Frits Staal: Sulbas Utras of The Last Millenium B.C
Frits Staal: Sulbas Utras of The Last Millenium B.C
Frits Staal: Sulbas Utras of The Last Millenium B.C
I. RITUAL GEOMETRY
106
FRITS STAAL
Some of these facts have long been known and can no longer be ignored
even by Euro- (or Indo-)centric historians of science. In such cases,
what J. D. Bernal called arrogant ignorance shifts ground. Unable to
deny that the Theorem of Pythagoras occurs worldwide, some historians
continue to believe (following, e.g., Sir Thomas Heath, the translator
of Euclids Elements: 1956, I: 363) that only the Greeks were capable
of abstraction or providing proofs. This can be disproved with the help
of Indian mathematics as Seidenberg (e.g., 1978, 1983) and van der
Waerden (1983: 26 sq., adding China on p. 36) have shown (for more
detailed discussion see Hayashi 1994: 122123 and 1995: 7277).
Van der Waerden has argued on heuristic grounds against independent invention as an explanation for such similarities. He prefers
diffusionism as a working hypothesis, i.e., the idea that mathematical
ideas arise only once and spread from a center. Much depends, of
course, on how specific the mathematical ideas we study are. There
may be no need to argue for the diffusion of elementary arithmetic
truths such as 2 2 = 4. But in the case of the geometrical constructions of ancient Greece and India, the degree of specificity is high and
independent invention therefore unlikely. Direct influences from Greece
to India or vice versa during that period are equally improbable. Can
we demonstrate a common origin?
It would seem natural to look for such an origin in Mesopotamia, situated between Greece and India, and especially among the Babylonians
who possessed a highly developed mathematics at a much earlier period.
Van der Waerden espoused this view in his well-known book Science
Awakening of 1954. But Babylonian mathematics is different from the
Indo-Greek variety: it was not constructive; it stressed algebraic and
computational methods including fractions expressed through the sexa-
107
108
FRITS STAAL
109
110
FRITS STAAL
Figure 1. Eurasia. AM: Asia Minor; MS: Mitanni Syria; T: Tocharian; BM: BactriaMargiana.
cube. Several legends dating from between the fifth and third century
B.C. refer to the duplication of the size of an altar or royal tomb of
that shape. The ritualists and architects realized that doubling the sides
would not do the trick: for when the sides are doubled, the area is
enlarged fourfold and the volume eightfold. Experts on geometry had
to be consulted. But why should an altar be doubled in the first place?
The oracle at Delos prescribed this duplication in order to fight a
plague. Seidenberg does not explain why this occurred to the oracle but
refers to a passage in which Plato expresses his scorn for priests who
use ritual for such purposes as expiating sins or harming an enemy:
Mendicant prophets go to rich mens doors and persuade them that they have a
power committed to them by the gods of making an attonement for a mans own
or his ancestors sins by sacrifices or charms, with rejoicings and feasts; and they
promise to harm an enemy, whether just or unjust, at small cost; with magic arts
and incantations binding heaven, as they say, to execute their will (Republic 364
bc; transl. B. Jowett).
111
Srauta
Sutra 17: 29, translated by Ikari and Arnold 1983 II: 668671;
cf. Pingree 1981: 3 sq. ubi alia). Several details are missing in these
texts and were transmitted orally. They are preserved by the Nambudiri
Brahmans of Kerala who have maintained, up to the present, three
traditions of bird altars: the Six-Tipped, the Five-Tipped and the
Square. For there is an important difference between ancient Greece
and India: ancient Greek civilization is no longer a living tradition
and there is nothing left of Greek ritual. In India, on the other hand,
some of the ancient ritual and geometric knowledge survives (albeit
in inaccessible corners) through unbroken chains of transmission as
is demonstrated by precisely these oral traditions. The following brief
description is based upon the Nambudiri tradition (cf. Staal 1982 and
for a complete account: Staal et al. 1983).
In the first Nambudiri tradition, the Bird Altar with Six-Tipped Wings,
the first layer looks as in Figure 2:
112
FRITS STAAL
113
In the first, third and fifth layers of the Five-Tipped Bird Altar, there
are 61 squares, 136 triangles (of two sizes) and three pentagons (of
two sizes and three shapes). The second and fourth layers consist of
72 squares and 128 triangles (of two sizes). The constructions of these
from the basic square is accordingly different. In the Square Bird Altar,
which seems trivial at first, the construction is different again and not
any simpler.
The mathematics of bricks is further constrained by other ritual
requirements. The total area of each layer of the altar must be sevenand-a-half times a square purus. a, i.e., a square of which the side is
the size of the Yajamana. Since the Yajamana was measured in five
units to arrive at the unit square of Figure 3, the square purus. a is 25
times that unit square. The numbers, shapes and areas of the bricks are
part of the oral tradition but the computation of the areas is not stated
anywhere in clear terms. This is not surprising since the matter is far
from obvious: for we should not forget that the Vedic Indians, like
their Greek cousins, lacked simple expressions for numbers, whether
integers or fractions. All calculation was done geometrically which is
more complicated, especially for us, than if it were done in our modern
notation. Only by adopting the latter can the number and area of bricks
be easily expressed. For the first and third layers of the Six-Tipped
Altar, it may be done as follows:
NUMBER
square
1 1/4
1 1/2
1/2
3/4
38
2
56
60
44
TOTAL
200
AREA
1
1.25
1.5
0.5
0.75
38
2.5
84
30
33
187.5 = 7 1/2
25
The fifth layer is similar but there are 205 bricks, 10 of half thickness.
The computation for the second and fourth layers are more complex
but basically similar.
None of the evidence from the Agnicayana we have so far considered
is in an obvious manner related to Greek geometry. Arrogant ignoramuses
will be quick to point out that this Agni does not look like the deductive
system of Euclid, which is true. The only Indian counterpart to Euclid
is the derivational system of Pan. inis Sanskrit grammar. It is in some
114
FRITS STAAL
I shall begin with Indian linguistics where the earliest works are the
Pratisakhyas. They are early because of their structure and function,
not because of the form in which they survive and which has already
been influenced by Pan. inis grammar of the fourth century B.C. (cf.
Cardona 1976: 273275). There is, as the name Pratisakhya indicates,
a)
of the Veda.
one of these treatises for each of the branches (sakh
They are practical manuals that provide the rules for converting the
. ha (word-for-word recitation) of their own branch into its
padapat
sam
. hita (continuous recitation). Some of these rules happened to be
general rules that pertain to the language of all the Vedas; though
variously formulated, they recur in most or all of the Pratisakhyas.
They also occur in Pan. inis grammar which deals primarily with the
language spoken during his time and in his part of country, a language
that happened to be more or less identical with the language of late
Vedic prose (i.e., the Brahman. as and Sutras) and came to be called
Classical Sanskrit.
Pan. ini included cases where (earlier) Vedic deviated from Classical
Sanskrit, but did not mention it when the same grammatical forms
are exhibited by both languages. Hence Patan~jalis remark on Pan. inis
. adam idam
grammar: sarvavedaparis
this science pertains to
. sastram,
all the Vedas the motto of Paul Thiemes 1935 book Pan. ini and
the Veda in which these distinctions were for the first time elucidated
clearly in a modern language.
The Sulba
Sutras are attached to the Srauta
Sutras, ritual sutras that
are practical manuals like the Pratisakhyas: they provide the officiating
priests with rules for the execution of their rites. Because of the proliferation of ritual, the subdivision into sutras is more precise than that into
115
s
akhas: there are several sutras within each sakha. All Sulbas
utras are
attached to Srautasutras of the Yajurveda, the ritual Veda par excellence
and the core of the Vedic tradition.1 It possesses two Pratisakhyas (one
for the Kr. s. n. a- and one for the Sukla-Yajurveda)
and ten Srautas
utras
to six of which the six known Sulbasutras are attached.
The Srautas
utras generally exhibit rites that are specific to their own
school. Others are shared by several manuals, e.g., many of the metarules
. a)
from which the science of ritual developed (Staal 1982).
(paribhas
The Sulbas
utras share even more of their contents with each other:
they deal with the geometry of altar construction which is the same
even if the altar shapes are different. The relation of the Sulbato the
Srautasutras is, therefore, similar to that of Pan. ini to the Pratisakhyas.
They pertain to all the Vedas also. Their compilers did not continue
to restrict themselves to their own ritual sutra, but began to develop
a more universal discipline. These manuals are therefore, as Thieme
wrote of Pan. ini, not merely practical but scientific. One could go further
and apply what David Hilbert wrote about the importance of scientific
work in general: it can be measured by the number of previous works
it makes superfluous to study (quoted by Neugebauer 1957: 145).
The emergence of geometry from ritual analysis is likely to have
happened earlier than that of linguistics from the analysis of recitation
because of the chronological priority of Baudhayana: Pan. ini probably
belonged to the fourth century B.C., but the Baudhayana Sulba
Sutra may
have been compiled before 500 B.C. (Pingree 1981: 4). The important
Baudhayana Srauta
Sutra is certainly earlier: 8th or 7th century B.C.
(see Staal 1989: 305 and the literature cited there).
The development of scientific geometry is exemplified by the construction mentioned by Seidenberg and van der Waerden, which is found in
the Sulbas
utras of Baudhayana, Apastamba
and Katyayana. It exemplifies both strengths and weaknesses of Greek and Vedic mathematics. I
mentioned that both traditions lacked simple expressions for numbers;
but they share deficiencies that go deeper. Both lacked a symbolic notation such as is used in modern algebra, a discipline Europe imported
from the Arabs who developed it from what they had inherited from
ancient Mesopotamia and discovered in India and China. The Arabs
combined and were accordingly in a position to disentangle the confusing
muddle of number references and notations they had inherited from
their predecessors: words which had been employed by all ancient
peoples; Greek letters; and Indian numerals which ultimately prevailed
and then led to the next step which combined them again with letters
viz., algebra. The only Greek scientist who used letters in such a manner
116
FRITS STAAL
as variables had been Aristotle and that was not in mathematics but in
logic.2
Instead of modern algebra, the ancient Greeks and Vedic Indians
possessed what has been called geometrical algebra, an algebra that
makes use of geometrical methods and that we can only understand
adequately if we desist from reading the expressions of modern algebra
into it. The situation is similar to Newtons equations which I discussed
in the 1995 issue of this Journal (23: 76) where I quoted the historian
of science C. Truesdell: It is true that we, today, can easily read them
into Newtons words, but we do so by hindsight. We similarly tend to
read modern equations into early medieval Indian algebra, which was
just beginning to develop a symbolic notation, and are predisposed to
do the same with respect to the much earlier phases of Vedo-Greek
ritual geometry.
I shall begin with geometrical algebra, following Seidenberg. In the
following figures, capital letters refer to points and line segments. I
shall use small letters to refer to lengths and express their relationships
through the language of modern algebra. In that language, the entire
ancient edifice of geometrical constructions is reduced to one line: if
the sides of the given rectangle are a and b, we are seeking a square
with side c such that:
c2 = a b and hence c =
a b.
117
squares AGKJ and FHKI is the area of the original rectangle. End of
Step 1.
To construct the difference of two squares geometrically (Step 2),
use is made of the Theorem of Pythagoras as in Figure 5:
Figure 5. Constructing a square equal to the difference between two squares with the
help of the Theorem of Pythagoras.
118
FRITS STAAL
opposite side [making the arc DK]. One should cut that side [IJ] at the
point [K] where it [AD] falls. By the line [IK] which has been cut off,
the small square is deducted.
Baudhayana divided the Theorem of Pythagoras into two complementary parts: proposition 1.51 relates to the difference between two
squares; the addition had been formulated in 1.50 (corresponding to
Euclid I.47). Note that Step 2 may be incorporated in Step 1 by draw
ing (in Figure 4) KG of the cut-off piece KGEI obliquely (aks. n. aya).
I have already expressed the problem in the simple language of modern
algebra. If we combine Seidenberg with van der Waerden (1983: 11),
we may depict its original Vedo-Greek complexity in algebraic terms,
as follows. Refer to the base AB of the original rectangle of Figure 4
as a and to its height AD = BC as b. The area of that rectangle is ab.
The area of square AEFD with sides equal to AD is b2 . The area of
the remainder EBCF is (a b)b. The area of each of the two equal
rectangles EGHF and GBCH into which it is divided is ((a b)b)/2.
After placing one of these on top, the area of the small square FHKI
which fills the empty space in the corner is ((a b)/2)2 or (a2 2ab +
b2 )/4. The area of the larger square AGKJ which has been created is
(b + (a b)/2)2 or ((a + b)/2)2 or (a2 + 2ab + b2 )/4. With the help of
algebra, it is easy to see that the difference between the two squares is
ab, i.e., the area of the original rectangle.
Now the deduction of the smaller square from the larger one. Call
the side of the second b, that of the first a and that of the difference
between the two that has to be found: c. If, in Figure 5, AI = b and
AK = AD = a, IK is the side of the square, c, since: c2 = a2 b2 an
algebraic expression of Pythagoras Theorem.
Seidenberg discovered that the same two steps were taken by Euclid
in Book II of his Elements. Van der Waerden concludes:
This is only one of several possible constructions. The fact that Euclid and the
Sulvas
utras
both use one and the same more complicated construction, based on the
Theorem of Pythagoras, is a strong argument in favour of a common origin.
119
The only two Vedas that concern us are the Rig- and Yajurveda. The
Rigveda was composed during the second millenium B.C. (the bulk
of it before 1100 B.C.: Witzel 1992: 614); but it does not mention the
Agnicayana and refers to altars only thrice. In each case, the altar seems
to be simple and part of the domestic ritual, performed within the home
and different from the Agnicayana ritual which is a (relatively) public
ritual. In the one (late) case in which the Rigveda description implies
a shape, it is quadrangular (RV 10.114.3; cf. Potdar 1953: 73). In the
domestic ritual of a few centuries later, about which more evidence is
available, the altar is circular. Though not referred to, it is likely that
this circular altar was already known at the time of the Rigveda.
The Yajurveda was composed between 1000 and 800 B.C. and a full
third of it is devoted to the Agnicayana. The language is still similar
to that of the Rigveda but begins to develop in different directions
(see Renou 1956: Ch. I; Witzel 1989). Whatever the difference in
language between the two Vedas with their numerous subdivisions, it
is likely that there were also social and ethnic differences among its
users. Most of the Rigveda was probably composed by members of the
semi-nomadic Indo-European pastoralists whose tribes and clans had
trickled in across the mountain ranges that separate Central Asia from
Iran and the Indian subcontinent; whereas parts of the Yajurveda are
more likely to have been composed by indigenous Indians who had
become bilingual by adopting the language of these incoming tribes
as a second language (cf. Deshpande 1993). Why they adopted this
alien language has not been satisfactorily explained; but the fact is not
in doubt and not confined to India: it holds true of Indo-European in
general (see Mallory 1989: 257261).
It is clear that nomads do not carry bricks across high mountain
passes and are unlikely to construct altars from them. It would stand
to reason, then, to assume that the bricks of the Agnicayana were a
product of Indian soil. This thesis was defended by H. S. Converse in
an important article of 1974. Her argument for the indigenous origin of
the Agnicayana was not only based upon the hypothesis that the art of
baking bricks is unlikely to have been known to nomads, but upon the
more important positive fact that it was common in Northwest India in
the early second millenium B.C., from the time of Indus Civilization
cities such as Harappa. Converse (1974: 83) writes: The Harappans
used millions of kiln-fired bricks as well as countless sun-baked ones.
(No bricks were fired in kilns in the urban civilizations of ancient
Mesopotamia.)
120
FRITS STAAL
121
122
FRITS STAAL
been one of the most important junctions on the trade routes between
West, East and South Eurasia. Contacts between the Indus and Oxus
regions started early and continued through historical times (Jettmar
1986: 145146; Brough 1965, republished 1996: 276312). Bactria does
not occur among the civilizations compared by van der Waerden (1983:
4445) or Joseph (1990: 10, 14) but Joseph refers in his account of
123
124
FRITS STAAL
NOTES
1
Bref, cest le Yajurveda qui demeure la base du culte et qui sans doute a determine
toute levolution du vedisme litteraire qui forme un tout coherent, ou chacque partie
sest organisee en liaison systematique avec les autres: Renou 1947: 9, 209210.
2
Lukasiewicz (1957: 78) expressed surprise that no philosopher or philologist
had paid attention to this discovery except the Oxford editor of most of Aristotles
works, Sir David Ross. Otto Neugebauer was more explicit: The origin of algebra is
totally independent of an algebraic notation in one of the most famous philosophical
works of antiquity (1957: 225).
3
Kashikar (1979) has argued against this conclusion (as formulated in Staal 1978)
and against any so-called pre-Vedic element; but see Staal (1982: 4247). That
the Harappans possessed a highly developed system of weights and measures was
shown by Chattopadhyaya 1986; but it is not reflected in the Vedas and nothing is
known about Harappan mathematics (Hayashi 1994: 126).
4
As in the case of germs, paucity of carriers or wandering geniuses does not
125
diminish the force of these conclusions. The further we go back in time, the smaller
the number of people who contributed decisively to the development of human skills
and knowledge. Greek mathematics itself consists of the fragments of writings of
about 10 or 20 persons scattered over a period of 600 years (Neugebauer 1957: 190).
The greatest, perhaps, of all human inventions or discoveries, the taming of fire,
may have been the single most important event in 10,000 years of human evolution.
REFERENCES
Bernal, J. D. (1969), Science in History, London: Penguin Books.
Brough, John (1965), Comments on third-century Shan-shan and the history of
Buddhism, Bulletin of the School of Oriental and African Studies 28: 582612.
Reprinted in Brough 1996: 276307.
Brough, John (1996), Collected Papers, eds. Minoru Hara and J. C. Wright, School
of Oriental and African Studies, London.
Burrow, T. (1955), The Sanskrit Language, London: Faber and Faber.
Burrow, T. (1973), The Proto-Indoaryans, Journal of the Royal Asiatic Society
123140.
Caillat, Colette (1989), Dialectes dans les litteratures indoaryennes, Paris: Institut
de Civilisation Indienne.
. ini. A Survey of Research, The Hague: Mouton
Cardona, George (1976 = 1980), Pan
= Delhi Varanasi Patna: Motilal Banarsidass.
Chattopadhyaya, D. (1986), History of Science and Technology in Ancient India: The
Beginnings, Calcutta: Firma KLM.
Converse, H. S. (1974), The Agnicayana Rite: Indigenous Origin? History of
Religion 14: 8195.
Deshpande, Madhav M. (1993), Sanskrit & Prakrit. Sociolinguistic Issues, Delhi:
Motilal Banarsidass Publishers.
Diakonoff, Igor M. (1995), Two Recent Studies of Indo-Iranian Origins, Journal
of the American Oriental Society 115: 473477.
Edwards, Paul (ed.) (1967), The Encyclopedia of Philosophy, New York: Macmillan.
Erdosy, G. (ed.) (1995), Language, Material Culture and Ethnicity, Berlin: W. de
Gruyter.
Friberg, J. (1990), Mathematik, Reallexikon der Assyriologie und Vorderasiatischen
Archaologie VII.7/8: 531585.
Wiesbaden: Otto Harrassowitz.
Gonda, Jan (1977), The Ritual Sutras,
Grattan-Guinness, I. (ed.) (1994), Companion Encyclopedia of the History and
Philosophy of the Mathematical Sciences, London and New York: Routledge.
Guthrie, W. K. C. (1967), Pre-Socratic Philosophy, in: Edwards: 441446.
Hayashi, Takao (1994), Indian Mathematics, in: Grattan-Guinness 118130.
Manuscript. An ancient Indian mathematical
Hayashi, Takao (1995), The Bhakshal
treatise, Groningen: Egbert Forsten.
Heath, Sir Thomas L. (1956), The Thirteen Books of Euclids Elements, IIII (second
edition), New York: Dover Publications.
Heesterman, J. C. (1962), Vratya and Sacrifice, Indo-Iranian Journal 1: 137.
Hiebert, Fredrik T. (1995), South Asia from a Central Asian Perspective, in: Erdosy,
ed., 192205.
Hiebert, Fredrik T. and Lamberg-Karlovsky (1992), Central Asia and the Indo-Iranian
Borderlands, Iran. Journal of Persian Studies 30: 115.
Hillebrandt, Alfred (1897), Ritual-Literatur. Vedische Opfer und Zauber, Strassburg:
K. J. Trubner.
126
FRITS STAAL
van den Hoek, A. W., Kolff, D. H. A. and Oort, M. S. (1992) (eds.), Ritual, State
and History in South Asia. Essays in Honour of J. C. Heesterman, Leiden: E.J.
Brill.
Hyrup, Jens (1994), Babylonian Mathematics, in: Grattan-Guinness 2129.
utra on the
Ikari, Yasuke and Arnold, Harold (1983), Baudhayanas Srautas
Agnicayana, in: Staal II: 478675.
Jarrige, Jean-Francois, Santonie, Marielle and Enault, Jean-Francois (1979) Fouilles
de Pirak, III, Paris: Diffusions de Boccard.
Jarrige, Jean-Francois (1985), Larcheologie de la Bactriane ancienne, Paris: Diffusions
de Boccard (especially pp. 105118: Les relations entre lAsie central meridionale,
la Baluchistan et la vallee de lIndus a la fin du 3e et au debut du 2e millenaire).
Jettmar, Karl (1986), The Religions of the Hindukush, I: The Religion of the Kafirs,
Warminster.
Joseph, George Gheverghese (1990), The Crest of the Peacock. Non-European Roots
of Mathematics, London: Penguin Books.
Jowett, B. (transl.; no date), The Dialogues of Plato, Vol. II: The Republic, New
York: Bigelow, Brown & Co.
utras,
Bombay: University of Bombay
Kashikar, C. G. (1968), A Survey of the Srautas
(= Journal of the University of Bombay 25 (New Series), Part 2, No. 41).
. n. n. a,
Annals of
Kashikar, C. G. (1979), Agnicayana: the Piling Up of svayamatr
the Bhandarkar Oriental Research Institute 60: 215218.
Lukasiewicz, Jan (1957), Aristotles Syllogistic from the Standpoint of Modern Formal
Logic, Oxford: Clarendon Press.
Mallory, J. P. (1989), In Search of the Indo-Europeans: Language, Archaeology and
Myth, London.
Masica, Colin P. (1989), The Indo-Aryan Languages, Cambridge: The University
Press.
Neugebauer, Otto (1957), The Exact Sciences in Antiquity, Providence: Brown
University Press.
Parpola, Asko (1993), Margiana and the Aryan Problem, Information Bulletin of the
International Association for the Study of the Cultures of Central Asia (Moscow:
Nauka) 19: 4162.
Pingree, David (1973), The Mesopotamian Origin of Early Indian Mathematical
Astronomy, Journal for the History of Astronomy 4: 112.
Pingree, David (1978), Mesopotamian Astronomy and Astral Omens in Other
Civilizations, Mesopotamien und seine Nachbarn. Politische und kulturelle Wechselbeziehungen im Alten Vorderasien vom 4. bis 1. Jahrtausend v. Chr., Berlin:
Dietrich Reimer 2: 613631.
Astral and Mathematical Literature, Wiesbaden:
Pingree, David (1981), Jyotih. sastra.
Otto Harrassowitz.
Pingree, David (1989), MUL.APIN and Vedic Astronomy, in: DUMU-E DUBBA-A: Studies in Honor of Ake W. Sjoberg. Editors Hermann Behrens, Darlene M.
Loding and Martha T. Roth, Philadelphia: The University Museum: 439445.
Potdar, K. R. (1953), Sacrifice in the R. gveda, Bombay.
vediques
et la formation du Veda, Paris: Imprimerie
Renou, Louis (1947), Les ecoles
Nationale.
Renou, Louis (1956), Histoire de la langue sanskrite, Lyon Paris: IAC.
Sarianidi, Viktor (1987), South-west Asia: Migrations, the Aryans and Zoroastrians,
Information Bulletin of the International Association for the Study of the Cultures
of Central Asia (Moscow: Nauka) 13: 4456.
Sulbas
utram
with Sanskrit
Satyaprakash and Sharma, R. S. (eds.) (1968), Baudhayana
Commentary by Dwarka Nath Yajvan and English Translation and Critical Notes
by G. Thibaut, New Delhi: The Research Institute of Ancient Scientific Studies.
127