Professional Documents
Culture Documents
38 Proofs III (MDS)
38 Proofs III (MDS)
38 Proofs III (MDS)
M ATHEMATICS IN I NDIA :
F ROM V EDIC PERIOD TO M ODERN TIMES
Lecture 38
Proofs in Indian Mathematics 3
M. D. Srinivas
Centre for Policy Studies, Chennai
Outline
Yuktibh
a.s
a estimate of the samagh
ata-sankalita
1k + 2k + . . . nk for large n.
Yuktibh
a.s
a estimate of V
arasankalita
Yuktibh
a.s
a derivation of M
adhava Series for
Yuktibh
a.s
a derivation of end-correction terms
Yuktibh
a.s
a derivation of M
adhava Rsine and Rcosine
Series
Yuktibh
as.
a of Jyes.t.hadeva
The most detailed exposition of upapattis in Indian mathematics is
found in the Malayalam text Yuktibh
a.s
a (1530) of Jyes.t.hadeva.
At the beginning of Yuktibh
a.s
a, Jyes.t.hadeva states that his purpose is
to present the rationale of the results and procedures as expounded
in the Tantrasangraha.
Variyar
(c.1500-1556) in his commentaries Kriy
akramakar (on Ll
avat) and
Yuktidpik
a (on Tantrasangraha)
Yuktibh
a.s
a has 15 chapters and is naturally divided into two parts,
Mathematics and Astronomy. In the Mathematics part, the first five
chapters deal with logistics, arithmetic of fractions, the rule of three
and the solution of linear indeterminate equations. Chapter VI
presents a detailed derivation of the M
adhava series for , his
estimate of the end-correction terms and their use in transforming the
series to ensure faster convergence. Chapter VII discusses the
derivation of the M
adhava series for Rsine and Rversine. This is
followed by derivation of various results on cyclic quadrilaterals and
the surface area and volume of a sphere.
Yuktibh
as.
a Estimation of Samagh
ata-Sankalita
which is
the sum of powers of natural numbers
(k)
Sn = 1k + 2k + . . . nk
Firstly, it is noted that the m
ula-sankalita
n2
n(n + 1)
for large n
2
2
Then, we are asked to write the varga-sankalita
as
(1)
Sn = 1 + 2 + . . . n =
(2)
Sn = n2 + (n 1)2 + . . . + 12
and subtract it from
(1)
n Sn = n [n + (n 1) + . . . + 1]
and get
(1)
(2)
n Sn Sn
= 1.(n 1)
= (n 1)
+2.(n 2)
+(n 2)
+(n 2)
+3.(n 3) + . . .
+(n 3) + . . .
+(n 3) + . . .
+(n 3) + . . .
+(n 1) .1
+1
+1
+1 + . . .
Estimation of Samagh
ata-Sankalita
Thus,
(1)
(2)
(1)
(1)
(1)
Sn
(1)
Sn
n Sn
(2)
(2)
it is argued that
(1)
n Sn
n2
2,
Therefore
(2)
Sn
n3
for large n.
3
Estimation of Samagh
ata-Sankalita
(3)
Sn
(k1)
n Sn
(k)
Sn
(k1)
(k1)
(k)
over Sn
(k1)
can be
Estimation of Samagh
ata-Sankalita
(k1)
Sn
has already been estimated to be,
k
(k 1)
Sn
nk , for large n, then the above relation leads to
(k1)
n Sn
(k )
Sn
(n 1)k
(n 2)k
(n 3)k
+
+
+ ...
k
k
k
1
(k)
Sn1
k
Sn
nk +1
for large n.
(k + 1)
Yuktibh
as.
a Estimation of V
arasankalita
or v
arasankalita)
of
natural numbers, given by
(1)
= 1 + 2 + 3 + ... + n =
(r )
= V1
Vn
Vn
(r 1)
(r 1)
+ V2
n(n + 1)
2
(r 1)
+ . . . + Vn
In Gan.itakaumud (c.1356) of N
ar
ayan.a Pan.d.ita, we find the
formula
n(n + 1) . . . (n + r )
(r )
Vn =
(r + 1)!
The above result is also known to the Kerala Astronomers, but
(r )
they prefer to derive the estimate for Vn , for large n, by
mathematical induction.
Estimation of V
arasankalita
Now,
(1)
Vn
(2)
We can express Vn
(2)
Vn
n2
n(n + 1)
for large n.
2
2
in the form
(1)
= Vn
(1)
+ Vn1 + . . .
(2)
n2 (n 1)2
S
+
+ ... = n
2
2
2
Sn
n2
,
3
we get
(2)
Vn
n3
6
Estimation of V
arasankalita
(r )
Vn = Vn
(r 1)
+ Vn1 + . . .
Vn
nr
,
(r )!
then we get,
(r )
Vn
(n 1)r
nr
+
+ ...
(r )!
(r )!
(r )
Sn
(r )!
nr +1
for large n.
(r + 1)!
Yuktibh
as.
a Derivation of M
adava Serires for
Yuktibh
a.s
a has presented the following derivation of the Madhava
series for the ratio of the circumference of a circle to its diameter. For
this purpose consider the quadrant OEAS of the square which
circumscribes the circle of radius r . The eastern side of the quadrant
is divided into a large number n of equal parts Ai Ai+1 = nr .
Join the hypotenuses (karn.as) O A1 , O A2 , . . . which meet the circle
at C1 , C2 , . . . Drop the perpendiculars, Ai Pi and Ci Qi onto OAi+1 .
Derivation of M
adhava Series for
If we note that the triangles Ai Pi Ai+1 and OEAi+1 are similar
and that the triangles OCi Qi and OAi Pi are similar, then we get
OE
OCi
OCi
= Ai Ai+1
Ci Qi = Ai Pi
OAi
OAi+1
OAi
Derivation of M
adhava Series for
We shall approximate the arc-bits Ci Ci+1 , by the corresponding
Rsines, Ci Qi . It is noted that larger the n the more accurate will be
the result.
If we denote the hypotenuse OAi as ki , then we get
r r 2 r 2
r2
C
+
+ ... +
8
n
k0 k1
k1 k2
kn1 kn
Derivation of M
adhava Series for
It is noted that when n is large,
#
"
1
1
1
1
+ 2
ki ki+1
2
ki2
ki+1
and that the earlier sum for the circumference can be replaced by
2
r r 2 r 2
r
C
+
+ ... +
2
2
8
n
k1
k2
kn2
If we note that
ki2
2
ir
=r +
n
2
then we get
2
2
r
C
r
r
+
2 + . . . +
r 2
8
n
2
2
r +
r + 2r
n
r
r2
nr 2
n
Derivation of M
adhava Series for
Each of the terms in the above sum for the circumference can be
expanded as a binomial series (which has been derived earlier in
Yuktibh
a.s
a) and we get, on regrouping the terms,
r
C
=
[1 + 1 + . . . . . . + 1]
8
n
"
#
r 1 r 2 2r 2
nr 2
+
+ ... +
n
r2
n
n
n
"
#
4
r 1
r 4
nr 4
2r
+
+
+ ... +
n
r4
n
n
n
......
Now, each of the sama-gh
ata-sankalita
Yuktibh
as.
a Derivation of the End-Correction Terms
The M
adhava series (or the so called Leibniz series) for the
circumference of a circle (in terms of odd numbers p = 1, 3, 5, . . . )
(p1) 1
1
2
+ ...
C = 4d 1 + . . . + (1)
3
p
is an extremely slowly convergent series. Adding fifty terms of the
series will give the value of correct only to the first decimal place.
In order to facilitate computation, M
adhava has given a procedure of
using end-correction terms (antya-sam
ara), of the form
. sk
(p1)
(p+1) 1
1
1
+ (1) 2
C = 4d 1 + . . . + (1) 2
3
p
ap
Both Yuktibh
a.s
a and Kriy
akramakar give a derivation of the
successive end correction terms given by M
adhava, which involve a
careful estimate of the inaccuracy (sthaulya) at each stage in terms of
inverse powers of the odd number p.
1
ap2
1
1
ap
p
E(p) =
=
1
1
1
+
(2p 2) (2p + 2) p
1
3
(p p)
A
2p + 2
4
(2p + 2)
4
(p5 + 4p)
16
(2p+2)
2304
+
+ 1792p3 2304p)
36
3
2
[(p p){(p 1) + 5}{(p + 1)2 + 5}]
(64p7
448p5
1
22
(2p + 2) +
42
(2p + 2) +
(2p + 2) +
62
(2p + 2) + . . .
Yuktibh
as.
a Derivation of the M
adhava Sine Series
Given an arc EC = s = Rx, divide it into n equal parts. The
pin.d.a-jy
as Bj = Cj Pj , kot.i-jy
as Kj = OPj and saras Sj = Pj E, with
j = 0, 1 . . ., are given by
h
i
Bj = R sin jxn , Kj = R cos jxn , Sj = Rvers jxn = R 1 cos jxn
Yuktibh
as.
a Derivation of the M
adhava Sine Series
Let Cj Cj+1 be the (j + 1)-th arc-bit. Let Mj+1 be the mid-point of the
arc-bit Cj Cj+1 and similarly Mj the mid-point of the previous (j-th)
arc-bit. Let the full-chord of the equal arc-bits ns be denoted . We
can easily see that the triangles Cj+1 FCj and Mj+1 GMj are similar to
OQj+1 Mj+1 and OPj Cj respectively.
Derivation of the M
adhava Sine Series
We can thus show,
Bj+1 Bj =
Kj+ 1 and Kj 1 Kj+ 1 = Sj+ 1 Sj 1 =
Bj
2
2
2
2
2
R
R
Therefore, the second order Rsine differences (jy
a-khan.d.
antaras) are
given by
(Bj Bj1 ) (Bj+1 Bj ) =
2
Sj+ 1 Sj 1 =
Bj
2
2
R
R
Hence
Sn 1 S 1
Bn n B1
(B1 + B2 + . . . + Bn1 )
R
2
S 1 + S 3 + . . . + Sn 1 nS 1
2
2
2
2
R
Derivation of the M
adhava Sine Series
The above relations are exact. Now, if B and S are the jy
a and sara of
the arc s, in the limit of very large n, we have
s
Bn B, Sn 1 S, S 1 0,
2
2
n
and hence
s
S
(B1 + B2 + . . . + Bn1 )
nR
s 2
B n B1
[B1 + (B1 + B2 ) + . . . + (B1 + B2 + . . . + Bn1 )]
nR
In the above relations, we first approximate the Rsines (jy
a-khan.d.as)
by the arcs (c
apas), Bj jsn , and make use of the estimates for sums
and repeated sums of natural numbers for large n, to get
2
s
s2
1
(1 + 2 + . . . + n 1)
S
R
n
2R
s 1 2 s 3
B n
[1 + (1 + 2) + . . . + (1 + 2 + . . . + n 1)]
n
R
n
s3
6R 2
Derivation of the M
adhava Sine Series
We now substitute the above second approximation for jy
a-c
ap
antara
3
js
n
js
Bj
n
6R 2
Then we get the next approximation
s4
s2
2R
24R 2
3
s
s5
B s
+
2
6R
120R 4
The above more refined approximation for jy
a-c
ap
antara is again fed
back into our original equations for B and S, and so on. In this way,
we are led to the series given by M
adhava for Rsine and Rversine
"
#
s 3
s 5
s
s
R
R
R sin
= R
+
...
R
R
3!
5!
" 2
#
s
s 4
s 6
s
R
R
R
R R cos
= R
+
...
R
2!
4!
6!
S
However vagaries of the external world were not by themselves responsible for the failure of Greek mathematics to
advance materially beyond Archimedes. There were also
internal factors that suffice to explain this failure. These
impeding factors centred on the rigid separation in Greek
mathematics between geometry and arithmetic (or algebra),
and a one-sided emphasis on the former. Their analysis dealt
solely with geometrical magnitudes lengths, areas, volumes
rather than numerical ones, and their manipulation of these
magnitudes was exclusively verbal or rhetorical, rather than
analytic (or algebraic as we would say today). ...
Ever since the seminal work of Needham, who showed that till
around the sixteenth century Chinese science and technology
seem to have been more advanced than their counterparts in
Europe, it has become fashionable for historians of science to
wonder Why modern science did not emerge in non-western
societies?
In the work of the Kerala School, we notice clear anticipations
of some of the fundamental discoveries which are associated
with the emergence of modern science, such as the mathematics of infinite series and the development of new
geometrical models of planetary motion.
References
1. Gan.itayuktibh
a.s
a of Jyes.t.hadeva (in Malayalam), Gan.it
adhy
aya,
Ed., with Notes in Malayalam, by Ramavarma Thampuran and
A. R. Akhileswara Aiyer,Trichur 1948.
2. Works of C. T. Rajagopal and Collaborators:
2.1 K. Mukunda Marar, Proof of Gregorys series, Teachers
Magazine 15, 28-34, 1940.
2.2 K. Mukunda Marar and C. T. Rajagopal, On the Hindu
quadrature of the circle, J.B.B.R.A.S. 20, 65-82, 1944.
2.3 K. Mukunda Marar and C. T. Rajagopal, Gregorys series in
the mathematical literature of Kerala, Math. Student 13,
92-98, 1945.
2.4 A. Venkataraman, Some interesting proofs from Yuktibh
a.s
a,
Math Student 16, 1-7, 1948.
2.5 C. T. Rajagopal, A neglected chapter of Hindu
mathematics, Scr. Math. 15, 201-209, 1949.
2.6 C. T. Rajagopal and A. Venkataraman, The sine and cosine
power series in Hindu mathematics, J.R.A.S.B. 15, 1-13,
1949.
References
2.7 C. T. Rajagopal and T. V. V. Aiyar, On the Hindu proof of
Gregorys series, Scr. Math. 17, 65-74, 1951.
2.8 C. T. Rajagopal and T. V. V. Aiyar, A Hindu approximation to Pi,
Scr. Math. 18, 25-30, 1952.
2.9 C. T. Rajagopal and M. S. Rangachari, On an untapped source
of medieval Keralese mathematics, Arch. for Hist. of Ex. Sc. 18,
89-101, 1978.
2.10 C. T. Rajagopal and M. S. Rangachari, On medieval Kerala
mathematics, Arch. for Hist. of Ex. Sc. 35(2), 91-99, 1986.
3. Gan.itayuktibh
a.s
a of Jyes.t.hadeva (in Malayalam), Ed. with Tr. by
K. V. Sarma with Explanatory Notes by K. Ramasubramanian,
M. D. Srinivas and M. S. Sriram, 2 Volumes, Hindustan Book
Agency, Delhi 2008.
nkara
4. Kriy
akramakar of Sa
V
ariyar on Ll
avat of Bhaskaracarya
II: Ed. by K. V. Sarma, Hoshiarpur 1975.
nkara
5. Tantrasangraha
V
ariyar, Ed. K. V. Sarma, Hoshiarpur 1977.
References
6. C. H. Edwards, The Historical Development of the Calculus,
Springer, New York 1979.
7. S. Parameswaran, The Golden Age of Indian Mathematics,
Swadeshi Science Movement, Kochi 1998.
8. M. D. Srinivas, Proofs in Indian Mathematics, in G. G. Emch,
R. Sridharan and M. D. Srinivas Eds. Contributions to the
History of Indian Mathematics, Hindustan Book Agency, Delhi
2005, pp.209-248.
9. C. K. Raju, Cultural Foundations of Mathematics: The Nature of
Mathematical Proof and the Transmission of the Calculus from
India to Europe in the 16th c.CE, Pearson Education, Delhi 2007.
10. G. G. Joseph, A Passage to Infinity, Sage, Delhi 2009.
11. K. Ramasubramanian and M. D. Srinivas, Development of
Calculus in India, in C. S. Seshadri (ed), Studies in History of
Indian Mathematics, Hindustan Book Agency, Delhi 2010,
pp.201-286.
Thanks!
Thank You