Sie sind auf Seite 1von 9

The Recto of the Rhind Mathematical Papyrus How Did the Ancient Egyptian Scribe Prepare

It?
Author(s): R. J. Gillings
Source: Archive for History of Exact Sciences, Vol. 12, No. 4 (21.VIII.1974), pp. 291-298
Published by: Springer
Stable URL: http://www.jstor.org/stable/41133399 .
Accessed: 17/06/2014 05:24

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Springer is collaborating with JSTOR to digitize, preserve and extend access to Archive for History of Exact
Sciences.

http://www.jstor.org

This content downloaded from 62.122.78.62 on Tue, 17 Jun 2014 05:24:14 AM


All use subject to JSTOR Terms and Conditions
TheRectooftheRhindMathematicalPapyrus
EgyptianScribePrepareIt?
How Did theAncient
R. J. GlLLINGS

byM. Kline
Communicated

The RectooftheRMP oftheBritishMuseum, nownearly4,000yearsold,


occupiesaboutone-third ofthewholeofthe18-foot roll,andis themostextensive
of all the arithmetical tablesto be foundamongthe ancientEgyptianpapyri.
It is inscribedin hieraticcharacters, the cursiveformof hieroglyphics, and
normally readsfromrightto left.The tablegivesthevaluesof2 dividedby the
fifty odd numbers from3 to 101,all expressed as thesumsofunitfractions; for
example,2 -r- 7 is ' ¿. Unitfractions haveunityfornumerators, whichwiththe
solitary exception ofthefraction §, weretheonlyfractions theEgyptians used,
orevercoulduse,becauseoftheirnotation.Thusthenumber 7 inhieratic was ^,
and thesamenumber witha strokeaboveit, >?,represented thefraction *.
TheRectotablewasofgreatimportance to thescribes, becauseofitsfrequent
usein ordinary multiplication and division,comparable perhapsin modern days,
witha fullset ofmultiplication tables.But one asks,whyalways2 dividedby
oddnumbers? The reasonis thatin Egyptianmultiplication and division, only
the twicetimestable was used, by constantdoubling,and by occasionally
finding | of a fraction fromtheirrule,(RMP61 B), or theirtables.It was the
very useful circumstance, ofwhichthescribeswereaware,thatthepowersof2,
namely, 1,2,4, 8, 16,32, . , willproduce,
. . byproperly choosing andaddingthem,
everypossibleinteger, entirely uniquely.Thus to multiply 147by 43, thescribe
wrote147,doubledit (147X2), doubledthe answer(147x4), doubledagain
(147X 8), and so on,so thattheadditionoftheproducts for147X (1,2, 8, 32),
gives the product147x43- Now if the multiplicand contained unitfractions,
thentheirconstantdoubling presented difficultieswhich the Recto tablehelped
to solve.Therewas no needto includethedoubling ofevenunitfractions in the
Rectotable,because,forexample,¿ x 2 is obviously', but a seconddoubling
wouldpresent ' x 2, or 2 dividedby9, whichthetableprovides as (' ¿), butnot
as (I I), sincetwoequalunitfractions wereneverwritten together,exceptperhaps
as partofa calculation. Divisionwasmuchthesame,becauseifthescribewished
to divide147by 43, he didit by finding outwhat43 mustbe multiplied by,to
obtain147,whichis multiplication againbut oftenmoredifficult. Additionof
the integralpartsof the variousmultipleproductswas easy enough,but the
additionofthevariousunitfractions was a differentproposition.
20 Arch.Hist. Exact Sci., Vol. 12

This content downloaded from 62.122.78.62 on Tue, 17 Jun 2014 05:24:14 AM


All use subject to JSTOR Terms and Conditions
292 R. J. Gillings

In the mathematical papyri, there are many examples, also tables, of the
sums of unit fractions,such as, . . . ,

(3 6) =2 (6 9 18) =3 (15 25 75 200) =8


(4 12) =3 (7 14 28) =4 (30 50 150 400) =16,
(5 20) =4 (11 22 33) =6 fromthe Egyptian
(7 42) =6 (14 21 42) =7 Mathematical Leather
(9 18) =6 (18 27 54) =9 Roll.
(21 42) =14 (22 33 66) =11
Plus signs have been omitted, as well as the unit numerators and vinculums,
where no confusion arises. So competent were the scribes at the addition of unit
fractions,that we find many additions like the above, done apparently mentally,
or perhaps memorised. In RMP 36, the 16 fractions,

12 20 30 53
53 106 106 159
212 265 318 3I8
530 636 795 1060

are totalled as 4, by the use of "red auxiliaries", using 1060 as what we call the
LCM. Further, in RMP 70, the 15 fractions,
6 12 14
21 21 42 63
84 126 126 168
252 336 504 1008
were totalled as 2, with no explanatory matter at all! Students of the mathematics
of ancient Egypt, endeavouring to explain the modus operandi of the scribe who
constructed the Recto table, have of course made use of the above evidence of
the scribes1remarkable ability in the summation of unit fractions. Since the first
translation of the RMP, there have been many attempts to explain what may
have been the scribe's method in preparing this table. Some of those responsible
for these attempts were,

EiSENLOHR 1877 Hultsch 1895 van der Waerden 1937


Sylvester 1880 Griffith 1911 Struik 1948
Tannery 1884 Peet 1923 Becker 1951
Mansion 1888 Neugebauer 1926 Bruins 1952
Loria 1892 Chace 1927 Vogel 1958.

Mansion wrote: Les décompositions sont toujours, à un point de vue ou à un


autre, plus simple que toute autre décomposition possible.
Hultsch has: Attempts to explain it have hitherto not succeeded.
Chace wrote: Of the discussions which I have seen, the clearest is that by Loria.
But no formula or rule has been discovered that will give all the results of the
table.
And Becker: Das Prinzip der Berechnung scheint kein einheitliches zu sein.

This content downloaded from 62.122.78.62 on Tue, 17 Jun 2014 05:24:14 AM


All use subject to JSTOR Terms and Conditions
TheRhindMathematical
Papyrus 293

I drawattentionto theverylastentryin theRMP Recto,whichis 2 divided


by 101,givenas, 101 202 303 606,Thisis interesting and important because
it is the only possibledecomposition into not morethan fourdifferent unit
fractionslessthan1,000.In thisrespectthedecomposition is unique!It is also
clearthat,followingthisspecific method, thewhole50 oftheRectodecomposi-
tionscouldhavebeenwritten as follows:
2-r3=3 _6 9 18
2-^5 = 5 To TŠ 30
2-f-7= 7 14 2Ï 42 etc,etc,
up to and including 2 -MOI. But although thescribewas clearlyawareofthis,
noteverdoeshe acceptthesevaluesin histable,(exceptforthelast,wherehe
had no otherchoice),becausethereare so manyfarmoreconvenient values
available,whichcan be foundwitha littleeffort. The precepts whichmusthave
guidedhimin his search,maybe thussummarised:
Canon
1. Prefersmallnumbers, andnoneas greatas 1,000.
2. The fewerthetermsthebetter, and nevermorethan4.
3. Prefereven,to oddnumbers.
4. The smallernumbers comefirst, and nevertwothesame.
5. A smallfirstnumbermaybe increased, if the size of othersis thereby
reduced.
I nowproposeto showhoweveryentryin the RMP Rectotablemaybe
deriveddirectly fromthe simplerelationexpressedin the finaluniqueentry,
whichforsimplicity maybe written in modernnotationas,
2+-n = (i 2 3 6) x«.
Fromnowon, vinculums willbe omittedon therighthandside of the equals
sign,sinceall numbers representunitfractions.
Firstconsiderthemultiplesof3:
2-^3 =(1 2 3 6)X3
= (3 6) (9 18)
- 6.
_2
2^-9 = (l 2 3 6)X9
= (9 18) (27 54)
- 18.
_6
2 -=-15 = (1 2 3 6)xi5
= (15 30) (45 90)
= 10 30.
This procedurecoversall the divisorsup to 99 whichare multiplesof 3. Further,
they could all have been obtained by the multiplicationof (2 6) by 3, 5, 7,
9, ..., 33, quite simply,if the scribechose,and this would accountfor17 of the
50 entries.Note that so far there are no odd numberfractions,and all are
binomial.
20*

This content downloaded from 62.122.78.62 on Tue, 17 Jun 2014 05:24:14 AM


All use subject to JSTOR Terms and Conditions
294 R. J. Gillings

Nextconsidermultiplesof5, except45 and 75 alreadytreated:


2 -f-5 = (1 2 3 6) x 5
= (5 10 15 30)
= (5 10 30) 15
3 15.
This procedurecovers the divisors25, 65 and 85, whichcould also have been
obtainedby the multiplication of (3 15) by 5, 13 an(i 17- Note that all are odd
numberfractions,and all are binomial:
2-f-35= (l 2 3 6)x35
= 35 70 105 210
= (35 210) (70 105)
= 30 42.
2^55 = (1 2 3 6)x55
= (55 110 165) 330
= 30 330.
2-4-95=(i 2 3 6)X95
= 95 190 285 570
= 95 (380 380) 285 570
= (95 285 380) (380 570)
= 60 228.
The scribe'sentryin this case is however(60 380 570) so that one mightthink
thescribemissedtheequality(380 570= 228),whichis obtainedfrom(10 15=6)
multipliedby 38. But obviouslythe scribenoted that 2-M9 was earliergiven
as (12 76 114), which multipliedby 5 gives (60 38O 570), since (19 X 5=95),
so that perhapswe shouldnot say thathe erredin his methodbut thathe missed
All fractionsfor 35, 55 and 95 are even. We have now
a neat simplification.
accountedfor24 entries,and therehave been 40 even fractionsand only 8 odd
ones,whileall entrieshave been binomial.
These are 7, 49, 77 and 91 :
Now considerthemultiplesof 7 notalreadytreated.
2-r7 = (l 2 3 6)X7
= 7 14 (21 42)
= 7 14 14
= 7 14 (28 28)
= (7 14 28) 28
= 4 28.
This procedurealso coversthedivisors49 and 77. It also covers91 giving(52 364),
but thescribeobtainsa bettervalue accordingto his canonofprecepts,as follows:
2-f-91= (l 2 (3 6))X91
= (1 (2 2) )X91
= (1 1 ) X 91
= 91 91
= (140 260) (140 260)
= (140 140) (260 260)
= 70 130.

This content downloaded from 62.122.78.62 on Tue, 17 Jun 2014 05:24:14 AM


All use subject to JSTOR Terms and Conditions
The RhindMathematical
Papyrus 295

We have now accountedfor28 entriesof the Recto, all binomial,and still only
8 of the 56 unit fractionsare odd numbers.Note that 2-1-95 is recordedas
trinomial.
We nowconsiderthetworemainingbinomials,11 and 23:

2-Ml=(l 2 3 6)xll
= (11 22 33) 66
= 6 66.
2 -=-23 = (1 2 3 6)X23
= (1 2 (4 12) 6)X23
= (1 2 4 6 12)X23
= (23 46 92 138) 276
= 12 276.
This completesthe 30 binomialvalues givenin the Recto,with 52 even fractions
and 8 odd ones.
We now considerthe12 trinomialentriesof theRecto,excluding95:

2-M3=(l 2 (3 6))xl3
= 13 26 26
= 13 26 (52 52)
= 13 26 52 (104 104)
= (13 26 104) 52 104
= 8 52 104.
2-M7 = (1 2 3 6)xl7
= 17 34 51 102
= 17 (68 68) 51 102
= (17 68 102) 51 68
= 12 51 68.
2-M9 = (i 2 3 6)X19
= 19 38 57 114
= 19 38 (114 114) (228 228)
= (19 38 228) (114 228) 114
= 12 76 114.
2-^31 =(1 2 3 6) X 31
= 31 62 (93 186)
= 31 (124 124) 62
= 31 124 (155 620) 62
= (31 62 620) 124 155
= 20 124 155.
2^37 = (1 2 3 6) X37
= (37 2x37 3X37 (8x37 24X37))
= (37 2X37 24X37) 3x37 8x37
= 24 111 296.

This content downloaded from 62.122.78.62 on Tue, 17 Jun 2014 05:24:14 AM


All use subject to JSTOR Terms and Conditions
296 R. J. Gillings

2 -Ml =41 2X41 3X41 6x41


= 41 (4X41 8x41 8X41) 3x41 6X41
= (41 3X41 4X41 8X41) 6X41 8x41
= 24 246 328.
2-M7= 47 2X47 3x47 6x47
= 47 2X47 3X47 (10X47 15X47)
= (47 2X47 15X47) 3x47 10X47
= 141 470.
^0
2 4-53= 53 2x 53 3X53 6x53
= 53 2X53 (6x53 6x53) (10x53 15x53)
= (53 2x53 6x53 10x53) 6x53 15x53
= 318 795.
^0
2^-59= 59 2X59 3X59 6x59
= 59 (4X59 4X59) 3x59 (9x59 18X59)
= (59 3x59 4x59 18x59) 4x59 9x59
36 236 531.
2 4-67= 67 2x67 (3x67 6x67)
= 67 (4 X 67 4 X 67) 2 X 67
= 67 (5 X 67 20 X 67) (8 X 67 8 X 67) 2 X 67
= (67 2 x 67 8 X 67 20 X 67) 5 X 67 8 X 67
= 40 335 536.
24-71=71 2X71 (4X71 12X71) (10X71 15X71)
= 71 2X71 4X71 12X71 10X71 (24X71 40X71)
= (71 2X71 4X71 40X71) (12x71 24x71) 10X71
= 40 8X71 10X71
= 40 568 710.
2 4-97=97 2X97 3X97 6X97
= 97 (4 X 97 8 X 97 8 X 97) 3 X 97 (7 X 97 42 X 97)
= (97 3X97 4X97 8x97 42X97) (7x97 8x97)
= 56 679 776.
At thisstage we have accountedfor42 entriesof the Recto, and the progressive
total of odd numberedfractionsis only 16.
Thereremainonly8 quadrinomialequalities,all primedivisors.
29 79
43 83
61 89
73 101.
2 4-29=29 2X29 3X29 6x29
= 29 2 X 29 X
(4 29 12 X 29) 6 X 29
= 29 2X29 (8X29 8x29) 12X29 6x29
= (29 8X29 12X29) 2X29 6X29 8x29
= 24 58 174 232.

This content downloaded from 62.122.78.62 on Tue, 17 Jun 2014 05:24:14 AM


All use subject to JSTOR Terms and Conditions
The RhindMathematical
Papyrus 297

2 -M3= 43 2 x 43 3 x 43 6 x 43
= 43 2 x 43 3 x 43 (7x 43 42x 43)
= (43 42X43) 2 x 43 3 x 43 7 x 43
= 42 86 129 301.
2-^61=61 2X61 3x61 6X61
= 61 (4x61 8x61 8x61) 3x61 (10x61 15x61)
= (61 3x61 8x61 15X61) 4X61 8x61 10x61
= 40 244 488 610.

2 + 73= 73 2 x 73 3 x 73 6 x 73
= 73 (4 x 73 4 x 73) 3 x 73 6 x 73
= 73 4 x 73 (5x 73 20X 73) 3 x 73 6 X 73
= (73 6 x 73 20X 73) 3 x 73 4 x 73 5 x 73
= 60 219 292 365.
2 + 79= 79 2 X79 3 X79 6 X79
= 79 (4 X 79 4 x 79) 3 x 79 (10 x 79 15 X 79)
= (79 4x79 15x79) 3X79 4X79 10x79
= 60 237 316 790.
2 ^-83 = 83 2 x 83 3 x 83 6 x 83
= 83 (4 x 83 4 x 83) 3 x 83 6 X 83
= 83 4 X 83 (5 x 83 20 X 83) 3 x 83 6 x 83
= (83 3x83 20X83) 4X83 5x83 6X83
= 60 332 498.
41j>
2-^89=89 2x89 3x89 6x89
= 89 (4x89 4x89) (6x89 6x89) (10x89 15x89)
= (89 4 X 89 6 x 89 15 X 89) 4 X 89 6 x 89 10 X 89
= 60 356 534 890.
2 -M01 =101 2X101 3X101 6x101
= 101 202 303 606.

This finalvalue is the only simplification possible,and the total of odd number
unit fractionsis only 24 out of a total of 128, evidenceof the scribe'sattention
to precept3 of the canon.
Chace (RMP, 1927) thoughtthat both Sylvester (1880) and Loria (1892)
discussedthe problem,"fromthe modernpointof view". Hultsch (1895) wrote
that "the Egyptian reckoningwas indeed a study of mystery ". Neugebauer
(The Exact SciencesIn Antiquity,1951) consideredthat the methodshe showed
formultiplesof 3 "were in essencethe procedurethat led to the rulesfor2/w",
while van der Waerden (ScienceAwakening,1954) said that the values for
non-multiples of 3, "were computedby different methods". Closer to an ac-
ceptabletheoryis probablythat of Bruins (Platon Et La Table Egyptienne2¡n,
Janus,Vol. 46, 1957) in whichthe scribewouldhave neededto use the following
transformation for 2'n. First he would need to choose a suitable composite

This content downloaded from 62.122.78.62 on Tue, 17 Jun 2014 05:24:14 AM


All use subject to JSTOR Terms and Conditions
298 R. J. Gillings

number, N =d1D1 = d2D2 = dzDZf

suchthat, 2N = n+ dx+ d2+ dz,

so that, 2 = -w + 1^ + 1¿r+ -^r

S0that ¥ = W+ ^ + ^r + ^¡ïr
therequired Whilethismayhavebeenscribally
decomposition. possible,critics
wouldhesitate to attribute
sucha modern techniqueto ancient Egyptianscribes.
Bruins himself notablyfor2/59and 2/97,as wellas
notescertaindifficulties,
thescribe'slapse,oroversight,with2/95.Thereference to Plato in thetitle,he
explainsin a notein whichhesays that,whileearlierreadingLes Lois de Platon
" Voilàla méthode
VII 819,hethought tohimself, de construction de la table2'rì'

14TheMall
Turramurra
New SouthWales 2074
Sydney,Australia

(ReceivedJanuary 21, 1974)

This content downloaded from 62.122.78.62 on Tue, 17 Jun 2014 05:24:14 AM


All use subject to JSTOR Terms and Conditions

Das könnte Ihnen auch gefallen