A Concise History of Science in India c7.s2.t3: Geometry
A Concise History of Science in India c7.s2.t3
- Language
- English
Geometry
c7.s2.t3.1
*Postulates*
c7.s2.t3.2
We have already stated that the various *śulba-sūtras* that have come down to us as parts of the *Śrauta-sūtras* are Brāhmaṇic geometrical manuals for the construction of sacrificial altars. In the various rules given, certain assumptions resembling Euclidean postulates are taken for granted. Thus a straight line is capable of division into an infinite number of equal parts; a circle can be divided into any number of parts by drawing diameters; each diagonal bisects a rectangle and both diagonals bisect each other, dividing the rectangle into four parts with the vertically opposite ones equal in all respects; a straight line joining the vertex with the middle point of the opposite side divides an isosceles triangle into two equal halves; parallelograms drawn on the same base and between the same parallels are equal to one another; a square inscribed within a circle and touching the circumference is the maximum, and so on. The texts give several rules as to how to construct a straight line perpendicular to another straight line, a square with a given side, a rectangle with given sides, an isosceles trapezium of a
c7.s2.t3.4
FIG. 3.2.
c7.s2.t3.5
given altitude, face and base. For constructing a square, Baudhāyana gives, among others, the following rule:
c7.s2.t3.6
‘If you intend to draw a square, take a cord of length equal to the side desired of the square. Make a tie at each end of the cord and mark its middle point. Draw the east-west line (*EW* of Fig. 3.2) and fix a pole at its middle (*O*). After fastening the two ties at that pole draw a circle with the mid-point of the cord; fix a pole at each end of the diameter (along the east-west line). Fasten one tie at the eastern pole and draw a circle with the other tie. Likewise, draw a circle round the western pole. Join the points of intersections of these circles to obtain the second (north-south) diameter; fix a pole at each end of this diameter. Fasten both ties at the east pole and describe a circle with the mid-point (of the cord). Draw similar circles around the southern, the western and the northern poles. The exterior points of intersection of these four circles will give the four corners (*ABCD*) of the required square.’a
c7.s2.t3.7
*Combinations and Transformations of Rectilinear Figures*
c7.s2.t3.8
Of greater importance are the rules given in the *śulbas* for the combination and transformation of rectilinear figures, specially the squares and the rectangles. In a series of rules for transforming a square into a rectangle, the simplest one is to draw the diagonal *AC* of the square *ABCD* to divide it into two equal parts, to divide again one of these parts into two equal parts by joining the vertex *D* with the mid-point *E* of the diagonal and then to fit them suitably one to each side *AB* and *BC* of the square. *AFGC* is then the required rectangle. In the *paitṛkīvedi*, it is often required to find one square double of another given square. This is done by joining the mid-points of the sides of a square. The thing is achieved by drawing on the other side of the diagonal in Fig. 3.3 a similar rectangle *AHIC*.
c7.s2.t3.9
When one side of a rectangle is given, the rule of transforming a square into a rectangle of the given side is given by Āpastamba as follows:b
c7.s2.t3.10
‘If you desire to make a rectangle out of a square (cut off from it a rectangle) of a side of the desired length. The excess remaining is to be added suitably as to fit (the rectangle cut off).’
c7.s2.t3.11
Baudhāyana also gives a similar rule.c The directions are incomplete but, according to interpretations given of the above rule by the commentators Sundararāja and Dvārakānātha Yajvā, the true process appears to be as illustrated in Fig. 3.4. *ABCD* is the given square. Its side *AB* is extended to *E*, so that *EB* equals the given side of the rectangle into which the square is to be transformed. Join *EC* cutting *AD* at *O*. Complete the rectangles *EFCB* and *GFCH*. Then *GFCH* is the required rectangle. From the postulate that a rectangle is bisected by its diagonal, it is easy to see that the rectangles *AOHB*, *GODF* are equal in areas and therefore the square *ABCD* equals the rectangle *GFCH*.
c7.s2.t3.12
a *Bśl.*, i. 22–28.
c7.s2.t3.13
b *Āśl.*, iii. 1.
c7.s2.t3.14
c *Bśl.*, i. 53.
c7.s2.t3.15
Baudhāyana, Āpastamba and Kātyāyana all teach methods of constructing a square equal to the sum of two different squares. The method
c7.s2.t3.16
FIG. 3.3.
c7.s2.t3.17
FIG. 3.4.
c7.s2.t3.18
FIG. 3.5.
c7.s2.t3.19
is to cut off from the larger of the two given squares, *ABCD*, a rectangular portion *ABEF* of which the side *BE* is equal to the side of the given smaller square. Then the diagonal *AE* is the side of the square equal to the sum of the two given squares. The proof will follow from Fig. 3.5, in which *AEGH* is the required square, and *ABCD* and *CKGI* are the given squares.a In other words, the proposed construction is based on what has generally come to be known as the theorem of Pythagoras.
c7.s2.t3.20
*Pythagorean Theorem*
c7.s2.t3.21
The theorem of the square of the diagonal is more explicitly given in more or less the same language in all the *śulba-sūtras* we know of. Here is Baudhāyana’s definition:b
c7.s2.t3.22
*dīrghacaturasrasyākṣṇayārajjuḥ pārśvamānī tiryaṅmānī ca*
*yatpṛthagbhūte kurutastadubhayaṁ karoti* |
‘The diagonal of a rectangle produces by itself both (the areas) produced separately by its two sides.’
c7.s2.t3.23
The same definition is given by Āpastamba. Kātyāyana’s definition given in the same language is characterized by the explicit mention of the knowledge of plane figures:
c7.s2.t3.24
*dīrghacaturasrasyākṣṇayārajjustiryaṅmānī pārśvamānī ca yat*
*pṛthagbhūte kurutastadubhayaṁ karotīti kṣetrajñānam* |
c7.s2.t3.25
Note the word *iti kṣetrajñānam* at the end, meaning ‘this is the knowledge of plane figures’.
c7.s2.t3.26
A question has often been asked whether such a definition resulted from empirical guesswork or was based on a proof of some kind. Although it is too much to expect elaboration of a proof in manuals of this type there is hardly any doubt that the Vedic *śulbavids* at this distant date possessed a valid proof of the theorem, of which the texts themselves provide reliable indications. For example, Baudhāyana, in the very next rule,a says that the rectangles for which the above is true have their sides 3 and 4, 12 and 5, 15 and 8, 7 and 24, 12 and 35, 15 and 36. The sum of squares of these numbers is also a square for each pair.
c7.s2.t3.27
a Datta (4), p. 77.
c7.s2.t3.28
b *Bśl.*, i. 48.
c7.s2.t3.29
3²+4² &= 5² / 12²+5² &= 13² / 15²+8² &= 17² / 7²+24² &= 25² / 12²+35² &= 37² / 15²+36² &= 39²
c7.s2.t3.30
FIG. 3.6.
c7.s2.t3.31
The above relationships between sides and diagonals or hypotenuse for rational rectangles or triangles have been freely used for finding the perpendicular directions (north-south) at the east or the west point of the east-west line in altar settings. The method is to take a cord of suitable length greater than the distance between the two poles one each at the east and the west point, fasten a loop at either end of the cord, fix it with the two poles by these loops and then stretch it by the *nirañchana* mark. The *nirañchana* mark divides the cord in two portions *x* and *y* such that a²+x²=y², where *a* is the distance between the east-west pole (Fig. 3.6). This finding of the properties of rational triangles is a strong argument in favour of the Vedic Hindus possessing proofs of the theorem of the square of diagonal or the hypotenuse. It is possible that by actually drawing the squares on the diagonal and the sides of a rational rectangle, dividing them into elementary unit squares and then counting them, they might have arrived at the truth of this theorem (Fig. 3.7).
c7.s2.t3.32
FIG. 3.7.
c7.s2.t3.33
*Question of Origin*
c7.s2.t3.34
The question of the Greek origin of the theorem and also whether Pythagoras himself was the discoverer of it and its proof has by no means been solved. The tradition attributing the theorem to Pythagoras is due to Cicero (*c.* 50 B.C.), Diogenes Laertius (second century A.D.), Athenaeus (*c.* A.D. 300), Heron (third century A.D.), and Proclus (*c.* A.D. 460), and therefore started about five centuries after the death of Pythagoras. Junge pointed out that the Greek literature of the first five centuries after Pythagoras contained no mention of the discovery of this or any other important geometrical theorem by the great philosopher and furthermore emphasized uncertainties in the statements of Plutarch and Proclus. Although various attempts have been made to justify the tradition and trace the proof to Pythagoras, no record of proof has come down to us earlier than that given by Euclid (Theorem 47, BK I). As to the relation 4²+3²=5² from which the theorem of rational triangle is derivable, very ancient Egyptian knowledge is attested by the Kahun papyrus of the twelfth dynasty (*c.* 2000 B.C.), but its association with rational triangles does not seem indicated in this or other Egyptian papyrii.a It is interesting to note that among the Egyptians, geometry of surveying was considered to be the science of the ‘rope-stretchers’ (*harpedonap'tae*) who thus appear to be the Egyptian counterpart of the Indian *śulbavids*. As to the antiquity of Pythagorean theorem in China, it is stated, though not proved, in the arithmetical classic *Chou Pei Suan Ching* (third or fourth century B.C.); the numerical relationship 4, 3 and 5 between the sides and the diagonal of a rational rectangle is also given in this text.
c7.s2.t3.35
a *Bśl.*, i. 49.
c7.s2.t3.36
*Areas and Volumes*
c7.s2.t3.37
Regarding other areas of geometry, the areas and volumes of a number of plane and solid figures have been given as follows:
c7.s2.t3.38
Area of a triangle = ½(base)×(altitude)
Area of a parallelogram = (base)×(altitude)
Area of a trapezium = ½(sum of two parallel sides)×(altitude)
Volume of a prism or cylinder = (base)×(height)
Volume of the frustum of a pyramid = ((a+a'))/2((b+b'))/2h,
c7.s2.t3.39
where a, b are the length and breadth of the rectangular base, a', b' the corresponding ones of the parallel face, and h is the height.
c7.s2.t3.40
*Squaring the Circle; π*
c7.s2.t3.41
Another type of problem which interested the *śulba* geometers was the circling of the square or its converse the squaring of the circle. Through these exercises, the Vedic Hindus were led to finding approximate values of π. For circling a square, Baudhāyana’s rule is first to draw the arc of a circle about the centre O of the given square ABCD with a radius equal to half the diagonal, intercepting the east-west line at P. Then the circle drawn about the centre O with a radius equal to half the side of the square plus one-third of the portion between the (previous) circle and the square (OQ=OR+QR where QR=PR/3) answers for the problem. If the side of the given square be 2a, the radius of the required circle works out to a × 1·1380718, taking the value of √2 = 1·4142156 as given in the *śulba* (*vide infra*). That the area of the circle 4·068987a² is somewhat higher than 4a² is fully reflected in the *śulba-sūtras*.
c7.s2.t3.42
a Heath, I, p. 352.
c7.s2.t3.43
FIG. 3.8.
c7.s2.t3.44
For squaring a circle, the same *śulba-sūtra* advises the diameter to be first divided into eight equal parts. The last 1/8th part is further divided into 29 parts and then 28 of these subdivisions are rejected (from this 1/8th part). Furthermore, from it is rejected 1/6th of the preceding subdivision (1/(8·29)) minus 1/8th part of the last subdivision. If d be the diameter of the circle, the side of the required square is given by $7d/8 + d/8 - 28d/(8·29) - (d/(8·29·6) - d/(8·29·6·8)) = 1224/1393 d (approx.).$
c7.s2.t3.45
From the foregoing operations, it is possible to calculate the value of π. In the example of circling of the square, the radius r in terms of a, half the side of the square is given by $r = a + a/3 (√2-1),$ where $π = (4a²)/([a+a/3(√2-1)]²) = 4/([1+⅓(√2-1)]²) = 3·0883$ accepting the *śulba* value of √2. Other results are 3·0885, 3·004. Baudhāyana's use of 3 as the value of π is therefore quite understandable. A more accurate value of π, e.g. 3·16049, is inferable from the data given in the *Mānava Śulba-sūtra*.
Cites (3)
- Āpastambaśulbasūtra — Āpastamba, Sanskrit
- Hiraṇyakeśiśulbasūtra — Sanskrit
- Baudhāyanaśulbasūtra — Baudhāyana, Sanskrit