Nash Transform in Algebraic & Analytic Geometry

• Nash transform is a canonical modification of an algebraic or analytic variety that captures the limiting behavior of tangent spaces near singular points.
• It constructs a closure of the Gauss map from the smooth locus into the Grassmannian, utilizing the universal quotient bundle to record embedded tangent spaces.
• The maximal dimension (d-1) of the special fibre at isolated singularities underpins its role in desingularization and advancing singularity theory.

The Nash transform, also known as the Nash blowup, is a construction in algebraic and analytic geometry associated to singularities of varieties, particularly capturing the limiting behavior of tangent spaces near a singular locus. Given a reduced, irreducible complex analytic or algebraic variety $(X,0)\hookrightarrow(\C^N,0)$ of dimension $d\ge2$ with an isolated singularity at the origin, the Nash transform formalizes the passage from the smooth locus $X^* = X \setminus \{0\}$ to the full variety, recording at each nonsingular point the corresponding embedded tangent space in the Grassmannian $G(d,N)$. Its key feature is providing a canonical modification of $X$ that often plays a central role in the study and potential resolution of singularities.

1. Foundational Definitions and Construction

Let $X \subset \C^N$ be as above. The Gauss map $\gamma : X^* \to G(d,N)$ assigns to each smooth point $x\in X^*$ its tangent space $T_{X,x}\subset \C^N$. The Nash transform $\nu:\widehat X\to X$ is defined as the closure $d\ge2$0, with projection $d\ge2$1 to the first factor. Thus, $d\ge2$2 is an analytic (or algebraic) subset of dimension $d\ge2$3, and $d\ge2$4 is a proper bimeromorphic (or birational) morphism, which is an isomorphism over the smooth locus $d\ge2$5.

The structure of the Nash transform is further elucidated via the universal quotient bundle $d\ge2$6 of rank $d\ge2$7 on $d\ge2$8. Over $d\ge2$9, the exact sequence

$X^* = X \setminus \{0\}$0

holds, with $X^* = X \setminus \{0\}$1 locally free of rank $X^* = X \setminus \{0\}$2 and canonically isomorphic to the dual of the sheaf of relative differentials of $X^* = X \setminus \{0\}$3 over $X^* = X \setminus \{0\}$4. If $X^* = X \setminus \{0\}$5 denotes the normalization and $X^* = X \setminus \{0\}$6, then $X^* = X \setminus \{0\}$7 is a normal analytic space of dimension $X^* = X \setminus \{0\}$8.

2. Special Fibre and Limiting Tangent Spaces

The special fibre of the Nash transform at $X^* = X \setminus \{0\}$9,

$G(d,N)$0

has geometric meaning as the set of limiting tangent spaces of $G(d,N)$1 at the singularity. For any $G(d,N)$2, the fibre $G(d,N)$3 records all tangent spaces that can arise as limits of tangent spaces at nearby smooth points approaching $G(d,N)$4.

The study of this fibre, particularly its dimension and structure, is crucial in understanding the local geometry and complexity of the singularity. It serves as a central object in questions about desingularization and in the classical Nash problem on the relationship between families of arcs and exceptional divisors.

3. Main Theorem: Dimension of the Special Fibre

Hennings (Hennings, 2014) establishes that, for an irreducible isolated singularity of dimension $G(d,N)$5,

$G(d,N)$6

i.e., the special fibre of the Nash transform achieves the maximal possible dimension. Equivalently, if $G(d,N)$7 is the normalization of the Nash transform, then $G(d,N)$8 is a pure $G(d,N)$9-dimensional analytic (or algebraic) subspace.

A direct corollary is that every irreducible isolated singularity with $X$0 admits a Nash special fibre of maximal possible dimension $X$1. Hennings’ proof proceeds by contradiction: assuming that a point $X$2 has codimension at least $X$3 would force certain properties of the canonical sheaf and differentials, but these ultimately contradict the isolated nature of the singularity or basic theorems in complex geometry (such as Zariski's Main Theorem).

4. Illustrative Examples

Two primary examples illustrate the application and scope of the result:

• A$X$4 surface singularities: $X$5 for $X$6, an isolated hypersurface singularity of dimension $X$7. The Nash special fibre $X$8 is a smooth conic if $X$9 is odd, and a union of two lines if $X \subset \C^N$0 is even, both cases yielding $X \subset \C^N$1.
• Affine cones over smooth curves: For a smooth projective curve $X \subset \C^N$2, let $X \subset \C^N$3 be its affine cone in $X \subset \C^N$4. The vertex $X \subset \C^N$5 is an isolated singularity; the limiting tangent spaces fill the dual projective curve, again with $X \subset \C^N$6 for $X \subset \C^N$7.

These examples confirm that the bound is both achievable and sharp for natural classes of singularities (Hennings, 2014).

5. Broader Implications and Open Problems

The dimension result settles a question posed by Simis–Smith–Ulrich: for irreducible isolated singularities, the Nash blowup’s special fibre achieves the maximal dimension. Previous analyses for Cohen–Macaulay singularities relied on results of Kunz–Waldi, but Hennings’ proof is direct and elementary, avoiding advanced cohomological tools.

The Nash blowup is a fundamental mechanism in attempts to resolve singularities by iterative normalization of Gauss-graph closures. Guaranteeing that each intermediate special fibre has full dimension facilitates tracking the emergence and behavior of new exceptional components in the desingularization process.

A notable remark extends the result: if the singular locus $X \subset \C^N$8 has positive dimension $X \subset \C^N$9, then

$\gamma : X^* \to G(d,N)$0

a uniform lower bound for the fibre dimension in the non-isolated case.

Open questions remain on whether iterated Nash blowups resolve isolated singularities in dimension $\gamma : X^* \to G(d,N)$1, the nature of the Nash problem in higher dimensions, and connections between the multiplicities of special fibre components and the topology or arc space of the singularity. A plausible implication is that the Nash transform encodes rich geometric data critical for advances in singularity theory and related fields (Hennings, 2014).