---
title: Kapranov Motivic Zeta Function
url: https://www.emergentmind.com/topics/kapranov-motivic-zeta-function
type: topic
---

# Kapranov Motivic Zeta Function

Kapranov’s motivic zeta function is the generating series of symmetric powers of an algebraic variety in the Grothendieck ring of varieties. For a field \(K\) of characteristic zero and a quasi-projective \(K\)-variety \(X\), it is defined by
\[
Z(X,t)=\sum_{n=0}^{\infty}[\Sym^n X]\,t^n \in K_0(\Var_K)[[t]],
\]
where \(\Sym^n X=X^n/S_n\) and \([X]\) denotes the class of \(X\) in \(K_0(\Var_K)\) [2508.15065]. The construction packages the geometry of all symmetric powers into a single formal series, interacts naturally with the \(\lambda\)-ring structure on \(K_0(\Var_K)\), specializes to the classical Weil zeta function under point counting over finite fields, and serves as a testing ground for rationality phenomena in motivic geometry [1707.03479].

## 1. Definition in the Grothendieck ring

The ambient object is the Grothendieck ring \(K_0(\Var_K)\), generated by symbols \([X]\) for \(K\)-varieties, subject to the scissor relation
\[
[X]=[Z]+[X\setminus Z]
\quad\text{for } Z\subset X \text{ closed},
\]
with multiplication induced by Cartesian product. The Lefschetz class is
\[
\mathbb{L}=[\mathbb{A}^1_K].
\]
Kapranov’s zeta function is then the formal power series obtained by taking classes of symmetric powers [2508.15065].

The symmetric-power operations define a \(\lambda\)-structure on \(K_0(\Var_K)\) via
\[
\lambda^n([X])=[\Sym^n X], \qquad
\lambda_t(x)=\sum_{n\ge 0}\lambda^n(x)t^n,
\]
so that \(Z(X,t)=\lambda_t([X])\) [2508.15065]. This places the construction in the general formalism of \(\lambda\)-rings and makes it compatible with motivic measures, Witt vectors, and various realization functors.

A broader version replaces the identity map on \(K_0(\Var_K)\) by a motivic measure \(\mu:K_0\Var(k)\to R\), where \(R\) is a commutative ring. One then defines
\[
Z_\mu(X;t)=\sum_{n=0}^{\infty}\mu([\Sym^nX])\,t^n,
\]
which specializes to Kapranov’s original series when \(\mu\) is the identity [1412.1795]. This generalized viewpoint is central in later structural results.

## 2. Curves as the basic rational case

The first major rationality theorem is the curve case. Kapranov showed that if \(X\) is a smooth projective curve with a \(K\)-rational point, then \(Z(X,t)\) is rational in \(K_0(\Var_K)[[t]]\); D. Litt later removed the need for a rational point, proving rationality for any geometrically connected curve over a field of characteristic zero [2508.15065].

For a smooth projective curve \(C\) of genus \(g\), one has the explicit form
\[
Z_{\mathrm{mot}}(C;t)=\frac{P(t)}{(1-t)(1-\mathbb{L}t)},
\]
where \(P(t)\) is a polynomial of degree \(2g\) [1907.05125]. This formula is the model for much of the subsequent theory: the denominator reflects the Jacobian-type geometry of symmetric powers, while the numerator encodes the finer motive.

The smooth unmarked curve case is also the point at which later generalizations reconnect with Kapranov’s original definition. For a smooth curve \(X\),
\[
\Div^+_X(d)\cong \Sym^d X,
\]
so the divisorial zeta function introduced for marked stable curves reduces exactly to
\[
Z_{\mathrm{mot}}(X;t)=\sum_{d\ge 0}[\Sym^d X]\,t^d
\]
in this case [1907.05125]. This identifies Kapranov’s series as the base case of a broader framework encompassing nodal and marked curves.

A common misconception is that the curve case is representative of the general situation. The later literature shows the opposite: curve rationality is exceptional rather than universal.

## 3. Motivic measures, Witt vectors, and formal properties

A decisive structural advance is the interpretation of motivic zeta functions in the big Witt ring \(W(R)\). As a set, \(W(R)\) may be identified with \((1+R[[t]])^\times\), with Witt multiplication determined by the Teichmüller rule
\[
(1-at)^{-1}\star (1-bt)^{-1}=(1-abt)^{-1}
\]
for \(a,b\in R\) [1412.1795]. If \(\mu\) is a motivic measure, the assignment
\[
X\longmapsto S_\mu(X;t)\in W(R)
\]
is always a group homomorphism, and \(\mu\) is called exponentiable when it is in fact a ring homomorphism [1412.1795].

Exponentiability yields the product formula
\[
S_\mu(X\times Y;t)=S_\mu(X;t)\star S_\mu(Y;t).
\]
This is the formal mechanism behind many multiplicativity statements for motivic zeta functions [1412.1795]. In particular, if \(Z_\mu(X;t)\) and \(Z_\mu(Y;t)\) are rational elements of \(W(R)\), then \(Z_\mu(X\times Y;t)\) is also rational [1412.1795].

In the Witt-ring setting, rationality means membership in the subring \(W^{\mathrm{rat}}(R)\), equivalently the existence of polynomials \(p,q\in R[t]\) with \(p(0)=q(0)=1\) such that
\[
G=p(t)\star q(t)^{-1}.
\]
This notion differs from the denominator-clearing definition used in \(K_0(\Var_K)[1/\mathbb{L}][[t]]\), and the distinction is significant in the literature [1412.1795].

A further consequence of exponentiability is Totaro’s formula:
\[
Z_\mu(X\times \mathbb{A}^n;t)=Z_\mu(X;\mu(\mathbb{L})^n t).
\]
This underlies many computations for affine and projective bundles [1412.1795]. The same paper shows that Gillet–Soulé’s measure is exponentiable and that any measure factoring through it, including Euler characteristic, Hodge–Deligne polynomial, Poincaré polynomial, Larsen–Lunts exotic measure, and Albanese, is also exponentiable [1412.1795].

## 4. Finite-field incarnation and MacDonald-type formulas

Over a finite field \(k=\mathbb{F}_q\), the counting measure
\[
\mu_\#([X])=\#X(\mathbb{F}_q)
\]
is exponentiable, and its associated Kapranov zeta function is the classical Hasse–Weil zeta function
\[
Z(X,t)=\prod_{x\in |X|}(1-t^{\deg x})^{-1}
      =\exp\!\left(\sum_{r=1}^{\infty}\#X(\mathbb{F}_{q^r})\,\frac{t^r}{r}\right)
      \in W(\mathbb{Z})
\]
[1707.03479]. Thus Kapranov’s construction is not merely analogous to the classical zeta function; under point counting it becomes exactly that function.

The Witt-vector formalism makes possible a MacDonald-type formula for the zeta functions of symmetric powers. If
\[
Z(X,t)=\sum_{i,j}(-1)^{i+1}[\alpha_{ij}]
\]
in \(W(\mathbb{Z})\), then in \(W(W(\mathbb{Z}))\) one has
\[
\sum_{n=0}^{\infty} Z(\Sym^n X,t)\,u^n
=
\sum_{i,j}(-1)^{i+1}[[[\alpha_{ij}]]],
\]
where \([[[\alpha]]]=(1-[\alpha]u)^{-1}\) is the double Teichmüller lift [1707.03479]. This realizes the generating series of the zeta functions of all symmetric powers as a closed Witt-theoretic expression.

The formalism admits explicit computations. For affine space,
\[
Z(\mathbb{A}^m,t)=\frac{1}{1-q^m t}=[q^m],
\]
and for projective space,
\[
Z(\mathbb{P}^m,t)=\frac{1}{\prod_{i=0}^m(1-q^i t)}=\sum_{i=0}^m[q^i]
\]
in \(W(\mathbb{Z})\) [1707.03479]. For an elliptic curve \(E/\mathbb{F}_q\) with
\[
Z(E,t)=\frac{(1-\alpha t)(1-\beta t)}{(1-t)(1-qt)},
\]
the Witt expression is
\[
Z(E,t)=[1]-[\alpha]-[\beta]+[q],
\]
and the symmetric-power generating series is obtained by replacing each Teichmüller term by its double lift [1707.03479].

These formulas are a finite-field analogue of the role played by Macdonald-type identities in topology and cohomology. This suggests that Kapranov’s zeta function is best understood as a universal symmetric-power generating series whose realizations recover more classical enumerative invariants.

## 5. Rationality, irrationality, and geometric constraints

The rationality problem for Kapranov’s zeta function is subtle because \(K_0(\Var_K)\) is not even an integral domain. One standard convention is that
\[
Z_X(t)=\sum_{n=0}^{\infty}[\Sym^n X]\,t^n
\]
is rational over \(K_0(\Var_K)[1/\mathbb{L}]\) if there exists a polynomial
\[
B(t)=1+b_1 t+\cdots+b_N t^N
\]
such that
\[
B(t)\,Z_X(t)\in K_0(\Var_K)[1/\mathbb{L}][t]
\]
[1802.03661]. In this form, the question was sharpened by the Denef–Loeser conjecture, which predicted rationality after inverting \(\mathbb{L}\) for every variety in characteristic zero [1802.03661].

Larsen and Lunts disproved this expectation by constructing a K3 surface \(X/\mathbb{Q}\) such that
\[
Z_X(t)=\sum_{n=0}^{\infty}[\Sym^n X]\,t^n
\]
is not rational in \(K_0(\Var_{\mathbb{Q}})[1/\mathbb{L}][[t]]\) [1802.03661]. Their argument uses a refined motivic measure
\[
\nu_\ell:K_0(\Var_{\mathbb{Q}})\to K_0^{\mathrm{split}}(G_{\mathbb{Q}(\zeta_\ell)})
\]
built from mod \(\ell\) étale cohomology, together with the Göttsche identity
\[
\sum_{n=0}^{\infty}[X^{[n]}]\,t^n
=
\prod_{i=1}^{\infty} Z_X(\mathbb{L}^{i-1}t),
\]
and a failure of the linear recurrences that rationality would force [1802.03661]. The result is the first unconditional counterexample to the Denef–Loeser conjecture in characteristic \(0\) [1802.03661].

For surfaces over \(\mathbb{C}\), the Larsen–Lunts criterion states: if \(X\) is a smooth projective surface, then \(Z(X,t)\) is rational in \(K_0(\Var_{\mathbb{C}})[[t]]\) if and only if the Kodaira dimension \(\kappa(X)\) is negative [2508.15065]. Recent work extends the irrationality direction to higher dimension: for a smooth complex projective variety \(X\) of dimension \(d>1\), rationality of \(Z(X,t)\) over \(K_0(\Var_{\mathbb{C}})\) implies
\[
\kappa(X)<0
\quad\text{and}\quad
H^0(X,\Omega_X^{2i})=0 \ \text{for every } i>0
\]
[2508.15065]. Thus nonnegative Kodaira dimension or the existence of nonzero even-degree differential forms obstruct rationality.

The same work proves that if a variety has \(\mathbb{L}\)-rational singularities, then all of its symmetric powers also have \(\mathbb{L}\)-rational singularities [2508.15065]. A plausible implication is that the singularity theory of \(\Sym^n X\) is not merely auxiliary: it is structurally tied to the rationality problem for \(Z(X,t)\).

## 6. Extension to marked stable curves

A natural generalization of Kapranov’s construction to singular and marked curves is the divisorial motivic zeta function of Brandt–Ulirsch. Let \(k\) be algebraically closed of characteristic zero, and let \((X,p_1,\dots,p_n)\) be a pointed quasiprojective curve. For each \(d\ge 0\), \(\Div^+_{(X,\vec p)}(d)\) is the coarse moduli scheme of tuples
\[
(X'\to \Spec k,\ p_1',\dots,p_n',\ D)
\]
such that \(X'\) is a connected nodal curve whose stabilization is \(X\), the \(p_i'\) lie over the marked points, \(D\) is an effective Cartier divisor of degree \(d\) avoiding nodes and marked points, and \(K_{X'}+\epsilon D+p_1'+\cdots+p_n'\) is ample with \(\epsilon=1/d\). One then sets
\[
Z_{\mathrm{div}}(X,\vec p;t)=\sum_{d=0}^{\infty}[\Div^+_{(X,\vec p)}(d)]\,t^d
\in 1+t\,K_0(\Var/k)[[t]]
\]
[1907.05125].

When \(X\) is smooth and unmarked, the identification
\[
\Div^+_X(d)\cong \Sym^d X
\]
shows that
\[
Z_{\mathrm{div}}(X;t)=\sum_{d\ge 0}[\Sym^d X]\,t^d=Z_{\mathrm{mot}}(X;t),
\]
so Kapranov’s motivic zeta function appears as the smooth case of the divisorial theory [1907.05125].

For any stable marked quasiprojective curve, \(Z_{\mathrm{div}}(X,\vec p;t)\) is rational [1907.05125]. The proof stratifies \(\Div^+_{(X,\vec p)}(d)\) by dual-graph combinatorics and multidegrees, identifies each stratum as a product of symmetric powers of smooth loci and copies of \(\mathbb{G}_m\), and then glues nodes to factor the zeta function into componentwise contributions with explicit correction terms [1907.05125].

If \(G=(V,E)\) is the dual graph of the stable curve \((X,\vec p)\), and \(\widetilde X_v\) is the normalization of the component corresponding to \(v\in V\), the master formula is
\[
Z_{\mathrm{div}}(X,\vec p;t)
=
\left(\frac{1-\mathbb{L}t}{1-\mathbb{L}t-t+t^2}\right)^{|E|+n}
(1-t)^{2|E|+n}
\prod_{v\in V} Z_{\mathrm{mot}}(\widetilde X_v;t).
\]
Here \(|E|\) is the number of nodes and \(n\) is the number of marked points [1907.05125]. Each factor
\[
\frac{1-\mathbb{L}t}{1-\mathbb{L}t-t+t^2}
\]
comes from introducing a new node marked by two halves, and each factor \((1-t)\) comes from closing a marked point to a node [1907.05125].

This formula makes precise how Kapranov’s zeta function changes under semistable degeneration. In particular, if \(|E|=n=0\), the correction factors disappear and one recovers the smooth case immediately. The example
\[
Z_{\mathrm{div}}(\mathbb{G}_m;t)=Z_{\mathrm{mot}}(\mathbb{G}_m;t)=\frac{1-t}{1-\mathbb{L}t}
\]
shows that the generalized theory still admits explicit calculations [1907.05125]. More broadly, it exhibits Kapranov’s zeta function as the smooth vertex-wise factor in a dual-graph factorization for nodal and marked curves.

Source: https://www.emergentmind.com/topics/kapranov-motivic-zeta-function