Subrings

Subrings appear throughout mathematics wherever one studies substructures preserving both additive and multiplicative operations. In algebraic number theory, rings of integers such as $\mathbb{Z}[\sqrt{2}]$ or $\mathbb{Z}[i]$ are subrings of larger fields, and understanding which subsets are subrings (versus merely additive subgroups) is foundational to studying integrality and factorization. In linear algebra and representation theory, rings of matrices such as upper-triangular or block-diagonal matrices are subrings of $M_n(F)$, and they organize the study of flags and invariant subspaces. In commutative algebra and algebraic geometry, coordinate rings of subvarieties arise as quotients, but the subrings of a polynomial ring (fixed subrings under group actions, for instance) drive invariant theory. In cryptography, subrings of finite rings such as $\mathbb{Z}/n\mathbb{Z}$ underlie constructions in coding theory and lattice-based cryptosystems. In computer algebra systems, recognizing subring structure allows symbolic computation software to restrict computations to smaller, more efficient substructures (e.g. working within $\mathbb{Z}[i]$ instead of all of $\mathbb{C}$). The subring/ideal distinction developed here is also the essential prerequisite for defining quotient rings and the Ring Homomorphisms whose kernels are always ideals, never merely subrings.

Ask the tutor: Why does every ideal automatically satisfy the subring criterion, but not every subring satisfy the ideal condition?

Think of a ring $R$ as a landscape with two overlaid grids: an addition grid (making $R$ an abelian group) and a multiplication grid (compatible with addition via distributivity). A subring is a sub-landscape $S$ that is self-contained under both grids: adding or multiplying two points of $S$ never leaves $S$, and every point of $S$ has its additive inverse already inside $S$. The subtlety is that "closed under $+$ and $\cdot$" is not quite the same as "is a ring." A subset can be closed under addition and multiplication yet fail to be a group under addition — for instance, the natural numbers $\mathbb{N}\subset\mathbb{Z}$ are closed under both operations but $\mathbb{N}$ has no additive inverses, so it is not a subring of $\mathbb{Z}$. This is exactly the gap the subring criterion is built to close: it forces closure under subtraction, which simultaneously guarantees $0\in S$ and that every element of $S$ has its negative in $S$.

For every $n\in\mathbb{Z}$, the set $n\mathbb{Z}=\{nk:k\in\mathbb{Z}\}$ is a subring of $\mathbb{Z}$ (in fact an ideal, as we will define elsewhere) — this is the family of examples the diagram depicts inside the ideal region. In contrast $\mathbb{N}$, though closed under both operations, fails to be a subring because it is not closed under subtraction (e.g. $2-5=-3\notin\mathbb{N}$).

Example 1: The Gaussian integers inside the complex numbers, checked against a near-miss. Let $R=\mathbb{C}$ with the usual addition and multiplication, and let

We verify the subring criterion. Nonempty: $0=0+0i\in S$. Closed under subtraction: let $z=a+bi$, $w=c+di$ with $a,b,c,d\in\mathbb{Z}$. Then

and $a-c,\,b-d\in\mathbb{Z}$ since $\mathbb{Z}$ is closed under subtraction. So $z-w\in S$. Closed under multiplication:

Since $a,b,c,d\in\mathbb{Z}$ and $\mathbb{Z}$ is closed under products and sums, $ac-bd\in\mathbb{Z}$ and $ad+bc\in\mathbb{Z}$. So $zw\in S$. By the subring criterion, $\mathbb{Z}[i]$ is a subring of $\mathbb{C}$. Notice also that $1=1+0i\in S$, so here the subring happens to contain the identity of the ambient ring — but this was a consequence we checked, not a requirement of the criterion. Now the near-miss, to sharpen the boundary: let

(Gaussian integers in the closed first quadrant). Is $T$ a subring of $\mathbb{C}$? Take $z=3+2i\in T$ and $w=1+5i\in T$. Then

which has negative imaginary part, so $z-w\notin T$. Subtraction closure fails, so $T$ is not a subring — despite $T$ being closed under addition and, one can check, under multiplication (products of quadrant-I Gaussian integers can be verified to stay in the quadrant only under further restriction, but even granting closure under $+,\cdot$, the failure of subtraction closure alone already rules out $T$ as a subring). This mirrors the $\mathbb{N}\subset\mathbb{Z}$ example from the intuition section and confirms that subtraction closure is doing real work, not a redundant formality.

Example 2: Upper triangular matrices, an asymmetric ring of matrices. Let $R=M_2(\mathbb{R})$, the ring of $2\times 2$ real matrices under matrix addition and multiplication (note $R$ is noncommutative and unital with identity $I=\begin{pmatrix}1&0\\0&1\end{pmatrix}$). Let

the upper triangular matrices. We check the subring criterion. Nonempty: the zero matrix is in $S$ (take $a=b=d=0$). Closed under subtraction: for $A=\begin{pmatrix}a&b\\0&d\end{pmatrix}$ and $A'=\begin{pmatrix}a'&b'\\0&d'\end{pmatrix}$ in $S$,

since the $(2,1)$ entry remains $0$. Closed under multiplication: compute

The $(2,1)$ entry is $0\cdot a' + d\cdot 0 = 0$, so $AA'\in S$. By the subring criterion, $S$ is a subring of $M_2(\mathbb{R})$, and $I\in S$ (taking $a=d=1,b=0$), so here too the identity survives — but note $S$ is not commutative: with $A=\begin{pmatrix}1&1\\0&0\end{pmatrix}$ and $A'=\begin{pmatrix}0&0\\0&1\end{pmatrix}$, both in $S$,

so $AA'\neq A'A$. This shows subrings need not inherit commutativity of multiplication from a commutative ambient ring assumption — although here $R=M_2(\mathbb{R})$ itself was already noncommutative, so nothing was violated; the point is simply that being a subring imposes no extra commutativity beyond what the ambient ring already has or lacks. Now consider whether $S$ is an ideal of $R$. Take $A=\begin{pmatrix}1&0\\0&0\end{pmatrix}\in S$ and an arbitrary matrix $M=\begin{pmatrix}0&0\\1&0\end{pmatrix}\in R$ (not necessarily in $S$). Then

which has a nonzero $(2,1)$ entry, so $MA\notin S$. Thus $S$ absorbs multiplication from $S$ into $S$, but fails to absorb multiplication from all of $R$: $S$ is a subring but not an ideal. This is the central distinction the article is building toward.