J 0 URN A 'L

OF
TAMIL STUDIES

DECEMBER 2009
 Meluhha and Agastya :
Alpha and Omega of the Indus Script
I ravatham Mahadevan

1. Introduction:

1.1 It is now generally accepted that Meluhha was the ancient name of
the Indus Civilisation as written in early cuneiform records 1. I
propose that the Harappan equivalent of the name and its
connection with ancient Indian historical tradition are revealed by
two signs which I shall designate here as the 'alpha and omega' of
the Indus Script.

1.2 There is as yet no agreed sign list for the Indus Script. Not only the
numbers, but also the order of signs vary widely in different lists.
However, I shall provide the following definition for the present
purpose, which should be acceptable to all scholars in the field,
whichever sign list they choose to follow2 :

<> 'alpha' : the most frequent initial sign in Indus texts .

U 'omega': the most frequent final sign in Indus texts .

1.3 The 'alpha' sign occurs 298 times at the commencement of texts,
more than double the number of times for the next most frequent
initial sign. The 'omega' sign occurs 971 times at the end of texts,
three times more than the next most frequent final sign. The relative
preponderance of the two signs in the initial and final positions
respectively becomes much more pronounced if one looks only at
the seals, the 'identity cards' of the Harappan ruling classes. In
short, the 'alpha' and 'omega' signs introduce and identify the
affiliation of the seal-holder whose name and/or titles stand
'bracketed' between them .

1.4 This is the second occasion when the 'alpha and omega' signs have
led to an important result. The first was way back in 1977, when
these two signs helped to clinch the argument in favour of a
generally right to left direction in the Indus texts. I had then pointed
out that, as the most frequent right-end sign (alpha) occurs at the

NO. 76 DECEMBER 2009

 92 JOURNAL OF TAMIL STUDIES

provided by Asko Parpola, that I could have more appropriately

interpreted the 'alpha' sign as Dr. meJ-akam, lit., 'the high (or great)
place (or house) inside (the citadel)'. While adopting this
nomenclature and interpretation in the present paper, I shall take
the opportunity to summarise and update my earlier studies on the
'alpha and omega' signs, reiterating how important it is to
understand them before proceeding with interpreting or deciphering
the Indus Script.

3. The 'alpha' sign

3.1 Place signs

In accordance with universal usage, the Indus seals depict owners'

names and titles. We know from historical inscriptions, especially
Dravidian, that place ~ames precede personal names. Most seal
texts commence with one of five frequent 'opening signs' listed
below in the order of frequency:

These are too few to denote personal names, but can be place
names preceding personal names. Since the same set of opening
signs occurs at all major Harappan sites, they cannot be identified
with the names of cities like Harappa or Mohenjodaro. The opening
signs must then refer to important places or institutions present in
every major Harappan settlement, like for example, 'temple, palace,
citadel, walled city' etc. Some of the frequent opening signs may
also represent important titles or offices, which would be much
fewer in number than personal names, like for example, 'ruler, chief,
priest, lord' etc. Judging from the extreme brevity of the Indus texts,
it is much more likely that place names and common titles would be
represented by single ideograms rather than by phonetic syllabic
writing. In particular, the most frequent opening 'alpha' sign appears
to depict the ground plan of a building with a forecourt inside a
fortified place, in other words, what is popularly known as the
'citadel'. I identify the 'alpha' sign with the Harappan Citadel and
interpret it as Dr. meJ-akam lit., 'high inner place', ' the 'address'
most members of the ruling class preferred to prefix to their
personal identification. Through constant use, the expression mel-

NO. 76 DECEMBER 2009

 93 Meluhha and Agastya: Alpha and Omega of the Indus Script

akam (melahha of the cuneiform records) came to represent the

people and the land of the Indus Civilisation.

3.2 Graphic evolution of place signs

Sign Rectangular Rhomboid Oval Forms

No. Forms Forms

261-373 D
5090
0 0 C)
2119 4379 1 4087 2

267 CJ ~ ()
8106 1057 1022

284
~
200e
~
2579
@
2522
Flg.2 Graphic evolution of Place Signs In the Indus SCript
(Schematic) (Cf. Mahadevan 1977:List of Sign Variants)

Notes:
(1) Oval form (373) now recognised as a variant of
rectangular or rhomboid forms (261) .

(2) The oval is sometimes split as 'brackets' to accommodate

one or two signs inSide to form compound signs.

(3) This form does not occur as a sign, but is inferred from
seal- motifs as in 2001 .

Enclosures serving as place signs in the Indus Script exhibit three

characteristic variant forms viz., rectangular, rhomboid or oval,
whose equivalence can be demonstrated from parallel texts. It
appears that acute pressure on space available on the seals
caused the variation in shape in order to conform to the general
space-saving pattern of tall and narrow signs in the Indus Script.
The graphic evolution of the three related place signs is shown in
Fig.2 with one example for each form from the texts.

NO. 76 DECEMBER 2009

 94 JOURNAL OF TAMIL STUDIES

3.3 Egyptian ideographic parallels to Indus place signs

Ideographic interpretation of place signs inferred from their shapes

is corroborated by a set of remarkably close parallels from the
Egyptian hieroglyphic script (Gardiner 1978: Sign List). The
comparison and the resulting broad interpretation of the Indus signs
is shown in Fig.3.

3.4 The Egyptian parallel goes beyond mere graphic

resemblances. 'Pharaoh ', the generic name of the Egyptian
rulers, is traced to the expression 'Great House'. Originally,
'Great House' referred only to the 'palace' or to the 'court', and not
to the person of the king . Later, the term 'pharaoh' became a
respectable designation for the king, "just as the head of the
Ottoman government was termed the Sublime Porte" (Gardiner
1978:75). As in the Egyptian script, the generic name of the rulers
of Harappan cities was also derived from the expression 'High
House' (conventionally called the 'citadel').

Egyptian Indus
-
Sign Sign Sign Sign (nearest Broad
No. I
No. variant with Interpreta tion
text No.) I

0.1 n 261-373 D 'ho~se' -

I, 5090
I
0.6
Q I
267 I C1 'fortified house'

~j
8106
-:--

~049 1 @ 284 I
@ 'city, town'
2522
Fig.3 Indus Place Signs and Egyptian Ideographic Parallels
(Schematic) (Egyptian : Gardiner 1978. Indus: Mahadevan 1977.)

NO. 76 DECEMBER 2009

 95 Meluhha and Agastya: Alpha and Omega of the Indus Script

3.5 The Egyptian parallel should not, however, be stretched too

far. The low-profile Harappan rulers (with no grandiose
palaces or rich tombs) can in no way be compared to the
vainglorious pharaohs. There is also no archaeological
evidence for contacts between the Egyptian and Indus
civilisations. It is, however, not unlikely that the two great
contemporary civilisations had at least indirect contacts
through the intermediary Sumerian-Akkadian city states in
West Asia.

3.6 Interpretation of place signs in the Indus Script

In the light of the foregoing discussion, we may interpret the three
related place signs in the Indus Script as follows:

: akam 'house, place, inside' (DEDR 7)

: meJ-akam 'High House (citadel), (DEDR 5086 & 7)

: paji'city' (probably, the 'lower city' in the

Harappan context).
Cf. paji'town, city'; (DEDR 4112)
paji'row, line, regular order' (DEDR 4113)
The expressionpajithus indicates a 'planned city'.

3.7 Strokes attached to 'opening' signs

The opening signs are typically followed by one of three short
superscript strokes, attached to them. They are shown below in the
order of frequency:
II ~
These strokes are conventional markers which cannot be 'read'. But
their function as attachments to nouns (names of places, institutions
or persons) can be broadly understood as suffixed case-endings.
Most probably, they represent the genitive or locative cases with the
meanings 'of, belonging to, among or in'. When followed by a case
suffix, the preceding noun may be either in direct or oblique form,
the latter with seemingly no overt marker. .

NO. 76 DECEMBER 2009

 96 JOURNAL OF TAMIL STUDIES

3.8 Role of Place Signs in the Indus Texts

Positional - statistical analysis of the three related place signs with
the attached case suffix yields significant results (Fig.4).

Sign Occurrences Dr. interpretation Meaning

f--
Pair Initial Total

(a) 1/0 254 291 me1-aka(-tt-Jigl-i1

'of / in the
High House'

(b)
"0 nil 2 aka(-tt-)igl-i1
'of / in the
house'

(c) I/~ 7 8 paji -ig l -i1

'of / in the
(lower) city'
FigA Interpretation of opening pairs with place signs

I interpret the data as follows:

(a) meJ-aka(-tt-)igl-i1 'of / in the High House' is the address
of the rulers of the Harappan City. This accounts for the
high frequency as well as initial occurrence of this pair.
It is of course not necessary that all members of the
ruling classes resided actually within the citadel. The
opening pair is more an assertion of their identity as
being associated with the citadel and the institutions
within . We know from archaeological evidence that there
were large houses in the lower city where we may
presume many of the ruling classes resided.

(b) aka(-tt-)igl-i/ 'of / in the house' relates to those who

served the rulers in institutions inside the citadel like
palace, temple, etc. , as guards, attendants and other
lower functionaries. This accounts for the absence of
the pair in the initial position and its very low frequency.
The data regarding frequency is, however, deceptive in
this case. A close study of the enclosure signs reveals
that there are more than 30 compound signs comprising
enclosures surrounding the basic signs within. The
enclosures are mostly oval in form, but may also be

NO. 76 DECEMBER 2009

 97 Meluhha and Agastya: Alpha and Omega of the Indus Script

rectangular or rhomboid ' and may also be split up into

two halves like 'brackets' to accommodate the signs
within as illustrated by the following pairs of signs::

~r (VI (~) i
I interpret the pairs as functionaries or institutions within,
or associated with the citadel as distinguished from those
outside.

(c) piji-ig l-iJ 'of I in the (lower) city'. This seems to be the
'address' of the residents of the city not directly involved
with the citadel either as rulers or as lower functionaries.
The low frequency of the pair especially at the
commencement of the texts can be quite simply
interpreted as showing that the authority in the Harappan
city did not vest with the citizens of the lower city but
with the rulers of the higher citadel, the seat of authority.

3.9 The discussion shows that the frequency of the three related place
signs is directly proportional to the importance of the institutions or
the persons concerned, and inversely proportional to the population.
It can hardly be doubted that the rulers would be less in number
than the functionaries who served them, who in turn would be less
in number than the common people of the lower city.
3.10 Parallels from Old Tamil traditions
Old Tamil literature contains several references to akam in the
sense of 'fort, palace or inner place'.
(e.g.) akam 'palace' (Peron.1.32.100)
aka-nakar 'the inner city' (CiJ. 2.15.109; Malli.1.72)
aka-p-pi'inner fortification' (Nar.14.4; Patif.22.26; 01.28.144)
aka-p-pi 'matJJ-u! uyar me.tai: high terrace inside the fort'
(Tivikaram 5.198)
matJJ-akam lit., 'fortified house'; (01.2.14.69); the palace of
the rulers of Kerala.
A clear distinction is drawn in Old Tamil literature between those
who ruled from inside the forts and those who served them, even
though the expressions for either group have the same base
aka-tt-u 'in the house'. The rulers of the forts were known as:

NO. 76 DECEMBER 2009

 98 JOURNAL OF TAMIL STUDIES

(e.g.) aka-tt-ar : ' (princes) of the palace' (Kah: 25.3)

aka-tt-ar 1hose inside the (impregnable) fortification' (KUaJ745)
aka-tt-or'those inside the fort' (PUTEI. 28.11)
aka-tt-og < he (king) inside the fort' (Tol. III: 68.4, 69.5)

Those who served as palace or temple attendants were known as

follows:
(e.g.) aka-tt-a.fimai , aka-t-toIJ.tar, aka-mp-a.fiyar etc.,
(Tamil Lexicon).
The palace or temple service was generally called:
(e.g.) aka-p-pa.tal: aka-p-paIJi, aka-p-parivaram etc.,
(Tamil Lexicon).
Another important set of Old Tamil expressions for palace and
temple attendants is derived from the root cui 'to surround' > iijiyam
'service, especially in palace or temple', uiiyar'palace or temple
servants' (DEDR 2698 > 758).
Cf. uii, uiai'place esp. about a king' (DEDR 684) which also
ultimately looks to cui 'to surround, surrounding area'. Note the
distinction between uJai-y-imnf1ig 'minister of state, companion of
the king' andujai-y-aj-ag 'attendant (in the palace)' (Tamil Lexicon).

3.11 From Etymology to History

The critical link between Dravidian etymology and history is

brought out by the following two sets of entries: DEDR 7:
aka-m 'inside, house, place'
aka-tt-u 'within, inside the house'
aka-tt-ag 'one who is in, a householder'.
C.W.Kathiraiver Pillai's Dictionary (1910) (gloss in English
added by me):
aka-tt-i: (1) akattiya mU[Jivag ('Agastya, the sage')
(2) u/!-irukkil"8-vag ('one who is in')
(3) oru maram ('Agasti grandiflora').

NO. 76 DECEMBER 2009

 99 Meluhha .and Agastya: Alpha and Omega of the Indus Script

Note how akatti in (1) and (3) get transformed to agasti in

Indo-Aryan loanwords.

3.12 I have suggested in my earlier papers (1981, 1986) that aka-tt-i 'he
of the (High) House' was the prototype of Indo-Aryan Agasti
(Agastya) as well as Dravidian Akatti (AkattiyaQ) of the Old Tamil
legends (who led the southern migration of the Vejir and other tribes
from Dviirakii in the Gujarat reg ion to the southern peninsula). I shall
revert to this theme when dealing with the 'omega ' sign in the next
section.

4. The 'Omega' SignV

4.1 The 'omega' or JAR sign is by far the most

frequent in the Indus texts, accounting for
about ten percent of the total sign
occurrences. The pictorial identification of the
sign as a 'vessel with handles and a tapering
bottom', is no longer in doubt, especially after
the publication of the pottery graffiti from
Kalibangan with a realistic depiction of the sign
(Lal 1979). See Fig.5.

Fig.S: JAR Sign incised on pottery

from Kalibangan

4.2 I admit that I have not been quite consistent in my earlier

attempts to interpret the JAR sign. I have sometimes
identified it as a grammatical suffix (1970), and at other
times, as an ideogram added to the names and titles of
priestly and ruling classes (1982, 1986). However, after nearly
three decades of intensive study, I have arrived at the
conclusion that both identifications are complementary, and
that the JAR sign has indeed a dual function in the Indus texts
(1998). As I have dealt with both aspects in detail in my
earlier writings, I shall present here only a very brief
summary of the essential evidence to harmonise the two

NO. 76 DECEMBER 2009

 100 JOURNAL OF TAM IL STUDIES

interpretations. However, I shall emphasise the ideographic

aspect more as it holds the key to the true import of the
alpha and omega signs of the Indus texts.

4.3 The JAR Sign as a Grammatical Suffix

The most marked characteristic of the JAR sign is its almost

constant final position in the texts. It can also be shown that
even when it occurs in the middle of a text, it acts as a terminal
sign. The high frequency and final position of the sign have led to
the reasonable presumption that it is a grammatical marker of some
kind . The most common supposition has been that the JAR sign
represents the genitive case with the meaning 'of, belonging to'.
However, the sign is found to be more closely attached to its
preceding signs than would be the case if it were a case-ending .
Further, early Dravidian inscriptions in Tamil, Telugu and Kannada,
do not have case-endings in text-final positions. That leaves only
the possibility that the sign is a nominal suffix marking the gender
and number of the preceding nouns. Considering the very high
frequency of the sign, it can only be equated with the masculine
singular nominal suffix in Dravidian, viz., -(a)nr. This interpretation
is corroborated by the evidence of earliest Dravidian inscriptions
which have the corresponding masculine singular suffix attached to
names and titles .
nd
(e.g.) ne.tiiicajiyag salakag ijaiica.tikag (Old Ta., 2 cent.
BCE) (Mahadevan 2003:No.2)
anikka-paSUl1J71 (Old Te., ca.600 CE) (Lockwood 200l:p.l99)
Nage1J1Jan (Old Ka., ca.675 CE) (Narasimhia 1941:No.3)

4.4 The phonetic value of the JAR sign can also be discovered
through the rebus method by comparing the likely 'vessel'
words in Dravidian languages:
Cf. Ka. aIJqige; Te.aIJqemu, aIJqiyamu, aqigamu 'pannier'
(DEDR 127).
Ka. aIJqu ' bottom of a vessel' (DEDR 129).
Ta. alJ.tai '(bamboo) squirt for festival occasions'; Ka. allqe
'bamboo vessel, generally with a handle'; Ko. t:llJqy 'milk pot,

.NO.76 DECEM BER 2009

 101 Meluhha and Agastya: Alpha and Omega of the Indus Script

bamboo pot'; To. aqy 'clay pot'; Tu. alJqe 'bamboo or

nutshell vessel' (DEDR 130).
Ta. antai (lex.) 'a weight' (?) (Tamil Lexicon), (included in a list of
weights and measures; prob., 'a measure' (Tol. eju. 170;
commentary).
Hence, by phonetic transfer, the JAR sign is equated with -(a)n-I
-(a)nt-, the masculine singular suffix added to names and titles in
the nominative case.

4.5 Paradigm of Gender-Number suffixes in the Indus script

The ARROW sign f
The ARROW sign is known to function exactly like the JAR
sign (including having dual functions) except for its much
lower frequency (about one-sixth of that of the JAR sign ). It
is therefore likely that the ARROW sign is also a grammatical
suffix indicating gender and number. Considering its much
lower frequency , it is probably the non-masculine (feminine
and / or neuter) singular suffix (Mahadevan 1998). The most
common word for 'arrow' in Dravidian is ampu (Ta., Ma.) or
ambu (Te., Ka.) (DEDR 178). This immediately leads, by the
rebus method, to the non-masculine singular suffix
-(a)mb(u) , attested in the earliest Old Telugu inscriptions (Mahadeva
Sastri 1969:135-138) . The gender suffixes -(a)nu and -(a)mbu are
attested in the Old Telugu inscriptions of Mahendra Pallava
assigned to ca .600 CE (Lockwood 2001: 199). See Fig.6.

I T' lJl1~
I BIilkkapliS ll1lf/J anikkapiisumbu
Fig.6 : Old Telugu inscriptions with gender suffixes -(a)n[u
and-(a)mbu

As Early Dravidian had only two genders in the singular number,

namely, masculine and non-masculine, they match the frequency-
distribution of JAR and ARROW signs respectively. The paradigm of

NO. 76 DECEMBER 2009

 102 JOURNAL OF TAMIL STUDIES

gender suffixes in the Indus Script including the plural marker

already recognised by Heras (1953:83) is shown below:

V Masculine singular -(a)nj"- / -(a)nt- (dialectally)

f Non-masculine singular -(a)mp(u)

I I Human (masc. & fern.) plural -aT/-Ii"

r--I (when combined with basic signs)

Fig. 7 Paradigm of Gender Suffixes in the Indus Script

When the Indus script became extinct, the memory of its

grammatical elements should have been lost; but it did not
quite happen that way. Since most of the Harappan ruling
classes had names or titles ending in -(aJllf, that sound was
borrowed by Indo-Aryan as an ethnic name to denote the
neighbouring Non-Aryan people. Thus, Dr. -(aJnr > IA andr >
andhra (attested in Aitareya Bra71malJa VII:18).

4.6 JAR Sign as an ideogram

The fact that the JAR sign has another value, apart from its function
as a grammatical suffix, is shown by its attachment to the top of the
BEARER sign, just like the ARROW sign, its functional twin:

BEARER JAR-BEARER ARROW-BEARER

There is no reason not to follow the normal practice of reading the

compound signs from top to bottom. We must therefore assume
that the JAR and ARROW signs have their literal pictographic values
in these compound signs, especially as such interpretation is
meaningful and productive:

NO. 76 DECEMBER 2009

 103 Meluhha and Agastya: Alpha and Omega of the Indus Script

JAR-BEARER: 'one who carries ceremonially a sacrificial

vessel with offerings'
ARROW-BEARER: 'one who carries the arrow' (a 'warrior').

The ideographic symbolism, especially of the JAR sign, would have

survived and continued to be associated with the descendants of
the Harappan ruling classes re-emerging in the new social order in
later periods.

4.7 Linguistic interpretation of JAR ideogram

In Vedic literature and ritual treatises, sata is mentioned as

some kind of a sacrificial vessel (VS.xix:27 ,88 ; SOB. xii :
7.2.1 3). Sabaraswamin identifies sata as a 'mleccha' " term
for a 'round wooden vessel with a hundred holes' Mimalnsa-
siitra-bha~ya 1.3.10). Numerous perforated jars have been
found at the Harappan sites. It is probable they served a
ritual purpose. My ongoing studies indicate that sata / sata
in Prakrit and later borrowed into Telugu and Tamil refer to
the food offerings as well as to the sacrificial vessel itself.
Cf.sata 'food' in a Pkt. cave inscription at Kanheri assigned
to ca .2nd cent. CE (Nagaraju 1979). Nagaraju has identified
sata as 'food' , contrasting with paniyaka 'drink' occurring in
the same inscription. I have connected the term sata
occurring here with Sata- / Sata-, names of the Andhra kings
as well as with Te. sadamu, Ta. catam food, lit., cooked
rice' (TamIl Lexicon). Cf. catI~ catam 'cooked rice '
(Pinkalantai 10:441, 10:463, ca.8 th cent. CE). As the word sata
in Vedic literature is identified as a 'mJeccha ' term, it may
be equated with Dr. cata. In the Harappan context, cata
may
be broadly interpreted as 'food or beverage in a sacrificial
vessel (offered to the deity)' .

4.8 'Jar -born' myths in Northern Traditions

The symbolism of 'water-pitcher' has always been closely

associated with priestly ritual. The legend of 'jar-born' sages

NO. 76 DECEMBER 2009

 104 JOURNAL OF TAMIL STUDIES

is very ancient and is found even in the -Rgveda (7.33.10-13).

There it is said that Vasi~tha and Agastya were generated by
Varul}a and Mitra in a 'sacred pitcher' or 'water-jar used in
sacrifice'. Agastya is especially known as the 'jar-born' sage
(kumbha-yoni, Jru.ta-muni etc.). The myth of miraculous
birth from jars was shared by priestly as well as royal
families. Drol}a, the priest-warrior, was generated in a
'wooden trough' by Bharadvaja (Mbh.) The Kauravas were
born from pots filled with clarified butter in which
Gandhari's foetus was stored (Mbh.)

4.9 'Jar-born' myths in Southern Traditions

The earliest reference to the myth in Old Tamil is found in

Puranigii..ru, a collection of four hundred poems compiled in the
early centuries CE, but containing much older oral bardic tradition, a
fact emphasised in this very poem (201). Kapilar, the poet, narrates
that 'fortynine generations' earlier, an ancestor of the Ve!ir ruled
over Tuvarai, a city surrounded by soaring bronze walls. Kapilar
also records that the VeFr arose in the tala vu of a 'northern sage'.
This extremely important historical reference has remained obscure,
as the medieval commentator did not identify the 'northern sage'
and misinterpreted ta.tavu as the 'sacrificial fire-pit' .
M.Raghavaiyangar (1907; 2004 reprint :26) has rightly interpreted
ta.ta vu (variant ta.ta) as a 'water pitcher', but missed the obvious
connection between the vessel and Agastya, the 'jar-born' sage.
The word tata occurs in a Tamil-Brahm! inscription incised on an
earthern storage jar excavated at Kodumanal, Tamilnadu , and
assigned to ca.2 nd cent. BCE (Y. Subbarayalu 1996:No.3) (see
Fig .8). The word is also attested in this sense in Tamil literature.
Cf. ta.tavu (Nar. 227.7); tala (Nicciyir Tirumoji: 9.6). Once the
meaning of ta.tavu (ta.ta) 'jar' is understood, it follows that the
reference in Puram 201 must be to Agastya, the 'jar born' sage par
excellence. U.V . Swaminathaiyar's identification of the 'northern
sage' with 'Campu mUlJi' (not known to Old Tamil tradition) from late
sources is unconvincing. I have attempted to set the record straight

NO. 76 DECEMBER 2009

 105 Meluhha and Agastya: Alpha and Omega of the Indus Script

by identifying the 'northern sage' in Pllfam 201 with Agastya, fa.fa vu

with his celebrated 'water-pitcher' and tuvarai with Dvaraka of the
Gujarat region (Mahadevan 1986). This re-interpretation links the
Puram 201 legend with another famous Old Tamil tradition
regarding Agastya and the Ve!ir (see para 4.11).

l F -
r

j~j\JC.1I~ I~ ~ .~t-;(~\
~ ~
;" ~~
Fig.S Pottery inscription from Kodumanal (ca.2 nd cent. BCE).
(ta/a 'jar' occurs as the last word at right)

The Pallavas of Kanchi belonging to the Bharadvaja gotra and

claiming DroI}.a to be one of their remote ancestors, traced their
descent from a water-pitcher (pattra- skhalifa- vrttlnain, Pallankoyil
th
Plates, ca. 6 cent. CE). According to tradition, the Chalukyas were
so-called as the dynasty sprang from a su.1uka' water-pot '
(VikramaJika-Canlra 1.318.8).

4.10 Agastya legend in the Southern tradition

It has generally been held that Agastya led the earliest Aryan
settlement of South India and introduced Vedic Aryanism
there. (For a comprehensive treatment of this view, see
Ghurye 1977.) This theory has, however, never been able to
explain satisfactorily how the Tamils, proud possessors of an
ancient culture of their own and a particularly strong
tradition of love for their language, came to accept Agastya, an
Aryan sage, as the founding father, not of the
Brahmanical religion or culture in the south, but of their
own Tamil language, literature and grammar. There is also

NO. 76 DECEMBER 2009

 106 JOURNAL OF TAMIL STUDIES

no linguistic evidence to support the theory of colonisation

of the Tamil country by speakers of Indo-Aryan languages
in pre-historic times. The interpretation of the Agastya
legend in terms of the Aryan acculturation of the south was
developed before the discovery of the Indus Civilisation,
which is considered by most scholars to be pre-Aryan and
probably Dravidian. It has now become possible to take a
fresh look at the Agastya legend and attempt an alternative
interpretation which would harmonise its two core features
which have hitherto remained irreconcilable, namely, the
northern origin of Agastya and his southern apotheosis as
the founder of Tamil language and grammar.

References to Agastya in early Tamil works have been collected

together in the essay on 'Akattiyar' by R.Raghavaiyangar (1941).
The secondary sources available in English are noticed and
succinctly summarised by Ghurye. While the Tamil Agastya shares
the basic myths of his northern counterpart, namely miraculous birth
from a pitcher and southern migration from the north across the
Vindhya, he is given a very different role by the Tamil tradition. Here
Agastya is so totally identified with Tamil that he is termed the Tamil
mugi (Tamil sage') and Tamil itself is named after him as agastyam.
Agastya received the Tamil language from Siva (or Skanda) and
gave it to the world. The Tamil Buddhists claimed that Agastya
learnt Tamil from Avalokitesvara (VlracoJiyam by Buddhamitra).
Agastya wrote the first Tamil grammar called Akattiyam (not extant
now). Even today, Tamilnadu has the largest number of Siva
temples dedicated to the 'Lord of Agastya' (Agastyesvara), a feature
almost unique Tamilnadu, as noted by Ghurye. According to most
competent scholars, it is from South India that the Agastya cult was
carried to the South-East Asian countries.

4.11 Agastya and the Southern migration of the Ve!ir

The story of the southern migration of the Velir from
Dvaraka under the leadership of Agastya is narrated by
Naccinarkkiniyar in his commentary on Tolkappiyam (payiram ;
Poruf.34). According to this legend, the gods congregated on Mount
Meru as a result of which the earth tilted, lowering Meru and raising
the southern quarter. The gods thereupon decided that Agastya was
the best person to remedy this situation and requested him to

NO. 76 DECEMBER 2009

 107 Meluhha and Agastya: AIpha and Omega of the Indus Script

proceed to the South. Agastya agreed and, on his way, visited

'Tuvarapati' (Dvaraka) and led the descendants of l1e.tu-mufi-alpJal
(Vi~Qu or Kr~qa) including 'eighteen kings, eighteen families of the
Ve!ir and the Aruva!ar' to the south, where they settled down 'clearing
the forests and cultivating the land' . The sage himself finally settled
down on the Potiyil hill. The fact of Agastya's leadership of the Ve!ir
clan rules out the possibility that he was even in origin an Aryan sage.
The Ventar-Ve!ir-Ve!a!ar groups constituted the ruling and land-
owning classes in the Tamil country since the beginning of recorded
history and betray no trace whatever of an Indo-Aryan linguistic
ancestry. The Tamil society had of course come under the religious
and cultural influence of the north even before the beginning of the
Cailkam Age, but had maintained its linguistic identity. From what we
now know of the linguistic prehistory of India, it is more plausible to
assume that the Yadavas were the Aryanised descendants of an
original non-Aryan people than to consider the Tamil Ve!ir to have
descended from the Indo-Aryan speaking Yadavas as suggested by
M. Raghavaiyangar (2004). As he has pointed out, ve/ means 'one who
performs a sacrifice' (namely a 'priest'). The Agastya legend itself can
be re-interpreted as non-Aryan and Dravidian even in origin and
pertaining to the Indus Civilisation.

4.12 Conclusion
The 'alpha and omega' signs have been so deSignated not only
because they respectively commence and end most of the Indus
texts, but also because they sum up the essence or most important
feature of the Indus seal-texts, namely, the identity of the Harappan
ruling class. This is shown below schematically (from left to right for
convenience) :

~"
' He of the (High) House
v
He with the JAR
aka-(t/)-(i) kumbha-mUIJi (Agastya)

akam ('High House') did not survive. But those who owned
allegiance to the meJ-akam, the akatt-u people, did survive and, in

NO. 76 DECEMBER 2009

 108 JOUR."NAL OF TAMIL STUDIES

course of time, re-emerged in the Vedic period as the 'jar-born'

priests typified by Agastya. A section of the Harappan ruling classes
did not stay on, but migrated under the leadership of the Akattiyar
clan to South India, where they founded the Early Historical
kingdoms (of Andhras and their successors in the Deccan, and the
triple kingdoms, Chera, Chola and Pandya, in the Tamil country).

Acknowledgements

I am grateful for the technical assistance extended by the The Indus

Research Centre, Roja Muthiah Library, Chennai, in the preparation
of this paper. Photographs of the seals are by courtesy of The
Archaeological Survey of India.

Notes

1. On identification of Meluhha with the Indus Civilisation,

see: Asko and Simo Parpola 1975, Romila Thapar 1975,
Simo and Asko Parpola and R. Brunswig 1977, Daniel
Potts 1982, Asko Parpola 1994, and Gregory Possehl
1996, 1997, & 2008.

2. Sign and Text Numbers cited in this paper are from

Mahadevan 1977.

NO. 76 DECEMBER 2009

 109 Meluhha and Agastya: Alpha and Omega of the Indus Script

References
Dravidian Etymological Dictionary 1984. Second Edn. Oxford
(DEDR).
Gardiner, A. 1957. Egyptian Grammar. Third Edn. (1978
reprint). Oxford.
Ghurye, G.S. 1977. Indian Acculturation : Agastya and Skanda.
Bombay.
Heras, Henry 1953. Studies in Proto-lndo-MeditelTanean
culture. Bombay.
Lal, B.B. 1979. On the most frequently used symbol in the
Indus script. East and West29 (1-4) : 27-35.
Lockwood, Michael 2001. Palla va Art Chennai.
Mahadevan, Iravatham 1970. Dravidian parallels in Proto-
Indian script. Journal of Tamil Studies 2 (1) : 157-276.
------ 1977. The Indus script: Texts, Concordance and Tables.
New Delhi.
------ 1981. Place Signs in the Indus script. Procds. of the V
Intemational Conference of Tamil Studies. VoU : 2.91-2.107.
------ 1982. Terminal Ideograms in the Indus script. (In)
Harappan Civilisation (ed.) Gregory Possehl. New Delhi;
pp 311-317.
------ 1986. Agastya Legend and the Indus Civilisation. Journal
of Tamil Studies 30: 24 - 37
------ 1995. An encycopaedia of the Indus script (Review of
Parpola 1994.) Book Review 19(6) : 9-12.
------ 1998. Phonetic value of the arrow Sign in the Indus script.
Journal of the Institute of Asian Studies 15 (2) : 69-74.
------2003. Early Tamil Epigraphy. Crea, Chennai & Harvard
University, USA.
Mahadeva Sastri, K. 19(i9. Historical Grammar of Telugu. Sri
Venkateswara University.

NO. 76 DECEMBER 2009

 110 JOURNAL OF TAMIL STUDIES

Nagaraju, S. 1979. An unpublished inscription from Kanheri :

Clue for the identification of an ancient almonry. Journal of the
Epigraphical Society of India. 6:46-49.
Narasimhia, A.N. 1941. A Grammar of the oldest Kanarese
inscriptions. Mysore.
Parpola, Asko 1994. Deciphering the Indus script. Cambridge
University Press.
Parpola, Asko and Simo Parpola 1975. On the relationship of
the Sumerian toponym Meluhha and Sanskrit mleccha. Studia
Orientalia 46:205:238.
Parpola, Simo; Asko Parpola and Robert Brunswig 1977. The
Meluhha Village: Evidence of acculturation of Harappan
traders in late third millennium Mesopotamia. Journal of
Economic and Social History of Orient 20(2) : 129-165.
Possehl, Greogory 1996. Meluhha. Julian Reade (ed.)
The Indian Ocean in Antiquity, pp. 133-208.
--------- 1997. Seafaring merchants of Meluhha. Bridget Allchin
(ed.) South Asian Archaeology 1995 : pp.87-100.
-------- 2008. An Akkadian Translator of the Meluhhan
Language: Some implications for the Indus Writing System.
AJi-avati:(Essays in honour of Iravatham Mahadevan)
pp.l39-147.
Potts, Daniel 1982. The road to Meluhha. Journal of Near
Eastem Studies 41 (4): 279-288.
Raghavaiyangar, M. 1907. Ve.Jir Varal<ifu. (2004 reprint).
Raghavaiyangar, R.1941. Tamil Varala..ru. Annamalai
University (1978 reprint).
Subbarayulu, Y. 1996. Catalogue of Pottery Inscriptions from
KodumanaI. Tamil University, Thanjavur.
Thapar, Romila 1975. A possible identification of Meluhha,
Dilmun and Makan. Joumal of Economic and Social History
of the Orient 18(1) :1-42.

NO. 76 DECEMBER 2009