Rings, ideals and modules
Introduced| English | Meaning |
|---|---|
| unit/ˈjuːnɪt/ | An element with a multiplicative inverse in the ring. |
| ideal/aɪˈdɪəl/ | An additive subgroup of a ring that absorbs multiplication by every ring element. |
| torsion/ˈtɔːʃn/ | The presence of a nonzero module element killed by a nonzero scalar from an integral domain. |
Two nonzero numbers, product zero
- In $\mathbb Z/6\mathbb Z$, $2\cdot3=0$. Also $2\cdot1=2\cdot4=2$, so cancelling the nonzero factor $2$ would discard a solution.
- Units 单位, ideals 理想 and modules explain when familiar algebra works. Until the Boolean-ring proof, assume a commutative ring with identity $1\ne0$; write $\mathbb Z/n\mathbb Z$ for integers modulo $n\geq2$.
Invertible elements and zero products
- A unit is an element with a multiplicative inverse in the ring. A zero divisor is a nonzero element whose product with some other nonzero element is zero. Multiplying a zero product by a unit's inverse shows that a unit cannot be a zero divisor.
- In $\mathbb Z/n\mathbb Z$, $a$ is a unit exactly when $\gcd(a,n)=1$. For any nonzero residue with $d=\gcd(a,n)>1$, the nonzero residue $n/d$ satisfies $a(n/d)=0$, proving that $a$ is a zero divisor.
How many units does $\mathbb Z/12\mathbb Z$ have?
Difficulty 2/5 · 140 coins + a small bonus on first completion today
The units of ℤ/12ℤ are the residues coprime to 12: 1, 5, 7, 11. That is φ(12) = 4.
Work the inverse or the annihilator
- In $\mathbb Z/12\mathbb Z$, $5\cdot5=25=1$, so $5$ is its own inverse. By contrast, $4\cdot3=0$, so $4$ is a zero divisor. The zero residue itself is excluded from the definition of zero divisor used here.
- In $\mathbb Z/6\mathbb Z$, the units are $1,5$ and the zero divisors are $2,3,4$. Solving $2x=2$ gives $x=1,4$; multiply by an inverse only when that inverse exists.
Nonzero residues in ℤ/12ℤ
Separate inverses from nonzero annihilators modulo 12.
In $\mathbb Z/6$, which $x$ satisfy $2x = 2$?
Difficulty 2/5 · 140 coins + a small bonus on first completion today
2x ≡ 2 (mod 6) means 6 divides 2x − 2, i.e. x ≡ 1 (mod 3). In ℤ/6 that gives x = 1 and x = 4 (2 · 4 = 8 ≡ 2). Two solutions appear because 2 is not a unit mod 6.
Under the definition used in this lesson, zero is a zero divisor.
Difficulty 1/5 · 100 coins + a small bonus on first completion today
A zero divisor must itself be nonzero and kill another nonzero element. Zero is excluded by this convention.
Domains, fields and valid cancellation
- An integral domain is a nonzero commutative ring with identity and no zero divisors. A field is such a ring in which every nonzero element is a unit. Every field is a domain; every finite domain is a field because multiplication by a nonzero element is an injective map of a finite set, hence onto.
- In any domain, $ab=ac$ with $a\ne0$ implies $a(b-c)=0$, hence $b=c$. A cancellable element need not be a unit: $2$ in $\mathbb Z$ cancels, but has no integer inverse. The unit-or-zero-divisor dichotomy above is specific to $\mathbb Z/n\mathbb Z$, not all rings.
Every finite integral domain is a field.
Difficulty 1/5 · 100 coins + a small bonus on first completion today
Multiplication by a nonzero element is injective because cancellation holds. On a finite set it is therefore onto, so 1 has a preimage and the element has an inverse.
Ideals absorb multiplication from outside
- An ideal is an additive subgroup of a ring that absorbs multiplication by every ring element. Thus $i\in I$ and $r\in R$ imply $ri\in I$. A proper ideal is an ideal different from the entire ring.
- The even integers $2\mathbb Z$ form an ideal of $\mathbb Z$. But $\mathbb Z$ is only a subring of $\mathbb Q$: multiplying $1\in\mathbb Z$ by $1/2\in\mathbb Q$ leaves $\mathbb Z$. If an ideal contains any unit, it contains $1$ and therefore every ring element.
$\mathbb Z$ is an ideal of $\mathbb Q$.
Difficulty 1/5 · 100 coins + a small bonus on first completion today
$\tfrac12 \cdot 1 \notin \mathbb Z$.
Why an ideal gives a quotient ring
- In $R/I$, two elements are equivalent when their difference lies in $I$. Define $(a+I)+(b+I)=(a+b)+I$ and $(a+I)(b+I)=ab+I$; the answer must be independent of the representatives.
- Replace $a,b$ by $a+i,b+j$, with $i,j\in I$. The product changes by $aj+ib+ij\in I$, using absorption and additive closure. A mere additive subgroup is insufficient: $0$ and $1$ represent the same class modulo $\mathbb Z$ in $\mathbb Q$, but multiplying them by $1/2$ gives different classes.
Why can $\mathbb Q/\mathbb Z$ be an additive quotient group but not a quotient ring under the usual product of representatives?
$\mathbb Z$ is an additive subgroup of $\mathbb Q$, so additive cosets are valid. However, $0+\mathbb Z=1+\mathbb Z$. Multiplying these equal classes by $1/2+\mathbb Z$ would give $0+\mathbb Z$ using representative $0$ and $1/2+\mathbb Z$ using representative $1$. The outputs differ because $1/2\notin\mathbb Z$. Multiplication is not well-defined: $\mathbb Z$ does not absorb multiplication by all rational numbers and is not an ideal.
Kernels identify the quotient
- For a ring homomorphism $\varphi$, its kernel is an ideal and $R/\ker\varphi\cong\operatorname{im}\varphi$. The quotient agrees with the full codomain when the map is onto; reduction $\mathbb Z\to\mathbb Z/5\mathbb Z$ has kernel $5\mathbb Z$.
- In $\mathbb R[x]$, evaluation at $i$ maps $p(x)$ to $p(i)\in\mathbb C$. Division by $x^2+1$ leaves a unique remainder $a+bx$, so the map is onto and its kernel is $(x^2+1)$. Hence $\mathbb R[x]/(x^2+1)\cong\mathbb C$; here $x^2=-1$ and $x$ has inverse $-x$.
In $\mathbb R[x]/(x^2+1)$, which class is the inverse of $x$?
Difficulty 1/5 · 100 coins + a small bonus on first completion today
The relation is x² = −1, so x(−x) = −x² = 1. The class x itself squares to −1 rather than 1.
Prime versus maximal ideals
- A proper ideal $I$ is prime when $ab\in I$ forces $a\in I$ or $b\in I$. This says exactly that $R/I$ has no zero divisors. A proper ideal is maximal when no ideal lies strictly between it and $R$; this is equivalent to $R/I$ being a field.
- Every maximal ideal is prime, but the converse fails: $(0)$ in $\mathbb Z$ is prime since $\mathbb Z$ is a domain, yet not maximal since $(0)\subsetneq2\mathbb Z\subsetneq\mathbb Z$. For positive $n\geq2$, $n\mathbb Z$ is prime and maximal exactly when $n$ is prime.
Which quotient ring is a field?
Difficulty 1/5 · 100 coins + a small bonus on first completion today
7ℤ is maximal because 7 is prime.
Which ideal is prime but not maximal in the stated ring?
Difficulty 1/5 · 100 coins + a small bonus on first completion today
ℤ/(0) is ℤ, a domain but not a field. The quotients by 2ℤ and 5ℤ are fields, while ℤ/6ℤ has zero divisors.
Test the quotient, not just its notation
- In $\mathbb R[x]/(x^2-1)$, the nonzero classes of $x-1$ and $x+1$ multiply to zero. Neither linear polynomial is a multiple of the quadratic, so these really are nonzero classes; the quotient is not a domain.
- By contrast, $\mathbb R[x]/(x^2+1)$ is the field $\mathbb C$. The quotient by the whole ring is the zero ring, where $1=0$; it is excluded from our field and domain definitions. Keep the word proper in the prime/maximal criteria.
Prime, maximal or neither?
Use the quotient ring to classify each stated ideal.
In $\mathbb R[x]/(x^2+1)$, $x^4+2x^2+3$ equals which real constant?
Difficulty 2/5 · 140 coins + a small bonus on first completion today
Use x² = −1 and x⁴ = 1. The expression is 1 − 2 + 3 = 2.
Compare $\mathbb R[x]/(x^2+1)$ with $\mathbb R[x]/(x^2-1)$. Decide whether each defining ideal is prime or maximal, and justify your answers.
Evaluation at $i$ gives an onto map $\mathbb R[x]\to\mathbb C$: each $a+bi$ is the image of $a+bx$. Dividing a polynomial by $x^2+1$ leaves $a+bx$; this remainder vanishes at $i$ only if $a=b=0$. Thus the kernel is $(x^2+1)$ and the quotient is a field, so this ideal is maximal and prime. In the other quotient, the nonzero classes of $x-1$ and $x+1$ multiply to zero. They are nonzero because neither linear polynomial is divisible by $x^2-1$. Therefore this quotient is not a domain, and $(x^2-1)$ is neither prime nor maximal.
Modules: scalars come from a ring
- An $R$-module is an abelian group with scalar multiplication by $R$, satisfying the distributive and associative scalar laws and $1m=m$. A vector space is the case where the scalar ring is a field.
- Every abelian group is a $\mathbb Z$-module: an integer acts by repeated addition, with negatives giving additive inverses. For example, $\mathbb Z/6\mathbb Z$ is a $\mathbb Z$-module, and the nonzero integer $6$ kills its nonzero element $1$.
Torsion prevents field-style cancellation
- Torsion 挠性 is the presence of a nonzero module element killed by a nonzero scalar from an integral domain. In a vector space this cannot happen: multiplying $av=0$ by $a^{-1}$ gives $v=0$. In the $\mathbb Z$-module $\mathbb Z/6\mathbb Z$, $2\cdot3=0$ disproves that implication.
- A free module has a basis giving unique finite linear combinations. The nonzero $\mathbb Z$-module $\mathbb Z/6\mathbb Z$ is not free: every element is killed by $6$, whereas a basis vector in a free $\mathbb Z$-module cannot be. Thus do not assign every module a vector-space dimension or apply the vector-space rank–nullity theorem without its hypotheses.
Injective need not mean surjective
- The $\mathbb Z$-linear map $T:\mathbb Z\to\mathbb Z$, $T(m)=2m$, has zero kernel, so is injective. Its image is $2\mathbb Z$, so $1$ has no preimage and the map is not surjective.
- Both modules have a basis with one element, but the map is not onto. Over a field, an injective linear map between spaces of the same finite dimension is onto: for example, multiplication by $2$ on $\mathbb Q$ reaches every target $q$ using the preimage $q/2$.
For the $\mathbb Z$-linear map $T:\mathbb Z\to\mathbb Z$, $T(m)=2m$, which statement is correct?
Difficulty 1/5 · 100 coins + a small bonus on first completion today
The equation 2m = 0 forces m = 0, but 2m = 1 has no integer solution. Thus T is injective and not onto.
Compare multiplication by $2$ on the $\mathbb Z$-modules $\mathbb Z$ and $\mathbb Z/6\mathbb Z$, and on the vector space $\mathbb Q$ over $\mathbb Q$. Find each kernel and image.
On $\mathbb Z$, the kernel is $\{0\}$ and the image is $2\mathbb Z$, so the map is injective but not onto. On $\mathbb Z/6\mathbb Z$, the kernel is $\{0,3\}$ and the image is $\{0,2,4\}$, so it is neither injective nor onto. On $\mathbb Q$, the kernel is $\{0\}$ and the image is all of $\mathbb Q$, since every $q$ has preimage $q/2$. The inverse scalar $1/2$ is available over $\mathbb Q$ but not over $\mathbb Z$; torsion creates the extra kernel in the finite module.
Boolean rings: derive commutativity
- For this proof, drop the assumption of commutativity. Let a ring satisfy $a^2=a$ for every element. Distributivity gives $(a+a)^2=4a^2=4a$, while idempotence gives $(a+a)^2=2a$. Subtracting $2a$ yields $2a=0$ for every $a$.
- Next expand $(a+b)^2=a^2+ab+ba+b^2$ in that order. Substituting idempotence gives $ab+ba=0$. Since $2(ba)=0$, we have $-ba=ba$, hence $ab=ba$. Commutativity follows; it was not used in the expansion.
In a Boolean ring, $2a =$ ____ for every $a$.
Difficulty 1/5 · 100 coins + a small bonus on first completion today
From (a + a)² = a + a.
What the Boolean proof does not say
- Write $\mathbb F_2=\mathbb Z/2\mathbb Z$. In $\mathbb F_2\times\mathbb F_2$, every element squares to itself, but $(1,0)(0,1)=(0,0)$. A nonzero Boolean ring therefore need not be a domain or a field.
- The property $2a=0$ alone does not make a ring Boolean. In $\mathbb F_2[t]/(t^2)$, every element added to itself is zero, but the nonzero class of $t$ has square zero rather than itself.
If $2a=0$ for every element of a ring, then every element must satisfy $a^2=a$.
Difficulty 1/5 · 100 coins + a small bonus on first completion today
In 𝔽₂[t]/(t²), every element doubles to zero, but the nonzero class of t has square zero. Thus characteristic 2 alone is insufficient.
Prove that a ring with $a^2=a$ for every $a$ is commutative without assuming multiplication commutes. Must such a nonzero ring be a field?
Apply idempotence to $a+a$. Distributivity gives $(a+a)^2=4a^2=4a$, while idempotence gives $(a+a)^2=2a$. Subtraction gives $2a=0$ for every $a$. Next expand $(a+b)^2$ as $a^2+ab+ba+b^2$. Idempotence then gives $ab+ba=0$. Since $2(ba)=0$, additive inverses satisfy $-ba=ba$, giving $ab=ba$. No multiplication was reordered before this conclusion. Such a ring need not be a field: $\mathbb F_2\times\mathbb F_2$ is Boolean and nonzero, but $(1,0)(0,1)=(0,0)$ exhibits zero divisors.
Check the structure before dividing
- Distinguish an inverse from cancellation, and an ideal from a subring. For a proper ideal, test whether the quotient is a domain or field to distinguish prime from maximal.
- Specify the scalar ring when discussing modules. Check invertibility, torsion and freeness before importing vector-space arguments; keep the Boolean proof independent of commutativity.