The Evolution of Modern Numerals from Ancient Tally Marks

Author(s): Charles E. Woodruff

Source: The American Mathematical Monthly , Aug. - Sep., 1909, Vol. 16, No. 8/9 (Aug. -
Sep., 1909), pp. 125-133
Published by: Taylor & Francis, Ltd. on behalf of the Mathematical Association of
America

Stable URL: https://www.jstor.org/stable/2970818

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 [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Taylor & Francis, Ltd. and Mathematical Association of America are collaborating with JSTOR
to digitize, preserve and extend access to The American Mathematical Monthly

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC:56 UTC
All use subject to https://about.jstor.org/terms
 THE
AMEIRICAN
MATHEMATICAL MONTHLY.
Entered at the Post-office at Springfield, Missouri, as second-class matter.

VOL. XVI. AUGUST-SEPTEMBER, 1909. NOS. 8-9.

THE EVOLUTION OF MODERN NUMERALS FROM

ANCIENT TALLY MARKS.

By MAJOR CHARLES E. WOODRUFF, A. M., M. D., U. S. Army.

No acceptable explanation has ever been given as to the origin of our

numerals though the literature on the subject is enormous. The multiplicity
of the theories is evidence of the inadequacy of each. Of course it is known
that the Arabs obtained the signs from Southern Asia or India before the
ninth century A. D. and brought them to Europe in the tenth, but the re-
mote origin has never been discovered. The later Greeks and Hebrews used
the first letters of their respective alphabets to represent units, the second
groups to represent tens and the third for the hundreds. Sometimes the
initial letter of the word for the number was used as a symbol for that num-
ber, as in the early Greek, and possibly the Roman C and M, though the
latter may have been evolved from earlier symbols. For these reasons quite
a number of scholars have sought for the original forms of our numerals in
the letters of some alphabet. There is certainly quite a remarkable resem-
blance between some of the old numerals and some of the letters of old al-
phabets, but that is no proof of common origin. Indeed it would be strange
if we could not find many such coincidences in the innumerable forms. which
both letters and numerals have taken. Indeed numerals were used where
there were no alphabets or before alphabets were evolved as in modern
China and ancient Egypt.
In addition, Sir E. Clive Bagley and George Buhler point out the fact
(Journal Royal Asiatic Society, p. 335, 1882) that there is no known reason
why certain letters should have been selected to represent the numerals four
to nine, which the former accepts as being derived from letters from several
alphabets widely separated. Bagley curiously enough also states that all
numerals at first were merely shorthand ways of expressing their names,
which wholly contradicts the idea of derivation from letters, and ignores the
fact that primitive tallies or numerals must have existed before they had
names.
It has also been frequently asserted that the numerals were inventions

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC
All use subject to https://about.jstor.org/terms
126

which sprang up in a very short time, although such a phenomenon is con-

trary to human experience. Sudden appearance generally means borrowing,
for all written symbols were slow in their evolution.
Reasoning from analogy, one would assume that if letters had their
origin in the first crude attempts to represent things and ideas, the numer-
als must necessarily have their origin in the first crude attempts to record
numbers.
The natural way of recording numbers is by tally marks, and it is the
universal custom of mankind, at least of all who were intelligent enough to
count. Historians of mathematics use the term tally-mark to refer to the
notched stick (French, tailler, to cut), but it is here meant to refer to any
simple marks or scores. The Maya numerals are described by George B.
AWTo equalize the plates it was necessary to include one Indian column on the first and two on the third plate.

CHINESE.

Probable Origin. x
(Chalfant.) i Evolutionary Forms.

1. One line or one J 2 \ i-X4

weapon.

2. Twolinesor two /1 \\ .V1 \\ Zy0 2

weapons.

3. Three lines. //) ,fl \\/// $

4. Four lines grad--___

ually connected liii 1171jj K7t7 c7
incursorystyle. AriJ
5. Five lines var- 7X i72 .
iouslyindicated. lX aXA a.7 o '-i
6. Probably six |L X -
lines united. J T L'
7. Probably seven R j j
lines united. 4 L

8. Eight lines A -2: it

9. Probably nine m X L LI-

lines united. =.7
10.Two contracted
signs for five
united. f i i

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC
All use subject to https://about.jstor.org/terms
 127

Gordon (The Century, January, 1898), as follows: "The numbers from one
to four are represented by dots; a bar signifies five; a bar and a dot six; ten
is written by two bars; and so on up to nineteen, the sign of which is three
bars and four dots; after this number the signs employed are in doubt."
The Egyptians used tally marks up to nine but ten is an inverted U. Cune-
iform numerals are also tally marks. Even as late as the third century B.
C. in India the Asoka edicts record numbers up to five as vertical tally marks.
In the accompanying chart, there is some evidence that, as a rule,
tally marks originally went only to five after which the one or five was re-
peated with extra markings. Professor Edwin S. Crawley (Popular Science
Monthly, August, 1897), says that all systems of numbers were originally
quinary from the use of the fingers of one hand in counting and became
decimal as soon as two hands were used, but others state that there is no
INDIAN.

C-4 z~~ 0. '1 *

CA \ \ _ CC _ cc _ _

7t~~ ff SX hjy fYYY( >

x t t A .- ~~- - X-
ri2

ZnJ i s 7? 7122Y7/ 2
4 fb t 7(3D7 v

4 X S x7oc?c2o(L"

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC
All use subject to https://about.jstor.org/terms
128

foundation for such a generalizetion. A few people have used both hands
and feet, a vigesimal system, whose remnants are seen in the "quatre-
vingts" of the French from 81 to 99 and our "score" of one finished tally
group. Duodecimal systems were sometimes grafted on as in Babylonia,
making also a sexagesimal system, side by side with the decimal. The
Chinese numerals for 11 to 20 are merely those from one to ten, with a ten
(X) over or before them, but the commercial forms are quinary, the six to
eight being based on the one to five with extra tally marks beneath.
The numerals in the accompanying chart are copied from Isaac
Taylor's "The Origin of the Alphabet," from the pamphlet on the subject
by Professor John C. C. Clark, and from a paper by Karl Mischke (Deutsche
Japan Post, Yokahama, Japan, December 29, 1906). Edward Clodd, "The
ARABIC. EUROPEAN.

0 0 mm 0

c 0 g S a A 88 s 80 W Z'

>~~~~ c0c 0 0 FJE335

~~ 3 ~~~ r~~"r 3 3)33 3

~~ g ~ut >I(hlLlqvtAj /f4 1\5

S E 6 A 5'~~ ~ 8YLb&<e'

Q' e 9 9 9 5o. oL /0
?~~~I lo l o0e w0 fr^ !OL /

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC
All use subject to https://about.jstor.org/terms
 129

Story of the Alphabet, " derived his from Taylor, who in 1882, described the
changes in Arabic numerals. The Chinese forms are copied mostly from the
monograph on early Chinese writing by Rev. Frank H. Chalfant, a mission-
ary in China (Memoirs of the Carnegie Museum, Pittsburg, Vol. IV, No. 1),
to which have been added a few old forms obtaned through the courtesy of
Mr. L. C. Solyom of the Congressional Library, who copied them from
"Luh-shoo-fun-luy," a dictionary of ancient forms and characters, by Foo
Lwan-tseang. Others were obtained through the kindness of Professor
Friedrich Hirth, Professor of Chinese in Columbia University in New York,
who searched through Chinese archeological inscriptions for the purpose.
Two columns of Nepal and Jaina manuscript and the "Bower" manuscript
are copied from George Buhler's "Indische Paleographie," in which there
are also a large number of variants, so closely resembling those copied else-
where that they have been omitted here. Identical forms are also found in
Burnell's "South Indian Paleography." There are also an enormous num-
ber of Medieval European forms in Cappelli's "Dizionario di Abbreviature
Latine ed Italiane" and in Chassant's "Dictionaire des Abreviations Latines
et Francaise" but they are almost identical with the ones here given. This
infinite variety is partly due to an occasional scribe's fondness for decorating

PROBABLE EVOLUTION OF THE NUMERALS.

.- N l / . 11

-XXK~'~Z ZAI'YbY

T`FS7
3-u 3bt+T mrwf
3'73?2 q
nS;3 iLtta
PS t) t u11SiSX A

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC
All use subject to https://about.jstor.org/terms
130

with flourishes, as well as due to the gradual evolution of simple forms from
more complex ones. In all of them there must also be considerable allow-
ance made for incorrect transcribing in the numerous times they have been
copied from one book to another since the original was taken from some
manuscript, coin, monument, or vase, and some were probably erroneously
translated.
Tally marks must be so distinctive that they can be read, no matter
how the coin or stick is held, vertically, horizontally, or inverted. The mod-
ern six and nine would have been useless, for they would have been indis-
tinguishable unless the top was known. Consequently we find many
instances in which the numerals have been inverted or inclined at various
angles as it was of no practical importance at first whether they were upside
down or not. Our two, for instance, as well as the four, five, six, or
seven have all been inverted and reversed at times. There may also have
been reversals of the forms due to the fact that the people from whom the
Arabs obtained the numerals, probably wrote from left to right while the
Arabs wrote from right to left, and the Europeans from left to right. In
addition to this, one Sanskrit language was written from right to left, and
another from left to right. It is remarkable that the old Arabic forms have
changed less in Europe than in Arabia. The final European forms of 6, 7,
8, 9, and 10 after many vicissitudes of change are nearly identical with old
Arabic forms and show strong relationship to the early Indian. The vast
majority of the old forms are like fossils of extinct species which have left
no descendants; but by picking out forms here and there, it is possible to
secure a gradation of changes from the first to the last, the zero, by the
way, being a very late development, possibly from the position dot of the
Arabic, which in turn was probably a modification of the Indian 0 or G
The use of position of some sort to indicate multiple value is a recent inven-
tion, G. R. Kaye stating that there is no evidence of position value prior to
1000 A. D. (Journal Royal Asiatic Society, July, 1907).
The numeral one, is always a single tally mark, straight or curved,
horizontal, vertical, or at an angle. Two was evidently two tally marks
and cursively became the modern 2 or the reversed or inverted form. Three
in all its forms is evidently derived from three parallels. But the higher
numerals have not so clear an origin. It is quite difficult to distinguish
groups of parallel lines of more than three. We use a cross tally for every
fifth stroke making groups of five, and probably we once grouped a "score"
together also to indicate each time all the fingers and toes were counted.
The Egyptians used groups of three strokes up to nine. Consequently some
other system than parallel groups must have been used for the original
numerals of four to ten. The modern Arabic has even returned to figures
of four and five strokes for these two numerals, after having evolved other
forms which they later abandoned, and there seems often to be an effort in
ancient Indian forms of four and five to make them of four and five strokes,
respectively.

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC
All use subject to https://about.jstor.org/terms
 131

The evolution of the Chinese numerals from tally marks is evident

from a mere glance at the chart, for in every case there is a gradation to the
modern. There are several facts which must be kept in mind in interpreting
these changes. In the first place there are two evolutions recorded side by
side -the oldest or quinary system which developed into the modern com-
mercial, and the later decimal which formed the common numerals. Both
of these are exceedingly ancient in origin, possibly a thousand years before
the Christian era or even much earlier still for they had developed into the
present decimal forms by 200 A. D. To represent six, seven, eight, and
nine, the Chinese evidently first used five vertical tally marks under which
they placed one to four horizontals, respectively, but the five verticals were
early contracted into one. The commercial forms for six, seven, and eight
are thus exceedingly old. The nine was evidently abandoned for the more
convenient decimal system. The X which represents the commercial four
is a direct descendant of the four vertical tally marks, the horizontal marks
were used for the numeral four as late as the Ham dynasty, 206 -220 A. D.
After this time only the new form appears though its origin was probably
very much more remote. The commercial five which looks like our eight, is
also a direct descendant of a very old form of five strokes which early
replaced the five parallel ones -the transition forms given by Chalfant be-
ing quite conclusive as to this point.
When a decimal system became necessary, the Chinese, or those from
whom they borrowed, had to invent other forms of tally marks for the
numerals from five to nine, as it was clearly impracticable to continue the
use of parallel strokes, but why they should have hit upon the curious forms
recorded is a mystery.
Another fact must be kept in mind in- explaining some of these
changes. The use of a brush for a pen made it impossible to write the old
curved forms which were carved on stone or written with a solid pencil.
Chinese writing thus very early abandoned curves for straight lines, the
squares and angles of the modern writing. This fact explains the change
from the rounded forms of 100 A. D. into the modern numerals at about that
date. The modern four, for instance, is sometimes thought to have been
four vertical strokes inclosed in a square, and that it changed into the mod-
ern form of a square containing two short strokes. The process was the
opposite, for the square is the manner in which the oval of the older form
had to be written with a brush, and the oval form is evidently a descendant
of the earlier vertical four strokes. The Chinese "running" characters for
two and three are the same as our own. The standard numerals (Cheng-
tzu) are not given on the chart as they are complicated late inventions to
prevent swindling, the commercial and common numerals being very easily
changed to higher ones by the addition of a stroke or two. The human fig-
ure or a weapon of some sort seems to have been used in addition to the tally
marks in the original decimal numerals. Buhler (Journal Royal Asiatic

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC
All use subject to https://about.jstor.org/terms
132

Society, 1882) says that there is no doubt that the numerals were introduced
into India from without, as they appear rather suddenly and in a well
evolved form. The original forms, then, are not found in India and the
oldest known Indian must be evolutionary forms of some nearby Mongolian
system, though Buhler and many others are inclined to believe that they
came from Egypt. Prior to their introduction the Hindoos used quite
a variety of methods of expressing numbers which do not concern us here.
Though letters and syllables were used, there are others having no such
derivation. Indeed one writer, Kern, is quoted by Btihler as believing the
old Indian four and five to be four and five strokes, respectively. The iden-
tity of the first three numerals in Chinese and Indian would lead us to sus-
pect identity of origin of the two systems, or more probably that the Chinese
forms were carried by traders into India. Indeed Dr. Fritz Hommel ("The
Civilization of the East") states that Indian "Culture is in the main an off-
shoot of the Chinese." The Chinese would also be suspected on account of
the identity of the old Nepal nine and an evolutionary Chinese nine and
there is a perfect gradation of forms from the oldest Chinese nine tally marks
to our present nine.
Doubtless there was a considerable interval between the date of intro-
duction of the numerals into India and the date of the oldest surviving forms
- an interval in which their origin was forgotten if it was ever known.
This fully accounts for the fact that in some cases considerable change had
taken place. Probably many of the ancient Indian evolutionary forms will
never be found, nevertheless those now known are within the bounds of or-
dinary variation of writi'ng. They were undoubtedly introduced while some
still retained an evident form of tally marks, which led to the frequent
writing of four and five with four and five strokes, respectively. Some of
the old Indian fives are almost identical with the Chinese original form, and
in the six there are undoubted affinities, the sevens are closer still and nines
identical. Moreover these resemblances are between forms of practically
the same period. The well evolved "western cave" forms are ascribed to
the first and second century A. D., and. the older Indian must be set at least
two centuries earlier which makes them contemporaneous with the later
Chinese evolutionary forms which they resemble. Indian dates prior to the
Christian era are very uncertain, for it is said the Brahma laws were not
written until about 325 B. C. As the Asoka edicts did not have the numer-
als in 250 B. C., and as the cave numerals were well evolved forms in the
first century A. D., we have a fairly definite period in which they were in-
troduced and evolved.
Among the old Indian forms given by Kaye (Journal Asiatic Society
of Bengal, July, 1907) there is an eight which is evidently an attempt
to write the Chinese eight cursively, and it makes this evolutionary series
almost as complete as the nine. The interesting thing about our eight and
two, is that each seems to be the result of an attempt to write two paraflel

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC
All use subject to https://about.jstor.org/terms
 133

strokes cursively, the two was two horizontal parallel bars and the eight two
curved vertical lines, and some of the old forms of eight are practically the
same as the two turned over 90?.
It is quite likely also that when forms of numerals are evidently tally
marks, the ancient tribes would not stick to any particular arrangement, but
form new ones provided they indicated numbers. This is the most reason-
able explanation of the very evident tally-mark nature of the numerals in
the Jaina manuscript. Ten is a nine with an extra stroke, and the eights
are sevens with an extra stroke. The Jaina four, five, and six are also
clearly derived from groups of marks. In course of time, by slurring, omis-
sion of strokes and adding embellishing flourishes, the Nepal and Bower
manuscript forms arose. Indeed in the seven there is a perfect gradation
of evolutionary forms to our present seven. In the four the resemblance is
seen by making an assumption. In the five there is more evidence of an at-
tempt to write cursively one of the X forms of the Chinese, but the six is
not so evident without making two assumptions.
The supremacy of the Chinese numerals is explained by the fact that
they were the first ideographs in the field. Egyptian pictographs evolved
in the direction of representing sounds and, besides, their tally marks
elsewhere were in groups of parallels, and not the fortunate Chinese group-
ings which lent themselves to change into ideographs. The invention
of position value of course killed all the numerals above nine.

ON THE NUMBER OF EQUAL REGULAR SPHERICAL POLYGONS

THAT CAN BE CONSTRUCTED SO AS TO COMP-
LETELY COVER A SPHERE.

By B. F. YANNEY, Alliance, Ohio.

Let N=the number of equal polygons required; n=the number of

sides in each, and k=the number of angles about each common vertex.

Then will [ 360n- (n-2)180]N=the area, in spherical degrees, of the

sum of all the spherical polygons completely covering the sphere; but this
area is also equal to 720.

Therefore, [360n (n-2)180]N=720; whence, N=-n 4 .*

k 2n-nkr+l2k
It remains to solve this equation for positive integers.
*This formula may be found on page 69, Vol. III, of Henrici and Treutline's Geometry, though developed by a
different method, and having different considerations in view.

This content downloaded from

78.201.168.241 on Sat, 26 Jun 2021 01:03:42 UTC
All use subject to https://about.jstor.org/terms