S.Duplij, "Positional numeral systems over polyadic rings", arxiv: 2506.12930

arXiv:2506.12930v1
[math.NT]
15
Jun
2025
POSITIONAL NUMERAL SYSTEMS OVER POLYADIC RINGS
STEVEN DUPLIJ
Center for Information Technology, University of Münster, 48149 Münster, Germany
and
Yantai Research Institute, Harbin Engineering University, 265615 Yantai, China
ABSTRACT. We construct positional numeral systems that work natively over nonderived polyadic pm, nq-
rings whose addition takes m arguments and multiplication takes n. In such rings, the length of an ad-
missible additive word and a multiplicative tower are not arbitrary (as in the binary case), but “quantized”.
Our main contributions are the following. Existence: every commutative pm, nq-ring admits a base-p place-
value expansion that respects the word length constraint in terms of numbers of operation compositions
ℓmult “ ℓaddpm ´ 1q ` 1. Lower bound: the minimum number of digits is greater than or equal to the
arity of addition m. Representability gap: for m, n ě 3 only a proper subset of ring elements possess finite
expansions, characterized by congruence-class arity shape invariants Ipmq
and Jpnq
. Mixed-base “polyadic
clocks”: allowing a different base at each position enlarges the design space quadratically in the digit count.
Catalogues: explicit tables for the integer rings Z4,3 and Z6,5 illustrate how ordinary integers lift to dis-
tinct polyadic variables. These results lay the groundwork for faster arity-aware arithmetic, exotic coding
schemes, and hardware that exploits operations beyond the binary pair.
CONTENTS
1. INTRODUCTION 2
2. PRELIMINARIES 3
3. POLYADIC RINGS 4
4. POLYADIZATION OF POSITIONAL NUMERAL SYSTEMS 7
5. CONCLUSIONS 12
REFERENCES 13
E-mail address: douplii@uni-muenster.de; sduplij@gmail.com; http://www.uni-muenster.de/IT.StepanDouplii.
Date: of start May 25, 2025. Date: of completion June 13, 2025.
Total: 11 references.
2010 Mathematics Subject Classification. 11-03, 11A07, 11A67, 20N15.
Key words and phrases. positional number system, arity, polyadic structure, polyadic number, congruence class, numeral,
polyadic ring, mixed-base, querelement.

1. INTRODUCTION
Classical positional numeral systems (for a review see, e.g. CONWAY AND GUY [1996], IFRAH [2000])
answer the question: how to encode quantities efficiently? The linked question, which discusses how to
operate on those encodings, has the narrow answer: the admissible operations are binary, that is addition
and multiplication decompose into chains of pairwise interactions. Nowadays in group theory, abstract al-
gebra, coding theory, ultra-metric analysis and physics n-ary (or polyadic) operations, those that swallow
n inputs at once, crop up with growing frequency (for a review, see DUPLIJ [2022] and refs therein), nev-
ertheless their foundational connections to numeral systems have remained largely unexplored (with the
exception of DUPLIJ [2017]). This paper fills that gap and bridges two seemingly disparate domains: the
positional numeral systems that represent integers MENNINGER AND BRONEER [1971], FOMIN [1974]
and polyadic rings that generalize ring theory to n-ary operations LEESON AND BUTSON [1980].
The main novel and unexpected phenomena which were uncovered in this paper are the following:
Double quantization of word length. As a starting point is the rich body of work on polyadic algebraic
structures, we observe that in an n-ary magma the fundamental operation µn combines n elements at
a time, such that closeness and total associativity now “quantizes” the admissible word lengths to the
arithmetic progression (opposed to the binary case, where word length is arbitrary). Adding a second m-
operation νm yields an pm, nq-ring, provided the two interact through a generalized distributive law and
obey “double quantization” of word length. The polyadic rings enjoy exotic features: multiple identities,
two kinds of querelements replacing inverses (of addition and multiplication), and the possibility that zero
or one may fail to exist at all.
Numeral polyadization. We introduce a novel systematic recipe that translates the classical place-value
expansion into a special genuine composition of m-ary additions and n-ary multiplications which reduces
to the standard positional numeral presentation in the binary p2, 2q case. The key feature of polyadic
numeral presentation is the unforeseen fact that the number of summands and the number of multiplicative
factors are no longer independent, but lock together through the arities, forcing numeral strings to respect
a built-in “polyadic uncertainty principle”.
Numeral representability is no longer guaranteed. Unlike the binary case (m “ n “ 2), not every
element of an pm, nq-ring admits a numeral expansion. We arrive to the theorem claiming that for m, n ě
3, the set of numbers with a finite positional expansion becomes a strict subset (characterized by the shape
invariants Ipmq
and Jpnq
) of the pm, nq-ring, unless m “ n “ 2.
Mixed-base and “polyadic clock” systems. Generalizing time-of-day notation, we allow each digit
position to carry its own n-ary tower of bases, yielding a combinatorial explosion of admissible encodings
and opening the door to cryptographic and coding-theoretic applications, as well as turn coin-change,
currency, and time-keeping problems to higher arity.
In general, polyadic numeral systems promise new tools wherever data are naturally grouped in arity-
specific chunks: multi-qubit gates in quantum circuits, multiary neural activations, and even higher-order
interaction terms in field theory.
The structure of the paper is as follows. SECTION 2 recaps the essentials of polyadic one set algebraic
structures with an emphasis on the quantization of admissible word lengths and the various exotic element
types (querelements, polyadic powers and idempotents, etc.). SECTION 3 develops the theory of pm, nq-
rings built from integer representatives of congruence classes, culminating in the arity-shape map that
dictates which classes admit the polyadic ring structure. SECTION 4 our main contribution constructs
positional numeral systems over the pm, nq-rings, proves the minimal-digit theorem, and works through
explicit examples in the p4, 3q and p6, 5q settings. We close with a catalogue of open problems, ranging
from polyadic floating-point arithmetic to potential links with higher-categories.
– 2 –

PRELIMINARIES
2. PRELIMINARIES
For self-consistency, we briefly recall polyadic structure notation and general properties of polyadic
rings (for details, see DUPLIJ [2022]).
Let X be a non-empty set and Xˆn
be its Cartesian product consisting of elements px1, . . . , xnq called
polyads (or n-tuples) and denoted by pxq P Xˆn
, and an n-tuple with equal elements is pxn
q. On Xˆn
the
polyadic (n-ary operation) is defined by µn : Xˆn
ÝÑ X and is denoted as µn rxs. A polyadic structure
@
X | µni
D
is the set X that is closed with respect to polyadic operations µni
(see, e.g., KUROSH [1963],
COHN [1965]).
The basic one-operation polyadic structure is called a (n-ary) magma M “ xX | µny. The (totally)
commutative n-ary magma M “ xX | µn | commy is defined by µn “ µn ˝ σ, where σ is an arbitrary
n-permutation. Additional requirements lead to various structures called group-like ones. For instance,
polyadic associative magma is an n-ary semigroup Sn “ xX | µn | assocy. The polyadic associativity
can be defined as the invariance DUPLIJ [2018]
µn rx, µn rys , zs “ invariant, (2.1)
where the internal multiplication can be on any of n ´ 1 places (giving n ´ 1 relations) and the polyads
x, y, z are of needed size, the total number of elements in (2.1) is fixed 2n ´ 1 (cf. with the binary
associativity px ¨ yq ¨ z “ x ¨ py ¨ zq). This allows us to omit internal bracket in compositions and define
the iterated multiplication
µ˝ℓµ
n rxs “
ℓµ
hkkkkkkkkkkikkkkkkkkkkj
µn rµn r. . . µn rxsss, x P Xℓµpn´1q`1
, (2.2)
where ℓµ is the “number” of operations (now multiplications). From (2.2) it follows the main and crucial
peculiarity of polyadic structures which differs them from ordinary binary ones (n “ 2): the length of a
word wµ pnq inside composition of an n-ary multiplications is not arbitrary, but it is “quantized”, obeying
the following admissible values
wadmiss
µ pnq “ ℓµ pn ´ 1q ` 1, (2.3)
which means that we can multiply wadmiss
µ pnq elements only.
Besides, one can treat the l.h.s. of (2.3) as another operation (the iterated multiplication) having higher
arity
µ̄ℓµpn´1q`1 “ µ˝ℓµ
n rxs . (2.4)
This allows us to divide all polyadic operations into two classes: iterated from lower arity operations
and noniterated, or derived and nonderived. Such division can also depend on the underlying set.
For simplest example, the ternary operation µ3 rx, y, zs “ xyz (product of three integers from Z) is
derived for positive integers X “ Z`
(because µ3 rx, y, zs “ µ2 rµ2 rx, ys , zs “ px ¨ yq ¨ z), but µ3 is
nonderived for negative integers X “ Z´
, because the product of two negative integers is positive, and so
the binary operation µ2 is not closed, i.e. the iteration does not exist. The ternary associativity is obvious
in both cases, and therefore the polyadic structures S`
3 “ xZ`
| µ3 | assocy and S´
3 “ xZ´
| µ3 | assocy
are ternary semigroups, nevertheless only the second one S´
3 is nonderived. Below we will consider
barely the nonderived polyadic operations and the corresponding polyadic structures. Recall definitions
of important elements in polyadic structures which will be used later on.
For each element x P X we can construct its ℓµ-polyadic power by
xxℓµy
“ µ˝ℓµ
n
“
xℓµpn´1q`1
‰
, (2.5)
and in the binary case n “ 2 the polyadic power differs by 1 from the ordinary power xxℓµy
“ xℓµ`1
.
– 3 –

PRELIMINARIES
A polyadic idempotent xid (if exists) is
x
xℓµy
id “ xid, xid P X. (2.6)
A polyadic zero is defined unambiguously by the n ´ 1 relations
µn rz, xs “ z, x P Xn´1
, (2.7)
where z can be on any of n ´ 1 places. A polyadic nilpotent element xnil is defined by one relation
x
xℓµy
nil “ z, xnil P X. (2.8)
The neutral pn ´ 1q-polyad e (in general non-unique) is given by
µn rx, es “ x, e P Xn´1
. (2.9)
If all elements in the neutral polyad are equal e “ en´1
, then
µn
“
x, en´1
‰
“ x, (2.10)
and e can be on any of n ´ 1 places and is called an identity of xX | µny.
If the identity e exists, we can define polyadic ℓµ-reflection by
x
xℓµy
refl “ e, xrefl P X. (2.11)
As follows from (2.7) with x “ zn´1
and (2.10) with x “ e, polyadic zero z and identity e are idem-
potents satisfying (2.6). There are exotic polyadic structures without idempotents, zero, identity at all or
with several identities.
For instance, the above ternary semigroup of negative integers S´
3 contains two neutral sequences
e` “ p`1, `1q, e´ “ p´1, ´1q and two ternary identities e` “ `1 and e´ “ ´1.
The invertibility in the polyadic case (n ě 3) is not connected with identity (2.10), but is guided by the
polyadic analogue of inverse, the querelement x̄ “ x̄ pxq defined by the n ´ 1 relations DÖRNTE [1929]
µn
“
x̄, xn´1
‰
“ x, x P X, (2.12)
which should be satisfied with x̄ be on all of n ´ 1 places, and such x is called polyadically invertible.
The polyadic solvability is defined as the unique solution for x0 with given u, y of the n ´ 1 equations
µn rx0, us “ y, x0, y P X, u P Xn´1
(2.13)
with x0 be on any of all n ´ 1 places. If each element of an n-ary semigroup Sn is polyadically in-
vertible, or equivalently, (2.13) is solvable on any place, then Sn becomes an n-ary (polyadic) group
Gn “ xX | µn | assoc | solvy. The existence of identity is not necessary for polyadic groups.
For example, the imaginary fractions ix{y with x, y P Zodd
(i2
“ ´1, operations are in C) form a
nonderived ternary group Godd
3 without identity, each element is polyadically invertible and has its unique
querelement, that is pix{yq “ ´iy{x.
3. POLYADIC RINGS
A set with two polyadic operations belongs to the so-called ring-like polyadic structures (see, e.g.,
LEESON AND BUTSON [1980], DUPLIJ [2022] and refs. therein).
A polyadic or pm, nq-ring Rm,n “ xX | νm, µny is a nonempty set X with m-ary addition νm : Xm
Ñ
X and n-ary multiplication µn : Xn
Ñ X, such that additively xX | νm | assoc | comm | solvy is an m-
ary commutative group, and multiplicatively xX | µn | assocy is a n-ary semigroup, while the operations
– 4 –

POLYADIC RINGS
νm and µn are not arbitrary, but connected by the following distributivity property having n relations
µn rνm rx1, . . . xms , y2, y3, . . . yns
“ νm rµn rx1, y2, y3, . . . yns , µn rx2, y2, y3, . . . yns , . . . µn rxm, y2, y3, . . . ynss (3.1)
µn ry1, νm rx1, . . . xms , y3, . . . yns
“ νm rµn ry1, x1, y3, . . . yns , µn ry1, x2, y3, . . . yns , . . . µn ry1, xm, y3, . . . ynss (3.2)
.
.
.
µn ry1, y2, . . . yn´1, νm rx1, . . . xmss
“ νm rµn ry1, y2, . . . yn´1, x1s , µn ry1, y2, . . . yn´1, x2s , . . . µn ry1, y2, . . . yn´1, xmss , (3.3)
where xi, yj P X, i “ 1, . . . , m, j “ 1, . . . , n.
If not all relations of distributivity (3.1)–(3.3), as well as associativity (2.1) are not satisfied, they are
called partial (as opposed to the total ones), which gives enormous different versions of polyadic rings.
For instance, the exotic p3, 2q-ring over rationals R3,2 “ xR |ν3, µ2y with the derived ternary addition
ν3 rx, y, zs “ x ¨ y ¨ z and the binary multiplication µ2 rx, ys “ xy
is noncommutative, because xy
‰ yx
,
nonassociative, since pxy
qz
‰ xpyzq
, and partial (left) distributive, because only the first relation (3.1)
holds pxyzqt
“ xt
yt
zt
, but txyz
‰ tx
ty
tz
, x, y, z, t P R.
The simplest ternary ring is the commutative, totally associative and totally distributive p2, 3q-ring of
negative integer numbers Rp´q
2,3 “ xZ´
|ν2, µ3y with the derived ternary addition ν2 rx, ys “ x ` y and
the nonderived ternary multiplication µ2 rx, ys “ x ¨ y ¨ z (the binary product x ¨ y gives result out of the
underlying set Z´
that is in positive integers Z`
).
More general polyadic number rings which are formed by representatives of congruence (residue)
classes were introduced in DUPLIJ [2017, 2019]. Recall that a congruence class of a non-negative in-
teger a P Z`
, modulo natural b P N, is defined by
rrassb “
!
x
ra,bs
k | k P Z
)
, (3.4)
where
x
ra,bs
k “ a ` bk, (3.5)
0 ď a ď b ´ 1, (3.6)
is a representative of the class being infinite. The set of representatives Xra,bs
”
!
x
ra,bs
k
)
(as elements
of the congruence class rrassb) was never considered before in details, because it does not form any
binary structure, while the “internal” operations are not simultaneously closed x
ra,bs
k1
` x
ra,bs
k2
R Xra,bs
,
x
ra,bs
k1
¨ x
ra,bs
k2
R Xra,bs
for arbitrary a, b (obviously, Xr01s
P Z are ordinary binary integers). This was a
reason, why only operations between classes (the binary class addition `1
and binary class multiplication
¨1
) were considered, we call them “external” operations DUPLIJ [2017].
Nevertheless, for special values of the congruence class parameters a and b polyadic (nonbinary) oper-
ations on Xra,bs
can be defined, and they are closed DUPLIJ [2017]. Indeed, the closeness of m additions
and n multiplications in Z
x
ra,bs
k1
` . . . ` x
ra,bs
km
P Xra,bs
Ă Z, (3.7)
x
ra,bs
k1
¨ . . . ¨ x
ra,bs
kn
P Xra,bs
Ă Z, ki, m, n P N, (3.8)
– 5 –

POLYADIC RINGS
hold valid, if the following “quantization” conditions
ma1
” a1
pmod b1
q ðñ
pm ´ 1q a1
b1
“ Ipmq
pa1
, b1
q ” I “ integer, (3.9)
a1n
” a1
pmod b1
q ðñ
a1n
´ a1
b1
“ Jpnq
pa1
, b1
q “ J “ integer, (3.10)
are satisfied with the special values of parameters as solutions ta1
u Ă tau, tb1
u Ă tbu, Xra1,b1s
Ă Xra,bs
.
The corresponding mapping is called the arity shape
Ψra1b1s
m,n : pa1
, b1
q ÝÑ pm, nq , (3.11)
and it is presented for lowest values in TABLE 1 of DUPLIJ [2019], however Ψ
ra1b1s
m,n cannot be presented
by a formula. The arity shape mapping (3.11) is injective and non-surjective (empty cells in TABLE
1): for the congruence classes rr2ss4, rr2ss8, rr3ss9, rr4ss8, rr6ss8 and rr6ss9 there are no solutions of the
“quantization” conditions (3.9)–(3.10), while, e.g., m “ 5, n “ 6 corresponds to different congruence
classes rr2ss5, rr3ss5, rr2ss10, and rr8ss10.
The closeness (3.7)–(3.8) and the “quantization” conditions (3.9)–(3.10) allow us to define two abstract
polyadic operations on the set of the abstract elements Xra1,b1s
“
!
x
ra1,b1s
k
)
” txku reflecting the represen-
tatives of the fixed congruence class rra1
ssb1 : m-ary addition and n-ary multiplication
νm rxk1 , . . . , xkm s “ xk1 ` . . . ` xkm “ xk0 , (3.12)
µn rxr1 , . . . , xrn s “ xr1 ¨ . . . ¨ xrn “ xr0 , (3.13)
where
k0 “ k1 ` . . . ` km ` Ipmq
pa1
, b1
q , (3.14)
r0 “ s pri, a1
, b1
q ` Jpnq
pa1
, b1
q , (3.15)
where the invariants Ipmq
pa1
, b1
q and Jpnq
pa1
, b1
q are in (3.9)–(3.10), and the integer s pri, a1
, b1
q is a special
polynomial of ri, a1
, b1
, which follows from the presentation (3.5). For instance, in the ternary case n “ 3
we have s pr1, r2, r3, a1
, b1
q “ a12
r1 ` a12
r2 ` a12
r3 ` a1
b1
r1r2 ` a1
b1
r1r3 ` a1
b1
r2r3 ` b12
r1r2r3.
Because the abstract polyadic operations νm and µn are closed, commutative, totally associative and
totally distributive (two latter ones follow from the binary associativity and distributivity), we can define
the abstract commutative polyadic pm, nq-ring of integers
Zra1,b1s
m,n “
A
Xra1,b1s
| νm, µn | comm
E
. (3.16)
It follows from (2.3), that we have “double quantization”: in the pm, nq-ring we can add
wadmiss
ν pmq “ ℓν pn ´ 1q ` 1 (3.17)
elements (ℓν is quantity of composed m-ary additions) and multiply
wadmiss
µ pnq “ ℓµ pn ´ 1q ` 1 (3.18)
elements, where ℓµ is quantity of composed n-ary multiplications, to be in the same underlying set Xra1,b1s
.
It is obvious, that p2, 2q-ring Zr0,1s
2,2 “ Z is the binary ring of ordinary integers, and wadmiss
ν pmq “ ℓν,
wadmiss
µ pnq “ ℓµ are any natural numbers N (without any “quantization”).
For example, in the polyadic p8, 7q-ring Z8,7 “ Zr5,7s
8,7 we can add only 7ℓν ` 1 “ 8, 15, 22 . . . elements
xi and multiply 6ℓµ ` 1 “ 7, 13, 19 . . . ones.
– 6 –

POLYADIC RINGS
The elements of the pm, nq-ring Zm,n “ Zra1,b1s
m,n (3.16) are abstract variables xk obeying the m-ary
addition (3.12) and n-ary multiplication (3.13) which inherit the “internal” operations (3.7)–(3.8) in the
congruence class rra1
ssb1 . Therefore, the elements of pm, nq-ring Zm,n should carry the arities as additional
indices to distinguish elements corresponding to the same representative (as decimal number) of the initial
congruence classes.
For instance, consider two infinite zeroless and unitless nonderived abstract polyadic rings
Z4,3 “ t. . . ´13, ´10, ´74,3, ´4, ´1, 2, 5, 84,3, 11, 14, 17 . . .u , (3.19)
Z6,5 “ t. . . ´22, ´17, ´12, ´76,5, ´2, 3, 86,5, 13, 18, 23, 28 . . .u , (3.20)
which are constructed from the congruence classes rr2s3s and rr3ss5, respectively. The intersection of
the generating classes for |k| ď 5 consists of two integer numbers U
r2,3s
r3,5s “ rr2s3s X rr3ss5 “ t´7, 8u
(in general, U
r2,3s
r3,5s this is an infinite set). We treat the elements of the rings (3.19)–(3.20) not as ordinary
decimal numbers, but as abstract variables (or abstract symbols, as, e.g., the letters A, B, C in the hexa-
decimal numeral system) x
r2,3s
k and x
r3,5s
k obeying (3.12)–(3.15) and carrying additional arity lower indices
(which are written manifestly in the needed cases only). In this way, we conclude that the elements corre-
sponding to the same numbers in the class intersection U
r2,3s
r3,5s are different p´74,3q ‰ p´74,3q, 84,3 ‰ 86,5,
in the sense that they obey different operations and their arities in distinct polyadic rings Z4,3 and Z6,5.
Indeed, e.g., 84,3 “ x
r2,3s
2 P Z4,3 and 86,5 “ x
r3,5s
1 P Z6,5, and their first polyadic powers (2.5) are different
p84,3qx1y
“ µ3
“
p84,3q3
‰
“ 5124,3 “ x
r2,3s
102 , while p86,5qx1y
“ µ5
“
p86,5q5
‰
“ 327686,5 “ x
r3,5s
6553.
4. POLYADIZATION OF POSITIONAL NUMERAL SYSTEMS
Let us remind the standard positional numeral system, as the presentation of a number by a special
sequence of its digits (for numerous extended versions and history, see, e.g., IFRAH [2000]).
In the manifest form, the presentation of a number over the (binary) ring of non-negative integers
Z`
“ 0, 1, 2 . . . is defined by the base (radix) p P Z`
and the digits y piq P Z`
as follows (place-value
notation)
Npℓq
ppq “
ℓ´1
ř
i“0
y piq pi
“ y pℓ ´ 1q pℓ´1
` y pℓ ´ 2q pℓ´2
` . . . ` y p1q p ` y p0q
ùñ py pℓ ´ 1q y pℓ ´ 2q . . . y p1q y p0qqpℓq
p , (4.1)
0 ď y piq ď p ´ 1, (4.2)
where the natural ℓ “ l ppq P N is the quantity of digits (function of the base p), being simultaneously
the amount of terms in the l.h.s., ℓ ´ 1 (in the binary case only) coincides with the quantity of additions
and the total amount of multiplications in the first term, and usually p0
“ 1 (to have the same summation
formula with i “ 0). Commonly, the leading zeroes in the r.h.s. of (4.1) are omitted. The change of
the base p ÞÑ p1
leads to another presentations of the same number, usually with the different quantity of
digits ℓ ÞÑ ℓ1
“ l pp1
q
Npℓ1q
pp1
q “
ℓ1´1
ř
i“0
y1
piq p1i
ùñ py1
pℓ ´ 1q y1
pℓ ´ 2q . . . y1
p1q y1
p0qq
pℓ1q
p1 , 0 ď y1
piq ď p1
´ 1. (4.3)
In the binary case, using ℓ digits one can describe E ppq “ pℓ
numbers: 0, . . . pℓ
´1 P Z`
. The efficiency
of a numeral system is the ability to represent as many numbers as possible using the smallest total number
of symbols s. In this case, the number of digits becomes s{p, and the quantity of the described numbers
– 7 –

POLYADIZATION OF POSITIONAL NUMERAL SYSTEMS
(the efficiency function) is
E ppq “ p
s
p . (4.4)
The function E ppq reaches its maximum (in p), when p “ e « 2.718... (the Euler’s number), and
therefore the most efficient numeral system (with integer p) is pmax “ 3.
For further details about positional numeral systems (over the binary ring Z`
), see, e.g., IFRAH [2000],
and refs. therein.
Now we generalize the above construction to the polyadic integer numbers DUPLIJ [2017, 2019], that
is we build an analog of positional numeral system over polyadic pm, nq-rings Zm,n considered in the pre-
vious section. The polyadic analog of binary numeral systems have complicated and nontrivial structure,
because the underlying pm, nq-rings obey unusual peculiarities, for instance, some of them can not have
zero or/and unity, absence of the natural ordering, etc. (see, for details, DUPLIJ [2022]).
First, we express the binary positional numeral system (presentation (4.1)) in the polyadic notation
from SECTION 2. Indeed, let the binary ring of integers is Z “ xν, µ | comm | assocy, where ν and
µ are the ordinary binary addition ν rx, ys “ x ` y and multiplication µ rx, ys “ y ¨ y, x, y P Z. In
this notation, each term in (4.1) can be written as the composition of i ` 1 multiplications (see (2.2))
µ˝pi`1q
«
y piq ,
i
hkkikkj
p, . . . , p
ff
” µ˝pi`1q
ry piq , pi
s, the first term becomes µ˝pℓµ´1q
“
y pℓµ ´ 1q , pℓµ´1
‰
, and the
sum as the composition of ℓν additions, such that the place-value presentation takes the form of operation
compositions
Npℓν ,ℓµq
ppq “ ν˝ℓν
»
—
–
ℓν `1
hkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkj
µ˝pℓµ´1q
“
‰
, µ˝pℓµ´1q
“
‰
, . . . , µ ry p1q , ps , y p0q
fi
ffi
fl
ùñ
¨
˝
ℓν `1
hkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkj
y pℓµ ´ 1q y pℓµ ´ 2q . . . y p1q y p0q
˛
‚
pℓv,ℓµq
p
, 0 ď y piq ď p ´ 1 P Z`
, (4.5)
where the important additional consistency condition
ℓµ “ ℓν ` 1 “ ℓ, (4.6)
which follows from the construction itself. Note that in the composition form we do not write multiplier
for the last degit y p0q at all for consistency with the higher arity generalizations (not all multiplicative
parts of polyadic rings contain unity, see SECTION 3, however we need the last digit in any case).
For example, in the case p “ 7 and ℓ “ ℓµ “ ℓν `1 “ 3 we have the ordinary binary positional numeral
presentation in the polyadic notation
p165qp2,2q
7 “ ν˝2
“
µ˝2
r1, 7, 7s , µ r6, 7s , 5
‰
“ 1 ¨ 7 ¨ 7 ` 6 ¨ 7 ` 5 “ p96qp1,1q
10 . (4.7)
Now the polyadization of the binary place-value presentation in the composition form (4.5) can be done
in the straightforward way. We assume that all ordinary numbers become polyadic numbers, that is the
initial number ring should be exchange to the polyadic pm, nq-ring: Z “ xν, µy ÝÑ Zm,n “ xνm, µny.
Thus, modifying (4.5) consistently by ν Ñ νm and µ Ñ µn, taking into account the admissible length of
words (3.17)-(3.18), we propose the direct polyadic generalization of the standard place-value presentation
– 8 –

(4.1) of a polyadic number N P Zm,n by the polyadic digits y piq P Zm,n and the polyadic base p P Zm,n
Npℓν ,ℓµq
m,n ppq “ ν˝ℓν
m r
ℓν pm´1q`1
hkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkj
µ˝pℓµ´1q
n
“
y pℓµ ´ 1q , pℓµpn´1q
‰
, µ˝pℓµ´1q
n
“
y pℓµ ´ 2q , pℓµpn´1q´1
‰
, . . . , µn
“
y p1q , pn´1
‰
, y p0q
fi
ffi
fl
ùñ
¨
˚
˝
ℓν pm´1q`1
hkkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkkj
˛
‹
‚
pℓv,ℓµq
m,n;p
, (4.8)
where the arity indices m, n, when they are obvious, can be omitted for conciseness. The polyadic analog
of the consistency condition (4.6) now is
ℓµ “ ℓν pm ´ 1q ` 1, (4.9)
and therefore the minimal quantity of multiplications in (4.8) is ℓµ ě m, which coincides with minimal
amount of digits in the place-value presentation over pm, nq-ring.
Theorem 4.1. In the polyadic numeral system over pm, nq-ring the minimal number of digits is more or
equal than the arity of addition m.
The m-ary addition of polyadic numbers N in the place-value presentation (4.8) can be done using the
total polyadic distributivity (3.1)–(3.3) in the general form by adding digits (as in the binary case, prime
sign is not a derivative, but numerates variables)
νm
”
N1pℓν ,ℓµq
m,n ppq , N2pℓν ,ℓµq
m,n ppq , . . . , N3pℓν ,ℓµq
m,n ppq
ı
“ Npℓν ,ℓµq
m,n ppq , (4.10)
νm ry1
piq , y2
piq , . . . , y3
piqs “ y piq , i “ 0, . . . , ℓµ. (4.11)
The n-ary multiplication of polyadic numbers N is more complicated and should be made in each
concrete case manifestly.
In general, the direct polyadization formula (4.8) for place-value presentation can be considered for any
commutative polyadic pm, nq-ring.
Here we will study some examples for polyadic rings of integer numbers from the previous SECTION 3.
Let us consider the abstract polyadic number ring Zr2,3s
4,3 (3.19) generated by the congruence class rr2ss3.
The polyadic integer numbers x
r2,3s
k are “symmetric” with respect to x
r2,3s
k“0 “ a “ 2 (playing the role of
zero in the binary case a “ 0, b “ 1 and xk “ k), therefore instead of (3.6), we use
0 ď kypiq ď kp ´ 1, (4.12)
where y piq “ xkypiq
, p “ xkp , and xk “ x
r2,3s
k P Xr2,3s
. In this way, for the simplest polyadic base
p “ x2 “ 8, ℓν “ 1 and ℓµ “ 4 (see (4.9)) we have the 4-digit (y piq “ 2, 5, i “ 0, 1, 2, 3) polyadic
numeral presentation for a number from Zr2,3s
4,3 as
N
p1,3q
4,3 p8q “ ν4
“
µ˝3
3 ry p3q , 8, 8, 8, 8, 8, 8s , µ˝2
3 ry p2q , 8, 8, 8, 8s , µ3 ry p1q , 8, 8s , y p0q
‰
ùñ py p3q y p2q y p1q y p0qqp1,3q
4,3;8 ” py p3q y p2q y p1q y p0qq8 . (4.13)
– 9 –

The polyadic numerals (polyadic numbers of the pm, nq-ring Zr2,3s
4,3 that can be presented in base-8
place-value form by 4 digits) are
p2, 2, 2, 2q8 “ 532610 “ x177536, p2, 2, 2, 5q8 “ 532613 “ x177537, p2, 2, 5, 2q8 “ 532802 “ x177600,
p2, 2, 5, 5q8 “ 532805 “ x177601, p2, 5, 2, 2q8 “ 544898 “ x181632, p2, 5, 2, 5q8 “ 544901 “ x181633,
p2, 5, 5, 2q8 “ 545090 “ x181696, p2, 5, 5, 5q8 “ 545093 “ x181697, p5, 2, 2, 2q8 “ 1319042 “ x439680,
p5, 2, 2, 5q8 “ 1319045 “ x439681, p5, 2, 5, 2q8 “ 1319234 “ x439744, p5, 2, 5, 5q8 “ 1319237 “ x439745,
p5, 5, 2, 2q8 “ 1331330 “ x443776, p5, 5, 2, 5q8 “ 1331333 “ x443777, p5, 5, 5, 2q8 “ 1331522 “ x443840,
p5, 5, 5, 5q8 “ 1331525 “ x443841, (4.14)
which correspond to the ordinary base-2 (or binary) numerals with two digits y piq “ 0, 1.
The more complicated case is the 7-digit (y piq “ 2, 5, 8, 11, i “ 0, . . . , 7) polyadic numeral presenta-
tion for numbers from Zr2,3s
4,3 with the polyadic base (see (3.19)) p “ x4 “ 14, composition of two m-ary
additions ℓν “ 2 and seven n-ary multiplications ℓµ “ 7 (see (4.9))
N
p1,3q
4,3 p14q “ ν˝2
4
“
µ˝6
3
“
y p6q , 1412
‰
, µ˝5
3
“
y p5q , 1410
‰
, µ˝4
3
“
y p4q , 148
‰
, µ˝3
3
“
y p3q , 146
‰
,
µ˝2
3 ry p2q , 14, 14, 14, 14s , µ3 ry p1q , 14, 14s , y p0q
‰
ùñ py p6q y p5q y p4q y p3q y p2q y p1q y p0qqp1,3q
4,3;14 ” py p6q y p5q y p4q y p3q y p2q y p1q y p0qq14 .
(4.15)
The first several numerals in (4.15) are
p2, 2, 2, 2, 2, 2, 2q14 “ 113 969 300 774 954,
p2, 2, 2, 2, 2, 2, 11q14 “ 113 969 300 774 963,
p2, 2, 2, 2, 2, 5, 2q14 “ 113 969 300 775 542,
p2, 2, 2, 2, 2, 8, 2q14 “ 113 969 300 776 130,
p2, 2, 2, 2, 2, 11, 11q14 “ 113 969 300 776 727,
p2, 5, 2, 2, 2, 2, 2q14 “ 114 837 064 739 882,
p2, 11, 2, 2, 2, 2, 2q14 “ 116 572 592 669 738,
p11, 2, 2, 2, 2, 2, 2q14 “ 624 214 512 152 618,
p11, 11, 11, 11, 11, 11, 11q14 “ 626 831 154 262 247. (4.16)
It follows from the examples (4.14) and (4.16), that not all polyadic integer numbers N can be repre-
sented by the polyadic numeral formula in the place-value form (4.8). We call such numbers numerally
representable tNrepsu Ď tNu.
Theorem 4.2. The set of polyadic numbers which are representable in the numeral form is a subset of the
set of all numbers, while the equality is reached in the binary case only Nreps “ N ðñ Nreps, N P
Zr0,1s
2,2 ” Z. In the polyadic nonbinary case m ě 3 _ n ě 3, the set of representable numbers is a proper
(strict) subset of all numbers tNrepsu Ă tNu.
Now we introduce the polyadic analog of the mixed-base (radix) positional numeral system. In the
polyadic notation, the mixed-base version of the binary (4.5) can be written as the sum as the composition
of ℓν additions having different bases p piq for each digit y piq, and we assume that for the last digit y p0q
the basis multiplier is absent (see note after (4.6)). In such assumption the connection between the quantity
– 10 –

of digits Ny and the quantity of bases Np is
Np “ Ny ´ 1. (4.17)
Then the place-value presentation takes the form of operation compositions
Npℓν ,ℓµq
ppq “ ν˝ℓν
r
ℓν `1
hkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkj
µ˝pℓµ´1q
ry pℓµ ´ 1q , p pℓµ ´ 1qs , µ˝pℓµ´1q
ry pℓµ ´ 2q , p pℓµ ´ 2qs , . . . , µ ry p1q , p p1qs , y p0q
fi
ffi
fl
ùñ
¨
˝
ℓν `1
hkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkj
˛
‚
pℓv,ℓµq
ppℓµ´1q,ppℓµ´2q...pp1q
, 0 ď y piq ď p piq ´ 1 P Z`
.
(4.18)
The simplest binary example of the mixed-base numeral system is computation of currency in different
banknote numbers y piq of the denomination p piq. Sometimes the bases are composed and connected
between them taking the “recurrent” form
p pi ` 1q “ p piq p pi ´ 1q . . . p p1q . (4.19)
A binary example of mixed-base recurrent numeral system is the time calculation in seconds
(“clock”/“timer”) using number of days y pdaysq, hours y phoursq, minutes y pminq, seconds y psecondsq,
and the lengths p pdaysq “ 24, p phoursq “ 60, p pminutesq “ 60. E.g. the polyadic/composition form
(4.18) of “2 days, 7 hours, 35 minutes, 48 seconds” (having m “ n “ 2, ℓν “ 2, ℓµ “ 4), is (in seconds)
ν˝2
“
µ˝3
r2, 24, 60, 60s , µ˝2
r7, 60, 60s , µ r35, 60s , 48
‰
“ 2¨24¨60¨60`7¨60¨60`35¨60`48 “ 200 148.
(4.20)
In general, the polyadic analog of the mixed-base (radix) positional numeral system can be defined by
the formula similar to (4.8), but the base in each term in the m-ary sum will depend of its place p ÝÑ p piq,
that is
Npℓν ,ℓµq
m,n pp p1q , . . . p pℓµ ´ 1qq “ ν˝ℓν
m r
µ˝pℓµ´1q
n
”
y pℓµ ´ 1q , p pℓµ ´ 1qℓµpn´1q
ı
, µ˝pℓµ´2q
n
”
y pℓµ ´ 2q , p pℓµ ´ 2qℓµpn´1q´1
ı
, . . .
. . . , µn
“
y p1q , p p1qn´1
‰
, y p0q
‰
ùñ
¨
˚
˝
ℓν pm´1q`1
hkkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkkj
˛
‹
‚
pℓv,ℓµq
m,n;pp1q,...ppℓµ´1q
, (4.21)
p p1q , . . . p pℓµ ´ 1q P Zm,n.
Here the connection between the quantity of digits Ny and the quantity of bases Np is the same as in
the binary case (4.17). Nevertheless, we can extend each ith term using in it different bases (which now
have two indices pj piq P Zm,n) as
µ˝pi´1q
n
“
y pi ´ 1q , p1 pi ´ 1q , p2 pi ´ 1q , . . . , pipn´1q pi ´ 1q
‰
. (4.22)
Summing up in (4.21), instead of (4.17), for the quantity of polyadic bases pj piq through the number
of digits Ny we will have
Np “
1
2
Ny pNy ´ 1q pn ´ 1q . (4.23)
– 11 –

For example, in the case ℓν “ 1 and ℓµ “ 4 (see (4.9)) we have the 4-digit (y piq, i “ 0, 1, 2, 3) polyadic
mixed-base numeral presentation for a number from the p4, 3q-ring Zr2,3s
4,3 as
N
p1,3q
4,3
`
pj piq
˘
“ ν4
“
µ˝3
3 ry p3q , p1 p3q , p2 p3q , p3 p3q , p4 p3q , p5 p3q , p6 p3qs ,
µ˝2
3 ry p2q , p1 p2q p2 p2q p3 p2q p4 p2qs , µ3 ry p1q , p1 p1q , p2 p1qs , y p0q
‰
ùñ py p3q y p2q y p1q y p0qq
p1,3q
4,3;pjpiq ” py p3q y p2q y p1q y p0qqpjpiq . (4.24)
Thus, the mixed-base polyadic 4-digit numerals over p4, 3q-ring with ℓν “ 1 and ℓµ “ 4 are described by
12 different polyadic bases pj piq P Zr2,3s
4,3 . Comparing with the “binary clock” (4.18)–(4.20), we can call
the presented construction (4.21)–(4.24) as the “polyadic clock”.
5. CONCLUSIONS
In this paper we showed that by transplanting the entire positional-numeral paradigm into the realm of
pm, nq-rings we uncovered a landscape where digit strings, carries, and even the very concept of “base”
behave in ways impossible in the ordinary arithmetic. Let us distil the main messages.
1. Polyadic rings admit a native positional calculus presented here. Replacing plus by an m-ary sum
and product by an n-ary multiplication forces a “double quantization” of word length.
2. Minimal-digit and representability theorems. We proved that in the positional expansion every
numeral must respect the constraint ℓmult “ ℓadd pm ´ 1q ` 1, and so the shortest admissible numeral
contains more positions (digits) than the additive arity ℓmult ě m (Theorem 4.1), and that for m, n ě 3
not every element of a polyadic ring possesses a finite positional expansion (Theorem 4.2). The gap is
governed by the arity-shape invariants Ipmq
and Jpnq
originating in the polyadic features of congruence-
class geometry.
3. Mixed bases are generalized naturally yielding a “polyadic clock”. Allowing each digit to sit atop
its own n-ary tower of bases replaces the linear count Np “ Ny ´ 1 by the quadratic formula
Np “
1
2
NypNy ´ 1qpn ´ 1q,
exploding the design space of numeral systems. Our exotic polyadic clock over p4, 3q-ring, specified by
twelve independent bases, hints at rich cryptographic and coding-theoretic applications.
4. Concrete catalogues. The worked-out examples in the p4, 3q and p6, 5q integer rings illustrate how or-
dinary integers lift to abstract polyadic variables that share a decimal ”shadow” yet obey distinct algebraic
laws. These catalogues provide ready test-beds for algorithmic experimentation.
Future directions. The present framework raises more questions than it settles:
Algorithms and complexity. What is the cost of addition, “carry” propagation, and comparison in a
polyadic positional machine? Can we find arities for which arithmetic accelerates relative to binary RAM
models?
Floating-point and analysis. A polyadic analogue of IEEE-754 would require rounding rules compatible
with m-ary adders and n-ary multipliers. How do error bounds scale with the arity pair pm, nq?
Hardware realization. Ternary logic became tangible with modern CMOS; could polyadic ALUs (arith-
metic logic units) unlock energy or area advantages?
Category-theoretic reformulation. Viewing pm, nq-rings as N2
-graded monoidal categories may clarify
the functorial behavior of numeral expansions.
Quantum and cryptographic angles. The inherent nonassociativity of many polyadic rings could serve
as a source of hard problems, while the enlarged digit alphabet suggests fresh qudit encodings.
– 12 –

Acknowledgments. The author would like to express his deep thankfulness to Vladimir Tkach for
productive discussions and to Qiang Guo and Raimund Vogl for valuable support.
REFERENCES
COHN, P. M. (1965). Universal Algebra. New York: Harper & Row. [Cited on page 3]
CONWAY, J. H. AND R. K. GUY (1996). The Book of Numbers. New York: Springer. [Cited on page 2]
DÖRNTE, W. (1929). Unterschungen über einen verallgemeinerten Gruppenbegriff. Math. Z. 29, 1–19. [Cited on page 4]
DUPLIJ, S. (2017). Polyadic integer numbers and finite pm, nq-fields. p-Adic Numbers, Ultrametric Analysis and Appl. 9
(4), 257–281. arXiv:math.RA/1707.00719. [Cited on pages 2, 5, and 8]
DUPLIJ, S. (2018). Polyadic algebraic structures and their representations. In: S. DUPLIJ (Ed.), Exotic Algebraic and
Geometric Structures in Theoretical Physics, New York: Nova Publishers, pp. 251–308. arXiv:math.RT/1308.4060. [Cited on
page 3]
DUPLIJ, S. (2019). Arity shape of polyadic algebraic structures. J. Math. Physics, Analysis, Geometry 15 (1), 3–56. [Cited on
pages 5, 6, and 8]
DUPLIJ, S. (2022). Polyadic Algebraic Structures. London-Bristol: IOP Publishing. [Cited on pages 2, 3, 4, and 8]
FOMIN, S. V. (1974). Number Systems. Chicago: Chicago Univ. Press. [Cited on page 2]
IFRAH, G. (2000). The Universal History of Numbers. New York: John Wiley & Sons. [Cited on pages 2, 7, and 8]
KUROSH, A. G. (1963). Lectures on general algebra. New York: Chelsea Pub. [Cited on page 3]
LEESON, J. J. AND A. T. BUTSON (1980). On the general theory of pm, nq rings. Algebra Univers. 11, 42–76. [Cited on
pages 2 and 4]
MENNINGER, K. W. AND P. BRONEER (1971). Number Words and Number Symbols. Cambridge: MIT Press. [Cited on
page 2]
– 13 –

S.Duplij, "Positional numeral systems over polyadic rings", arxiv: 2506.12930

More Related Content

Similar to S.Duplij, "Positional numeral systems over polyadic rings", arxiv: 2506.12930 (20)

More from Steven Duplij (Stepan Douplii) (14)

Recently uploaded (20)

S.Duplij, "Positional numeral systems over polyadic rings", arxiv: 2506.12930