Pramāṇa · Search · About · API

A Concise History of Science in India c7.s2.t4: Algebra

A Concise History of Science in India c7.s2.t4

Language
English

Algebra

c7.s2.t4.1

The altar geometry of the *Śulba-sūtras* does not fail to give us a glimpse of the beginnings of algebraic notions among the Vedic Hindus. These notions include quadratic equations, indeterminate equations, surds, conception of irrational numbers and determinations of their approximate values.

c7.s2.t4.2

FIG. 3.9a. *Mahāvedi.*

c7.s2.t4.3

Quadratic Equation

c7.s2.t4.4

The sacrificial ritual often necessitated enlargement or reduction of the altar in accordance with a number of plans. According to one plan, the shape is to remain the same, all sides being proportionally increased or decreased. The standard *mahāvedi*, for example, has the shape of an isosceles trapezium having for its face, base and altitude 24, 30 and 36 units respectively (area 972 sq. units). The standard *śyena-cit* (the falcon-shaped altar) consists of a central trunk of 4 sq. units, two rectangular wings each measuring 1 by 1⅕ or 1⅕ sq. units and a rectangular tail (*prādeśa*) 1 by 1 1/10 or 1 1/10 sq. units and has a total area of 7½ sq. units. By the first plan, the measures are left undisturbed but the unit of length is changed as required. According to another plan, some parts of the altar are proportionately increased while others are left undisturbed. Thus, in the case of *śyena-cit*, while the units for the complete square are proportionately changed, that of the fractional parts for the two *aratnis* and one *prādeśa* are left unchanged. If p represents the modified unit and α the change in area, we obtain the following quadratic equations for the two types of *vedis* mentioned above:

c7.s2.t4.5

FIG. 3.9b. *Śyena-cit.*

c7.s2.t4.6

Plan I: *Mahāvedi*: 36p× ((30p+24p))/2=(36×(30+24))/2+α

c7.s2.t4.7

972p²=α+972

c7.s2.t4.8

$Śyena-cit: 2p × 2p + 2[p(p+p/5)] + p(p+p/10) = 15/2 + α$ $15p² = 2α + 15$ $Plan II: Śyena-cit: 2p × 2p + 2[p(p+⅕)] + p[p+1/10] = 15/2 + α$ $14p² + p = 2α + 15$

c7.s2.t4.9

Specific examples given in the *śulba-sūtras* of Baudhāyana, Āpastamba and Kātyāyana agree with the solutions of these equations.

c7.s2.t4.10

Indeterminate Equations

c7.s2.t4.11

FIG. 3.10.

c7.s2.t4.12

The *śulbas* contain rules for the construction of a square *n* times a given square. The rule involves the application of the diagonal square theorem, for it says that the diagonal of the square produces the double square, the diagonal of the rectangle formed of the previous diagonal and the side of the given square gives a square thrice as large, the diagonal of the rectangle formed of the previous diagonal and the side of the given square produces a fourfold square and so on. Kātyāyana gives a general rule for finding a square *n* times a given square. According to this rule, the required square will be given by the altitude of an isosceles triangle whose base is (*n*−1) times and the sum of the two equal sides (*n*+1) times the side of the given square. That is, in Fig. 3.10, $BC = (n-1)a, AB = AC = ((n+1))/2a,$ where *a* is the side of the given square. It is clear that $AD² = [((n+1)a)/2]² - [((n-1)a)/2]² = na².$ If we put n = m², the above leads to $m²a² + ((m²-1)/2)² a² = ((m²+1)/2)² a²$ which is a solution of indeterminate equation of the second degree $x² + y² = z².$

c7.s2.t4.13

The questions relating to finding the sizes and number of bricks required for building the different layers of the altars envisage solutions of simultaneous indeterminate equations of the first degree. The *Baudhāyana Śulba-sūtra* prescribes that a *gārhapatya vedi* should be constructed with five layers of bricks, each layer containing 21 bricks so arranged that their rift in two consecutive layers do not coincide. For a square altar in which square bricks are used, Baudhāyana says that three kinds of bricks having their sides 1/6, 1/4 and 1/3 of the side (*vyāyāma*) of the altar should answer for the construction. How were the sizes as well as the number of bricks determined? Since 21 is not a square number and the rifts of consecutive layers must not coincide, square bricks of at least three different sides must be used. Suppose *a* to be the side of the square altar and l₁, l₂, l₃ the side of each of the three square bricks so that l₁p₁=l₂p₂=l₃p₃=a, where p₁, p₂, p₃ are rational integers. If *x* and *y* be the number of bricks of sides l₁ and l₂ for the first layer, and *x'* and *y'* those of sides l₃ and l₁ for the second layer, we have

c7.s2.t4.14

for the first layer, $x+y = 21$ $xl₁²+yl₂² = a²$ $x/(p₁²)+y/(p₂²) = 1$

c7.s2.t4.15

for the second layer, $x'+y' = 21$ $(x')/(p₃²)+(y')/(p₁²) = 1.$

c7.s2.t4.16

In other words, the problem leads to the following simultaneous indeterminate equations of the first degree: $x+y = 21$ $x/(m²)+y/(n²) = 1$

c7.s2.t4.17

where *m* and *n* are integral parts of *a*. Although the method of solution is not indicated, Baudhāyana gives correctly the values of *m* and *n* as 6 and 4 for x = 9, y = 12 and again as 3 and 6 for x = 5 and y = 16.

c7.s2.t4.18

Surds

c7.s2.t4.19

The technical expressions used in the *śulbas* for such quantities as √2, √3, √4, √18, etc., are *dvi-karaṇī*, *tṛ-karaṇī*, *catuṣkaraṇī*, *aṣṭādaśa-karaṇī*, etc. By using words for ordinal numbers along with *karaṇī*, quantities √(⅓), √(1/7) are expressed as *tṛtīya-karaṇī*, *saptama-karaṇī*, etc. Thus *karaṇī* has been used in the same sense as root or surd. Elementary operations with surds are clearly indicated in the various places of the texts. To draw a *sautramaṇikī vedi*, an isosceles trapezium of the same shape as, but one-third the area of, the *mahāvedi*, Āpastamba uses the surds as follows: 'For a unit *prakrama* 1/√3 of a *prakrama* is to be substituted;

c7.s2.t4.20

alternatively, the transverse sides should be 8 and 10 times √3 and the east-west line 12 times √3.'a Expressed in figures, $the first direction gives the area = 36/(√3) · ½(24/(√3) + 30/(√3))$ $= 324 sq. units;$ $the second direction gives the area = 12√3×½(8√3+10√3)$ $= 324 sq. units.$

c7.s2.t4.21

*Irrational Numbers and their Approximate Values*

c7.s2.t4.22

The Vedic Hindus have been credited further with the notion of irrationality of the quantities √2, √3. They have even given the values of such irrational numbers up to a high degree of approximation. For the value of √2, the exact wordings of some of the *śulbakāras* are quoted below:

c7.s2.t4.23

*pramāṇaṃ tṛtīyena vardhayettacca caturthenātmacatus-*
*triṃśonena | saviśeṣaḥ |*b
*karaṇīm tṛtīyena vardhayettacca svacathurthenātma-*
*catustriṃśonena saviśeṣa iti viśeṣaḥ*c

c7.s2.t4.24

Both versions having used practically the same wordings, these may be rendered as follows:

c7.s2.t4.25

'Increase the measure by its third and this third by its own fourth less the thirty-fourth part of that fourth. This is the value with a special quantity in excess.'d

c7.s2.t4.26

If we take 1 for the measure of a square side, the above formula gives the diagonal as follows: $√2 = 1 + ⅓ + 1/3.4 - 1/3.4.34 = 1·4142156.$

c7.s2.t4.27

In the same manner, the value of *tṛ-karaṇī* is given by $√3 = 1 + ⅔ + 1/3.5 - 1/3.5.52 = 1·7320513.$

c7.s2.t4.28

The method by which the values of these irrationals were obtained is not indicated in the texts. They did not obviously know how to extract the square root up to so many decimal places at this distant time. Thibaut, Bürk, Müller, Datta and others have reconstructed possible methods out of the very elements and procedures dealt with in the *śulbas*. For √2, one possible method appears to be as follows:

c7.s2.t4.29

b *Bśl.*, i. 61–62; *Āśl.*, i. 6.

c7.s2.t4.30

Two squares of unit side are taken. One of them is divided into three equal rectangular strips, of which the first and the second are marked 1 and 2. The third strip is subdivided into three squares of which the first is marked 3. The remaining two squares are each divided into four equal strips marked by 4, 5, 6, 7 and 8, 9, 10, 11. These eleven strips are added to the other square in the manner shown in Fig. 3.11 to obtain a square less a small square (shaded) at the corner. The side of the square equals 1+⅓+1/3.4 and the area of the shaded square (1/3.4)², so that the former is greater than the sum of the two original squares by the amount (1/3.4)². To find the approximate value, suppose x be the thickness of the smallest strips, 4, 5, ... 11. Then $2x(1+⅓+1/3.4)-x²=(1/3.4)²$ or $x=1/3.4.34, neglecting x² which is too small.$

c7.s2.t4.31

a *Āśl.*, v. 8.

c7.s2.t4.32

c *Kśl.*, ii. 13.

c7.s2.t4.33

d Thibaut (1), p. 21; Datta (4), pp. 188 ff.

c7.s2.t4.34

FIG. 3.11.

c7.s2.t4.35

The side of the square equalling the sum of the two original squares or the diagonal of each of the original square is given by $√ 2=1+⅓+1/3.4-1/3.4.34.$

c7.s2.t4.36

One of the cuneiform tablets from the old Babylonian times (1600 B.C.), now in the Yale Babylonian collection, shows a square with its two diagonals and values in sexagesimal system, from which √ 2 in the same system works out as $√ 2=1; 24, 51, 10.$

c7.s2.t4.37

Neugebauera has shown that the sexagesimal equivalent of the value of √ 2 as given in the *Sulba-sūtras* is 1; 24, 51, 10, 37, He has further hazarded a guess that both the main term and the subtractive correction of the Hindu value were based ultimately on Balylonian approximations. The more ancient character of the Babylonian cuneiform texts cannot, however, be doubted. Serious objections have been raised in accepting any date for the earliest of the *Sulba-sūtras* beyond 500 B.C. But that itself does not appear to be a reliable ground for making such a suggestion in view of a more complete and clear textual statement and several indications of the derivations of approximate values, embedded in the very texts themselves.

c7.s2.t4.38

a Neugebauer (1), p. 34.

c7.s2.t4.39

Permutations, Combinations and Pascal Triangle

c7.s2.t4.40

Another favourite mathematical pursuit of the Vedic Hindus was in the field of permutations and combinations, which was also very popular among the ancient Jainas about whose mathematical interests we shall speak presently. This interest was undoubtedly activated by the considerations of the Vedic meters and their variations. There are several Vedic meters, e.g. *Gāyatri*, *Anuṣṭubh*, *Bṛhatī*, *Triṣṭubh*, *Jagatī*, to mention a few, with 6, 8, 9, 11, 12 syllables. The Vedic meter specialists were concerned with the problem of producing different possible types of meters from those of varying syllables by changing the long and short sounds within each syllable group. In this effort, they were led invariably to laying the foundation of the mathematics of permutations and combinations. Emerging from the intricacies of the Vedic meters, these rules found immediate applications, e.g. finding possible combinations of the six tastes taking one, two, three, etc., at a time,a possible numbers of philosophical categories through combinations of *n* fundamental categories taken one, two, three, etc., at a time,b total number of perfumes that can be made from sixteen different substances taken one, two, three or four substances at a time,c and so on. In all these examples, the results are given correctly.

c7.s2.t4.41

Special importance attaches to Piṅgala’s *Chandaḥ-sūtra* (200 B.C.) which contains a method called *meru-prastāra* for finding the number of combinations of *n* syllables taken 1, 2, 3 . . . *n* at a time. The method, as explained by the commentator Halāyudha (tenth century A.D.), is as follows:

c7.s2.t4.42

‘After drawing a square on the top, two squares are drawn below (side by side) so that half of each is extended on either side. Below it three squares, below it (again) four squares are drawn and the process is repeated till the desired pyramid is attained. In the (topmost) first square the symbol for one is to be marked. Then in each of the two squares of the second line figure one is to be placed. Then in the third line figure one is to be placed on each of the two extreme squares. In the middle square (of the third line) the sum of the figures in the two squares immediately above is to be placed; this is the meaning of the term *pūrṇa*. In the fourth line one is to be placed in each of the two extreme squares. In each of the two middle squares, the sum of the figures in the two squares immediately above, that is three, is placed. Subsequent squares are filled in this way. Thus the second line gives the expansion of combinations of (short and long sounds forming) one syllable; the third line the same for two syllables, the fourth line for three syllables, and so on.’d

c7.s2.t4.43

a *SS. Rasabheda-vikalpādhyāya*, 63.

c7.s2.t4.44

b *Bhag. Sū.*, 314.

c7.s2.t4.45

c *Bṛh. S.*, lxxvii. 13–14.

c7.s2.t4.46

d Bag, pp. 72–73.

c7.s2.t4.47

The *meru-prastāra* is the same as the triangular array known in Europe as Pascal's triangle. The triangle first appeared in Europe on the title page of the arithmetic of Apianus (A.D. 1527) and subsequently in the works of Stifel (A.D. 1544), Scheubel (A.D. 1545), Tartaglia (A.D. 1556),

c7.s2.t4.48

Fig. 3.12.

c7.s2.t4.49

Bombelli (A.D. 1572) and other Renaissance methematicians and in Pascal's posthumous work *Traité du triangle arithmétique* (1665). Apart from giving a quick method of finding the number of combinations, the method in its association with meters clearly envisaged the computation of the binomial terms an, an-1b, . . . abn-1, bn of the expression (a+b)n.

Cites (10)