Contraction Morphism¶

Map a projective variety onto another without a nontrivial finite Stein factor, so the fibers remain connected.

Version: v1 · 2026-10-04 · History
Domain-specific #: 13722
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomain: Algebraic Geometry → Mathematics
Aliases: Algebraic contraction

Core Idea¶

In the projective-variety convention used here, a contraction morphism is a surjective projective morphism f:X→Y between normal projective varieties for which f_*O_X = O_Y. The sheaf equality says that regular functions over the target are not supplemented by a nontrivial finite intermediate algebra from the source. In the proper Stein-factorization setting, it leads to connected geometric fibers. The definition is algebraic: the fact that a drawing shows something “collapsed” is not enough.^[1]^[2]

Contractions may be birational, such as the blow-down of an exceptional curve, or fiber type, such as projection of a projective-line bundle over a curve. A contraction need not be indexed by a canonical-class-negative extremal ray; that is an important special construction in the minimal model program, not the general identity.^[3]^[4]

Structural Signature¶

Sig role-phrases:

Projective source and target — Algebraic varieties X and Y are considered under a stated normal/projective convention.
Surjective projective morphism — Every target point is reached by an algebraic map with proper projective structure.
Structure-sheaf equality — The condition f_*O_X = O_Y is the defining algebraic test.
Connected geometric fibers — The Stein picture has no disconnected finite stage after the contraction; this follows in the cited proper setting.^[1]
Possible exceptional locus — A birational contraction can send a curve or divisor to a lower-dimensional locus, but this is not required in fiber-type examples.
Optional extremal geometry — Cone-of-curves faces and canonical-class negativity enter when the contraction is selected by a minimal-model-program theorem.^[4]

What It Is Not¶

Not merely any surjective morphism. A nontrivial finite cover is surjective and projective but generally has a larger direct-image algebra than O_Y.
Not only a blow-down. The projection P¹×C→C has a full P¹ fiber over every base point rather than one exceptional curve.^[3]
Not unrestricted equivalence with set-theoretically connected fibers. The pushforward condition is the defining one here. Geometric connectedness is an associated conclusion in the proper Stein setting; converse formulations need their exact normality, field and scheme hypotheses.^[1]
Not automatically an MMP step. Contracting a K-negative extremal ray requires the relevant cone and contraction theorems.^[4]
Not a rational map with indeterminacy. It is an everywhere-defined morphism under this convention.

Scope of Application¶

The structural role of contractions is clarified by Stein factorization. The Stacks Project states that a proper morphism factors through an intermediate Y' so that the first map has connected geometric fibers and pushforward sheaf O_{Y'}, while Y'→Y is finite. Thus one can separate the connected part of a map from a finite part. For a map already satisfying f_*O_X=O_Y, the intermediate finite algebra is trivial in the indicated sense.^[1]

On surfaces, blowing down a suitable (-1)-curve gives a birational contraction: a curve is sent to a point and the map is an isomorphism elsewhere. Ruled surfaces give a different shape. A projection P¹×C→C contracts every P¹ fiber to its base point, so it is fiber type and not birational. Minimal-model work uses still more restrictive contractions of specified curve classes, subject to theorem hypotheses.^[2]^[3]^[4]

Clarity¶

Consider Bl_p(P²)→P², the blow-down of the exceptional curve above a point p. Most target points have one-point fibers; over p the fiber is the exceptional P¹. This illustrates why one curve can be “contracted” without the map splitting into a nontrivial finite covering. By contrast, a generic degree-two finite cover has two points over most target points and an enlarged pushforward algebra; it is not the connected Stein part.^[2]^[1]

Now consider P¹×C→C. Every fiber is a copy of P¹. There is no unique exceptional curve and the map lowers dimension everywhere. Yet regular functions on a projective line fiber are constant, which is the intuition behind the pushforward condition in this standard bundle case. The two examples show that “contraction” is not synonymous with “birational blow-down.”^[3]

Manages Complexity¶

Many projective maps blend connected fibers with finite multiplicity or covering behavior. The sheaf condition and Stein factorization separate those two structures: first account for the connected-fiber part, then the finite target map. This lets geometers compare birational and fiber-type maps under a common algebraic criterion without conflating them.^[1]

The simplification carries a proof obligation. A visual collapse does not establish f_*O_X=O_Y; connectedness of underlying point sets without the proper scheme/field qualifications may not establish it either. The source and target category, field, normality and morphism hypotheses must be stated before transferring theorems.

Abstract Reasoning¶

Suppose a surjective projective f:X→Y has a nontrivial finite intermediate Y' in its Stein factorization. Then f has not yet isolated the contraction part: the finite map Y'→Y still contributes additional algebraic information. If the intermediate target already agrees with Y and the structure-sheaf equality holds, that finite stage disappears. This is the meaning of the algebraic test rather than simply counting geometric points in a schematic fiber.^[1]

In minimal-model arguments, one may seek a map that sends curves in a specified extremal ray to points. Such a theorem can produce a contraction with a special numerical characterization. But beginning with an arbitrary f_*O_X=O_Y map does not by itself identify a K-negative ray or prove it is a permitted step of a minimal model program.^[4]

Knowledge Transfer¶

The Stein-factor criterion transfers between surface blow-downs and projective-bundle projections: both have a surjective projective morphism with no nontrivial finite remainder in the stated setting. What changes is fiber dimension, birationality and exceptional-locus structure. MMP consequences transfer only after checking the relevant canonical divisor, cone and existence assumptions.^[1]^[3]^[4]

Examples¶

Blow-down of an exceptional curve¶

On a suitable smooth projective surface, a (-1)-curve can be contracted to a point. The resulting morphism is birational and is an isomorphism away from the exceptional curve. The pushforward sheaf agrees with the normal target's structure sheaf in the standard blow-down setting.^[2]^[3]

Mapped back: Source → blown-up surface; target → original normal surface; map → projective birational blow-down; sheaf test → no nontrivial finite Stein remainder; fibers → exceptional curve at one point, singletons elsewhere.

Projection of a ruled surface¶

For a smooth projective curve C, the projection P¹×C→C is projective and has connected P¹ fibers. It models a fiber-type contraction rather than a birational one; the base receives no extra regular functions from the projective-line fibers.^[3]^[1]

Mapped back: Source → projective ruled surface; target → base curve; map → projective surjection; sheaf test → fiberwise regular functions descend; fibers → P¹ over every base point.

Structural Tensions¶

The defining sheaf equality is an exact mathematical condition, not an optimization with opposed costs. A visible exceptional curve is neither required nor sufficient as a general test: a ruled-surface projection has positive-dimensional fibers throughout, while the proper Stein-factorization theorem identifies the connected geometric-fiber part through its sheaf and finite-factor conditions under the theorem's locally Noetherian hypotheses.^[1]^[3] Likewise, an extremal-ray contraction is a specially justified MMP instance, not the desirable side of a tradeoff against ordinary contractions.^[4]

The useful decision is therefore classificatory: verify f_*O_X=O_Y in the declared projective-variety setting, then separately ask whether the numerical cone and negativity hypotheses of an MMP theorem apply. Calling these proof obligations a “tension” would suggest that one may exchange correctness for convenience; one may not.

Structural–Framed Character¶

This is a domain-specific algebraic-map identity. Its stable core is the projective surjection with the structure-sheaf equality; connected-fiber geometry and Stein factorization explain the core's effect. Surface and MMP examples instantiate the identity but do not replace its defining algebraic test.

The map condition is a formal structural relation, not a favorable judgment about simplifying a space. It depends on mathematicians' definitions of projective morphism and structure sheaf for its formulation, but not on a policy institution or a particular construction practice; once the objects and map are fixed, the condition is determinate. “Contraction” travels to analysis and other fields, yet their shortened-distance uses cannot be imported into this sheaf-theoretic criterion. Its character: a highly structural identity with a precise algebraic-geometric frame, not a general contraction prime.

Structural Core vs. Domain Accent¶

Skeletal relation. A map has a connected structural part with no nontrivial finite factor left between it and the target.

Domain-bound condition. Structure sheaves, projective morphisms, geometric fibers and finite Stein factors make that relation precise in algebraic geometry.

Prime bar. The word “contraction” appears elsewhere, but this sheaf-theoretic projective morphism has not been shown to be the same cross-domain abstraction.

Parent check. The broader genus is Morphism of Algebraic Varieties: every contraction under this convention is such a map, with projectivity, surjectivity and the sheaf equality added. A cross-domain connected-factor-map skeleton is a future question, not an assumed existing prime.

This entry is a kind of Morphism of algebraic varieties.

Contraction morphism is a special case of Morphism of Algebraic Varieties: an algebraic morphism that is projective and surjective and satisfies f_*O_X=O_Y. Morphism of Schemes is the broader categorical setting under a more general convention, and Projective Bundle supplies one family of examples.

Relationships to Other Abstractions¶

Current abstraction Contraction Morphism Domain-specific

Parents (1) — more general patterns this builds on

Contraction Morphism is a kind of Morphism of algebraic varieties Domain-specific

A projective contraction is a variety morphism with surjectivity and structure-sheaf conditions.

Hierarchy path (1) — routes to 1 parentless root

Contraction Morphism → Morphism of algebraic varieties → Function (Mapping)

Neighborhood in Abstraction Space¶

Contraction Morphism sits in a sparse region of the domain-specific corpus (62^nd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Sheaves & Birational Geometry (22 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With¶

Stein factorization decomposes a proper map and helps recognize the contraction part; it is not the map itself. Finite cover may be projective and surjective but generally retains a nontrivial finite stage. Blow-up constructs an exceptional locus; its blow-down can be a contraction. MMP extremal contraction is a specially certified instance, not an alternate definition of all contraction morphisms.^[1]^[4]

References¶

[1] The Stacks Project, Theorem 76.36.4, “Stein factorization; Noetherian case”, clauses 1–4. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k

[2] Patrick Brosnan, “Castelnuovo’s Contractibility Criterion”, Algebraic Surfaces notes, blow-down construction. registry ↩a ↩b ↩c ↩d

[3] Ravi Vakil, Complex Algebraic Surfaces, Class 9, ruled-surface and Castelnuovo-contraction sections. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h

[4] Sean Keel and James McKernan, “Contractible Extremal Rays on M̄₀,ₙ”, 1996, §1 pp. 1–2 and §2.1 p. 4, on cone/contraction theorems and theorem-conditioned log-extremal-ray contractions. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h