---
title: Larsen–Lunts Conjecture
url: https://www.emergentmind.com/topics/larsen-lunts-conjecture
type: topic
---

# Larsen–Lunts Conjecture

The Larsen–Lunts Conjecture is a central and deeply consequential statement in the interplay between algebraic geometry and the theory of motivic invariants, specifically concerning the relationship between the Grothendieck ring of varieties and piecewise isomorphism. Formulated in 2003, it posits that two smooth projective varieties over an algebraically closed field with the same class in the Grothendieck ring are necessarily piecewise isomorphic. This property, if true in general, would imply a powerful connection between the algebraic structure of $K_0(\mathrm{Var}_k)$ and the geometry of varieties, tying cut-and-paste phenomena to deep birational and stable birational invariants.

## 1. The Grothendieck Ring and Piecewise Isomorphism

Let $k$ denote an algebraically closed field. The Grothendieck ring $K_0(\mathrm{Var}_k)$ is the free abelian group generated by the isomorphism classes $[X]$ of reduced, separated, finite-type $k$-schemes (varieties) modulo the scissor relations:
\[
[X] = [Z] + [X \setminus Z],\quad Z \subset X \text{ closed}.
\]
Multiplication is given by
\[
[X]\cdot[Y] = [X \times Y],
\]
endowing $K_0(\mathrm{Var}_k)$ with a commutative ring structure. 

Two varieties $X$ and $Y$ are called **piecewise isomorphic** if there exist finite decompositions into locally closed subvarieties,
\[
X = \bigsqcup_{i=1}^n X_i, \quad Y = \bigsqcup_{i=1}^n Y_i, \quad X_i \cong Y_i\;\;\forall i,
\]
possibly after reindexing. Clearly, piecewise isomorphic varieties have the same Grothendieck class, but the converse, which is the assertion of the Larsen–Lunts Conjecture, is highly nontrivial and generally fails in high dimensions [2301.13040, 1411.6084, 1311.1736].

## 2. Precise Forms of the Larsen–Lunts Conjecture

**Original Formulation (2003):**
If $X,Y$ are smooth projective varieties over $k$ such that $[X]=[Y]$ in $K_0(\mathrm{Var}_k)$, then $X$ and $Y$ are piecewise isomorphic.

This can be equivalently stated as: the semiring of piecewise isomorphism classes is cancellative under disjoint union [1311.1736].

**Isomorphism Modulo Lefschetz Motive:**
There is a refined version involving the Lefschetz class $\mathbb{L} = [\mathbb{A}^1]$:
\[
K_0(\mathrm{Var}_k)/(\mathbb{L}) \cong \mathbb{Z}[\mathbf{SB}],
\]
where $\mathbf{SB}$ denotes the stable birational equivalence classes of irreducible varieties. Every smooth projective $X$ maps to its stable birational class in this setting [1506.06200].

## 3. Verified Cases and Counterexamples

Significant progress has been made for special classes:
- For **rational normal projective surfaces** admitting a desingularisation by trees of smooth rational curves (e.g., Du Val surfaces), it is proven that $[S_1]=[S_2]$ in $K_0(\mathrm{Var}_k)$ if and only if $S_1$ and $S_2$ are piecewise isomorphic [2301.13040].
- The proof in this case utilizes explicit combinatorial classification of such surfaces via $\operatorname{Pic}(X) \cong \mathbb{Z}^n$ and explicit decompositions into $\mathbb{P}^2 \sqcup \{\text{pt}\} \sqcup (\mathbb{P}^1)^{\sqcup m}$, reducing piecewise isomorphism to matching these summands.

However, **counterexamples** in higher dimension show the conjecture fails in general:
- Karzhemanov constructs smooth projective threefolds $X, \tilde X$ with $[X]=[\tilde X]$ in $K_0(\mathrm{Var}_k)$ but which are not piecewise isomorphic by any decomposition [1411.6084].
- The counterexample uses pencils of cubic surfaces, techniques from birational rigidity theory, and careful manipulation in the Grothendieck ring.

## 4. Structural and Homotopical Approaches

A homotopical framework has been developed to study the Grothendieck ring via the **spectrum of varieties** $K(\mathcal{V}_k)$, with $\pi_0 = K_0(\mathrm{Var}_k)$ and higher homotopy groups encoding deeper geometric invariants [1506.06200]. Notably:
- The spectral sequence arising from filtrations by dimension relates the relations in $K_0(\mathrm{Var}_k)$ to birational and piecewise automorphism obstructions.
- The kernel of multiplication by $\mathbb{L}$ in $K_0(\mathrm{Var}_k)$ can be explicitly represented by pairs $X, Y$ with $[X\times \mathbb{A}^1]=[Y\times \mathbb{A}^1]$ but $X\times \mathbb{A}^1$ and $Y\times \mathbb{A}^1$ not piecewise isomorphic, thus capturing the failure of $\mathbb{A}^1$-cancellation [1506.06200].

## 5. Connections to Birational Geometry and Stable Birationality

The quotient $K_0(\mathrm{Var}_k)/(\mathbb{L})$ is isomorphic to the free abelian group generated by stable birational equivalence classes. This isomorphism is closely related to the structure theorem of Larsen–Lunts [1506.06200]. The implication is that, modulo $\mathbb{L}$, all relations in the Grothendieck ring come from stable birational geometry:
- If $[X]=[Y]$ modulo $\mathbb{L}$, then $X$ and $Y$ are stably birational.
- This underpins the phenomenon that nontrivial geometric differences (e.g., types of singularities, or invariants like Kodaira dimension) are invisible in $K_0(\mathrm{Var}_k)/(\mathbb{L})$.
- Quotient singularities for $X/G$ behave according to the stable rationality of the classifying stack $BG$ [2208.14313].

## 6. Motivic Zeta Functions and Rationality Criteria

Kapranov’s motivic zeta function,
\[
Z(X, t) = \sum_{n=0}^\infty [\operatorname{Sym}^n X] t^n \in K_0(\mathcal{V}_K)[[t]],
\]
links to the Larsen–Lunts program via rationality considerations.

- For surfaces, rationality of $Z(X,t)$ in $K_0(\mathcal{V}_{\mathbb{C}})[[t]]$ characterizes negative Kodaira dimension; this criterion is conjectured to extend to all dimensions [2508.15065].
- The irrationality direction—if $Z(X,t)$ is rational, then $X$ must have negative Kodaira dimension and admit no global nonzero differential forms of even degree—now holds for arbitrary dimension [2508.15065].
- The technical advance in this context is the theory of $\mathbb{L}$-rational singularities, ensuring that all symmetric powers of a variety inherit this property, stabilizing the passage between $[\operatorname{Sym}^n X]$ and its resolutions in the Grothendieck ring modulo $(\mathbb{L})$.

## 7. Limitations, Open Problems, and Broader Implications

Despite substantial progress, the Larsen–Lunts conjecture fails in general. Known limitations and open directions include:
- General piecewise isomorphism classification is tractable only in dimension one and for certain rational surfaces; there is no general classification in higher dimension [2301.13040].
- Counterexamples hinge on the coexistence of equal Grothendieck classes with birational rigidity or other geometric obstructions to piecewise isomorphism.
- The status of the Lefschetz class $\mathbb{L}$ as a potential zero divisor in $K_0(\mathrm{Var}_k)$ remains unresolved [1311.1736].
- The interplay with Gromov’s extension problem for birational self-maps is now known to be equivalent to the Larsen–Lunts conjecture [1311.1736].
- Motivic measures and finer invariants are required to distinguish varieties beyond the detecting power of the Grothendieck ring, especially for cut-and-paste phenomena and singularities invisible in $K_0(\mathrm{Var}_k)$ or its stable-birational quotient [1411.6084].

A plausible implication is that while the Grothendieck ring is a powerful invariant, it does not, in general, encode the full piecewise or birational geometry of varieties, especially for complex or rigid structures in higher dimension. The failure of the conjecture motivates ongoing development of refined cohomological, λ-ring, and homotopical invariants within the study of motives and birational geometry.

Source: https://www.emergentmind.com/topics/larsen-lunts-conjecture