See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/374583495

RADICALS OF PARAGRADED RINGS by M. VUKOVIĆ

Article in Journal of Mathematical Sciences · October 2023

CITATIONS READS
0 10

1 author:

Mirjana Vukovic
Academy of Science and Arts of Bosnia and Herzegovina
188 PUBLICATIONS 160 CITATIONS

SEE PROFILE

All content following this page was uploaded by Mirjana Vukovic on 10 October 2023.

The user has requested enhancement of the downloaded file.

DOI 10.1007/s10958-023-06689-3
Journal of Mathematical Sciences, Vol. 275, No. 4, September, 2023

RADICALS OF PARAGRADED RINGS

M. Vuković UDC 512.552.12

Abstract. This paper is concerned with the theory of paragraded rings, which begins with a series of
Krasner and Vuković’s notes in Proceedings of the Japan Academy, which first appeared in late 1980s. We
present prime and Jacobson radicals, discuss the general Kurosh–Amitsur theory of radicals of paragraded
rings, establish that the theorem of Anderson, Divinsky, and Suliński holds for paragraded rings, charac-
terize paragraded normal radicals, and prove that all special paragraded radicals of paragraded rings can
be described by appropriate classes of their graded modules.

I wholeheartedly dedicate this paper

to one of the greatest Lomonosov algebraists
A. V. Mikhalev in honor of his 80th birthday

1. Introduction
In this paper, we are going to present some results on the different types of radicals introduced and
studied in joint papers with my student E. Ilić-Georgijević [13, 14] (see also [15, 29]) in the class of the
paragraded rings, the most general graded rings introduced by M. Krasner and myself [20–23].
It is worth noting that abstract radical theory has its origin in the theory of rings going back to the
1940s, as well as that in the theory of rings many structure results were obtained with the use of radicals.
In generalizing the classical notion of the radical in a ring, different kinds of radicals have been defined
by a number of authors including S. A. Amitsur, G. Azumaya, R. Baer, B. Brown, N. Jacobson, G. Köthe,
J. Levitzki, and N. H. McCoy, but a significant contribution was made by the entire constellation of Russian
and USSR mathematicians as well A. G. Kurosh, A. I. Mal’cev, V. A. Andrunakievich, Yu. M. Ryabukhin,
A. I. Shirshov, L. A. Skornjakov, A. I. Kostrikin, A. V. Mikhalev, and their students.
It should be emphasized that an essential role in developing theory of radicals was played by Kurosh
and Amitsur, who, between 1952 and 1954, independently introduced the general theory of radicals in
various algebraic structures, and Andrunakievich and Ryabukhin, who introduced the special theory of
radicals. In addition, they have shown that every Kurosh–Amitsur radical for associative rings can be
described in the language of modules in much the same way as the Jacobson radical.
Although the genesis of abstract radical theory begins with rings, and other algebraic structure
(groups, algebras) mentioned only in the fundamental papers of Kurosh and Amitsur in the 1950s, an
examination of the mathematical literature of the USSR reveals that group theory exerted a major
influence on the development of the abstract radical concept. Namely, Kurosh and Amitsur observed that
the general theory of radicals can be developed in any algebraic system in which the concept of the kernel
homomorphism with usual properties takes place, i.e., in the “good” categories.
Similarly to the case of graded rings, an abstract concept of radical class for paragraded rings has
been introduced and investigated in [13, 14]. In particular, in this paper we present some of our results
from the radical theory of paragraded rings beginning from the Baer and Jacobson radicals through
general Kurosh–Amitsur theory of radicals to normal radicals and Andrunakievich–Ryabukhin special
radicals [13–15, 29].
Now, we will briefly say something about the content of this paper.

Translated from Fundamentalnaya i Prikladnaya Matematika, Vol. 24, No. 2, pp. 3–22, 2022.

1072–3374/23/2754–0379 
c 2023 Springer Nature Switzerland AG 379
 The first section begins with an introduction to the subject. In the second section, we recall some
notions and basic facts about paragraded rings and paragraded modules [20–23].
In Secs. 3 and 4, we examine properties of some concrete radicals: the Baer radical and the Jacobson
radical of a paragraded ring and their paragraded versions. In particular, in Sec. 3, we prove that
the paragraded Baer radical of a paragraded ring coincides with the largest homogeneous ideal that is
contained in the classical Baer radical of a paragraded ring. Similar results are obtained in Sec. 4, for the
Jacobson radical, but following Halberstadt’s results on graded rings [11], we establish a relation between
the paragraded Jacobson radical and the Jacobson radical of a ring that corresponds to an idempotent,
under assumptions that the underlying paragraded ring is regular and that each element of its set of
grades is either idempotent or nilpotent of degree two with respect to minimal multiplication.
In Sec. 5, we study the general radical theory of the category of paragraded rings. Further, we discuss
the general Kurosh–Amitsur theory of radicals of paragraded rings and establish that the theorem of
Anderson, Divinsky, and Suliński for associative paragraded rings holds.
Note that all general concepts and statements of a Kurosh–Amitsur radical theory of paragraded
rings are defined and formulated by the aid of quasihomogeneous homomorphisms and ideals [23].
In Sec. 6, we define special paragraded radicals of paragraded rings. It is known that any special
radical of a ring can be defined by means of some class of modules over that ring. Namely, we study
special paragraded radicals of paragraded rings. Our aim is to show that all special paragraded radicals of
paragraded rings can be described by an appropriate class of their paragraded modules, which is already
done for graded rings [4].

2. Preliminaries
This section is devoted to the paragraded rings and paragraded modules. As very little is known
about them, the beginning is historical in nature, offering a short outline of Krasner’s general graded
rings [11] and modules, which lead to the Krasner–Vuković’s paragraded rings [21, 23].
Krasner’s investigating of valued fields and observing a connection between them and their valuation
rings using equivalence of valuations came to the abstract notion of a corpoid [18], which represents the
origin of the development of the general graded theory [7, 11, 19], a theory of homogroupoids, anneids
and moduloids, more general than Bourbaki’s [5], since neither the associativity, nor the commutativity
nor the existence of a neutral element is assumed in the set of grades, where corpoid, as a special case of
an anneid, is viewed as a homogeneous part of a graded field. Besides, Krasner’s graded structures were
studied from three different points of view: non-homogeneous, semi-homogeneous, and homogeneous.
Since the category of graded structures (groups, rings, modules) has no property of closure with respect
to the direct product and the direct sum, and the direct product and the direct sum of homogroupoids,
anneids, and moduloids does not have to be a homogroupoid, anneid, and moduloid, respectively, it was
motivation for Krasner and myself to focus on this problem and we were the first who succeeded to solve
it. In this way, we discovered a theory of paragraded structures (groups, rings, modules) [20–23], which
generalize, not only, the theory of graded structures as it is exposed in Bourbaki [5], but also the previous
results of Krasner [19] and his students M. Chadeyras [7] and E. Halberstadt [11].
In this section, we are particularly interested in the structure of paragraded rings and paragraded
modules, which appeared for the first time, in the late 1980s, in Krasner–Vuković’s three notes [20–22]
and monograph [23].
The connection between graded and paragraded rings can be found in [23, p. 70, 71].
Let us recall the notions of a paragraded ring and module [20–23].
Definition 2.1. A ring R is called paragraded if there exists a mapping
π : Δ → Sg(R, +), π(δ) = Rδ (δ ∈ Δ),
of a partially ordered set (Δ, <) that is a semi-lattice that is from below complete and from above
inductively ordered to the set Sg(R, +) of subgroups of (R, +), called a paragraduation, and the following
axioms are satisfied.

380
 (r1) π(0) = R0 = {0}, where 0 = inf Δ, δ < δ  implies Rδ ⊆ Rδ .

Remark 2.1. A = h(R) = Rδ is called the homogeneous part of R with respect to π, and elements
δ∈Δ
of A are called homogeneous elements.
Remark 2.2. If x ∈ A, then we say that δ(x) = inf{δ ∈ Δ | x ∈ Rδ } is the grade or degree of x. We
have δ(x) = 0 if and only if x = 0. Elements δ(x), x ∈ A, are called principal degrees, and they form a set
which is denoted by Δp .

(r2) θ ⊆ Δ implies Rδ = Rinf θ .
δ∈θ
(r3) The homogeneous part A is a generating set of (R, +) with the set of A-inner relations x + y = z.
(r4) Let B ⊆ A be a subset such that, for all x, y ∈ B, there exists an upper bound for δ(x), δ(y).
Then there exists an upper bound for all δ(x), x ∈ B.
(r5) For all ξ, η ∈ Δ, there exists ζ ∈ Δ such that Rξ Rη ⊆ Rζ .
From this definition, a binary operation on Δ
ξη = sup{δ(x) | x ∈ Rξ Rη } (ξ, η, ζ ∈ Δ)
called the minimal multiplication [22, 23] is obtained, and so, Rξ Rη ⊆ Rξη .
Definition 2.2. An ideal I of a paragraded ring R is called homogeneous if I is generated by I ∩ A by
A-inner relations, where A is the homogeneous part of R.
We discuss prime and Jacobson radicals of paragraded rings and as main results we obtain that
paragraded counterparts of prime and Jacobson radicals are the largest paragraded ideals contained in
them. The notion of the Jacobson radical, introduced in [14], is more general than that from [15] where
we considered the category of paragraded rings with the same set of grades.
In this paper, we will use the notions of a quasianneid [23], a paragraded module, and of a quasimod-
uloid [23], which we recall in the sequel.
If R is a paragraded ring with homogeneous part A, then we may observe restrictions of operations
from R to A. The induced addition is partial and we write x # y if and only if x + y ∈ A. The obtained
structure is called a paraanneid [23]. If x ∈ A, let A(x) = {y ∈ A | x # y}.
A paraanneid certainly satisfies the following axioms.
(a1) There exists an element 0 ∈ A such that A = A(0) and for all x ∈ A, we have 0 + x = x.
(a2) If a ∈ A, x + y is always defined on g(a) = {x ∈ A | A(x) ⊇ A(a)} and (g(a), +) is an Abelian
group.
(a3) If B ⊆ A is such that for all x, y ∈ B we have x # y, then there exists G ⊆ A such that x + y ∈ G
for all x, y ∈ G; x ∈ G implies g(x) ⊆ G and B ⊆ G.
(a4) A2 ⊆ A.
(a5) x # x and y # y  imply xy # x y  .
A structure (A, +, ·) that satisfies axioms (a1)–(a5) is called a quasianneid [23]. A quasianneid does
not have to be a paraanneid, but it is under a few more assumptions (see [23]), in which case A can be
linearized [23] to a paragraded ring, denoted by Ā, whose homogeneous part it is.
First, let us say what is the linearization Ā.
Let A be a paraanneid, i.e., the homogeneous part of some ring R. If A is also the homogeneous part
of some other ring R , the rings R and R are A-isomorphic, i.e., a ring with the homogeneous part A is
determined (up to an A-isomorphism), and when it occurs will be denoted by Ā and called a linearization
of the paraanneid A.
If R is a paragraded ring with paragraduation π : Δ  δ → Rδ and M a right R-module, then M is
called a paragraded R-module [22, 23] if (M, 
 +) satisfies axioms (r1)–(r4) for π instead of π, D instead
of Δ, M instead of R, and N = h(M ) = Md instead of A = h(R), and if, moreover, for all d ∈ D and
d∈D
δ ∈ Δ, there exists t ∈ D such that π  (d)Rδ ⊆ π  (t).

381
 If we observe restrictions of addition from M to N and of external multiplication M × R → M to
N ×A → N , then we obtain a structure called a paramoduloid [23]. A paramoduloid N over a paraanneid A
certainly satisfies the following axioms:
(m1) x(ab) = (xa)b (a, b ∈ A, x ∈ N );
(m2) if a, a ∈ A and x, x ∈ N are such that a # a and x # x , then xa # x a ;
(m3) if a # a (a, a ∈ A) and x ∈ N , then x(a + a ) = xa + xa ;
(m4) if x # x (x, x ∈ N ) and a ∈ A, then (x + x )a = xa + x a.
A quasimoduloid N over a quasianneid A is a structure that satisfies axioms (m1)–(m4). It does not
have to be a paramoduloid. It will be a paramoduloid under a few more assumptions (see [22, 23]), in
which case N can be linearized to a paragraded module, which we denote by N̄ , whose homogeneous part
it is.
If N is an A-quasimoduloid, then K ⊆ N is called a subquasimoduloid [23] if
(1) x ∈ K implies −x ∈ K;
(2) x, y ∈ K and x # y imply x + y ∈ K;
(3) a ∈ A and x ∈ K imply xa ∈ K.
A subquasimoduloid of a quasianneid A, observed as an A-quasimoduloid, is called a right ideal of
a quasianneid A [23].
Factor structures are defined as usual (for more details, see [23, 29]).

3. On the Prime Radical

The prime spectrum of a paragraded ring is introduced in [12]. Here we reformulate this definition
for quasianneids.
Definition 3.1. Let A be a quasianneid. A proper ideal P of A is said to be prime if for all ideals I, J
of A, IJ ⊆ P implies I ⊆ P or J ⊆ P .
As in the case of rings, one may prove that an ideal P of A is prime if and only if xAy ⊆ P implies
x ∈ P or y ∈ P , where x, y ∈ A.
The set of all prime ideals of A is denoted by Spec(A) and it is called the prime spectrum of a quasian-
neid A.  
The prime radical or the Baer radical of a quasianneid A is defined to be the set P | P ∈ Spec(A)
and is denoted by β(A).
Proposition 3.1. Let A be a paraanneid, that is, let A be linearizable to a paragraded ring Ā.
(a) If P̄ ∈ Spec(Ā), then P̄ ∩ A ∈ Spec(A), where Spec(Ā) denotes the spectrum of Ā regarded as
a ring.
(b) If Q ∈ Spec(A), then there exists P̄ ∈ Spec(Ā) such that Q = P̄ ∩ A.
(c) Prime radical of A coincides with β(Ā) ∩ A, where β(Ā) denotes the prime radical of Ā regarded
as a ring.
Proof. Statements (a) and (b) can be proved by the same means as in the case of ordinary graded rings
(see [8] or [25]), while (c) follows directly from (a) and (b).
Let us now observe a paragraded ring R with the set of grades Δ. Inspired by the well-known notion
of an u.p.-semigroup (unique-product semigroups, see [17]), a set of grades with respect to minimal
multiplication will be called uniquely presented if, for any two nonempty subsets Θ and Λ of Δ, one of
which is not a singleton, there exists an element uniquely presented as θλ (minimal multiplication), where
θ ∈ Θ, λ ∈ Λ.
Theorem 3.1. The prime radical β(R) of a paragraded ring R, with uniquely presented set of grades Δ,
is the homogeneous part of R.

382
Proof. Following the proof of Lemma 1.1 from [17], we, first, prove that if I = 0 is a prime ideal of R and
if R is not prime, the I contains a nonzero homogeneous ideal. Indeed, let a and b be nonzero elements
of R such that aRb = 0. Then there exist elements aξ ∈ Rξi , bηj ∈ Rηj , i = 1, . . . , n, j = 1, . . . , m, such
that
a = aξ1 + . . . + aξn , b = nη1 + . . . + aηm .
These representations, since R is a paragraded ring, do not have to be unique, but we may assume
that they are of minimal lengths, i.e., that n and m are minimal. With that in mind, we may also
chose elements a and b of minimal lengths such that aRb = 0, and let us assume that this is the case.
Since Δ is uniquely presented, for every δ ∈ Δ, there exists a uniquely presented element ξi δη j of the
set {ξ1 δ, . . . , ξn δ} · {η1 , . . . , ηm }, with respect to the minimal multiplication. Now, as in the proof of
Lemma 1.1 from [17], we arrive to aξ1 Rbη1 = 0, which implies aξ1 ∈ I or bη1 ∈ I, whence, I contains
a homogeneous ideal.
We immediately have that any minimal prime ideal of R is homogeneous, and so, the claim follows.

4. On the Jacobson Radical

The Jacobson radical of an anneid, the homogeneous part of a graded ring with induced operations,
throughly examined E. Halberstadt in [11].
In this section, we will see that the similar theory may be established for quasianneids under some
appropriate modifications in formulations and in proofs of relevant results.
For convenience, quasimoduloids are here denoted by M .
Let A be a quasianneid, M a right A-quasimoduloid, N a subquasimoduloid of M , and S a subset of
M . If (N : S) denotes the set of elements a ∈ A such that Sa ⊆ N , then it can be easily proved that
(N : S) is a right ideal of A. Particularly, if x ∈ M , then (0 : x) is a right ideal of A.
Definition 4.1. The heart of an A-quasimoduloid M is defined to be the set
C = {x ∈ M | | (∀y ∈ M ) x # y}.
Definition 4.2. An A-quasimoduloid M is called regular if, for every a, b ∈ A and x ∈ M , xa, xb ∈ / C,
where C is the heart of M , and xa # xb, imply a # b.
A quasianneid A is called right (left) regular if it is regular as a right (left) A-quasimoduloid.
Definition 4.3. If M and M  are two A-quasimoduloids, then the mapping f : M → M  is called a quasi-
homomorphism if for all x, y ∈ M and a ∈ A
(a) x # y implies f (x) # f (y) and f (x + y) = f (x) + f (y);
(b) f (xa) = f (x)a.
/ C  imply x # y, where C  denotes the heart of M  , then f is
If, moreover, f (x) # f (y) and f (x), f (y) ∈
called a homomorphism.
Lemma 4.1. If M is a regular A-quasimoduloid and 0 = x ∈ M , then A/0 : x) ∼
= xA.
Proof. Let f : A → xA be the mapping defined by f (a) = xa (a ∈ A). If a, b ∈ A and a # b, then
xa # xb and x(a + b) = xa + xb, so f (a + b) = f (a) + f (b). Also, f (ab) = x(ab) = (xa)b = f (a)b.
Hence, f is a quasihomomorphism of A-quasimoduloids. It is a homomorphism, since, if f (a) # f (b) and
f (a), f (b) ∈
/ C, then xa # xb, xa, xb ∈
/ C, and since M is regular, we have a # b.
The kernel of f is a right ideal of A, namely, (0 : x), while its image is obviously xA. According to
the first isomorphism theorem [23], A/(0 : x) ∼ = xA.
Definition 4.4. An A-quasimoduloid M is called irreducible if M A = 0 and if 0 and M are its only
subquasimoduloids.
Definition 4.5. The Jacobson radical of a quasianneid (paraanneid) A, denoted by J(A), is defined to
be the intersection of annihilators of all irreducible regular A-quasimoduloids (A-paramoduloids).

383
Definition 4.6. The intersection of annihilators of all irreducible A-quasimoduloids (A-paramoduloids)
is called the large Jacobson radical of a quasianneid (paraanneid) A, and we denote it by Jl (A).
Remark 4.1. It is clear that the large Jacobson radical of a quasianneid is contained in the Jacobson
radical of a quasianneid, as well as that the prime radical of a quasianneid is contained in the large
Jacobson radical of an anneid.
Definition 4.7. A right ideal I of a quasianneid A is called modular if there exists an element u ∈ A
such that a ∼ ua (mod I), for all a ∈ A, and we call such an ideal I a modular ideal with respect to u.
Remark 4.2. Let us notice that, if I is a proper modular ideal with respect to u, then δ(u) and δ(u2 )
have a common upper bound. Indeed, since I is proper, u ∈/ I, and so, u # u2 and u − u2 ∈ I. Hence,
2
u ∈ 2
/ I, and so, u = 0, and the claim follows.
Definition 4.8. A quasimoduloid M over a quasianneid A is called strictly cyclic if there exists x ∈ M
such that M = xA. Such an element x is called a strict generator of M .
Definition 4.9. An A-quasimoduloid M is said to be without heart if
C = {x ∈ M | (∀y ∈ M ) x # y} = 0.
Lemma 4.2. A regular A-quasimoduloid M without heart that is strictly cyclic is isomorphic to A/I,
where I is a modular right ideal of A. Every modular right ideal I of A is of the form (0 : x), where x is
a strict generator of an A-quasimoduloid M .
Proof. Let M be a strictly cyclic regular A-quasimoduloid without heart, and let x be its strict generator.
Then, according to Lemma 4.1, M = xA ∼ = A/(0 : x). We hence need to prove that (0 : x) is modular.
Since x ∈ xA and M is regular, there exists u ∈ A such that x = xu.
For an arbitrary a ∈ A, we have xa = xua. If xa = 0, then both a and ua belong to (0 : x). If
xa = xua = 0, then, since M is regular and without heart, we have a # ua and 0 = xa − xua = x(a − ua),
and so, a − ua ∈ (0 : x). Therefore, u is a left identity modulo (0 : x), and so, (0 : x) is modular.
The second assertion is clear.
Similarly to the case of rings, it is possible to prove the following theorem.
Theorem 4.3.
(1) If M is an irreducible regular A-quasimoduloid without heart, then M ∼ = A/I, where I is a max-
imal modular right ideal of A.
(2) Conversely, if I is a maximal modular right ideal of a regular quasianneid A without heart, then
A/I is an irreducible regular A-quasimoduloid without heart.
Theorem 4.4. The Jacobson radical of a regular quasianneid A without heart coincides with the inter-
section of all maximal modular right ideals of A.
Proof. If A is a regular quasianneid A without heart, and I is a maximal modular right ideal of A, then
A/I is a regular A-quasimoduloid without heart, and so, the claim follows from the previous theorem.
Although the proof of the following theorem is analogous to the proof of the corresponding theorem
for anneids [11], we will give it here, for the sake of completeness.
Theorem 4.5. Let A be a paraanneid, Ā its linearization. If Jl (A) is the large Jacobson radical of A and
J(Ā) the ordinary Jacobson radical of the ring Ā, then Jl (A) = J(Ā) ∩ A.
Proof. Since every irreducible Ā-module M may be regarded as an irreducible A-paramoduloid M and
since (0 : M )A ⊆ (0 : M )Ā , we have Jl (A) ⊆ J(Ā) ∩ A. Now, let M be an irreducible A-paramoduloid
and let a ∈ J(Ā) ∩ A. It is enough to prove that a ∈ (0 : M ).

384
 If a ∈/ (0 : M ), then there exists x = 0 such that xa = 0. Since M is irreducible, xa is a strict
generator of M , and hence, M̄ = xaĀ, where M̄ is the linearization of M . Let b̄ ∈ Ā such that x = xab̄.
Then, for all ȳ ∈ Ā, x(ȳ − ab̄ȳ) = 0, i.e., ȳ − ab̄ȳ ∈ (0 : x), for all ȳ ∈ Ā.
Since a ∈ J(Ā), we see that ab̄ also belongs to J(Ā). If z̄ is a quasi-inverse of ab̄ in Ā, then, since
z̄ − ab̄z̄ ∈ (0 : x), we have ab̄ ∈ (0 : x), which implies xab̄ = 0, and therefore x = 0, a contradiction.
Hence, J(Ā) ∩ A ⊆ Jl (A), and so J(Ā) ∩ A = Jl (A).
Lemma 4.6. Let A be a regular paraanneid without heart such that for each element δ of the corresponding
set of grades Δ∗ we have either δ 2 = 0 or δ 2 = δ with respect to the minimal multiplication. Then degrees
of all unities modulo a proper modular right ideal of A have a common idempotent upper bound.
Proof. Let I be a proper modular right ideal of A and let u and u be two unities modulo I. Observe
the canonical mapping f : A → A/I. Clearly, 0 = f (x) = f (ux) = f (u x), for all x ∈ A \ I. Since f
is a homomorphism and A is without heart, it follows that ux # u x. A is regular, without heart and
both ux and u x are nonzero together imply u # u . This means that δ(u) and δ(u ) have a common
upper bound, denote it by ξ. Then 0 < δ(u) < ξ and 0 < δ(u ) < ξ, and also u # u2 = 0. Therefore,
0 < δ(u) = δ(u)δ(u) < ξ 2 . This, together with our assumption, imply ξ 2 = ξ.
Remark 4.3. A paraanneid satisfying the conditions from the previous lemma exists as following example.
It is the ring  
A A
R=
0 A
of upper triangular matrices with respect to the mapping
π : Δ → Sg(R, +), π(δ) = Rδ (δ ∈ Δ),
where
         
A 0 0 A 0 0 A A 0 0
Rδ1 = , Rδ2 = , Rδ3 = , Rδ4 = , R0 = ,
0 0 0 0 0 A 0 0 0 0
and Δ = {δi | i = 0, 1, 2, 3, 4}.
The set Δ is a partially ordered set, which is from below complete semi-lattice and from above
inductively ordered, with respect to
δi < δj if and only if Rδi ⊆ Rδj , i, j = 0, 1, 2, 3, 4.
It is easy to see that R is a paragraded ring with respect to π : δi → Rδi , δi ∈ Δ.
The following table shows that every element δ ∈ Δ is either an idempotent or nilpotent of degree of
nilpotency 2 with respect to the minimal multiplication.
· 0 δ1 δ2 δ3 δ4
0 0 0 0 0 0
δ1 0 δ1 δ2 0 δ4
δ2 0 0 0 δ2 0
δ3 0 0 0 δ3 0
δ4 0 δ1 δ2 δ2 δ4
Definition 4.10. Let I be a proper modular right ideal of a regular paraanneid without heart such that
for each element δ of the corresponding set of grades Δ∗ we have either δ 2 = 0 or δ 2 = δ with respect to
the minimal multiplication. The least common upper bound of all unities modulo I is called the degree
of I.
Theorem 4.7. Let A be a regular paraanneid without heart such that for each element δ of the correspond-
ing set of grades Δ∗ we have either δ 2 = 0 or δ 2 = δ with respect to the minimal multiplication and let ξ
be an idempotent element of Δ∗ . Also, assume that Δ∗ is a cancellative partial groupoid with respect to

385
the minimal multiplication. Then there exists an one-to-one correspondence between the maximal modular
right ideals of A of degree ξ and the maximal modular right ideals of the ring A(ξ) = Āξ .
Proof. With these assumptions, we may proceed as in [11] for regular anneids. Namely, correspondence is
given in the following way: if I is a maximal modular right ideal of A(ξ), let Iˆ = {x ∈ A | xA ∩ A(ξ) ⊆ I}.
Now, as in [11], it may be proved that Iˆ is a maximal modular right ideal of A of degree ξ. Conversely, if I
is a maximal modular right ideal of A of degree ξ, then it can be easily verified that I ∩ A(ξ) is a maximal
modular right ideal of A(ξ).
Remark 4.4. A paraanneid satisfying the conditions from the previous theorem exists according to the
following example: clearly, a subring  
A A
0 0
of the paragraded ring  
A A
0 A
from the Remark 4.3 is also paragraded with respect to paragraduation
   
0 A A A
Rδ1 = , Rδ2 = .
0 0 0 0
This paragraduation induces the minimal multiplication, which makes the set of grades Δ a regular partial
groupoid.
Definition 4.11. An element a of a paraanneid A is called right (left) quasi-regular if a is not a left
(right) unity modulo any proper right ideal of A. A right ideal of A is called quasi-regular if each of its
elements is right quasi-regular.
The following characterization of a right quasi-regular element is a corollary of the previous theorem
for Halberstadt’s analogous result for regular anneids (see [11]).
Theorem 4.8 ([14]). Let A be a regular paraanneid without heart such that for each element δ of the
corresponding set of grades Δ∗ we have either δ 2 = 0 or δ 2 = δ, with respect to the minimal multiplication
and assume that Δ∗ is a cancellative partial groupoid with respect to the minimal multiplication. An
element a ∈ A is right quasi-regular if and only if one of the following two conditions is satisfied :
(1) the degree of a is not an idempotent element of Δ∗ ;
(2) if the degree of a is an idempotent element ξ of Δ∗ , then a is a right quasi-regular element of the
ring A(ξ).
As in the classical case [16] and in the case of regular anneids [11], one may now prove that, under the
assumptions of Theorem 4.7, the Jacobson radical of a regular paraanneid A is a quasi-regular ideal that
contains all right quasi-regular ideals of A. This and Theorem 4.7 together imply the following theorem
(for the case of regular anneids, see [11]).
Theorem 4.9. Under the assumptions of Theorem 4.7, we have
 
J A(ξ) = J(A) ∩ A(ξ),
 
where J A(ξ) and J(A) denote the Jacobson radical of the ring A(ξ) and A, respectively.

5. On the General Radical Theory

In this section, we study the general radical theory of the category K of paragraded rings defined by
quasihomogeneous homomorphisms and ideals (see [23]).
In this context, one always deals with rings contained in a fixed universal class K, i.e., a non-empty
subclass γ of all paragraded rings that also contains all quasihomogeneous homomorphic images and ideals
of each ring A ∈ γ.

386
 Following Kurosh–Amitsur’s definition and well-known theory of a ring (see, e.g., [10]), we introduced
a general radical theory for paragraded rings.
Let γ be a non-empty class of rings from K such that
(a) γ is closed under quasihomogeneous homomorphic images, i.e., if A ∈ γ and B is homomorphic
image of A under a quasihomogeneous homomorphism,
 then B ∈ γ;
(b) for every paragraded ring A, the sum γ(A) = {I  A | I ∈ γ} of homogeneous ideals is in the
class
 γ; 
(c) γ A/γ(A) = 0, for every paragraded ring A.
Definition 5.1. A class γ of paragraded rings from K that satisfies (a), (b), and (c) is called a paragraded
radical class or a radical class of paragraded rings in Kurosh–Amitsur’s sense or, briefly, a paragraded
radical. γ(A) is called a paragraded γ-radical of A. A paragraded ring A is called a paragraded γ-radical
ring if A ∈ γ, i.e., γ(A) = A.
One easily proves the following theorem.
Theorem 5.1. A non-empty class γ of paragraded rings is a paragraded radical class if and only if
(a ) γ is closed under quasihomogeneous homomorphic images;
(b ) γ has the inductive property: if I1 ⊆ · · · ⊆ Iλ ⊆ .
. . is an ascending chain of homogeneous ideals
of a paragraded ring A and each Iλ is in γ, then Iλ ∈ γ;
λ
(c ) γ is closed under extensions: if I is a homogeneous ideal of a paragraded ring A and if both I
and A/I are in γ, then A ∈ γ.
Lemma 5.2. Let A ∈ K, K be a homogeneous ideal of I, I a homogeneous ideal of A, and a a homogeneous
element of A. Then
(a) aK + K  I;
(b) (aK + K)2 ⊆ K;
(c) the mapping f : K → (aK + K)/K defined by f (x) = ax + K, for any x ∈ K, is a surjective
quasihomogeneous homomorphism;
(d) Ker f  I.
Proof. As in the case of associative rings, it is straightforward to verify (a), (b), and (d), as well as to
prove that the mapping f from (c) is a surjective homomorphism. Since clearly, a homogeneous element
is mapped onto a homogeneous element under f , this mapping is also quasihomogeneous, which concludes
the proof.

As a consequence, we have that the theorem of Anderson, Divinsky, and Suliński [2] holds in K.
Theorem 5.3. For any paragraded radical γ and any A ∈ K, if I is a homogeneous ideal of A, then
γ(I)  A.
The proof of this theorem is analogous to that for rings (see, e.g., [10]).

6. Special Radicals
Andrunakievich and Ryabukhin showed that any radical of a ring can be defined in terms of the
appropriate class of modules over that ring [3], which holds for group graded rings too [4].
The main aim of this section is to show that any paragraded radical of a paragraded ring can be defined
by means of some class of paragraded modules, over that ring, particularly, any special paragraded radical,
using methods presented in [3, 4].
For an associative paragraded ring A with a set of grades Δ, we denote a class of paragraded A-modules
by ΣA .

387
Definition 6.1. The homogeneous ideal of a paragraded ring A generated by the set
{x ∈ Aδ | M x = 0}
δ∈Δ

is called the paragraded annihilator AnnA (M ) of a paragraded A-module M .

In the rest of the article, the paragraded annihilator of a paragraded A-module will be simply called
annihilator, since there is no threat of confusion with the classical annihilator.
Definition 6.2. The set
Ker ΣA = {AnnA (M ) | M ∈ ΣA }
is called the kernel of the class ΣA . If ΣA = ∅, then we define Ker ΣA = A.
Definition 6.3. The class Σ consisting of all the classes ΣA for paragraded A-modules is called a general
class of the paragraded modules if the following axioms are satisfied:
(P1) if M ∈ ΣA/B for some homogeneous ideal B of A, then M ∈ ΣA ;
(P2) if M ∈ ΣA and B ⊆ AnnA (M ), then M ∈ ΣA/B ;
(P3) Ker ΣA = 0 if and only if ΣB = ∅, for all nonzero homogeneous ideals B of A.
Definition 6.4. The Σ-paragraded radical γ(Σ, A) of a paragraded ring A is defined to be the set
γ(Σ, A) = Ker ΣA ,

where Σ = ΣA is a general class of paragraded modules.
Theorem 6.1. Let Σ be a general class of paragraded modules. Then the Σ-paragraded radical is a radical
in the category of paragraded rings. Conversely, if γ is a radical in the category of paragraded rings, then
there exists a general class of paragraded modules Σ such that γ coincides with the Σ-radical.
Proof. The first part can be proved analogously to the proof in the ungraded case (see [3]), as well as the
second part but after we establish what the Dorroh extension of a paragraded ring is.
If A is a paragraded ring with set of grades Δ, then we define the Dorroh extension A1 of A as
the Dorroh extension of A, regarded as a ring. Namely, A1 = A × Z, with componentwise addition and
multiplication
(a, n)(b, m) = (ab + ma + nb, nm),
but with paragraduation defined as follows:
A1δ = {(a, 0) | a ∈ Aδ }, A11 = {(0, n) | n ∈ Z}.
It is easy to verify that A1 is then a paragraded ring with set of grades Δ∪ {1}, where 1 acts as an identity
with respect to the minimal multiplication.
Now we turn our attention to special paragraded radicals.
Let M be a class of paragraded rings. A homogeneous ideal I of a paragraded ring A is called an
M-ideal if A/I ∈ M.
Definition 6.5. A class M of paragraded rings is said to be a special class of paragraded rings if the
following conditions are satisfied:
(M1) if B is a homogeneous ideal of a paragraded ring A and P is an M-ideal of A that does not
contain B, then P ∩ B is a proper M-ideal of B;
(M2) if Q is a proper M-ideal of B and B is a homogeneous ideal of A, then there exists a unique
M-ideal P of A such that P ∩ B = Q.
The following lemma can be proved as in the case of graded rings [4].

388
Lemma 6.2.
(a) Let M be a special class of paragraded rings, Σ a general class of paragraded modules such that
the special radical γM coincides with Σ-radical. Then a homogeneous ideal P of a paragraded ring
A is an M-ideal if and only if P = AnnA (M ) for some paragraded A-module M ∈ ΣA .
(b) The largest special class of paragraded rings coincides with the class of all paragraded prime rings.
Therefore, a class M of paragraded rings is special if and only if the following axioms are satisfied.
A1. All rings that belong to M are paragraded prime rings.
A2. If A ∈ M and I is a nonzero homogeneous ideal of A, then I ∈ M.
A3. If B is a homogeneous ideal of a paragraded prime ring A and B ∈ M, then A ∈ M.
A class of special paragraded modules that defines the special radical of a given paragraded ring is
determined analogously to the ungraded case [3] (see also [10]).
Example 6.1. A paragraded A-module M is called paragraded prime if for every nonzero homogeneous
submodule N of M and every homogeneous ideal I of A, N I = 0 implies I ⊆ AnnA (M ).
In [14], the paragraded prime radical of a paragraded ring A is defined as the intersection of all
homogeneous prime ideals of A. Now, it is straightforward to verify that the class Σ of all paragraded
prime A-modules defines the paragraded prime radical of A.
In [15], the paragraded Jacobson radical of a paragraded ring is defined as the intersection of anni-
hilators of all paragraded irreducible modules over that ring with the same set of grades. This notion
is generalized in [14], namely, the paragraded Jacobson radical (large Jacobson radical) of a paragraded
ring A is defined as the intersection of annihilators of all regular (not necessarily regular) paragraded
irreducible A-modules.
In [14] it is proved that the large Jacobson radical of a paragraded ring is the largest homogeneous
ideal contained in the ordinary Jacobson radical of that ring using the method presented in [11] for graded
rings.
The same is proved for the prime radical [14] by means known for group graded rings [25].
Here, we give the following theorem.
Theorem 6.3. Let Σp be a general class of special paragraded modules that is contained in the class of
paragraded irreducible modules, and let Σ be the corresponding unparagraded class of modules. If γ p (A)
is the Σp -radical of a paragraded ring A and γ(A) the Σ-radical of A considered as an ordinary ring, then
γ p (A) is the largest homogeneous ideal of A contained in γ(A).
Proof. Since every A-module M may be regarded as a paragraded A-module, we have that γ p (A) is
contained in the largest homogeneous ideal contained in γ(A). Now, let a be an arbitrary element from
the largest homogeneous ideal of A contained in γ(A), and let M be a paragraded A-module belonging
to Σp . Let us prove that a ∈ AnnA (M ). Assume that a ∈ / AnnA (M ). Then there exists x = 0 such that
xa = 0. Since Σp is contained in the class of paragraded irreducible modules, M is irreducible. Therefore,
xa is a strict generator of M , and hence, M = xaA. Let b ∈ A be such that x = xab. Now, since a ∈ γ(A),
we also have that ab ∈ γ(A). This means that {y − aby | y ∈ A} is not contained in any γ-ideal of A.
Therefore, the largest homogeneous ideal contained in {y − aby | y ∈ A} is not contained in AnnA (M ).
Therefore, there exists y ∈ A and k ∈ AnnA (M ) such that y − aby + k = −ab. Now,
0 = x − xab − (x − xab)y = x − x(ab + y − aby) = x + xk,
which implies x = 0, a contradiction. Hence, a is contained in γ p (A).

7. Paragraded Normal Radicals

In [9], group-graded normal radicals were observed in order to generalize results from [26]. To define
paragraded normal radicals, we need some preliminary notions.

389
Definition 7.1. If A is a paragraded ring, a paragraded radical γ is called paragraded left (right) strong
if I ∈ γ implies I ⊆ γ(A) for all left (right) ideals I of A.
Let A and B be paragraded rings with sets of grades Δ and Δ , respectively, and also, let V and
W be a paragraded (A, B)-bimodule and a paragraded (B, A)-bimodule with sets of grades D and D ,
respectively.
Definition 7.2. A Morita context (A, V, W, B) is called a paragraded Morita context if
(∀ d ∈ D) (∀ d ∈ D ) (∃ δ ∈ Δ) Vd Wd ⊆ Aδ
and
(∀ d ∈ D) (∀ d ∈ D ) (∃ δ  ∈ Δ ) Wd Vd ⊆ Bδ .
Theorem 7.1. Let the quadruple (A, V, W, B) be a Morita context. Then the set of matrices
 
A V
R=
W B
is a paragraded ring under matrix addition and multiplication. In particular, if a given Morita context is
paragraded, then R is a paragraded ring.
Proof. We may use the proof of Theorem 2.1 from [14] and make R paragraded in a similar manner. We
take        
A 0 A V A 0 A 0
R1 = , R2 = , R3 = , R4 = ,
0 0 0 0 W 0 0 B
and so on.
Definition 7.3. A radical γ of paragraded rings is called paragraded normal if
V γ(B)W ⊆ γ(A),
for every paragraded Morita context (A, V, W, B).
Using the paragraded Dorroh extension from the proof of Theorem 6.1, the following result can be
proved the same way as in the case of rings [26].
Proposition 7.1. If γ is a paragraded normal radical, then γ is left and right strong.
Now we will first introduce a paragraded principally left (right) hereditary radical.
Definition 7.4. A paragraded radical γ is paragraded principally left (right) hereditary if A ∈ γ implies
Aa ∈ γ (aA ∈ γ), for every homogeneous element a ∈ A.
Proposition 7.2. Every paragraded normal radical γ is paragraded principally left hereditary and prin-
cipally right hereditary.
Proof. Let A ∈ γ, and let a ∈ A be its homogeneous element. As in the case of rings we observe
the paragraded Morita context (Aa, A1 , Aa, A) and establish that A2 a ⊆ γ(Aa). The mapping f : A →
Aa/γ(Aa) given by f (x) = xa + γ(Aa), for all x ∈ A, is a quasihomogeneous homomorphism, and the
rest of the proof is similar to the proof given in the case of rings.
Finally, taking into account previous results, with the appropriate modification of a proof given in [26]
for rings, the following result holds.
Theorem 7.2. A paragraded radical γ is paragraded normal if and only if γ is paragraded left strong
and paragraded principally left hereditary (or paragraded right strong and paragraded principally right
hereditary).

390
 REFERENCES
1. S. A. Amitsur, “A general theory of radicals, II,” Amer. J. Math., 76, Radicals in Rings and Bi-
categories, 100–125 (1954).
2. R. Anderson, N. Divinsky, and A. Suliński, “Hereditary radicals in associative and alternative rings,”
Can. J. Math., 7, 594–603 (1965).
3. V. A. Andrunakievich and Yu. M. Ryabuhin, “Special modules and special radicals,” Dokl. Akad.
Nauk SSSR, 147, No. 6, 1274–1277 (1962).
4. I. N. Balaba, “Special radicals of graded rings,” Bul. Acad. Ştiinţe Mold. Mat., 44, No. 1, 26–33
(2004).
5. N. Bourbaki, Algèbre, Chap. I, Herman, Paris (1962).
6. B. Brown and N. H. McCoy, “Radicals and subdirect sum,” Am. J. Math. Soc., 67, 46–48 (1947).
7. M. Chadeyras, “Essai d’une théorie noetherienne pour les anneaux commutatifs, dont la graduation
est aussi générale que possible,” Bull. Soc. Math. Fr. Suppl. Mém., No. 22, 3–143 (1970).
8. M. Cohen and S. Montgomery, “Group-graded rings, smash products, and group actions,” Trans.
Am. Math. Soc., 282, No. 1, 237–258 (1984).
9. H. Fang and P. Stewart, “Radical theory for graded rings,” J. Austral. Math. Soc. Ser. A, 52, 143–153
(1992).
10. B. J. Gardner and R. Wiegandt, Radical Theory of Rings, Pure Appl. Math., Vol. 261, Marcel Dekker
(2004).
11. E. Halberstadt, Théorie artinienne homogène des anneaux gradués à grades non commutatifs
réguliers, PhD Thesis, Univ. Pierre and Marie Curie, Paris (1971).
12. E. Ilić-Georgijević and M. Vuković, “Sheaves of paragraded rings,” Sarajevo J. Math., 7 (20), 153–161
(2011).
13. E. Ilić-Georgijević and M. Vuković, “A note on general radicals of paragraded rings,” Sarajevo J.
Math., 12 (25), No. 2, 317–324 (2016).
14. E. Ilić-Georgijević and M. Vuković, “A note on radicals of paragraded rings,” Sarajevo J. Math.,
12 (25), No. 2, 307–316 (2016).
15. E. Ilić-Georgijević and M. Vuković, “The Wedderburn–Artin theorem for paragraded rings,” J. Math.
Sci., 221, No. 3, 391–400 (2017).
16. N. Jacobson, “The radical and semi-simplicity for arbitrary rings,” Am. J. Math., 78, 300–320 (1945).
17. E. Jespers, J. Krempa, and E. R. Puczylowski, “On radicals of graded rings,” Commun. Algebra,
10 (17), 1849–1854 (1980).
18. M. Krasner, “Une généralisation de la notion de corps-corpoı́de. Un corpoı́de remarquable de la
théorie des corps valués,” C. R. Acad. Sci. Paris., 219, 345–347 (1944).
19. M. Krasner, “Anneaux gradués généraux,” Colloq. d’Algèbre Rennes, pp. 209–308 (1980).
20. M. Krasner and M. Vuković, “Structures paragraduées (groupes, anneaux, modules), I,” Proc. Japan
Acad. Ser. A, 62, No. 9, 350–352 (1986).
21. M. Krasner and M. Vuković, “Structures paragraduées (groupes, anneaux, modules), II,” Proc. Japan
Acad. Ser. A, 62, No. 10, 389–391 (1986).
22. M. Krasner and M. Vuković, “Structures paragraduées (groupes, anneaux, modules), III,” Proc.
Japan Acad. Ser. A, 63, No. 1, 10–12 (1987).
23. M. Krasner and M. Vuković, Structures paragraduées (groupes, anneaux, modules), Queen’s Papers
Pure Appl. Math., No. 77, Queen’s University, Kingston, Ontario, Canada (1987).
24. A. G. Kurosh, “Radical of rings and algebras,” Mat. Sb., 33, No. 1, 3–26 (1953).
25. C. Nǎstǎsescu and F. van Oystaeyen, Methods of Graded Rings, Lect. Notes Math., Vol. 1836, Springer
(2004).
26. A. D. Sands, “On normal radicals,” J. London Math. Soc., 11, 361–365 (1975).
27. M. Vuković, Structures graduées et paragraduées, Prepublication de l’Institut Fourier, Université de
Grenoble I, No. 536 (2001).

391
28. M. Vuković, “From Krasner’s corpoid and Bourbaki’s graduations to Krasner’s graduations and
Krasner–Vuković’s paragraduations,” Sarajevo J. Math., 14 (27), No. 2, 175–190 (2018).
29. M. Vuković and E. Ilić-Georgijević, “Paragraded rings and their ideals,” J. Math. Sci., 191, No. 5,
654–660 (2013).

Mirjana Vuković
Academy of Sciences and Arts of Bosnia and Herzegovina,
Bistrik 7, 71000 Sarajevo, Bosnia and Herzegovina
E-mail: mvukovic@anubih.ba

392

View publication stats