Numeral system

A numeral system is a writ ing syst em for expressing numbers; t hat is, a mat hemat ical not at ion for
represent ing numbers of a given set , using digit s or ot her symbols in a consist ent manner.

Numbers written in different numeral systems

The same sequence of symbols may represent different numbers in different numeral syst ems. For
example, "11" represent s t he number eleven in t he decimal or base-10 numeral syst em (t oday, t he
most common syst em globally), t he number three in t he binary or base-2 numeral syst em (used in
modern comput ers), and t he number two in t he unary numeral syst em (used in t allying scores).

The number t he numeral represent s is called it s value. Not all number syst ems can represent t he
same set of numbers; for example, Roman numerals cannot represent t he number zero.

Ideally, a numeral syst em will:

Represent a useful set of numbers (e.g. all int egers, or rat ional numbers)

Give every number represent ed a unique represent at ion (or at least a st andard represent at ion)

Reflect t he algebraic and arit hmet ic st ruct ure of t he numbers.

For example, t he usual decimal represent at ion gives every nonzero nat ural number a unique
represent at ion as a finit e sequence of digit s, beginning wit h a non-zero digit .

Numeral syst ems are somet imes called number systems , but t hat name is ambiguous, as it could
refer t o different syst ems of numbers, such as t he syst em of real numbers, t he syst em of complex
numbers, various hypercomplex number syst ems, t he syst em of p-adic numbers, et c. Such syst ems
are, however, not t he t opic of t his art icle.

History

The first t rue writ t en posit ional numeral syst em is considered t o be t he Hindu–Arabic numeral
syst em. This syst em was est ablished by t he 7t h cent ury in India,[1] but was not yet in it s modern
form because t he use of t he digit zero had not yet been widely
Western Arabic 0 1 2 3 4 5 6 7 8 9
accept ed. Inst ead of a zero somet imes t he digit s were marked
Eastern Arabic ٠ ١ ٢ ٣ ٤ ٥ ٦ ٧ ٨ ٩
wit h dot s t o indicat e t heir significance, or a space was used as a
Persian ۰ ۱ ۲ ۳ ۴ ۵ ۶ ۷ ۸ ۹
placeholder. The first widely acknowledged use of zero was in
876.[2] The original numerals were very similar t o t he modern ones, Devanagari ०१ २३४ ५६७८९

even down t o t he glyphs used t o represent digit s.[1]

The digits of the Maya

numeral system

By t he 13t h cent ury, West ern Arabic numerals were accept ed in European mat hemat ical circles
(Fibonacci used t hem in his Liber Abaci). They began t o ent er common use in t he 15t h cent ury.[3] By
t he end of t he 20t h cent ury virt ually all non-comput erized calculat ions in t he world were done wit h
Arabic numerals, which have replaced nat ive numeral syst ems in most cult ures.

Other historical numeral systems using digits

The exact age of t he Maya numerals is unclear, but it is possible t hat it is older t han t he Hindu–
Arabic syst em. The syst em was vigesimal (base 20), so it has t went y digit s. The Mayas used a shell
symbol t o represent zero. Numerals were writ t en vert ically, wit h t he ones place at t he bot t om. The
Mayas had no equivalent of t he modern decimal separat or, so t heir syst em could not represent
fract ions.

The Thai numeral syst em is ident ical t o t he Hindu–Arabic numeral syst em except for t he symbols
used t o represent digit s. The use of t hese digit s is less common in Thailand t han it once was, but
t hey are st ill used alongside Arabic numerals.

The rod numerals, t he writ t en forms of count ing rods once used by Chinese and Japanese
mat hemat icians, are a decimal posit ional syst em able t o represent not only zero but also negat ive
numbers. Count ing rods t hemselves predat e t he Hindu–Arabic numeral syst em. The Suzhou numerals
are variant s of rod numerals.
Rod numerals (ver tical)

0 1 2 3 4 5 6 7 8 9

–0 –1 –2 –3 –4 –5 –6 –7 –8 –9

Main numeral systems

The most commonly used syst em of numerals is decimal. Indian mat hemat icians are credit ed wit h
developing t he int eger version, t he Hindu–Arabic numeral syst em.[4] Aryabhat a of Kusumapura
developed t he place-value not at ion in t he 5t h cent ury and a cent ury lat er Brahmagupt a int roduced
t he symbol for zero. The syst em slowly spread t o ot her surrounding regions like Arabia due t o t heir
commercial and milit ary act ivit ies wit h India. Middle-East ern mat hemat icians ext ended t he syst em
t o include negat ive powers of 10 (fract ions), as recorded in a t reat ise by Syrian mat hemat ician Abu'l-
Hasan al-Uqlidisi in 952–953, and t he decimal point not at ion was int roduced by Sind ibn Ali, who also
wrot e t he earliest t reat ise on Arabic numerals. The Hindu–Arabic numeral syst em t hen spread t o
Europe due t o merchant s t rading, and t he digit s used in Europe are called Arabic numerals, as t hey
learned t hem from t he Arabs.

The simplest numeral syst em is t he unary numeral syst em, in which every nat ural number is
represent ed by a corresponding number of symbols. If t he symbol / is chosen, for example, t hen t he
number seven would be represent ed by ///////. Tally marks represent one such syst em st ill in
common use. The unary syst em is only useful for small numbers, alt hough it plays an import ant role
in t heoret ical comput er science. Elias gamma coding, which is commonly used in dat a compression,
expresses arbit rary-sized numbers by using unary t o indicat e t he lengt h of a binary numeral.

The unary not at ion can be abbreviat ed by int roducing different symbols for cert ain new values. Very
commonly, t hese values are powers of 10; so for inst ance, if / st ands for one, − for t en and + for
100, t hen t he number 304 can be compact ly represent ed as +++ //// and t he number 123 as + −
− /// wit hout any need for zero. This is called sign-value not at ion. The ancient Egypt ian numeral
syst em was of t his t ype, and t he Roman numeral syst em was a modificat ion of t his idea.

More useful st ill are syst ems which employ special abbreviat ions for repet it ions of symbols; for
example, using t he first nine let t ers of t he alphabet for t hese abbreviat ions, wit h A st anding for "one
occurrence", B "t wo occurrences", and so on, one could t hen writ e C+ D/ for t he number 304 (t he
number of t hese abbreviat ions is somet imes called t he base of t he syst em). This syst em is used
when writ ing Chinese numerals and ot her East Asian numerals based on Chinese. The number syst em
of t he English language is of t his t ype ("t hree hundred [and] four"), as are t hose of ot her spoken
languages, regardless of what writ t en syst ems t hey have adopt ed. However, many languages use
mixt ures of bases, and ot her feat ures, for inst ance 79 in French is soixante dix-neuf (60 + 10 + 9)
and in Welsh is pedwar ar bymtheg a thrigain (4 + (5 + 10) + (3 × 20)) or (somewhat archaic) pedwar
ugain namyn un (4 × 20 − 1). In English, one could say "four score less one", as in t he famous
Get t ysburg Address represent ing "87 years ago" as "four score and seven years ago".

More elegant is a positional system, also known as place-value not at ion. The posit ional syst ems are
classified by t heir base or radix, which is t he number of symbols called digits used by t he syst em. In
base 10, t en different digit s 0, ..., 9 are used and t he posit ion of a digit is used t o signify t he power
of t en t hat t he digit is t o be mult iplied wit h, as in 304 = 3×100 + 0×10 + 4×1 or more precisely
3×102 + 0×101 + 4×100. Zero, which is not needed in t he ot her syst ems, is of crucial import ance
here, in order t o be able t o "skip" a power. The Hindu–Arabic numeral syst em, which originat ed in India
and is now used t hroughout t he world, is a posit ional base 10 syst em.

Arit hmet ic is much easier in posit ional syst ems t han in t he earlier addit ive ones; furt hermore, addit ive
syst ems need a large number of different symbols for t he different powers of 10; a posit ional
syst em needs only t en different symbols (assuming t hat it uses base 10).[5]

The posit ional decimal syst em is present ly universally used in human writ ing. The base 1000 is also
used (albeit not universally), by grouping t he digit s and considering a sequence of t hree decimal
digit s as a single digit . This is t he meaning of t he common not at ion 1,000,234,567 used for very large
numbers.

In comput ers, t he main numeral syst ems are based on t he posit ional syst em in base 2 (binary
numeral syst em), wit h t wo binary digit s, 0 and 1. Posit ional syst ems obt ained by grouping binary
digit s by t hree (oct al numeral syst em) or four (hexadecimal numeral syst em) are commonly used. For
very large int egers, bases 232 or 264 (grouping binary digit s by 32 or 64, t he lengt h of t he machine
word) are used, as, for example, in GMP.

In cert ain biological syst ems, t he unary coding syst em is employed. Unary numerals used in t he
neural circuit s responsible for birdsong product ion.[6] The nucleus in t he brain of t he songbirds t hat
plays a part in bot h t he learning and t he product ion of bird song is t he HVC (high vocal cent er). The
command signals for different not es in t he birdsong emanat e from different point s in t he HVC. This
coding works as space coding which is an efficient st rat egy for biological circuit s due t o it s inherent
simplicit y and robust ness.

The numerals used when writ ing numbers wit h digit s or symbols can be divided int o t wo t ypes t hat
might be called t he arit hmet ic numerals (0, 1, 2, 3, 4, 5, 6, 7, 8, 9) and t he geomet ric numerals (1, 10,
100, 1000, 10000 ...), respect ively. The sign-value syst ems use only t he geomet ric numerals and t he
posit ional syst ems use only t he arit hmet ic numerals. A sign-value syst em does not need arit hmet ic
numerals because t hey are made by repet it ion (except for t he Ionic syst em), and a posit ional
syst em does not need geomet ric numerals because t hey are made by posit ion. However, t he spoken
language uses both arit hmet ic and geomet ric numerals.

In some areas of comput er science, a modified base k posit ional syst em is used, called biject ive
numerat ion, wit h digit s 1, 2, ..., k (k ≥ 1), and zero being represent ed by an empt y st ring. This
est ablishes a biject ion bet ween t he set of all such digit -st rings and t he set of non-negat ive
int egers, avoiding t he non-uniqueness caused by leading zeros. Biject ive base- k numerat ion is also
called k-adic not at ion, not t o be confused wit h p-adic numbers. Biject ive base 1 is t he same as
unary.

Positional systems in detail

In a posit ional base b numeral syst em (wit h b a nat ural number great er t han 1 known as t he radix or
base of t he syst em), b basic symbols (or digit s) corresponding t o t he first b nat ural numbers
including zero are used. To generat e t he rest of t he numerals, t he posit ion of t he symbol in t he
figure is used. The symbol in t he last posit ion has it s own value, and as it moves t o t he left it s value
is mult iplied by b.

For example, in t he decimal syst em (base 10), t he numeral 4327 means

(4×103) + (3×102) + (2×101) + (7×100), not ing t hat 100 = 1.

In general, if b is t he base, one writ es a number in t he numeral syst em of base b by expressing it in

t he form anbn + an − 1bn − 1 + an − 2bn − 2 + ... + a0b0 and writ ing t he enumerat ed digit s
anan − 1an − 2 ... a0 in descending order. The digit s are nat ural numbers bet ween 0 and b − 1, inclusive.

If a t ext (such as t his one) discusses mult iple bases, and if ambiguit y exist s, t he base (it self
represent ed in base 10) is added in subscript t o t he right of t he number, like t his: numberbase. Unless
specified by cont ext , numbers wit hout subscript are considered t o be decimal.

By using a dot t o divide t he digit s int o t wo groups, one can also writ e fract ions in t he posit ional
syst em. For example, t he base 2 numeral 10.11 denot es 1×21 + 0×20 + 1×2−1 + 1×2−2 = 2.75.

In general, numbers in t he base b syst em are of t he form:

The numbers bk and b−k are t he weight s of t he corresponding digit s. The posit ion k is t he logarit hm
of t he corresponding weight w, t hat is . The highest used posit ion is close t o
t he order of magnit ude of t he number.

The number of t ally marks required in t he unary numeral syst em for describing the weight would have
been w. In t he posit ional syst em, t he number of digit s required t o describe it is only
, for k ≥ 0. For example, t o describe t he weight 1000 t hen four digit s are
needed because . The number of digit s required t o describe the position is
(in posit ions 1, 10, 100,... only for simplicit y in t he decimal example).

A number has a t erminat ing or repeat ing expansion if and only if it is rat ional; t his does not depend on
t he base. A number t hat t erminat es in one base may repeat in anot her (t hus
0.310 = 0.0100110011001...2). An irrat ional number st ays aperiodic (wit h an infinit e number of non-
repeat ing digit s) in all int egral bases. Thus, for example in base 2, π = 3.1415926...10 can be writ t en
as t he aperiodic 11.001001000011111...2.

Put t ing overscores, n, or dot s, ṅ, above t he common digit s is a convent ion used t o represent
repeat ing rat ional expansions. Thus:

14/11 = 1.272727272727... = 1.27 or 321.3217878787878... = 321.32178.

If b = p is a prime number, one can define base- p numerals whose expansion t o t he left never st ops;
t hese are called t he p-adic numbers.

It is also possible t o define a variat ion of base b in which digit s may be posit ive or negat ive; t his is
called a signed-digit represent at ion.

Generalized variable-length integers

More general is using a mixed radix not at ion (here writ t en lit t le-endian) like for
, et c.

This is used in Punycode, one aspect of which is t he represent at ion of a sequence of non-negat ive
int egers of arbit rary size in t he form of a sequence wit hout delimit ers, of "digit s" from a collect ion
of 36: a–z and 0–9, represent ing 0–25 and 26–35 respect ively. There are also so-called t hreshold
values ( ) which are fixed for every posit ion in t he number. A digit (in a given posit ion in
t he number) t hat is lower t han it s corresponding t hreshold value means t hat it is t he most -
significant digit , hence in t he st ring t his is t he end of t he number, and t he next symbol (if present ) is
t he least -significant digit of t he next number.

For example, if t he t hreshold value for t he first digit is b (i.e. 1) t hen a (i.e. 0) marks t he end of t he
number (it has just one digit ), so in numbers of more t han one digit , first -digit range is only b–9 (i.e. 1–
35), t herefore t he weight b1 is 35 inst ead of 36. More generally, if tn is t he t hreshold for t he n-t h
digit , it is easy t o show t hat . Suppose t he t hreshold values for t he second and
t hird digit s are c (i.e. 2), t hen t he second-digit range is a–b (i.e. 0–1) wit h t he second digit being most
significant , while t he range is c–9 (i.e. 2–35) in t he presence of a t hird digit . Generally, for any n, t he
weight of t he (n + 1)-t h digit is t he weight of t he previous one t imes (36 − t hreshold of t he n-t h
digit ). So t he weight of t he second symbol is . And t he weight of t he t hird symbol is
.

So we have t he following sequence of t he numbers wit h at most 3 digit s:

a (0), ba (1), ca (2), ..., 9a (35), bb (36), cb (37), ..., 9b (70), bca (71), ..., 99a (1260), bcb (1261), ..., 99b
(2450).

Unlike a regular n-based numeral syst em, t here are numbers like 9b where 9 and b each represent 35;
yet t he represent at ion is unique because ac and aca are not allowed – t he first a would t erminat e
each of t hese numbers.

The flexibilit y in choosing t hreshold values allows opt imizat ion for number of digit s depending on t he
frequency of occurrence of numbers of various sizes.

The case wit h all t hreshold values equal t o 1 corresponds t o biject ive numerat ion, where t he zeros
correspond t o separat ors of numbers wit h digit s which are non-zero.

See also

List of numeral syst ems

Comput er number format s

Non-st andard posit ional numeral syst ems

Hist ory of ancient numeral syst ems

Hist ory of numbers

 List of numeral syst em t opics

Number names

Repeat ing decimal

Residue numeral syst em

Long and short scales

Scient ific not at ion

-yllion

Numerical cognit ion

Number syst em

References

1. O'Connor, J. J. and Robert son, E. F. Arabic Numerals (ht t p://www-hist ory.mcs.st -andrews.ac.uk/H
ist Topics/Arabic_ numerals.ht ml) . January 2001. Ret rieved on 2007-02-20.

2. Bill Casselman (February 2007). "All for Nought " (ht t ps://www.ams.org/feat urecolumn/archive/in
dia-zero.ht ml) . Feature Column. AMS.

3. Bradley, Jeremy. "How Arabic Numbers Were Invent ed" (ht t ps://www.t heclassroom.com/how-t o
-ident ify-numbers-on-brass-from-india-12082499.ht ml) . www.theclassroom.com. Ret rieved
2020-07-22.

4. David Eugene Smit h; Louis Charles Karpinski (1911). The Hindu–Arabic numerals (ht t ps://archiv
e.org/det ails/hinduarabicnume05karpgoog) . Ginn and Company.

5. Chowdhury, Arnab. Design of an Efficient Multiplier using DBNS (ht t ps://books.google.com/book

s?id=WXn-mT3K6dgC&q=Arit hmet ic+is+much+easier+in+posit ional+syst ems+t han+in+t he+ea
rlier+addit ive+ones;+furt hermore,+addit ive+syst ems+need+a+large+number+of+different +sy
mbols+for+t he+different +powers+of+10;+a+posit ional+syst em+needs+only+t en+different +
symbols+(assuming+t hat +it +uses+base+10).&pg=PA2) . GIAP Journals. ISBN 978-93-83006-
18-2.

6. Fiet e, I. R.; Seung, H. S. (2007). "Neural net work models of birdsong product ion, learning, and
coding". In Squire, L.; Albright , T.; Bloom, F.; Gage, F.; Spit zer, N. New Encyclopedia of
Neuroscience.
Sources

Georges Ifrah. The Universal History of Numbers : From Prehistory to the Invention of the
Computer, Wiley, 1999. ISBN 0-471-37568-3.

D. Knut h. The Art of Computer Programming. Volume 2, 3rd Ed. Addison–Wesley. pp. 194–213,
"Posit ional Number Syst ems".

A.L. Kroeber (Alfred Louis Kroeber) (1876–1960), Handbook of t he Indians of California, Bullet in 78
of t he Bureau of American Et hnology of t he Smit hsonian Inst it ut ion (1919)

J.P. Mallory; D.Q. Adams, Encyclopedia of Indo-European Culture, Fit zroy Dearborn Publishers,
London and Chicago, 1997.

Hans J. Nissen; Pet er Damerow; Robert K. Englund (1993). Archaic Bookkeeping: Early Writing and
Techniques of Economic Administration in the Ancient Near East. Universit y of Chicago Press.
ISBN 978-0-226-58659-5.

Schmandt -Besserat , Denise (1996). How Writing Came About. Universit y of Texas Press.
ISBN 978-0-292-77704-0.

Zaslavsky, Claudia (1999). Africa counts: number and pattern in African cultures . Chicago Review
Press. ISBN 978-1-55652-350-2.

External links

Media relat ed t o Numeral syst ems at Wikimedia Commons