Skip to main content

Ordering Operations in Square Root Extractions, Analyzing Some Early Medieval Sanskrit Mathematical Texts with the Help of Speech Act Theory

  • Chapter
  • First Online:
Texts, Textual Acts and the History of Science

Part of the book series: Archimedes ((ARIM,volume 42))

Abstract

Procedures for extracting square roots written in Sanskrit in two treatises and their commentaries from the fifth to the twelfth centuries are explored with the help of Textology and Speech Act Theory. An analysis of the number and order of the steps presented in these texts is used to show that their aims were not limited to only describing how to carry out the algorithm. The intentions of authors of these Sanskrit mathematical texts are questioned by taking into account the expressivity of relationships established between the world and the text.1

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 39.99
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Hardcover Book
USD 54.99
Price excludes VAT (USA)
  • Durable hardcover edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Notes

  1. 1.

    This study was undertaken within the History of Science, History of Text Seminar in Rehseis in 2007. It was completed with the help of the algo-ANR. I would like to thank J. Virbel, K. Chemla, C. Proust, F. Bretelle-Establet, J. Ritter, C. Singh, A. Bréard, K. Vermeir, M. Keller, C. Montelle, K. Plofker, C. Singh and R. Kennedy: their thoughtful comments and encouragement have been woven into this article and have brought it into existence.

  2. 2.

    For a more general study on how Austin’s work could help contextualize Sanskrit scholarly knowledge see Ganeri (2008).

  3. 3.

    Starting with the procedure given by Āryabhaṭa (499), remaining virtually unchanged in Bhāskarācārya’s (fl. 1114) Līlāvatī and Bījagaṇita, and thus still in use in later commentaries of these texts.

  4. 4.

    Keller (2010).

  5. 5.

    Pingree (1981, p. 13).

  6. 6.

    Shukla (1976).

  7. 7.

    Shukla (1976, pp. 52–53). A translation of Bhāskara’s commentary on Āryabhaṭa’s verse on root extraction can be found in Keller (2006 Volume 1, pp. 20–21), and an explanation of the process in Keller (2006, Volume 2, pp. 15–18).

  8. 8.

    Sarma (1976) A translation of his commentary on Āryabhaṭa’s verse for square root extractions is given in Appendix C.

  9. 9.

    Shukla (1959) A translation of the anonymous and undated commentary on the Pāṭīgaṇita’s rule for extracting square roots is given in Appendix D). An explanation of this rule is given in Shukla’s translation.

  10. 10.

    Pollock (2006, Part I).

  11. 11.

    Thus Śrīdhara starts his treatise with the following statement, Shukla (1959, Sanskrit: i, English: 1):

    PG. 1cd (aham) lokavyavahārārthaṃ gaṇitaṃ saṃkṣepato vakṣye‖

    I will briefly state mathematics aiming at wordly practices

  12. 12.

    This diagram should not be seen as an attempt to formalize the algorithm: it is only a heuristic illustration.

  13. 13.

    Austin (1962, p. 76).

  14. 14.

    Searle (1969, pp. 57–61).

  15. 15.

    Singh (1927).

  16. 16.

    Plofker (2009, pp. 123–125).

  17. 17.

    Renou (1963).

  18. 18.

    Incidentally, this study shows that these mathematical rules do not correspond either to Group A or Group B as defined by Renou in Renou (1963), part C. Features of group A, such as the use of the optative, are combined here with the prescriptive norms of group B.

  19. 19.

    The cryptic character of this rule has been analyzed in Keller (2006, p. xvii; Keller 2010, pp. 235–236) and is noted in Plofker (2009, pp. 123–125). Some of its characteristics are described in Singh (1927).

  20. 20.

    The Ab and the PG provide rules for extraction in a verse form that counts the number of syllabic units, the āryā. This is a very common verse form for prescriptive texts.

  21. 21.

    The difficulty of actually singling out the steps in Āryabhaṭa’s verse, addressed in the next section, can be seen when this enumeration is compared with Fig. 5.4. With less contrast, the same can be seen for Śrīdhara’s rule as well.

  22. 22.

    To apply in this context the concept that Severi (2007) uses to denote pictorial mnemonic artifacts mostly used by North American Indians: the important idea is stamped into the artifact by relating two things that normally should not be connected. This association works like a knot in a handkerchief: something should be remembered here. In this case: where a digit is noted down and a square quantity are given a common name. They are associated in a confusing way in the verse, creating such a chimera.

  23. 23.

    Note that in grammar ( vyakaraṇa), according to an oral communication by Jan Houben, the optative belongs essentially to the commentary.

  24. 24.

    Renou (1984, § 292):

    L’optatif exprime les nuances variées d’un optatif propre- souhait, hortatif, délibératif, éventualité, prescriptif, hypothétiques (…). La coexistence de ces divers emplois n’est relevable que dans la poésie littéraire; dans les textes techniques prédomine la valeur prescriptive.

    That is, in English (my translation):

    The optative expresses diverse nuances of a true optative: wish, hortative, deliberative, possibility, prescriptive, hypothetical (…) voices. The coexistence of these various uses are only found in literary poetry; in technical texts a prescriptive value prevails.

  25. 25.

    tam bhāgam haret gṛhṇīyāt.

  26. 26.

    avargasthānād bhāgam haret.

  27. 27.

    This cross-reference may refer to the merging of both forms as referred to by Renou (1963) who considers that real sūtras are non-versified. The etymology of kārik ā, derived from the verb kṛ-.‚ “to make”, can maybe be understood in this context as “(verse) for action”.

  28. 28.

    vargamūlānayanāyāha.

  29. 29.

    vargamūlānayanam āryayāha.

  30. 30.

    vargamūle karaṇasūtram āryadvayam.

  31. 31.

    Standard vocabulary is used throughout Sanskrit mathematical texts to refer to computations, methods and algorithms. We do not know if there was any difference in meaning between these different words, if their meaning changed over time, according to authors. We have adopted the following translations here: “computation” for words derived from ānī; “method” for karman; algorithm or procedure for karaṇa.

  32. 32.

    The fact that the PG’s process provides a doubled root that needs to be halved is highlighted (by mistake?) in SYAB. Indeed, this commentator on the Āryabhaṭīya notes in Sarma (1976):

    labdhe mūlarāśau dviguṇī kṛ taṃ dalayet

    When the root quantity has been obtained, having multiplied it by two, it should be halved.

    However, Āryabhaṭa’s rule does not provide a double root and therefore does not request a halving at the end.

  33. 33.

    Note that the Indian subcontinent’s diversity of manuscripts presents a great variety of material settings; its scholarly texts, a large number of lists. However, there seems to have been no specific typographical layout for lists in mathematical manuscripts in the Indian subcontinent.

  34. 34.

    dviguṇ ena ca tena tatraiva sthitena upariṣṭāt bhāgam apaharet labdhaṃ paṇ ktau viniveśayet tatas tad vargam upariṣṭāc chodhayet tac ca dviguṇīkuryāt tasmin dviguṇe kṛte yadi sthānam adhikaṃ jāyet tat prāglabdhe yojayet (…) tam utsārayet tato vibhajet labdhaṃ paṅktau viniveśayed ity ādi pūrvavat yāvat utsarpaṇasambhavaḥ / samāptau sarvaṃ labdhaṃ dalayed/.

  35. 35.

    In the corpus looked at here, this is the only use of an optative in a non-prescriptive form.

  36. 36.

    Except for two ambiguous elements: the semi-tacit use of decimal place-value notation, and when APG considers the case of a two digit result.

  37. 37.

    śuddhe varge vargagaṇita iti arthaḥ.

  38. 38.

    We studied this aspect of the process, and what it means for the concept of decimal place-value notation in Keller (2010), we will thus not dwell on this aspect here.

  39. 39.

    kasmāt sthānāt prabhṛtīty āha – avargāt (…) atra gaṇite viṣamaṃ sthānaṃ vargaḥ (…) avarga iti samam. sthānam, yato hi viṣamaṃ samaṃ ca sthānam/.

  40. 40.

    Although, even in this case, there is still a great disparity between the representation of the action given in Fig. 5.5 and in Table 5.2.

  41. 41.

    if my interpretation of the use of “both” and “three” in this text is correct, as noted in the footnotes of Appendix C.

  42. 42.

    taṃ bhāgam, haret gṛhṇīyāt/

    kasmāt sthānāt prabhṛtīty

    āha – avargāt (…) atra gaṇite viṣamaṃ sthānaṃ vargaḥ

    (…) kena bhāgaṃ haret ity

    āha – nityaṃ dviguṇena vargamūlena

    (…)

    kathaṃ punas tat vargamūlaṃ labhyate ity

    āha – vargāt varge śuddhe labdhaṃ sthānāntare mūlam

  43. 43.

    (Austin 1962, p. 100).

  44. 44.

    Thus Sūryadeva needs to explain:

    taṃ saṅkhyāviśeṣaṃ mūlatvena gṛ hnīyāt / tadatra vargamūlaphalam ity ucyate /

    “This special number is referred to (in the rule) as a root. Consequently, here, the result

    which is a square root (vargamulaphala) has been mentioned (ucyate).”

  45. 45.

    The difficulty of pinpointing exactly how the iteration is given in the verses explains why it appears and disappears in the previous illustrations we have given of the algorithm.

  46. 46.

    etat eva sūtraṃ punaḥ punararāvartaye yāvatparisamāptaṃ gaṇitakarmeti.

  47. 47.

    tribhiḥ patanāt caturviṃśatauśuddhāyāṃ upari dvau śeṣaḥ.

  48. 48.

    adhaḥ ṣāḍaṣītijayīte / eśa rāśiḥ sarpati, paṅktyāṃ dvayor adhaḥ (ṣāṭkaṃ) bhavati, aṣṭakaṃ saptādhāh / nyāsaḥ- \(\begin{array}{*{20}{c}} 1&7&2&4\\ {}&8&6&{} \end{array}\).

  49. 49.

    He also uses once each labh (to obtain), bhu (to be, have, produce), vidyate (to exist, discern). The three only other conjugated verbs of this part of his commentary are: (1) the optative used for division, and (2) verbs used while solving examples at the end of the commentary.

  50. 50.

    yadatra labdhaṃtat sthānāntare mūlasaṃjñaṃ bhavati/ (…) tasmin sthānāntare tasya labdhasya mūlasaṃjñā/ yatra worldplayaḥ sthānāntaram eva na vidyate, tatra tasya tatraiva mūlasaṃjñā.

  51. 51.

    tasmin dvigune kṛte yadi sthānam adhikaṃ jāyet tat prāglabdhe yojayet, tayor ubhayor ekarāśitājñeyā / tasya rāśer labdhasaṃjña.

  52. 52.

    taṃ saṃkhyāviśeṣaṃ mūlatvena gṛhnīyāt

  53. 53.

    tat purve vargasthāne mūlaṃ bhavati.

  54. 54.

    saṃkhyāvinyāsasthāneṣu viṣamasthānāni vargasamjñāni / samasthānāny avargasaṃjñāni.

  55. 55.

    Keller (2007).

  56. 56.

    Le terme de sūtra ou “fil” désigne tantôt une règle énoncée sous la forme d’une proposition (…) plus ou moins brêve, tantôt un ensemble de propositions concourant à constituer un même recueil. (…) Le genre du sūtra se définit par sa relation plutôt que par son contenu: un sūtra (au sens de “règle” ou “aphorisme”) est d’abord un élément dépendant du contexte, même s’il est grammaticalement autonome; il est déterminé par le système et (…) corrélatif au groupe qui l’environne.

  57. 57.

    The “reconstruction” of these variants of the different steps of the process is not discussed here. Hopefully this issue will be tackled in a forthcoming article.

  58. 58.

    Although this is never mentioned in the ancient texts, the quotient obtained needs to be sufficiently small. This sometimes requires a subtraction by 1 or 2 (and a change in the remainder of the division) to find the adequate digit.

  59. 59.

    * mark non-attested forms.

  60. 60.

    For a translation into English of BAB.2.4, see Keller (2006).

  61. 61.

    (Sarma 1976, pp. 36–37).

  62. 62.

    This long sentence has an equivocal expression: is sthānātare (in a different place) glossed into pūrve vargasthāne (in the next square place), or should one understand that two actions are prescribed, first setting aside the quotient as a digit of the root, and then that its square enters an operation in the next square place?

  63. 63.

    This is a mysterious cryptic expression, it is thus my interpretation that the three here, as the “both” ( ubhaya) used in the sentence before, refers to the number of digits of the square root being extracted.

  64. 64.

    Reading labdham instead of the misprinted ladhdham.

  65. 65.

    Shukla (1959, pp. 18–19).

  66. 66.

    Reading here as in the manuscript rather than the utsarpaṇa suggested by Shukla (1959), p. 18, footnote 9.

Abbreviations

Ab:

Āryabhaṭa I ’s Āryabhaṭīya (fifth century)

APG:

The Anonymous and undated commentary on the Pāṭīgaṇita of Śrīdhara

BAB:

Bhāskara I ’s commentary on the Āryabhaṭīya: Āryabhaṭīyabhāṣya (seventh century)

PG:

The Pāṭīgaṇita of Śrīdhara (tenth century)

SYAB:

Sūryadeva Yajvan’s commentary on the Āryabhaṭīya: Bhaṭaprakāśikā(twelfth century)

References

  • Austin, J. L. 1962. How to do things with words. Clarendon Press, 1971, Oxford: Oxford University Press. paperback; second edition published 1975 by the Clarendon press; second edition published 1976; new edition 1980 with new index.

    Google Scholar 

  • Ganeri, Jonardon. 2006. Contextualism in the study of Indian intellectual cultures. Journal of Indian Philosophy 36:551–562.

    Google Scholar 

  • Keller, Agathe. 2006. Expounding the mathematical seed, Bhāskara and the mathematical chapter of the Āryabhaṭa. 2 volumes. Birkhaüser: Basel.

    Google Scholar 

  • Keller, Agathe. 2007. Qu’est ce que les mathématiques? les réponses taxinomiques de Bhāskara Sciences et Frontières, ed. Marcel Hert, Philippe Paul-Cavalier, 29–61. Bruxelles: Echanges, Kimé.

    Google Scholar 

  • Keller, Agathe. 2010. On Sanskrit commentaries dealing with mathematics (fifth-twelfth century). In Looking at it from Asia: the Processes that Shaped the Sources of History of Science, ed. F. Bretelle-Establet, 211–244. Boston Studies in the Philosophy of Science, volume 265. Dordrecht: Springer. URL http://halshs.archives-ouvertes.fr/halshs-00189339. Accessed June 2015.

  • Pingree, David. 1981. Jyotih&astra: Astral and mathematical literature. Wiesbaden: Harrassowitz.

    Google Scholar 

  • Plofker, Kim. 2009. Mathematics in India. Princeton: Princeton University Press.

    Google Scholar 

  • Pollock, Sheldon. 2006. The language of the gods in the world of men: Sanskrit, culture and power in premodern India. University of California Press.

    Google Scholar 

  • Renou, Louis. 1963. Sur le genre du sūtra dans la littérature sanskrite. Journal Asiatique 251: 165–261.

    Google Scholar 

  • Renou, Louis. 1984. Grammaire Sanskrite. Paris: J. Maisonneuve.

    Google Scholar 

  • Sarma, K. V. 1976. Āryabhaṭīya of Āryabhaṭa with the commentary of Sūryadeva Yajvan. New-Delhi: INSA.

    Google Scholar 

  • Searle, John R. 1969. Speech acts, an essay in the philosophy of language. Cambridge: Cambridge University Press.

    Google Scholar 

  • Severi, Carlo. 2007. Le Principe de la chimère, une anthropologie de la mémoire. Paris: Aesthetica. Aesthetica, Editions de la rue d’Ulm, musée du quai Branly, Paris.

    Google Scholar 

  • Shukla, K. S. 1959. Patıganita of Śrıdharacarya. Lucknow: Lucknow University.

    Google Scholar 

  • Shukla, K. V., and K. S. Sharma. 1976. Āryabhaṭīya of Āryabhaṭa, critically edited with translation. New-Delhi: Indian National Science Academy.

    Google Scholar 

  • Singh, Avadhesh Narayan. 1927. On the Indian method of root extraction. Bulletin of the Calcutta Mathematical Society XVIII (3): 123–140.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Agathe Keller .

Editor information

Editors and Affiliations

Appendices

A Appendix 1: Different Steps in the Algorithm for Extracting Square Roots as Spelled Out in the Corpus

Taking into account all the steps detailed by the authors considered here (with an arbitrary filter—the mesh of the net may at times seem too small and at others too large- that is underlined in paragraph 5.2.3), thirteen steps for extracting a square root can be listed. Step 3, 6 and 12 state common tacit steps. The algorithm may be more efficiently illustrated in Fig. 5.2, although they are not equivalent renderings of the process.Footnote 57

  1. 1.

    The number whose square root is to be extracted is noted down in decimal place-value notation. Places are categorized with a grid that enables one to identify square powers of ten. Either positions for square and non-square powers of ten are listed or the series enumerating positions starting with the place with the lowest power of ten is considered. This list categorizes places as even or odd places.

  2. 2.

    The highest odd/square place is identified.

  3. 3.

    Consider (tacitly) the number made by all the digits to the left of the digit noted down in that place, that digit included.

  4. 4.

    Find the largest square contained in the number noted down to the left in the last/highest odd place.

    From here, onwards, one could also start by considering step 8, before turning to step 5 to 7.

  5. 5.

    Subtract the square from the number under consideration.

  6. 6.

    Replace (tacitly) the minuend by the remainder of the subtraction.

  7. 7.

    The root of the subtracted square is the first digit of the square root being extracted.

  8. 8.

    The root of this square (Ab family)/The double of the root of this square (PG family) is noted on the same line, to the left of the whole number/ on a line below the line of the number whose root is being extracted. In the PG family then, the doubling of the digit is a separate step in the process. The doubling does not necessarily need to take place immediately, one can note down the digit, and then double it just before it enters the division described in Step 10. This is what the APG recommends.

  9. 9.

    Consider the number whose highest digits are the previously noted remainder and which includes the next digit to its right.

  10. 10.

    Divide this number by (twice) the partial square root from Step 8.Footnote 58 In the following,

  11. 11.

    Replace the dividend with the remainder of the division. Then one should consider the next place on the right, which is a square/uneven place.

  12. 12.

    The quotient is the next digit in the partial square-root. It (Ab), or its double (PG), is thus noted down next to the previously found digit, as in Step 8. Its square is what will be subtracted from the number with the next digit as the process is iterated here from Step 5.

  13. 13.

    When there is no place on the right, the algorithm is finished. Examples only consider a process that extract a perfect square, consequently, either the square-root, or its double is obtained, according to the procedure followed. If we are in the latter case, the number obtained is halved.

B Appendix 2: Extracting the Square Root of 186 624

This is a numerical example addressed in APG. Footnotes and asterisks indicate non-attested forms.

  1. 1.

    The number whose square root is extracted is noted in decimal place-value notation. These decimal places are categorized using a grid: Square (varga), non-square (avarga) powers of ten (Ab), or even (sama, abr. sa) and odd (visama, abr. vi) place ranks - counted starting with the lowest power of ten- (BAB, PG, SYAB, APG).

    \(\begin{array}{cccccc} avarga & varga & avarga & varga & avarga & varga \\ \\ sa & vi & sa & vi & sa & vi\\ \\ 10^{5} & 10^{4} & 10^{3} & 10^{2} & 10^{1} & 10^{0} \\ \\ {\bf 1}& {\bf 8} & {\bf 6} &{\bf 6} & {\bf 2} & {\bf 4}\\ \end{array}\)

    $$186624= \mathbf{1}.10^{5}+\mathbf{8}.10^{4}+\mathbf{6}.10^{3}+\mathbf{6}.10^{2}+\mathbf{2}. 10^{1}+\mathbf{4}. 10^{0}$$
  2. 2.

    Subtract the square from the highest odd place

    The highest “odd” (viṣama) place or “square” (varga) place is 104. The process starts by finding, by trial and error, the highest square number contained in the number noted to the left of this place. In this example, one looks for the highest square that will go into 18. And thus 18 − 42 is the operation carried out.

  3. 3.

    Replace the minuend with the reminder, 4, (BAB) or place the reminder, 4, below (APG).

  4. 4.

    The root, 4, of the subtracted square (16) is the first digit of the square root being extracted. The root of this square (Ab family-4)/ The double of the root of this square (PG family- 8) is noted down on the same line (BAB)/ or a separate line (PG).

    Bhāskara might have written:Footnote 59

    $$\begin{array}{cccccc} \ast 10^{5}& 10^{4} & 10^{3} & 10^{2} & 10^{1} & 10^{0} \\ & {\bf 4}/2 &6 & 6 & 2 & 4\end{array}$$

    While the APG writes:

    Because \(1.10^{5}+ 8.10^{4}=[4.10^{2}]^{2}+2.10^{4}\),

    \(186624= [\mathbf{4.10^{2}}]^{2}+2.10^{4}+6. 10^{3}+6.10^{2}+2.10^{1}+4. 10^{0}\).

  5. 5.

    Moving one place to the right, one should divide by twice the root.

    In this example, 26 is divided by 8: \(26=8\times 3+2\). The quotient is 3, 2 is the remainder.

    This is then set down. Bhāskara’s style

    $$\begin{array}{cccccc} \ast & {\bf 4} &{\bf 3}/2 & 6 & 2 & 4\end{array}$$

    APG style

    $$\begin{array}{ccccc} \ast 2& 6 & 2 & 4 \\ & {\bf 8} & 3 & & \\ \end{array}$$

    In other words, because \(2.10^{4}+6.10^{3}=8\times3.10^{3}+2.10^{3}\) \(186624= [\mathbf{4.10^{2}}]^{2}+[2\times (4. 10^{2}) (3.10^{1})]+2.10^{3}+6.10^{2}+2.10^{1}+4. 10^{0}\).

  6. 6.

    Moving one place to the right, iterate. That is “subtract the square” again. This time the square of the quotient is subtracted. In this example \(3^{2}\) is subtracted from 26: \(26-9=17\). The remainder is 17. This is noted down again:

    Bhāskara style:

    In other words, writing \(26-9=17\) according to the corresponding powers of ten.

    \(186624= [\mathbf{4.10^{2}}]^{2}+[2\times (4.10^{2}) (3. 10^{1})] +[\mathbf{3. 10^{1}}]^{2} - [3. 10^{1}]^{2}+2.10^{3}+6.10^{2}+2.10^{1}+4. 10^{0}= [\mathbf{4.10^{2}}]^{2}+[2\times (4.10^{2}) (3. 10^{1})] +[\mathbf{3. 10^{1}}]^{2} +1.10^{3}+7.10^{2}+2.10^{1}+4.10^{0}\)

  7. 7.

    Moving one place to the right, divide by twice the root. In this example one should divide 172 by \(2\times43=86\): \(172= 2\times 86\). The quotient is 2 and there is no remainder. This is noted

    Bhāskara style:

    \(186624= [\mathbf{4. 10^{2}}]^{2}+[2\times (4. 10^{2}) (3. 10^{1})] +[\mathbf{3.10^{1}}]^{2} +2\times (2. 10^{0})(4.10^{2}+3. 10^{1})+4. 10^{0}= (\mathbf{4.10^{2}+3.10^{1}+2.10^{0}})^{2}\)

  8. 8.

    The square root is 432. To end the procedure, moving one step to the right, one can “subtract the square of the quotient” (\(2^{2}\)):

    Bhāskara style: 4 3 2

    APG style:

    8 6 4

    \(186624= (432)^{2}\)

C Appendix 3: SYAB.2.4

Footnote 60He states (āha) a square root computation (vargamūlānayana) with an āryaFootnote 61:

One should divide, repeatedly, the non-square [place] by twice the square root |

When the square has been subtracted from the square [place], the result is a root in a different place ||

In places where numbers are set-down (vinyāsa), the odd places have the technical name (samjñā) “square”. Even places have the technical name “non-square”. In this verse, when a square quantity is chosen (uddiṣṭa), having initially started by marking (cihnayitvā) the square and non-square places, when one is able to subtract (śodhayituṃ śakyate) the square of a special number- among those [squares of the digits] beginning with one and ending with nine- from the last square place, having subtracted (apāsya) that square; this special number is referred to [in the rule] as a root (mūlatvena gṛhṇīyāt). Consequently, here, the result which is a square root (vargamūlaphala) has been mentionned (ucyate). One should divide (bhāgaṃ haret) the next adjacent non-square place by twice that [root]. In this verse, when the square of this quotient has been subtracted (śuddhe) from the next adjacent square place, that quotient from the non-square place, in a different place, in the next square place, that [quotient] becomes (bhavati) the root.Footnote 62 Also, when one has multiplied it (the quotient) by two (dviguṇīkṛtya), dividing (bhāgahāraṇa) in due order both [digits] from its adjacent non-square, as before, the computation of the third root [is accomplished]. Once again with three [digit numbers, the process is carried out]. In this way, one should perform (kuryād) [the process] until no square and non-square [place] remain (bhavanti).Footnote 63 When the root quantity has been obtained (labdhe), having multiplied it by two (dviguṇī kṛtaṃ), it should be halved (dalayet). Concerning fractions also, having divided (vibhajya) the square root of the numerator by the square root of the denominator the quotientFootnote 64 becomes (bhavanti) the root. One states (āha) in this way:

When the square root of the numerator has been extracted, and the root born from the denominator [also,] the root [is obtained](PG 34) |

In order to obtain the roots of the square which were previously explained (in the commentary of verse 3 which is on squares), setting down: 15 625. The result is the square root 125. Setting down the second: \(\frac{4}{9}\). The root of the numerator 2, the root of the denominator 3, having divided (vibhajya) the numerator by that, the result is the square root of the fraction: \(\frac{2}{3}\). Thus the fourth rule [has been explained].

D Appendix 4: APG

An algorithmic rule (karaṇasūtra) of two āryas for square roots:Footnote 65

PG.25. Having removed the square from the odd term, one should divide the remainder by twice the root that has dropped down to a place [and] insert the quotient on a line||

PG.26. Having subtracted the square of that, having moved the previous result that has been doubled, then, one should divide the remainder. [Finally] one should halve what has been doubled.||

What is the root of a given quantity whose nature is a square? This is the aim of that procedure. One should subtract (tyajet) a possible (saṃ bhavina) square, from the odd (viṣama) of the square quantity, from what is called odd (oja), that is from the first, third, fifth, or seventh etc., ; the places for one, one hundred, ten thousand, or one million, etc.; from the last term (pada), that is from among other places. This should provide (syāt) the root of that square which one should place (sthāpayed) beneath the place of decrease, the place the possible square is subtracted (śodhita) from that, one, a hundred, ten or thousands, etc., the last among the other places. And one should divide (bhāgam apaharet) from above (upariṣṭāt) by twice that, just there. The result should be inserted (viniveśayet) on a line, one should subtract (ś) the square of that from above that, and this should be doubled (dviguṇīkūryāt). If when this is doubled (dviguṇe kṛte) an additional place is created (jāyet), then it should be used (yojayet) as before when it is a result. Both have the quality of being a unique quantity (raśita). This quantity has the name “result”. One should repeat (utsārayet) this, thus one should divide (vibhajet), one should insert (viniveśayet) the result on a line, etc. as before in as much as the serpentine is possible (sarpaṇasaṃbhava)Footnote 66. When finished (samāpta) one should halve (dalayet) the whole result, thus obtaining the square root.

Thus for 186624, for which quantity is this a square?

In due order starting from the first place which consists of four, making (karaṇa) the names: “odd viṣama), even (sama), odd (viṣama), even (sama)”.

$$\begin{array}{*{20}{c}} {{\rm{Setting}}\,{\rm{down}}:}\\ {\begin{array}{*{20}{l}} {{\rm{sa}}}&{{\rm{vi}}}&{{\rm{sa}}}&{{\rm{vi}}}&{{\rm{sa}}}&{{\rm{vi}}}\\ 1&8&6&6&2&4 \end{array}} \end{array}$$

In this case, the odd terms which are the places for the ones, hundreds, and ten thousands, consist of four, six and eight. Therefore the last odd term is the ten thousand place which consists of eight. Then, the first quantity is 18. Having subtracted (apāsya) 16 since it is a possible square for these quantities, the remainder is two. That last quantity is placed separately (vyavatiṣthate) above. Thus, where it is placed (sthite sati) the root of 16, 4, times 2, 8, is to be led (neyaḥ) below the place where the square was subtracted (vargaśuddhi), which consists of six for the place of decrease. And then division (bhāgāpahāraḥ) of twenty six led above (uparitanyā). Setting down:

$$\begin{array}{*{20}{c}}2&6&6&2&4\\ {}&8&{}&{}&{} \end{array}.$$

When 24, which is eight multiplied by 3, is subtracted from below (that is from 26), above two remains. Below, the quotient which is three should be inserted (niveśya) on a line, they (e.g., these three units) should be placed (sthāpya) under consisting of six. Its square is nine. Having subtracted (śodhayitvā) this from above, these (1724) , three is multiplied by two, six is to be made (kartavya). Below, eighty six is produced. This quantity slithers (sarpati) on a line. Below two, there is [six], below seven, eight. Setting down:

$$\begin{array}{*{20}{c}} 1&7&2&4\\ {}&8&6&{}\end{array}$$

Division above of a 100 increased by 72 by that 86. Decreasing from above the dividend without remainder by two, the result is two, having inserted (niveśya) those two on a line, having placed (sthāpyau) both below four, its square is four; having subtracted (śodhayitvā) from above, those two multiplied by two should be made (kartavyā) four, therefore 800 increased by 64 is produced (jayite).

Since above the quantity subtracted has no remainder, there is no sliding like a snake etc. process, it remains just the halving of the quantity obtained which should be carried out. Thus, when that is done (kṛta), the result is 432. Its square is 186624.

Rights and permissions

Reprints and permissions

Copyright information

© 2015 Springer International Publishing Switzerland

About this chapter

Cite this chapter

Keller, A. (2015). Ordering Operations in Square Root Extractions, Analyzing Some Early Medieval Sanskrit Mathematical Texts with the Help of Speech Act Theory. In: Chemla, K., Virbel, J. (eds) Texts, Textual Acts and the History of Science. Archimedes, vol 42. Springer, Cham. https://doi.org/10.1007/978-3-319-16444-1_5

Download citation

Publish with us

Policies and ethics