like the horse (sic). The read-ingclearlywouldbepointless iftheimagerepresentedaunicornbull.Rajaramclaims thattherearelinksbetween thisdeciphered text and alaterVedicreligious document,theShukla Yajurveda. This again pushestheRigveda,whichis linguisticallymuchearlierthanthattext,toanabsurdly early period. As we have seen, Rajaram claimsthat the language of Harappa waslateVedicSanskrit.This conflicts withcountlessfactsfrom archaeology, lin-guistics, and other fields. Indeed, lateVedicdidnotexistuntilsometwothousand yearsafter the start of matureHarappan culture! Let uslook at a little linguistic evi-dence. Some of it isa bit technical, butit isuseful since it showshow datesareassigned to partsof ancient Indian his-tory. The Rigveda isfull of descriptionsof horses(asva), horse races, and theM-595aFi gure 7.1a: The Horse Seal (Mackay 453)Raj aram s comput er enhancement of Mackay 453 on t he l eft ; t he arrow poi nt s t o anobj ectapparent l y st uck i nt o t he ori gi nali mage. On t he ri ght , pi ct ures of Mohenj o-darocopper pl at es showi ng si mi l ar t el ephone-l i ke feedi ng t roughs.11swift spoke-wheeled chariot (ratha). We have already seen thatnone of these existed anywhere in the Old World until around2000 BCE or so. In most places, they did not appear until muchlater. The introduction of chariotsand horsesisone marker forthe earliest possible datesof the Rigveda. Linguistic evidence providesother markers. In both ancientIran and Vedic India, the chariot iscalled a ratha, from the pre-historic (reconstructed) Indo-European word for wheel *roth2o-(Latin rota, German Rad). (A chariot = wheels, just asin themodern slang expression my wheels = my automobile.) Wealso have shared Iranian and Vedic wordsfor charioteer theVedic ratheSTha or old Iranian rathaeshta, meaning standingon the chariot. Indo-European, on the other hand the ances-tor of Vedic Sanskrit and most European languages doesnothave a word for chariot. This is shown by the fact that manyEuropean languagesuse different wordsfor the vehicle. In thecase of Greek, for example, a chariot isharmat(-os). The implication isthat the ancient Iranian and Vedic wordfor chariot wascoined sometime around 2000 BCE about whenchariotsfirst appeared but before those languagessplit into two.A good guessisthat thisoccurred in the steppe belt of Russia andKazakhstan, which iswhere we find the first remainsof chariots.That area remained Iranian-speaking well into the classical peri-od, a fact reflected even today in northern river names all theway from the Danube, Don, Dnyestr, Dnyepr and the Ural (Rahaa= Vedic Rasaa) riversto the Oxus(Vakhsh). These are only a few piecesof evidence confirming what lin-guistshave known for 150 years: that Vedic Sanskrit wasnot nativeto South Asia but an import, like closely related old Iranian. Theirusual assumed originsare located in the steppe belt to the northof Iran and northwest of India. Thisview issupported by recent linguistic discoveries. One isthat approximately 4 per cent of the wordsin the Rigveda do notfit Indo-Aryan (Sanskrit) word patterns but appear to be loansfrom a local language in the Greater Panjab. That language iscloseto, but not identical with, the Munda languagesof Central andEast India and to Khasi in Meghalaya. A second finding pertainsto shared loan wordsin the Rigveda and Zoroastrian textsrefer-ring to agricultural products, animals, and domestic goodsthat weknowfromarchaeologyfirstappearedinBactria-Margianac.2100-1700 BCE. These include, among others, wordsfor camel(uSTra/ushtra), donkey (khara/xara), and bricks(iSTakaa/ishtiia,ishtuua). The evidence suggeststhat both the Iraniansand Indo-Aryansborrowed these wordswhen they migrated through thisregion towardstheir later homelands.5A third find relatesto Indo-Aryan loan wordsthat show up in the non-Aryan Mitanni of north-ern Iraq and Syria c.1400 BCE. These loan wordsreflect slightlyolder Indo-Aryan formsthan those found in the Rigveda. Thisevi-dence isone reason why Indologistsplace the composition of theRigveda in the last half of the second millennium. Thisevidence, and much more like it, showsthat the claim byRajaram that mature Harappansspoke late Vedic Sanskrit thelanguage of the Vedic sutras(dating to the second half of the firstmillennium) isoff by at least two thousand years! At best, a fewadventurousspeakersmay have existed in Harappa of some earlyancestor of old Vedic Sanskrit the much later language of theRigveda trickling into the Greater Panjab from migrant Aryantribes. These early Indo-Aryan speakerscould have mingled withothersin the townsand citiesof Harappan civilisation, which wereconceivablyjustas multilingualas anymoderncityinIndia.(Indeed, Rigvedic loan wordsseem to suggest several substrate lan-guages.) But to have all, or even part, of Harappansspeaking lateVedic ispatently absurd. But thisevidence pertainsto what Rajaram representsasthepetty conjectural pseudo-science called linguistics. By rejectingthe science wholesale, he giveshimself the freedom to invent Indianhistory at hiswhim. Consonant s CountLit t le, Vowels Not hing! According to Rajaram and Jha, the Induswriting system wasa proto-alphabetical system, supposedly derived from a complex(now lost) system of pre-Induspictorial signs. Faced with a mul-titude of Harappan characters, variously numbered between 400and 800, they select a much smaller subset of charactersand readthem asalphabetical signs. Their adoption of these signsfollowsfromtheallegedresemblances ofthesesigns tocharacters inBrahmi, the ancestor of later Indian scripts. (Thiswasthe scriptadopted c. 250 BCE by Asoka, whom Jhas1996 book assignstoc. 1500 BCE!) Unlike Brahmi, which letsyou write Indian wordsphonetically, the alphabet imagined by Jha and Rajaram ishigh-ly defective, made up only of consonants, a few numbers, and somespecial-purpose signs. The hundredsof left-over pictorial signsnormally stand for single words. Whenever needed, however andthisgoesfor numbersaswell they can also be tapped for theirsupposed sound values, giving Rajaram and Jha extraordinary free-dom in making their readings. The only true vowel that Jha andRajaram allow isa single wildcard sign that standsfor anyinitialvowel as in A-gni or I-ndra or sometimes for semi-vowels.Vowelsinside wordscan be imagined at whim. Vowelswere lacking in some early Semitic scripts, but far fewervowelsare required in Semitic languagesthan in vowel-rich Indianlanguageslike Sanskrit or Munda. In Vedic Sanskrit, any writingsystem lacking vowelswould be so ambiguousthat it would beuseless. In the fictional system invented by Jha and Rajaram, forexample, the supposed Induska sign can be read kaa, ki, ku, ke,ko, etc., or can also represent the isolated consonant k. A script likethisopensthe door to an enormousnumber of alternate readings. Supposing with Jha and Rajaram that the language of Harappawaslate Vedic, we would find that the simple two-letter inscrip-tion mnmight be read:mana ornament; manaH mind (since Rajaram letsusadd the Visarjaniya or final -H at will); manaa zealoraweight;manuManu;maanaopinionorbuilding or thinker; miina fish; miinein a fish;miinau two fish; miinaiH with fish; muni Muni,Rishi, ascetic; mRn- made of clay; menaa wife;meni revenge; menehe hasthought; mauna silence;and so on. There are dozensof other possibilities. How isthe poor read-er, presented with our two-character seal, supposed to decide ifit refersto revenge, a sage, the great Manu, a fish, or hiswife?The lordsof Harappa or Dholavira, instead of using the scripton their seals, would have undoubtedly sent itsinventor off tofinish hisshort and nasty life in the copper minesof the Aravallis!If all of thiswere not enough to drive any reader mad, Rajaramand Jha introduce a host of other devicesthat permit even freerreadingsof inscriptions. The most ridiculousinvolvestheir claimthat the direction of individual inscriptionsfollowsno hard andfast rules. Thismeansthat if tossing in vowelsat will in our mninscriptiondoes notgiveyouthereadingyouwant,youcanrestart your reading (again, with unlimited vowel wildcards) fromthe opposite direction yielding further alternativeslike namaH5 For linguistic details, seeM. Witzel, SubstrateLanguagesin Old Indo-Aryan (Rigvedic, Middleand LateVedic), ElectronicJournal of VedicSanskrit, Vol. 5 (1999), Issue1 (September),availablein PDF format from http://www1.shore.net/~india/ejvs/ejvs0501/ejvs0501article.pdf. Seealso F. Staal in TheBook Review, Vol. XXIV, Jan.-Feb., 2000, p.17-20.or namo honour to..., naama name, and so on. There are other principles like this. A number of signsrep-resent the samesound, while conversely the same sign canrepresent different sounds. With some 400-800 signsto choosefrom, thisgivesyou unlimited creative freedom. AsRajaram putsit deadpan, Harappan isa rough and ready script. Principleslike thisgave itsscribesseveral waysin which to expressthe samesounds, and write wordsin different ways. All thisisstated insuch a matter-of-fact and scientific manner that the non-spe-cialist getshardly a clue that he isbeing had. In other words, figure out what reading you want and fill inthe blanks! AsVoltaire supposedly said of similar linguistic tricks:Consonantscount little, and vowelsnothing.A little guidance on writing direction comesfrom the wild-card vowel sign, which Rajaram tellsususually comesat the startof inscriptions. Thisiswhy such a large number of messagesonthe Indussealshave thisvowel symbol asthe first letter. WhatJha and Rajaram refer to asa vowel (or semi-vowel) sign istheHarappan rimmed vessel or U-shaped symbol. Thisisthe mostcommon sign in the script, occurring by some countssome 1,400timesin known texts. It ismost commonly seen on the left sideof inscriptions. Back in the 1960s, B.B. Lal, former Director-General of theArchaeological Survey of India, convincingly showed, partly bystudying how overlapping characterswere inscribed on pottery,that the Harappan script wasnormally read from right to left.Much other hard evidence confirming thisview hasbeen knownsince the early 1930s. Thismeansthat in the vast majority ofcasesthe U-sign isthe last sign of an inscription. But here, assooften elsewhere, Rajaram and Jha simply ignore well-establishedfacts, since they are intent on reading Harappan left to right toconform to late Vedic Sanskrit. (In timesof interpretive need,however, any direction goes including reading inscriptionsver-tically or in zig-zag fashion on alternate lines.)The remarkable flexibility of their system issummarised instatementslike this:First, if the word beginswith a vowel then the geneticsign has to be given the proper vowel value. Next theintermediate consonantshave to be shaped properly byassigning the correct vowel combinations. Finally, theterminal letter may also have to be modified accordingto context. In the last case, a missing visarga or anusvaaramay have to be supplied, though thisisoften indicated. How, the sceptic might ask, can you choose the right wordsfrom the infinite possibilities?The problem callsfor a little Vedicingenuity:In resolving ambiguities, one is forced to fall back onones knowledge of the Vedic language and the liter-12INtheirattempts toforce fitHarappanscriptintoSanskritmoulds, Rajaram and his collaborator ignore many known factsabout Harappan inscriptions. Oneof themost glaring conflictswiththeevidencecomesin their claim that in most casesthescript istoberead from left to right, likeSanskrit.Much evidencehasaccumulated over seven decadesthat thisisthereverseof thecase. Indeed, oneof thefew thingsthat all Harappanresearchers agree onconcerns the usualright-leftdirectionofthescript. Writing direction in ancient scriptsoften varied in differentcontexts, but evidenceof many sortssuggeststhat Harappan deviat-ed from right-left patternsin lessthan sevenper cent of inscriptions.Someof thisevidencearisesfrom studiesof inscriptionson potsherds. AsB.B. Lal showed in the1960s, examination of overlappinglineson thoseinscriptionsshowsthat thescript wasnormally inscribedfrom right to left. Other evidenceisapparent to theuntrained eye.Below, wegivetwo examplesfrom imagesin theCorpusof IndusSealsand Inscriptionscompiled by Asko Parpola and hiscollaborators. Theevidencein both caseshasbeen known sincetheearly 1930s.Onekind of evidenceinvolvesthespacing of characters. In sealimpression M-66a(using Parpolasnumbers), shown below, weseeoneof many caseswherean engraver ran out of room when engraving theseal, causing abunching of letterson theleft. In theseal, no room atall wasleft for thejar sign often found at theend of inscriptions. Thisforced theengraver to placeit below therest of theinscription, on thefar left. Itsplacement would beinconceivableif thejar sign wereawildcard vowel beginning inscriptions, asRajaram and Jhaclaim.Other evidenceshowsup in Parpolasseal H-103a, shown below.Theunusually long inscription in thiscaserunsaround threesidesofthe seal, with the top of the symbolspointing towardsthe nearestedge. Thissuggeststhat theinscription wasto beread by turning itaround in thehand to read itsthreeparts. Only thetop sideof theinscription isfilled with symbols,indicating that thisisthefirst line.The inscriptionwas hence toberead right to left, turning it clock-wiseto seetherest.Further evidence comesfromstudies ofinitialandfinalsignsequences, from studiesof repeat-ing sign combinations, and otherdata.Allthis evidencehas beendiscussedbyalonglineofresearchersstretching from G.C.Gadd in 1931 to Gregory Possehlin 1996. None of thisevidence ismentioned in Jha and Rajaramsbook. IThe direction of Harappan writingMICHAEL WITZELSTEVE FARMERM-66a H-103a13ary context. For example: when the common compos-ite letter r + k is employed, the context determines ifit is to be pronounced as rka (as in arka) or as kra asin kruura. The context Rajaram wantsyou to use to fill in the blanksisthe one that he wantsto prove: any reading isproper that illus-trates the(imaginary)links betweenlateVediccultureandIndusCivilisation. Once you tossin wildcard vowels, for exam-ple, any rk or kr combination providesinstant Harappan horse-playgivingyouaVedic-Harappanhorse(recallingtheirequationthat arka sun = horse) long before the word (or ani-mal) appeared in India. Why did the Indusgeniuswho invented the alphabet notinclude all basic vowel signs like those in Asokasscript whichwould have made thingsunambiguous?It certainly could not bebecause of a lack of linguistic knowledge, since Rajaram claimsthat the Harappanshad an advanced state of knowledge of gram-mar, phonetics, and etymology, just asthey had modern scien-tific knowledge of all other kinds. But vowels, of course, wouldrob Rajaram of hischancesto find Vedic treasure in Harappaninscriptions where he discoverseverything from horse thievesto Rigvedic kingsand advanced mathematical formulae. Peculiarly, in contrast to the lack of vowel signs, Jha andRajaram give usa profusion of special signsthat stand for finegrammatical detailsincluding word-final -H and -M (Visarjaniyaand Anusvaara; if these are missing, you can just tossthem in);special verb endingslike -te; and noun endingssuch as-su. Allof these are derived from Paninian grammar more than two thou-sand years before Panini! They even find special phonologicalsignsfor Paninian guNa and vRddhi (that is, ubecomesoor au)and for Vedic pitch accents(svara). Althoughthescribes lackedvowels,theythus hadsignsapplicable only to vowel combination (sandhi) which isremark-able indeed, given the absence of the vowelsthemselves. A Hundred Noi sy CrowsIt isclear that the method of Rajaram and Jha isso flexiblethatyoucansqueezesomepseudo-Vedicreadingoutofanyinscription. But, with all thisfreedom, what a motley set of read-ingsthey hand us! Moreover, few of their readingshave anythingto do with Harappan civilisation. WhatwereIndus seals usedfor? Weknowthatsome(aminority) were stamped on balesof merchandise; many were car-ried around on strings, perhapsasamuletsor ID cards. Many ofthem were lost in the street or were thrown out asrubbish whenno longer needed. Sometimesa whole set of identical inscrip-tionshasbeen found tossed over Harappan embankment walls. In their usual cavalier way, Rajaram and Jha ignore all thewell-known archaeological evidence and claim that the inscrip-tions representrepositories ofVedicworks liketheancientNighantu word lists, or even the mathematical formulae of theShulbasutras. The main object of Harappan seals, they tell us,wasthe preservation of Vedic knowledge and related subjects.How many merchantsin the 5000-odd year history of writ-ing would have thought to put mathematical formulae or geo-metric sloganson their sealsand tokens?Or who would be likelyto wear sloganslike the following around their necks?It isthe rainy season; House in the grip of cold; Adog that stayshome and doesnothing isuseless whichRajaram and Jha alternately read as: There israw meaton the face of the dog; Birdsof the eastern country;One who drinksbarley water; A hundred noisy crows;Mosquito; The breathing of an angry person; Ramathreatened to use agni-vaaNa (a fire missile); A shorttemperedmother-in-law;Thoseabouttokillthem-selveswith sinfulnesssay; or, best of all, the refreshing-ly populist: O! Moneylender, eat (your interest)!By now, we expect lotsof horse readings, and we are not dis-appointed. What use, we wonder, would the Harappanshave forseal inscriptionslike these?Water fit for drinking by horses; A keeper of horses(paidva) by name of VarSaraata; A horsekeeper by nameof Asra-gaura wishesto groom the horses; Food for theowner of two horses; Arci who brought under controleight loose horses; and so on. The most elaborate horse reading showsup in the most famousof Indusinscriptions the giant signboard hung on the wallsofthe Harappan city of Dholavira. The deciphered inscription isanother attack on the no horse in Harappa argument:I wasa thousand timesvictoriousover avariciousraidersdesirousof my wealth of horses!In the end, readersof Jha and Rajaram are likely to agree withonly one deciphered message in the whole book:apa-yaso hamahaat A great disgrace indeed!Vedi c Sanskri t ?Before concluding, we would like to point out that the linewe just quoted containsan elementary grammatical error a read-ing of mahaat for mahat. The frequency of mistakeslike thissaysa lot about the level of Vedic knowledge (or lack thereof) of theauthors. A few examplesat random: on p. 227 of their book we find adma eat! But what formisadma?admaH we eat?At best, adma food, not eat! on p. 235, we find tuurNa ugra svasruuH. No feminine adjec-tivesappear in the expression (tuurNaa, ugraa), asrequired bythe angry mother-in-law (read: svasruuH!). on p. 230, we read apvaa-hataa-tmaahuH, where hataatmamight mean one whose self isslain, or the self of a slain (per-son),but not those about to kill themselves. In the samesentence, apvaa doesnot mean sinfulness (which is, in anycase, a non-Vedic concept) but mortal fear. on p. 232, we have amasaityaarpaa, supposedly meaningHouse in the grip of cold. But amaa (apparently what theywant, not ama force) isnot a word for house, but an adverbmeaning at home. The word saitya cold isnot late Vedicbut post-Vedic, making the reading even more anachronisticthan the other readingsin the book. on p. 226, we find paidva for horses, in a passage referringto horse keepers. But in Vedic literature thisword doesnot referto an ordinary but a mythological horse. Many similar errorsare found in the 1996 pamphlet by Jha,billed by Rajaram asone of the worldsforemost Vedic scholarsand palaeographers.Noneof thoseerrorscan beblamed on ignorant Harappan scribes.Hist ory and Hindut va Propaganda It might be tempting to laugh off the Indusscript hoax astheharmless fantasy of an ex-engineer who pretends to be a worldexpert on everything from artificial intelligence to Christianity toHarappan culture. What beliesthisreading isthe ugly subtext of Rajaramsmes-sage, which isaimed at millionsof Indian readers. That messageis anti-Muslim,anti-Christian,anti-Indological,and(despiteclaimsto the opposite) intensely anti-scientific. Those viewspre-sent twisted imagesof Indiaspast capable of inflicting severe dam-age in the present. FRONTLI NE, OCTOBER 13, 2000 14Rajaramswork isonly one example of a broader reactionarytrend in Indian history. Movementslike thiscan sometimesbeseen more clearly from afar than nearby, and we conclude with afew commentson it from our outside but interested perspective. In the past few decades, a new kind of history hasbeen prop-agated by a vocal group of Indian writers, few of them trained his-torians, who lavishly praise and support each othersworks. Theiraim isto rewrite Indian history from a nationalistic and religiouspoint of view. Their writingshave special appeal to a new middleclassconfused by modern threatsto traditional values. With alarm-ingfrequencytheirmovementis backedbypowerfulpoliticalforces, lending it a mask of respectability that it doesnot deserve. Unquestionably, all sidesof Indian history must be repeated-ly re-examined. But any massive revisionsmust arise from the dis-coveryofnewevidence,notfromdesires toboostnationalorsectarian pride at any cost. Any new historical modelsmust be con-sistent with all available data judged apart from parochial concerns. Thecurrentrevisionistmodels contradictwell-knownfacts:theyintroducehorse-drawnchariots thousands ofyearsbefore their invention; imagine massive lost literaturesfilled withscientificknowledgeunimaginableanywhereintheancientworld; project the Rigveda into impossibly distant eras, compiledin urban or maritime settingssuggested nowhere in the text; andimagine Vedic Sanskrit or even Proto Indo-European rising inthe Panjab or elsewhere in northern India, ignoring 150 yearsofevidence fixing their originsto the northwest. Extreme out-of-India proponentseven fanaticise an India that isthe cradle ofall civilisation, angrily rejecting all suggestionsthat peoples, lan-guages, or technologiesever entered prehistoric India from for-eign soil asif modern conceptsof foreign had any meaningin prehistoric times. Ironically, many of those expressing these anti-migrationalviewsare emigrantsthemselves, engineersor technocratslike N.S.Rajaram, S. Kak, and S. Kalyanaraman, who ship their ideastoIndia from U.S. shores. They find alliesin a broader assortmentof home-grown nationalistsincluding university professors, bankemployees, and politicians(S. S. Misra, S. Talageri, K.D. Sethna,S.P. Gupta, Bh. Singh, M. Shendge, Bh. Gidwani, P. Chaudhuri,A. Shourie, S.R. Goel). They have even gained a small but vocalfollowing in the West among New Age writersor researchersoutsidemainstreamscholarship,includingD.Frawley,G.Feuerstein,K.Klostermaier,andK.Elst.Wholepublishingfirms, such asthe Voice of India and Aditya Prakashan, are devot-ed to propagating their ideas.Thereareadmittedly no universal standardsfor rewriting histo-ry. But afew demandsmust bemadeof anyoneexpecting hisor herscholarship to betaken seriously. A short list might include: (1) open-nessin theuseof evidence; (2) arespect for well-established facts; (3)awillingnessto confront datain all relevant fields; and (4) indepen-dencein making conclusionsfrom religiousand political agendas.N.S. Rajaram typifiesthe worst of the revisionist movement,and obviously failson all counts. TheDeciphered IndusScript isbased on blatantly fake data (the horse seal, the free-form deci-pherments); disregardsnumerouswell-known facts(the datesofhorsesand chariots, the usesof Harappan seals, etc.); rejectsevi-dence from whole scientific fields, including linguistics(a strangeexclusion for a would-be decipherer!); and isdriven by obviousreligious andpoliticalmotives inclaimingimpossiblelinksbetween Harappan and Vedic cultures. Whatevertheirpretensions,Hindutvapropagandists likeRajaram do not belong to the realm of legitimate historical dis-course. They perpetuate, in twisted half-modern ways, medievaltendenciesto use every meanspossible to support the authority ofreligioustexts. In the political sphere, they falsify history to bol-ster national pride. In the ethnic realm, they glorify one sector ofIndia to the detriment of others. It isthe responsibility of every seriousresearcher to opposethese tendencieswith the only sure weapon available hard evi-dence. If reactionary trendsin Indian history find further politi-cal support, we risk seeing violent repeatsin the coming decadesof the fascist extremesof the past. The historical fantasiesof writerslike Rajaram must be exposedfor what they are: propaganda issuing from the ugliest cornersofthe pre-scientific mind. The fact that many of the most unbeliev-able of these fantasiesare the product of highly trained engineersshould give Indian educational plannersdeep concern. In a recent online exchange, Rajaram dismissed criticismsofhis faked horse seal and pointed to political friends in highplaces,boastingthattheUniongovernmenthadrecentlyadvised the National Book Trust to bring out my popularbook, FromSarasvati River to theIndusScript, in English andthirteen other languages. We fear for India and for objective scholarship. To quoteRajarams Harappan-Vediconelasttime:Agreatdisgraceindeed! I Michael Witzel & Steve Farmer, 2000Michael Witzel isWalesProfessor of Sanskrit at Harvard Universityand theauthor of manypublications, includingtherecent monograph Early SourcesforSouth Asian SubstrateLanguages, Boston: ASLIP/Mother Tongue1999. A collection of hisVedicstudieswill bepublished in India byOrient Longman laterthisyear. Heisalsoeditor of TheElectronic Journal of Vedic Studies, accessiblethrough hishomepageat http://www.fas.harvard.edu/~witzel/mwpage.htm.SteveFarmer, whoreceived hisdoctoratefromStanford University, hasheld anumber of academicpostsin premodern historyand thehistoryof science.Amonghisrecent worksishisbook Syncretism in theWest, which developsacross-cultural model of theevolution of traditional religiousand philosophicalsystems. Heiscurrentlyfinishinga new book on brain and theevolution of culture. Hecan becontacted at india@safarmer.com.Graphics source credits:Frontlineand theauthorsthank Asko Parpola, Professor of Indology, Universityof Helsinki, Finland, for permission to reproducethephotographsof M-1034a,M-772A, M-595a, M-66a, H-103ain thisarticle.M-1034a, Vol. 2 of A. Parpolasphotographic corpus(**) = DK 5582, MohenjoDaro Museum 778, P 694; photographed by S.M. Ilyas. Courtesy Department ofArchaeology and Museums, Government of Pakistan.M-772A,Vol.2(**),DK6664,MohenjoDaroMuseum742,JL884;pho-tographed by Jyrki Lyytikk. Courtesy Department of Archaeology and Museums,Government of Pakistan.M-595a, Vol. 2 (**), HR 4601a, LahoreMuseum, P-1815; photographed by S.M.Ilyas. Courtesy Department of Archaeology and Museums, Government of Pakistan.M-66a, Vol. 1 (*), HR 5629, ASI 63.10.371, HU 441; photographed by ErjaLahdenper. Courtesy ASI, Government of India.H-103a,Vol.1(*),2789,ASI63.11.116,HU601;photographedbyErjaLahdenper. Courtesy ASI, Government of India.(*) Jagat Pati Joshi & A. Parpola, Corpusof IndusSealsand Inscriptions1. Collectionsin India, Helsinki 1987.(**) Sayid Ghulam MustafaShah & A. Parpola, Corpusof IndusSealsand Inscriptions2. Collectionsin Pakistan, Helsinki 1991. All other photographsarefrom N. Jhaand N.S. Rajaram, TheDeciphered IndusScript, cited earlier, except for thethreeanimalson theright in thephotograph onpage 10,whichare takenfromJohnMarshall,Mohenjo-Daroand theIndusCivilization, Vol. III, platescxvii-cxviii, London 1931.