Alphabeta Math
Pipeline-generated
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Fibre Products Base Change and Scheme Theoretic Fibres — Examples

1 · Prerequisites

2 · Summary

These calculations use the tensor-product and fibre formulas from the theory page. The families xy=t and x2=t illustrate reducible, nonreduced, empty and field-valued fibres. Scalar extension can split an integral scheme or introduce nilpotents; the examples distinguish actual scheme points from points valued in a specified field.

Polynomial graphs and quadratic self fibre products give explicit equations. The multiplicity comparison uses nilradical powers to distinguish schemes with the same one-point topology and residue field. All calculations retain the full coordinate rings, in every characteristic stated.

3 · Logical flowchart

4 · Definitions, theorems and proofs

None yet.

5 · Examples, counterexamples and false statements

ExampleConstruction: AI-generatedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

The fibres of xy=t

Example

Over any field k, consider X=Spec⁡k[x,y,t]/(xy−t)→Spec⁡k[t]. At t=a∈k the fibre is Spec⁡k[x,y]/(xy−a). At zero it is the union of two reduced axes; at a≠0 it is Gm,k=Spec⁡k[x,x−1]. The generic fibre is Spec⁡k(t)[x,x−1].

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

Let A→B be a ring map, p∈Spec⁡A, and M=A∖p, acting on B through the ring map. The fibre over p is canonically Spec⁡(B⊗Aκ(p))≅Spec⁡(M−1B/pM−1B). The residue field is κ(p)=Ap/pAp. No reduction of the tensor ring is taken. (Coordinate ring of an affine fibre)

[F2]

Let A→C be a unital ring map. For any set of variables (ti) and any ideal I⊆A[ti], (A[ti]/I)⊗AC≅C[ti]/IC[ti]. Here the extended ideal is generated by the coefficient images of all elements of I. For a multiplicative subset M⊆A, (M−1A)⊗AC≅M‾−1C. These are ring isomorphisms; no flatness, finite-generation or nonzero-ring hypothesis is required. (Presentations and localization under base extension)

Verification

1.1givenF1F2

F1 computes a fibre by tensoring with its residue field. F2 substitutes t=a to give k[x,y]/(xy−a), and at the generic point it extends coefficients to k(t) with relation xy=t.

2.1step 1.1algebra

At a=0, (xy)=(x)∩(y): a polynomial divisible by both x and y has every monomial divisible by xy. The ideals (x),(y) are prime, so their intersection is radical and the quotient is reduced. Every prime containing xy contains x or y; these two incomparable minimal primes give exactly the two axes, meeting at (x,y).

3.1step 1.1algebra∎

At a≠0, the inverse maps send x↦x, y↦ax−1 in one direction and x−1↦a−1y in the other. They identify the ring with k[x,x−1]. The same formulas with the nonzero unit t∈k(t) give the generic fibre. In particular a=1 has the same form; all three rings are nonzero, so none of these fibres is empty.

ExampleConstruction: Literature-sourcedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

Quadratic fibres over rational points and the generic point

Example

For any field k, the family Spec⁡k[x,t]/(x2−t)→Spec⁡k[t] has fibre Spec⁡k[x]/(x2) at zero, nonreduced in every characteristic. Over Q, the fibre rings at t=1,0,−1 and at the generic point are respectively Q×Q,Q[ϵ]/(ϵ2),Q(i),Q(x), where the generic base field Q(t) embeds in Q(x) via t=x2, with degree two.

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

Let A→B be a ring map, p∈Spec⁡A, and M=A∖p, acting on B through the ring map. The fibre over p is canonically Spec⁡(B⊗Aκ(p))≅Spec⁡(M−1B/pM−1B). The residue field is κ(p)=Ap/pAp. No reduction of the tensor ring is taken. (Coordinate ring of an affine fibre)

[F2]

Let A→C be a unital ring map. For any set of variables (ti) and any ideal I⊆A[ti], (A[ti]/I)⊗AC≅C[ti]/IC[ti]. Here the extended ideal is generated by the coefficient images of all elements of I. For a multiplicative subset M⊆A, (M−1A)⊗AC≅M‾−1C. These are ring isomorphisms; no flatness, finite-generation or nonzero-ring hypothesis is required. (Presentations and localization under base extension)

[F3]

Let R be a commutative ring and let I1,…,Ir be pairwise comaximal ideals, where r≥1. Then the canonical map R⟶∏i=1rR/Ii,x⟼(x+I1,…,x+Ir) is surjective, its kernel is ⋂i=1rIi, and ⋂i=1rIi=∏i=1rIi. Equivalently, R/∏i=1rIi≅∏i=1rR/Ii. (Chinese remainder theorem for pairwise comaximal ideals)

Verification

1.1givenF1F2

F1 and F2 give k[x]/(x2−a) at t=a. At a=0, the classes of 1,x are a basis, so x is nonzero with square zero, in every characteristic. The ideal (x) is the only prime and the residue field is k.

1.2F1F2algebra

The generic fibre is Q(t)[x]/(x2−t), equivalently Q[x] with all nonzero polynomials in x2 inverted. For nonzero g(x), the product g(x)g(−x) is a nonzero even polynomial, hence a polynomial in x2. Thus g(x)−1=g(−x)/(g(x)g(−x)) belongs to this ring. It is exactly Q(x). Its dimension over Q(t) is two, because division by the monic polynomial x2−t leaves unique remainders of degree less than two.

2.1F3step 1.1algebra

Over Q, at a=1 the factors x−1,x+1 generate comaximal ideals since their difference is 2. F3 therefore gives Q×Q. At a=−1, x2+1 has no rational root, hence is irreducible of degree two and its quotient is the field Q(i).

3.1F3step 1.1step 2.1step 1.2∎

The displayed fibres are nonempty: each coordinate ring contains 1 nonzero. Over an algebraic closure of any residue field in this rational family, x2−a splits into two distinct factors when a≠0, since the characteristic is zero and a root then has nonzero derivative 2x. At a=0 it remains a doubled point. CRT proves the two-point assertion just as at a=1.

ExampleConstruction: AI-generatedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

A real conic acquires complex points

Example

The real scheme X=Spec⁡R[x,y]/(x2+y2+1) has no R-valued points over R. Its complex base change is Spec⁡C[x,y]/(x2+y2+1) and has the complex point (i,0). Absence of real-valued points does not mean X is empty.

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

For every field K and scheme X, morphisms Spec⁡K→X correspond bijectively to pairs (x,ι) with x∈X and a field embedding ι:κ(x)→K. The identity embedding gives a canonical morphism Spec⁡κ(x)→X, compatible with all scheme morphisms. More generally, for a nonzero local ring (R,m), morphisms Spec⁡R→X correspond to pairs (x,φ) with a local homomorphism φ:OX,x→R. Assuming Choice, two field-valued points have the same image in X if and only if they are dominated by a common field-valued point, by compatible embeddings of their fields into a third field. (Field-valued points and local-ring points)

[F2]

For a field extension K/k and a k-scheme X, the inverse image of every affine open U=Spec⁡A in XK is Spec⁡(A⊗kK). These affine charts cover XK and are compatible on overlaps and with coefficient localizations. (Affine charts after extension of the ground field)

[F3]

Let A→C be a unital ring map. For any set of variables (ti) and any ideal I⊆A[ti], (A[ti]/I)⊗AC≅C[ti]/IC[ti]. Here the extended ideal is generated by the coefficient images of all elements of I. For a multiplicative subset M⊆A, (M−1A)⊗AC≅M‾−1C. These are ring isomorphisms; no flatness, finite-generation or nonzero-ring hypothesis is required. (Presentations and localization under base extension)

Verification

1.1givenF1algebra

A real-valued point over R gives a real algebra map to R, hence images a,b∈R satisfying a2+b2+1=0. This is impossible since the left side is at least 1. This agrees with the residue-field description of F1; the requirement that the morphism be over R is retained.

2.1F2F3algebra∎

F2 and F3 give the displayed complex ring. Evaluation x↦i,y↦0 kills its defining polynomial and sends 1 to 1, so it is a complex point. Its composition with the projection gives a point of X, proving X is nonempty. The same evaluation directly shows neither coordinate ring is zero.

CounterexampleConstruction: Literature-sourcedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

An integral real scheme splits over the complex numbers

Statement refuted

False claims: an integral real scheme must stay irreducible or connected after extension to C; an injective morphism of schemes must remain injective after arbitrary base change. The morphism Spec⁡C→Spec⁡R refutes all these assertions.

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

For a field extension K/k and a k-scheme X, the inverse image of every affine open U=Spec⁡A in XK is Spec⁡(A⊗kK). These affine charts cover XK and are compatible on overlaps and with coefficient localizations. (Affine charts after extension of the ground field)

[F2]

Let A→C be a unital ring map. For any set of variables (ti) and any ideal I⊆A[ti], (A[ti]/I)⊗AC≅C[ti]/IC[ti]. Here the extended ideal is generated by the coefficient images of all elements of I. For a multiplicative subset M⊆A, (M−1A)⊗AC≅M‾−1C. These are ring isomorphisms; no flatness, finite-generation or nonzero-ring hypothesis is required. (Presentations and localization under base extension)

[F3]

Let R be a commutative ring and let I1,…,Ir be pairwise comaximal ideals, where r≥1. Then the canonical map R⟶∏i=1rR/Ii,x⟼(x+I1,…,x+Ir) is surjective, its kernel is ⋂i=1rIi, and ⋂i=1rIi=∏i=1rIi. Equivalently, R/∏i=1rIi≅∏i=1rR/Ii. (Chinese remainder theorem for pairwise comaximal ideals)

Counterexample

1.1givenF1F2

The source has one point with field local ring C, so it is integral and connected, and the map to the one-point spectrum of R is injective. By F1 and F2 its complex base change has ring C⊗RC≅C[z]/(z2+1).

2.1F3step 1.1algebra∎

The factors z−i,z+i are comaximal because 2i is a unit. F3 gives C[z]/(z2+1)≅C×C. The two prime ideals are C×0 and 0×C; the complementary idempotents exhibit them as two disjoint nonempty clopen points. Thus the base change is reduced but disconnected and reducible, and its map to Spec⁡C is not injective.

ExampleConstruction: Literature-sourcedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

The affine plane and the pair of generic points

Example

For any field k, Ak1×kAk1≅Ak2 functorially. Nevertheless the product has distinct points over the pair of generic points: in k[x,y], both (0) and (y−x) contract to (0) in each of k[x] and k[y].

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

Let A→B and A→C be maps of commutative unital rings, allowing the zero ring. In the category of all schemes, Spec⁡B×Spec⁡ASpec⁡C≅Spec⁡(B⊗AC). The projections correspond to b↦b⊗1 and c↦1⊗c. (Affine fibre products are spectra of tensor products)

[F2]

For scheme morphisms f:X→S and g:Y→S, points of P=X×SY are in bijection with quadruples (x,y,s,r) where f(x)=g(y)=s and r∈Spec⁡(κ(x)⊗κ(s)κ(y)). The residue field at the corresponding point of P is canonically κ(r). (Points of a fibre product via residue-field tensors)

Verification

1.1givenF1algebra

F1 identifies the product ring with k[x]⊗kk[y]≅k[x,y]: the maps send x⊗1 to x and 1⊗y to y, with inverse on every polynomial specified by these images. Consequently morphisms from every k-scheme T into the plane are compatible pairs of morphisms into the two lines. Empty test schemes are allowed.

2.1F2step 1.1algebra∎

The ring k[x,y] is a domain, so (0) is prime. The quotient by (y−x) is k[x], also a domain, so this is another prime, distinct because y−x≠0. Substitution y=x is injective on either single-variable subring, giving the asserted contractions. F2 explains these as distinct residue-tensor primes over the same point pair. This works over every field, including characteristic two.

ExampleConstruction: AI-generatedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

A closed-immersion fibre is one residue point or empty

Example

For an ideal I⊆A and p∈Spec⁡A, the fibre of Spec⁡(A/I)→Spec⁡A at p is Spec⁡κ(p) if I⊆p, and is empty otherwise.

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

Let A→B be a ring map, p∈Spec⁡A, and M=A∖p, acting on B through the ring map. The fibre over p is canonically Spec⁡(B⊗Aκ(p))≅Spec⁡(M−1B/pM−1B). The residue field is κ(p)=Ap/pAp. No reduction of the tensor ring is taken. (Coordinate ring of an affine fibre)

[F2]

Let A→C be a unital ring map. For any set of variables (ti) and any ideal I⊆A[ti], (A[ti]/I)⊗AC≅C[ti]/IC[ti]. Here the extended ideal is generated by the coefficient images of all elements of I. For a multiplicative subset M⊆A, (M−1A)⊗AC≅M‾−1C. These are ring isomorphisms; no flatness, finite-generation or nonzero-ring hypothesis is required. (Presentations and localization under base extension)

Verification

1.1givenF1F2

F1 and F2 give the fibre ring (A/I)⊗Aκ(p)≅κ(p)/Iκ(p).

2.1step 1.1algebra∎

If I⊆p, every element maps to zero, so the quotient is the residue field. Otherwise some element of I maps to a nonzero element of the field, hence a unit, so the quotient is zero and its spectrum empty. These cases exhaust all ideals, including I=0 and I=A. The argument applies to any prime, not just a maximal one.

ExampleConstruction: AI-adaptedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

The ideal of a polynomial graph

Example

Let k be a field, n,m≥0, and let u:Akn→Akm be given by polynomials f1,…,fm∈k[x1,…,xn]. Its graph is the closed subscheme of Akn+m with ideal (y1−f1(x),…,ym−fm(x)).

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

For an S-morphism u:X→Y (as in def-scheme-over-base), the graph morphism is Γu=(id⁡X,u):X→X×SY, supplied by thm-fibre-products-of-schemes-exist. Its first projection is the identity and its second projection is u. The definition alone does not assert that its image is closed. (The graph morphism over a base)

[F2]

For an S-morphism u:X→Y, put H=(upr⁡X,pr⁡Y):X×SY→Y×SY. The square with top arrow Γu:X→X×SY, bottom arrow ΔY/S:Y→Y×SY, left arrow u, and right arrow H is Cartesian. (The graph is a pullback of the diagonal)

[F3]

For a ring A, closed immersions Z→Spec⁡A are, up to unique isomorphism over Spec⁡A, precisely the morphisms Spec⁡(A/I)→Spec⁡A for ideals I⊆A. (Closed immersions into affine schemes are quotient spectra)

[F4]

Let A→C be a unital ring map. For any set of variables (ti) and any ideal I⊆A[ti], (A[ti]/I)⊗AC≅C[ti]/IC[ti]. Here the extended ideal is generated by the coefficient images of all elements of I. For a multiplicative subset M⊆A, (M−1A)⊗AC≅M‾−1C. These are ring isomorphisms; no flatness, finite-generation or nonzero-ring hypothesis is required. (Presentations and localization under base extension)

Verification

1.1givenF1F3algebra

By F1 the graph has coordinate map sending xi to xi and yj to fj(x). This is a surjection to k[x1,…,xn]. In the quotient by the displayed ideal, successive substitution replaces every yj by fj(x), giving inverse ring maps with k[x]. Thus the displayed ideal is exactly the kernel; F3 makes the morphism a closed immersion.

2.1F2F4step 1.1∎

The diagonal in Akm×kAkm has ideal (zj−yj) by the same substitution calculation. Pulling it back along (upr⁡X,pr⁡Y) sends those generators to fj(x)−yj. By F2 this pullback is the graph; F4 computes its quotient ideal and gives the identical presentation. If m=0 the list is empty and the graph is the identity; if n=0 it is the point given by the constants fj.

CounterexampleConstruction: AI-generatedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

Equal fibre points can have different multiplicities

Statement refuted

False claim: the underlying set of a scheme fibre together with its pointwise residue fields determines that fibre up to scheme isomorphism. For any field k, the zero fibres of x2=t and x3=t have the same one-point underlying space and residue field k, but are not isomorphic as schemes. Their coordinate rings have k-dimensions two and three.

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

Let A→B be a ring map, p∈Spec⁡A, and M=A∖p, acting on B through the ring map. The fibre over p is canonically Spec⁡(B⊗Aκ(p))≅Spec⁡(M−1B/pM−1B). The residue field is κ(p)=Ap/pAp. No reduction of the tensor ring is taken. (Coordinate ring of an affine fibre)

[F2]

For commutative unital rings A,B, the assignment φ↦Spec⁡(φ) gives a natural bijection Hom⁡CRing(A,B)≅Hom⁡LRS(Spec⁡B,Spec⁡A). Consequently A↦Spec⁡A is a contravariant equivalence from commutative rings to affine schemes, with quasi-inverse global sections. (Affine schemes are contravariantly equivalent to commutative rings)

Counterexample

1.1givenF1algebra

F1 identifies the two fibre rings as R2=k[x]/(x2) and R3=k[x]/(x3). Each has exactly one prime, (x), since every prime contains the nilpotent class of x and the quotient is the field k. Hence their underlying spaces and pointwise residue fields agree.

2.1F2step 1.1algebra∎

The monomial classes 1,x,…,xr−1 form a k-basis of Rr by division by xr. Thus their dimensions are two and three, excluding a k-algebra isomorphism. To exclude even an abstract ring isomorphism, note that the nilradical of R2 has square zero, whereas the nilradical of R3 has nonzero square generated by x2. Every ring isomorphism preserves the nilradical and its powers. F2 therefore excludes any scheme isomorphism. These rings are nonzero and the argument works in every characteristic.

ExampleConstruction: AI-generatedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

An empty fibre from a zero tensor ring

Example

For the open immersion Spec⁡k[t,t−1]→Spec⁡k[t], the fibre at t=0 is empty: its tensor coordinate ring is the zero ring.

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

Let A→B be a ring map, p∈Spec⁡A, and M=A∖p, acting on B through the ring map. The fibre over p is canonically Spec⁡(B⊗Aκ(p))≅Spec⁡(M−1B/pM−1B). The residue field is κ(p)=Ap/pAp. No reduction of the tensor ring is taken. (Coordinate ring of an affine fibre)

[F2]

Let A→C be a unital ring map. For any set of variables (ti) and any ideal I⊆A[ti], (A[ti]/I)⊗AC≅C[ti]/IC[ti]. Here the extended ideal is generated by the coefficient images of all elements of I. For a multiplicative subset M⊆A, (M−1A)⊗AC≅M‾−1C. These are ring isomorphisms; no flatness, finite-generation or nonzero-ring hypothesis is required. (Presentations and localization under base extension)

Verification

1.1givenF1F2

F1 computes its ring as k[t,t−1]⊗k[t]k, with t acting as zero on k. By F2 this is the localization of k at the image of the powers of t.

2.1step 1.1algebra∎

In this ring t is simultaneously zero and invertible, forcing 1=tt−1=0. Thus the ring is zero and has no prime ideals, so the fibre is empty. This agrees with the absence of (t) from D(t). At t=1 the same formula instead gives k, since 1 is already invertible; the emptiness depends on the specified point.

CounterexampleConstruction: AI-adaptedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

A reduced field spectrum becomes nonreduced

Statement refuted

False claim: reducedness survives algebraic extension of the ground field. Let p be prime, k=Fp(v) with v transcendental, and L=k[u]/(up−v). Then L is a field, but L⊗kL≅L[ϵ]/(ϵp), so the base change of the reduced k-scheme Spec⁡L along the finite algebraic extension L/k is nonreduced.

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

For a field extension K/k and a k-scheme X, the inverse image of every affine open U=Spec⁡A in XK is Spec⁡(A⊗kK). These affine charts cover XK and are compatible on overlaps and with coefficient localizations. (Affine charts after extension of the ground field)

[F2]

Let A→C be a unital ring map. For any set of variables (ti) and any ideal I⊆A[ti], (A[ti]/I)⊗AC≅C[ti]/IC[ti]. Here the extended ideal is generated by the coefficient images of all elements of I. For a multiplicative subset M⊆A, (M−1A)⊗AC≅M‾−1C. These are ring isomorphisms; no flatness, finite-generation or nonzero-ring hypothesis is required. (Presentations and localization under base extension)

[F3]

Let F have characteristic p>0, let a∈F not be a pth power in F, and let n≥1. Then xpn−a is irreducible in F[x]. (If a is not a pth power in a characteristic-p field, then xpn−a is irreducible for every n≥1)

[F4]

For every field F, the polynomial ring F[x] is a unique factorisation domain. (For every field F, F[x] is a unique factorisation domain)

Counterexample

1.1givenF3F4

By F4, Fp[v] is a UFD. If v=(a/b)p for nonzero polynomials a,b, comparison of the exponent of the prime polynomial v in vbp=ap gives 1+pord⁡v(b)=pord⁡v(a), impossible modulo p. Thus v is not a pth power in k, and F3 with exponent parameter 1 proves up−v irreducible. Its quotient L is a field of degree p over k.

2.1F1F2step 1.1algebra

By F1 and F2 the base-change ring is L[z]/(zp−v). In L one has v=up, and the characteristic-p binomial formula gives zp−up=(z−u)p. Substitution ϵ=z−u supplies the claimed ring isomorphism, with inverse z=u+ϵ.

3.1step 1.1step 2.1algebra∎

The classes 1,ϵ,…,ϵp−1 form a basis by division by the monic polynomial ϵp. Since p≥2, ϵ is nonzero but nilpotent. Thus this ring, whose unique prime is (ϵ), is nonreduced whereas L is reduced. The argument includes p=2; no integer-only Eisenstein criterion is used.

ExampleConstruction: Literature-sourcedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

The self fibre product of the quadratic cover

Example

For Ak1→Ak1 given by t↦x2, the self fibre product is Spec⁡k[x,y]/(x2−y2). If char⁡k≠2 it is a reduced union of two distinct line components; if char⁡k=2 it is the doubled diagonal.

Facts & Assumptions

Given: The objects, hypotheses and conventions in the statement above.

[F1]

Let A→B and A→C be maps of commutative unital rings, allowing the zero ring. In the category of all schemes, Spec⁡B×Spec⁡ASpec⁡C≅Spec⁡(B⊗AC). The projections correspond to b↦b⊗1 and c↦1⊗c. (Affine fibre products are spectra of tensor products)

[F2]

Let A→C be a unital ring map. For any set of variables (ti) and any ideal I⊆A[ti], (A[ti]/I)⊗AC≅C[ti]/IC[ti]. Here the extended ideal is generated by the coefficient images of all elements of I. For a multiplicative subset M⊆A, (M−1A)⊗AC≅M‾−1C. These are ring isomorphisms; no flatness, finite-generation or nonzero-ring hypothesis is required. (Presentations and localization under base extension)

Verification

1.1givenF1F2

Use F1 with the two maps from k[t]. F2 eliminates t from the presentation k[x,y,t]/(t−x2,t−y2) and obtains k[x,y]/(x2−y2), with the two required projections.

2.1step 1.1algebra

If 2≠0 in k, the invertible linear substitution u=x−y,v=x+y turns the ring into k[u,v]/(uv). Every prime over (uv) contains u or v. The ideals (u),(v) are incomparable primes and their intersection is (uv) by monomial comparison. Thus the ring is reduced with exactly two line components meeting at the origin. They are not comaximal: their sum is (u,v), so this is not a product-ring decomposition.

3.1step 1.1algebra∎

If char⁡k=2, then x2−y2=(x−y)2. Put ϵ=x−y to obtain k[y,ϵ]/(ϵ2). Its class ϵ is nonzero with square zero, and its reduced support is the diagonal. Both characteristic cases give nonzero rings and exhaust all fields.

Sources