We let
$${\mathfrak{P}}_{1}^{\u266f}(X,F)$$
be the set of Gmetrics
$P$
on
$F$
whose base locus
${\mathcal{Z}}_{P}$
is empty, i.e.,
$P$
has no nontrivial zeros. A metric
$P\in {\mathfrak{P}}_{1}^{\u266f}(X,F)$
will also be called a strict Gmetric. Eventually we will be interested in Gmetrics with nontrivial zero loci. An important generalization of strict Gmetric is the notion of a Gmetric
$P$
with only basic zeros.
Definition 2.4.
We say that a point
$x\in X$
is a basic zero of
$P\in {\mathfrak{P}}_{1}(X,F)$
if
$x\in {\mathcal{Z}}_{P}$
and for any
$s\in {V}_{P}$
the function
$$\begin{array}{ccc}{F}^{*}\supset {\pi}^{1}\left(U\right)\ni v\mapsto \frac{\langle s,v\rangle {}^{2}}{P(v,\overline{v})}& & \end{array}$$ 
(4)

is bounded in some neighborhood
$U$
of
$x$
in
$X$
. (Here
$\pi :{F}^{*}\to X$
denotes the projection map.) We denote by
${\mathfrak{B}}_{P}\subset X$
the locus of basic zeros of
$P$
. We say that a Gmetric has only basic zeros if the function in ( 4 ) is bounded on
$X$
. The set of Gmetrics for a line bundle
$F\to X$
having only basic zeros will be denoted
$${\mathfrak{P}}_{1}^{\u266e}(X,F).$$
Remark.
Observe that the function defined by ( 4 ) depends only on the fiber of
$F$
containing
$v$
, and thus gives rise to a welldefined function on
$X$
. We will denote this function by
$s{}^{2}/P$
.
Clearly
${\mathfrak{B}}_{P}\subset {\mathcal{Z}}_{P}$
. Unlike
${\mathcal{Z}}_{P}$
,
${\mathfrak{B}}_{P}$
is always an analytic set. It is also clear that
$P$
has only basic zeros if and only if
${\mathfrak{B}}_{P}={\mathcal{Z}}_{P}$
.
Remark.
Observe that a Gmetric
$P$
has only basic zeros if and only if the function
$v\mapsto \mathit{\forall}P\mathit{\forall}(v,\overline{v})/P(v,\overline{v})$
is bounded on
${F}^{*}$
. There is an alternative way of phrasing the condition that
$P$
has only basic zeros. If we take a distinguished basis
$\{{f}^{1},...,{f}^{k},{g}^{1},...,{g}^{\ell}\}$
of
${V}_{P}$
, then
$P$
has only basic zeros if and only if there is a constant
$\lambda \in [0,1)$
such that, with
$f=({f}^{1},...,{f}^{k})$
and
$g=({g}^{1},...,{g}^{\ell})$
,
$$\left\rightg{}^{2}\le \lambda \leftf\right{}^{2}.$$
Note that since
$P=\left\rightf{}^{2}\leftg\right{}^{2}$
, this way of stating that
$P$
has only basic zeros has the advantage that automatically
$P\in {\mathfrak{P}}_{1}(X,F)$
.
Factoring basic zeros in codimension 1
Remark.
If all of the sections lying in the support space
${V}_{P}$
of a Gmetric
$P\in {\mathfrak{P}}_{1}(X,F)$
vanish on a divisor
$D\subset X$
, then one can replace
$F$
with the line bundle
$F\otimes {L}_{D}^{*}$
and
$P$
with the Gmetric
$\stackrel{~}{P}=P/{s}_{D}{}^{2}$
. By taking
$D$
to be the largest divisor in
$X$
such that every section of
${V}_{P}$
vanishes on
$D$
, we would obtain in this way a Gmetric
$\stackrel{~}{P}$
whose basic zero locus has no codimension 1 component.
Thus we will assume from here on out, without loss of generality, that
All of our Gmetrics have no codimension 1 basic zero sets.
Associated singular Hermitian metric
To every Gmetric
$P\in {\mathfrak{P}}_{1}(X,F)$
we can associate a singular Hermitian metric
${e}^{{\phi}_{P}}$
of
$F$
, defined as follows. If
$\xi $
is a local section of
${F}^{*}$
with no zeros, then
$${\phi}_{P}\left(x\right):=logP({\xi}_{x},{\xi}_{x}).$$
Gcurvature
Given a Gmetric
$P\in {\mathfrak{P}}_{1}(X,F)$
, we can define its Gcurvatures (see [
DV]
) as follows. For each integer
$N\ge 1$
, the
${N}^{\text{th}}$
Gcurvature of
$P$
along
$({v}_{1},...,{v}_{N})$
is the Hermitian matrix
$${\Theta}_{N}P({v}_{1},...,{v}_{N}):=\left(\begin{array}{ccc}P({v}_{1},{\overline{v}}_{1})& \cdot \cdot \cdot & P({v}_{1},{\overline{v}}_{N})\\ \cdot & \cdot \cdot \cdot & \cdot \\ \cdot & \cdot \cdot \cdot & \cdot \\ \cdot & \cdot \cdot \cdot & \cdot \\ P({v}_{N},{\overline{v}}_{1})& \cdot \cdot \cdot & P({v}_{N},{\overline{v}}_{N})\\ \end{array}\right)$$
We let
$${\mathfrak{P}}_{N}(X,F)$$
denote the set of Gmetrics
$P$
such that for any
${v}_{1},...,{v}_{N}\in {F}^{*}$
the curvature matrix
$${\Theta}_{N}P({v}_{1},...,{v}_{N})$$
is positive semidefinite. Clearly
$${\mathfrak{P}}_{N+1}(X,F)\subseteq {\mathfrak{P}}_{N}(X,F).$$
We write
$${\mathfrak{P}}_{\infty}(X,F):={\bigcap}_{N\in \mathbb{N}}{\mathfrak{P}}_{N}(X,F).$$
In [
DV]
it is shown that

(i)
${\mathfrak{P}}_{\infty}(X,F)$
consists of all those Gmetrics
$P$
whose associated Hermitian form
$\mathsf{\u212d}$
is positive semidefinite, and

(ii)
there exists
${N}_{0}$
(depending on
$F$
) such that
${\mathfrak{P}}_{{N}_{0}}(X,F)={\mathfrak{P}}_{\infty}(X,F)$
. (This conclusion uses the compactness of
$X$
.)
In this paper we will only be concerned with the classes
${\mathfrak{P}}_{1}$
,
${\mathfrak{P}}_{2}$
and
${\mathfrak{P}}_{\infty}$
.
Sums of squared norms and quotients of squared norms
As we have just said,
${\mathfrak{P}}_{\infty}(X,F)$
consists of those Gmetrics
$P$
whose associated Hermitian form is positive semidefinite. It follows that there is a basis
$\{{s}^{1},...,{s}^{k}\}$
of
${V}_{P}$
such that
$$P={\sum}_{j=1}^{k}{s}^{j}{}^{2}.$$
For this reason, we refer to the elements of
${\mathfrak{P}}_{\infty}$
as sums of squared norms.
Definition 2.5.
A Gmetric
$P\in {\mathfrak{P}}_{1}(X,F)$
is said to be a quotient of squared norms if there exist sums of squared norms
${P}_{1}\in {\mathfrak{P}}_{\infty}(X,{F}_{1})$
and
${P}_{2}\in {\mathfrak{P}}_{\infty}(X,{F}_{2})$
such that
$${F}_{1}=F\otimes {F}_{2}and{P}_{1}=P\otimes {P}_{2}.$$
The set of Gmetrics on
$F$
that are quotients of squared norms will be denoted
$\mathcal{Q}(X,F)$
.
In the language of this section, we can rephrase the following beautiful result of Catlin and D'Angelo [
CD2]
.
Theorem 2.6.
For any holomorphic line bundle
$F\to X$
,
$${\mathfrak{P}}_{1}^{\u266f}(X,F)\subset \mathcal{Q}(X,F).$$
Theorem 2.6 is a consequence of Theorem 2.8 given below. To state the latter, we need the notion of Global CauchySchwarz metrics.
Global CauchySchwarz metrics
The elements of the class
${\mathfrak{P}}_{2}$
are also called Global CauchySchwartz metrics, or simply GCSmetrics, because the class
${\mathfrak{P}}_{2}$
is easily seen to consist precisely of those Gmetrics
$R$
that satisfy
$$\leftR\right(v,\overline{w}){}^{2}\le R(v,\overline{v}\left)R\right(w,\overline{w}).$$
D'Angelo [
D2]
proved that if
$R\in {\mathfrak{P}}_{2}(X,E)$
then the associated singular Hermitian metric
${e}^{{\phi}_{R}}$
of
$E$
has semipositive curvature current.
Definition 2.7.
Let
$E\to X$
be a holomorphic line bundle. A Gmetric
$R\in {\mathfrak{P}}_{1}(X,E)$
is said to be a strong GCS metric (or simply SGCS metric) if the following conditions hold.

(S1)
$R\in {\mathfrak{P}}_{1}^{\u266f}(X,E)\cap {\mathfrak{P}}_{2}(X,E)$
.

(S2)
The matrix
${\Theta}_{2}R(v,w)$
degenerates if and only if either
$v=0$
,
$w=0$
, or
$v$
and
$w$
lie in the same fiber of
$E$
.

(S3)
The (smooth) metric
${e}^{{\phi}_{R}}$
of
$E$
associated to
$R$
has strictly positive curvature.
The class of SGCS metrics on
$E$
will be denoted
${\mathfrak{P}}_{2}^{S}(X,E).$
Definition 2.7 is due essentially to Catlin and D'Angelo [
CD2]
. The main difference is that they do not include property (S3) in the definition, but make it a hypothesis in the following theorem.
Theorem 2.8.
(CatlinD'Angelo [
CD2]
) Let
$P\in {\mathfrak{P}}_{1}^{\u266f}(X,F)$
and
$R\in {\mathfrak{P}}_{2}^{S}(X,E)$
. Then there is an integer
${N}_{0}$
such that for all
$m\ge {N}_{0}$
,
$${R}^{m}\otimes P\in {\mathfrak{P}}_{\infty}(X,{E}^{m}\otimes F).$$
Remark.
By applying Theorem 2.8 with
$P=1\in {\mathfrak{P}}_{1}^{\u266f}(X,{\mathcal{O}}_{X})$
we see that for all sufficiently large
$m$
,
${R}^{m}\in {\mathfrak{P}}_{\infty}(X,{E}^{m})$
, and thus Theorem 2.6 follows.
To prove Theorem 1 , we will need a version of the notion of strong GCS metric that is better suited to dealing with resolution of singularities. We now introduce such a notion.
Definition 2.9.
Let
$E\to X$
be a holomorphic line bundle. A Gmetric
$R\in {\mathfrak{P}}_{1}(X,E)$
is said to be a pseudostrong GCS metric (or simply
$\Psi $
GCS metric) if in addition to satisfying (S1) and (S2) of definition 2.7 ,
$R$
also satisfies the following condition.

(
$\Psi $
)
The (smooth) metric
${e}^{{\phi}_{R}}$
associated to
$R$
has strictly positive curvature away from an analytic subvariety of
$X$
.
The class of
$\Psi $
GCS metrics on
$E$
will be denoted
${\mathfrak{P}}_{2}^{\Psi}(X,E)$
.
Remark.
We emphasize that because a
$\Psi $
GCS metric has no nontrivial zeros, its curvature is a smooth
$(1,1)$
form that, by assumption, is positive away from an analytic subset of
$X$
.
The following examples give a bit of a flavor for what is going on.
Example 2.10.
It was observed in [
DV]
that
${\mathfrak{P}}_{2}({\mathbb{P}}_{n},\mathbb{H})={\mathfrak{P}}_{\infty}({\mathbb{P}}_{n},\mathbb{H})$
, and thus it follows that
${\mathfrak{P}}_{2}^{\Psi}({\mathbb{P}}_{n},\mathbb{H})={\mathfrak{P}}_{2}^{S}({\mathbb{P}}_{n},\mathbb{H})$
.
Example 2.11.
Let
$E\to X$
be a very ample line bundle. Choose sections
${s}^{1},...,{s}^{k}\in {H}^{0}(X,E)$
whose common zero locus is empty, and such that
$$\sqrt{1}\partial \overline{\partial}log\sum {s}^{j}{}^{2}$$
is a strictly positive
$(1,1)$
form. For example, one could take
${s}^{1},...,{s}^{k}$
to be a basis of
${H}^{0}(X,E)$
. Then the Gmetric
$$R:={\sum}_{j=1}^{k}{s}^{j}{}^{2}\in {\mathfrak{P}}_{1}^{\u266f}(X,E)$$
lies in
${\mathfrak{P}}_{2}^{S}(X,E)\subset {\mathfrak{P}}_{2}^{\Psi}(X,E)$
.
Nevertheless in this paper metrics do arise that are
$\Psi $
GCS but not SGCS. These metrics come up due to our use of the resolution of singularities.
Example 2.12.
There exist
$\Psi $
GCS metrics that are not SGCS metrics. To see this, fix a line bundle
$E\to X$
and an SGCS metric
$R\in {\mathfrak{P}}_{2}^{\u266f}(X,E)$
. Let
$\mu :\stackrel{~}{X}\to X$
be the blowup of
$X$
at one point. Then
${\mu}^{*}R\in {\mathfrak{P}}_{2}^{\Psi}(\stackrel{~}{X},{\mu}^{*}E)$
but
${\mu}^{*}R\notin {\mathfrak{P}}_{2}^{S}(\stackrel{~}{X},{\mu}^{*}E)$
, since the curvature of
${\mu}^{*}R$
degenerates on the exceptional divisor of
$\mu $
.
Remark.
Although we deal with Gmetrics that have zero loci in this paper, we always resolve these metrics and divide out the resulting zeros. Ultimately, the analytic method we develop in the next section works only for those Gmetrics whose unit disk bundle is bounded.
3 Asymptotics of powers of
$\Psi $
GCS metrics
In this section we prove the following theorem.
Theorem 3.
Let
$X$
be a complex projective algebraic manifold and
$F\to X$
,
$E\to X$
holomorphic line bundles. Suppose given two Gmetrics
$P\in {\mathfrak{P}}_{1}^{\u266f}(X,F)$
and
$R\in {\mathfrak{P}}_{2}^{\Psi}(X,E)$
. Then there is an integer
$N$
such that for all
$m\ge N$
,
$${R}^{m}\otimes P\in {\mathfrak{P}}_{\infty}(X,{E}^{m}\otimes F).$$
Remark.
Since
$R\in {\mathfrak{P}}_{2}^{S}(X,E)\subset {\mathfrak{P}}_{2}^{\Psi}(X,E)$
, Theorem 3 generalizes the CatlinD'Angelo Theorem 2.8 above.
The technical approach used in this section is inspired by the paper [
B]
.
An outline of the approach
Let
$R$
and
$P$
be as in Theorem 3 , and denote by
${e}^{\phi}$
and
${e}^{\psi}$
the associated smooth Hermitian metrics of
$R$
and
$P$
respectively.
Theorem
3 claims that for all sufficiently large
$m$
, the Hermitian matrix associated to
${R}^{m}\otimes P$
(in some and hence any basis of
${H}^{0}(X,{E}^{m}\otimes F)$
) is positive definite. To get our hands on the matrix of
${R}^{m}\otimes P$
, we invoke the method of orthogonality: if we consider the Hilbert space
${\mathfrak{A}}^{2}\left({e}^{(m\phi +\psi )}\right)$
, and represent
${R}^{m}\otimes P$
in terms of an orhonormal basis
$\left\{{s}^{\alpha}\right\}$
by
$${R}^{m}\otimes P={\mathsf{\u212d}}_{\alpha \overline{\beta}}^{\left(m\right)}{s}^{\alpha}{\overline{s}}^{\beta},$$
then with
$\pi :{E}^{m}\otimes F\to X$
denoting the projection map,
$${\mathsf{\u212d}}_{\alpha \overline{\beta}}^{\left(m\right)}={\int}_{X}{\int}_{X}\frac{{R}^{m}\otimes P(z,\overline{w}){s}^{\beta}\left(w\right)\overline{{s}^{\alpha}\left(z\right)}}{{R}^{m}\otimes P(z,z){R}^{m}\otimes P(w,w)}\Omega \left(\pi z\right)\Omega \left(\pi w\right).$$
(Note that although
$z$
and
$w$
are points in
$({E}^{m}\otimes F{)}^{*}$
, the integrand is bihomogeneous of degree
$0$
in
$({E}^{m}\otimes F{)}^{*}\times (({E}^{m}\otimes F{)}^{*}{)}^{\u2020}$
, and thus well defined on
$X\times X$
.) Now, the Bergman kernel
${K}_{{E}^{m}\otimes F}$
is the Gmetric in
${\mathfrak{P}}_{\infty}(X,{E}^{m}\otimes F)$
associated to the identity matrix. To prove that
${\mathsf{\u212d}}^{\left(m\right)}$
is positive definite, one could try to show that as
$m\to \infty $
,
${R}^{m}\otimes P$
is asymptotic to
${m}^{n}{K}_{{E}^{m}\otimes F}$
. In fact, this asymptotic statement is not true unless one chooses the Kähler form of
$X$
to be the curvature of the Hermitian metric
${e}^{\phi}$
.
In view of the reproducing property of the Bergman kernel, the key to understanding the asymptotics of the Bergman kernel lies in studying the asymptotics of the Bergman function, i.e., the restriction of the Bergman kernel to the diagonal. This has been done by numerous authors.
Here we reproduce an elementary approach that we learned from the paper of Berman [
B]
.
The conditions defining a
$\Psi $
GCS metric (see Definition 2.9 ) enter in the analysis of the asymptotics of
${R}^{m}\otimes P$
. One is forced to break up
$X$
into three regions, where different properties of
$\Psi $
GCS metrics are helpful in controlling the integrals that must be studied.
The delicate point in the argument appears because the curvature of the metric
${e}^{\phi}$
is degenerate on an analytic set. For this reason, we begin with a Kähler form and then degenerate this form to the curvature of
${e}^{\phi}$
. Since the curvature is smooth and also positive away from an analytic set, careful analysis may be used to overcome the technical difficulties.
The Bergman kernel of
${H}^{0}(X,{E}^{m}\otimes F)$
.
Let
${\omega}_{o}$
be a Kähler form on
$X$
, and set
$${\omega}_{\varepsilon}=\sqrt{1}\partial \overline{\partial}\phi +\varepsilon {\omega}_{o}.$$
Then for each
$\varepsilon >0$
${\omega}_{\varepsilon}$
is also a Kähler form.
Let
$E\to X$
,
$F\to X$
be holomorphic line bundles, and fix Gmetrics
$R\in {\mathfrak{P}}_{2}^{\Psi}(X,E)$
and
$P\in {\mathfrak{P}}_{1}^{\u266f}(X,F)$
. We write
${e}^{\phi}$
for the smooth metric of
$E$
associated to
$R$
and
${e}^{\psi}$
for the smooth metric of
$F$
associated to
$P$
. In a similar fashion, given local nonvanishing holomorphic sections
$\xi ,\eta $
of
$E\to X$
near points
${x}_{o},{y}_{o}\in X$
respectively, we write
$$r(x,y):=R({\xi}_{x},{\eta}_{y}),$$
and likewise
$p(x,y)$
for the local expression of
$P$
. Thus
$$\phi \left(x\right):=logr(x,x)\text{and}\psi \left(x\right)=logp(x,x)$$
provided we use
$\xi =\eta $
when
$x$
and
$y$
are close. We endow
$X$
with the volume element
$${\Omega}_{\varepsilon}:=\frac{{\omega}_{\varepsilon}^{n}}{(2\pi {)}^{n}n!}.$$
Recall the Hilbert spaces
${L}^{2}\left({e}^{(m\phi +\psi )}\right)$
and
${\mathfrak{A}}^{2}\left({e}^{(m\phi +\psi )}\right)$
from Section 1 . Because
$R$
and
$P$
are strict Gmetrics,
$${\mathfrak{A}}^{2}\left({e}^{(m\phi +\psi )}\right)={H}^{0}(X,{E}^{m}\otimes F),$$
i.e., every global holomorphic section is
${L}^{2}$
. We write
$${K}_{\varepsilon}^{\left(m\right)}:={K}_{{E}^{m}\otimes F},$$
for the Bergman kernel with respect to the Kähler form
${\omega}_{\varepsilon}$
as defined in Section 1 .
Let
${X}_{o}$
be the set of points
$x\in X$
such that
$\phi $
is strictly plurisubharmonic in a neighborhood of
$x$
. By property
$(\Psi )$
of Definition 2.9 ,
${X}_{o}$
is the complement of an analytic set.
Let
${\mathfrak{D}}_{\varepsilon}:X\to [0,\infty )$
be the function defined by
$$\frac{1}{(2\pi {)}^{n}n!}\sqrt{1}\partial \overline{\partial}\phi \wedge {\omega}_{\varepsilon}^{n1}={\mathfrak{D}}_{\varepsilon}\cdot {\Omega}_{\varepsilon}.$$
As is well known,
${\mathfrak{D}}_{\varepsilon}\left(x\right)$
is the product of the eigenvalues of
$\sqrt{1}\partial \overline{\partial}\phi \left(x\right)$
with respect to the form
${\omega}_{\varepsilon}$
.
The function
${\mathfrak{D}}_{\varepsilon}$
has the following properties.

(1)
${\mathfrak{D}}_{\varepsilon}$
is smooth.

(2)
For each
$\varepsilon >0$
,
${\mathfrak{D}}_{\varepsilon}\left(x\right)\ne 0$
if and only if
$x\in {X}_{o}$
.

(3)
$${\mathfrak{D}}_{\varepsilon}{\Omega}_{\varepsilon}=\frac{1}{(2\pi {)}^{n}n!}\left({\left(\sqrt{1}\partial \overline{\partial}\phi \right)}^{n}+{\sum}_{k=1}^{n}{\varepsilon}^{k}\left(\begin{array}{c}n\\ k\end{array}\right){\left(\sqrt{1}\partial \overline{\partial}\phi \right)}^{k}\wedge {\omega}_{o}^{nk}\right).$$
Let us fix a sequence
${c}_{m}$
of positive numbers such that
$${c}_{m}\nearrow +\infty \text{and}\frac{{c}_{m}}{\sqrt{m}}\to 0.$$
For example, one can take
${c}_{m}={m}^{1/4}$
. Fix
$x\in {X}_{o}$
and choose local coordinates
$z$
near
$x$
such that
$z\left(x\right)=0$
and
${z}^{*}{\omega}_{\varepsilon}$
agrees with the Euclidean metric
$\sqrt{1}\partial \overline{\partial}z{}^{2}$
at
$x$
. Next choose near
$x$
local sections
$\xi $
of
$E$
and
$\gamma $
of
$F$
such that, with
$\phi \left(x\right)=logR({\xi}_{x},{\xi}_{x})$
and
$\psi \left(x\right)=logP({\gamma}_{x},{\gamma}_{x})$
,
$${z}^{*}\phi ={\phi}_{o}\left(z\right)+o\left(\rightz{}^{2})and{z}^{*}\psi ={\psi}_{o}(z)+o(\leftz{}^{2}\right)$$
where
${\phi}_{o}=\sqrt{1}\partial \overline{\partial}\phi \left(x\right)$
and
${\psi}_{o}=\sqrt{1}\partial \overline{\partial}\psi \left(x\right)$
are the
$(1,1)$
parts of the quadratic terms in the Taylor series of
$\phi \circ z$
and
$\psi \circ z$
.
Lemma 3.1.
Let
$${A}_{\varepsilon}\left(m\right):={\int}_{\leftz\right\le {c}_{m}/\sqrt{m}}{e}^{\left(m{\phi}_{o}\right(z)+{\psi}_{o}(z\left)\right)}{\Omega}_{\varepsilon}.$$
Then

(1)
there exist
${\varepsilon}_{m}\to 0$
such that
$${lim}_{m\to \infty}{A}_{\varepsilon}\left(m\right)=\frac{{m}^{n}}{{\mathfrak{D}}_{\varepsilon}\left(x\right)}(1+{\varepsilon}_{m}),$$
and

(2)
${A}_{\varepsilon}\left(m\right)$
is uniformly bounded below, independent of
$x$
,
$\varepsilon $
and
$m$
.
Proof. Making the change of variables
$\zeta =z\sqrt{m}$
in the integral
${A}_{\varepsilon}\left(m\right)$
, we see that
$$\begin{array}{ccc}{A}_{\varepsilon}\left(m\right)& \simeq & {m}^{n}{\int}_{\left\zeta \right\le {c}_{m}}{e}^{\left({\phi}_{o}\right(\zeta )+\frac{1}{m}{\psi}_{o}(\zeta \left)\right)}{\Omega}_{\varepsilon}\end{array}$$  
$$\begin{array}{ccc}& \simeq & {m}^{n}{\int}_{\left\zeta \right\le {c}_{m}}{e}^{{\phi}_{o}\left(\zeta \right)}{\Omega}_{\varepsilon}.\end{array}$$  
After a unitary transformation in
$\zeta $
, we may assume that
${\phi}_{o}\left(\zeta \right)$
is diagonal, i.e.,
$${\phi}_{o}\left(\zeta \right)={\lambda}_{1}\left(x\right){\zeta}^{1}{}^{2}+\cdots +{\lambda}_{n}(x\left)\right{\zeta}^{n}{}^{2},$$
where
$0<{\lambda}_{1}\left(x\right)\le {\lambda}_{2}\left(x\right)\le \cdots \le {\lambda}_{n}\left(x\right)$
are the eigenvalues of
$\sqrt{1}\partial \overline{\partial}\phi \left(x\right)$
. After a scaling of the last integral, we see that for some constants
${\varepsilon}_{m}\to 0$
,
$${A}_{\varepsilon}\left(m\right)=(1+{\varepsilon}_{m})\frac{{m}^{n}}{{\pi}^{n}{\mathfrak{D}}_{\varepsilon}\left(x\right)}{\int}_{{\mathbb{C}}^{n}}{e}^{w{}^{2}}dV\left(w\right).$$
This formula shows statement (1), and the lower bound claimed in statement (2) is also easily read off. The proof is complete. □ We shall need the following wellknown lemma. We give here the proof in [
B]
.
Lemma 3.2.
For each
$x\in X$
,
$$\begin{array}{c}{m}^{n}{K}_{\varepsilon}^{\left(m\right)}(x,x){e}^{\left(m\phi \right(x)+\psi (x\left)\right)}\le {\mathfrak{D}}_{\varepsilon}\left(x\right)+o\left(1\right)asm\to \infty .\end{array}$$ 
(5)

Moreover,
${m}^{n}{K}_{\varepsilon}^{\left(m\right)}(x,x)$
is uniformly bounded on
$X$
, independent of
$m$
and
$\varepsilon $
.
Remark.
It is not hard to show, using Hormander's Theorem, that in fact equality holds in ( 5 ). However, we shall not need this additional information.
Proof. Fix
$x\in X$
. Let
$z$
denote the local coordinates on
$X$
that were chosen before the statement of Lemma 3.1 and let
$\xi $
and
$\eta $
be the local sections chosen there. Finally, let
${c}_{m}$
be the constants defined in that same paragraph.
We are going to use the variational definition of
${K}_{\varepsilon}^{\left(m\right)}$
, i.e.,
$${K}_{\varepsilon}^{\left(m\right)}(x,x)={sup}_{\left\rights\left\right=1}\lefts\right(x){}^{2}.$$
Let
$s$
be an element of
${H}^{0}(X,{E}^{m}\otimes F)$
having unit norm. Writing
$s=f\left(z\right){\xi}^{m}\otimes \gamma $
, for
${c}_{m}/\sqrt{m}$
sufficiently small the submean value property shows that
$$\begin{array}{ccc}\lefts\right(x){}^{2}{e}^{\left(m\phi \right(x)+\psi (x\left)\right)}& =& \leftf\right(0){}^{2}\end{array}$$ 
(6)

$$\begin{array}{ccc}& \le & \frac{1}{{A}_{\varepsilon}\left(m\right)}{\int}_{\leftz\right\le {c}_{m}/\sqrt{m}}\leftf\right(z){}^{2}{e}^{\left(m{\phi}_{o}\right(z)+{\psi}_{o}(z\left)\right)}{\Omega}_{\varepsilon},\end{array}$$  
where
${A}_{\varepsilon}\left(m\right)$
is as in Lemma 3.1 . Then the integral on the right hand side of the estimate ( 6 ) is bounded above by
$$(1+{\delta}_{m}){\int}_{X}s{}^{2}{e}^{(m\phi +\psi )}{\Omega}_{\varepsilon}=1+{\delta}_{m},$$
where
${\delta}_{m}\to 0$
as
$m\to \infty $
. Thus if
$x\in {X}_{o}$
then ( 5 ) holds by Lemma 3.1 . On the other hand, if
$x\notin {X}_{o}$
then
${A}_{\varepsilon}\left(m\right)$
diverges, and thus
${K}_{\varepsilon}^{\left(m\right)}(x,x)\to 0$
. Since
${\mathfrak{D}}_{\varepsilon}\left(x\right)=0$
, we see again that ( 5 ) holds. Finally, the boundedness statement follows from the lower bound for
${A}_{\varepsilon}\left(m\right)$
established in Lemma 3.1 . □ In particular, we have the following corollary.
Corollary 3.3.
$$\begin{array}{ccc}{\int}_{X}{m}^{n}{K}_{\varepsilon}^{\left(m\right)}(x,x){\Omega}_{\varepsilon}& \le & {\int}_{X}\frac{{\left(\sqrt{1}\partial \overline{\partial}\phi \right)}^{m}}{(2\pi {)}^{n}n!}+{O}_{\varepsilon}\left(\varepsilon \right)+{o}_{m}\left(1\right)\end{array}$$  
$$\begin{array}{ccc}& =& {\int}_{X}{\Omega}_{0}+{O}_{\varepsilon}\left(\varepsilon \right)+{o}_{m}\left(1\right).\end{array}$$  
Powers of a
$\Psi $
GCS metric
Lemma 3.4.
Let
$y\in {X}_{o}$
. Let
$${C}_{o}\in \left(0,\frac{3}{4}{\lambda}_{1}\left(y\right)\right),$$
where
${\lambda}_{1}\left(y\right)$
is the smallest eigenvalue of
$\sqrt{1}\partial \overline{\partial}\phi \left(y\right)$
with respect to
${\omega}_{\varepsilon}\left(y\right)$
. Then there exists
$\delta >0$
,
${M}_{1}$
,
${M}_{2}$
, independent of
$y$
, such that as
$m\to \infty $
,
$$\begin{array}{ccc}& & {\int}_{X}\frac{{m}^{n}\leftr\right(x,y\left){}^{2m}\rightp(x,y){}^{2}{\Omega}_{\varepsilon}\left(x\right)}{r(x,x{)}^{m}p(x,x\left)r\right(y,y{)}^{m}p(y,y)}\end{array}$$  
$$\begin{array}{ccc}& & =\frac{1}{{\mathfrak{D}}_{\varepsilon}\left(y\right)}+{M}_{1}{m}^{n}{e}^{\delta m}+{M}_{2}{m}^{n}{e}^{{C}_{o}\sqrt{m}}+o\left(1\right),\end{array}$$  
where the term
$o\left(1\right)$
is uniform in
$y$
.
We emphasize that the only constant depending on
$y$
is
${C}_{o}$
.
Proof. Let
$${f}_{y}\left(x\right)=logr(x,x)+logr(y,y)log\leftr\right(x,y){}^{2}.$$
Then
${f}_{y}\left(y\right)=0$
and the condition that
$R$
is GCS implies that
${f}_{y}$
has a minimum at
$y$
. Thus
$D{f}_{y}\left(y\right)=0$
. Moreover, the
$(1,1)$
part of
$\left({D}^{2}{f}_{y}\right)\left(y\right)$
is exactly
$\sqrt{1}\partial \overline{\partial}\phi \left(y\right)$
.
Choose local coordinates
$(z,U)$
with
$y\in U$
, such that
$z\left(y\right)=0$
and
${z}^{*}{\omega}_{\varepsilon}$
is Euclidean at
$y$
. As in the proof of Lemma 3.2 , we may choose the local expression of
$r(x,y)$
so that the quadratic term in the Taylor expansion of
${f}_{y}\left(z\right)$
near
$z=0$
has only
$(1,1)$
part. It is quite important for us here that the neighborhood
$U$
need not lie in
${X}_{o}$
.
We obtain the Taylor series
$$\begin{array}{ccc}{f}_{y}\left(z\right)=\sqrt{1}\partial \overline{\partial}\phi \left(y\right)(z,\overline{z})+...,& & \end{array}$$ 
(7)

and thus
$$\begin{array}{ccc}& & {lim}_{m\to \infty}{m}^{n}{\int}_{\leftz\right<{m}^{1/4}}{e}^{m{f}_{y}\left(z\right)}\frac{\leftp\right(z,y){}^{2}}{p(z,z)p(y,y)}{z}^{*}{\Omega}_{\varepsilon}\end{array}$$  
$$\begin{array}{ccc}& =& {lim}_{m\to \infty}{m}^{n}{\int}_{\left\zeta \right<{m}^{1/4}}{e}^{\left({\lambda}_{1}\right(y\left)\right{\zeta}^{1}{}^{2}+\cdots +{\lambda}_{n}\left(y\right)\left{\zeta}^{n}{}^{2}\right)}\frac{\leftp\right(\zeta /\sqrt{m},y){}^{2}}{p(\zeta /\sqrt{m},\zeta /\sqrt{m})p(y,y)}dV\left(\zeta \right)\end{array}$$  
$$\begin{array}{ccc}& =& \frac{1}{{\mathfrak{D}}_{\varepsilon}\left(y\right)},\end{array}$$  
where
$\zeta $
is obtained from
$\sqrt{m}z$
by a unitary transformation, and
$$0<{\lambda}_{1}\left(y\right)\le \cdots \le {\lambda}_{n}\left(y\right)$$
are the eigenvalues of
$\sqrt{1}\partial \overline{\partial}\phi \left(y\right)$
with respect to
${\omega}_{\varepsilon}$
.
Next we choose our neighborhood
$U$
of
$y\in X$
such that the Taylor series ( 7 ) converges well. By the compactness of
$X$
and the smoothness of
${f}_{y}$
, we may do this so that the neighborhood
$U$
does not get too small, independent of
$y$
.
By the hypothesis (S2) for
$\Psi $
GCS metrics, there is a number
$\delta >0$
such that
$$x\notin U\Rightarrow {f}_{y}\left(x\right)\ge \delta .$$
Since our neighborhood
$U$
in
$X$
is not too small regardless of
$y$
, the number
$\delta $
may be taken independent of
$y$
. We then have
$${\int}_{XU}{e}^{m{f}_{y}\left(x\right)}\frac{\leftp\right(x,y){}^{2}}{p(x,x)p(y,y)}{\Omega}_{\varepsilon}\left(x\right)\le {e}^{m\delta}{\int}_{XU}\frac{\leftp\right(x,y){}^{2}}{p(x,x)p(y,y)}{\Omega}_{\varepsilon}\left(x\right)={M}_{1}{e}^{m\delta}.$$
Finally, in view of the Taylor expansion ( 7 ) we have
$${f}_{y}\left(z\right)\ge {C}_{o}\left\rightz{}^{2},z\in U,$$
and thus
$$\begin{array}{ccc}& & {\int}_{z\left(U\right)\left\{\rightz<{m}^{1/4}\}}{e}^{m{f}_{y}\left(z\right)}\frac{\leftp\right(z,y){}^{2}}{p(z,z)p(y,y)}dV\left(z\right)\end{array}$$  
$$\begin{array}{ccc}& \le & {\int}_{z\left(U\right)\left\{\rightz<{m}^{1/4}\}}{e}^{m{C}_{o}\left\rightz{}^{2}}\frac{\leftp\right(z,y){}^{2}}{p(z,z)p(y,y)}dV\left(z\right)\end{array}$$  
$$\begin{array}{ccc}& \le & {e}^{{C}_{o}\sqrt{m}}{\int}_{U}\frac{\leftp\right(z,y){}^{2}}{p(z,z)p(y,y)}dV\left(z\right)\end{array}$$  
$$\begin{array}{ccc}& \le & {e}^{{C}_{o}\sqrt{m}}{sup}_{y\in X}{\int}_{X}\frac{\leftp\right(x,y){}^{2}}{p(x,x)p(y,y)}\Omega \left(x\right)\end{array}$$  
$$\begin{array}{ccc}& =& {M}_{2}{e}^{{C}_{o}\sqrt{m}}.\end{array}$$  
Corollary 3.5.
Let
${X}^{\prime}\subset \subset {X}_{o}$
. Then
$$\begin{array}{ccc}{\int}_{{X}^{\prime}}{\int}_{X}\frac{{m}^{n}\leftr\right(x,y{)}^{m}p(x,y){}^{2}{\Omega}_{\varepsilon}\left(x\right){\Omega}_{\varepsilon}\left(y\right)}{r(x,x{)}^{m}p(x,x\left)r\right(y,y{)}^{m}p(y,y)}={\int}_{{X}^{\prime}}\frac{1}{{\mathfrak{D}}_{\varepsilon}\left(y\right)}{\Omega}_{\varepsilon}\left(y\right)+o\left(1\right).& & \end{array}$$ 
(8)

Remark.
In equation 8
$\left(\right)$
the term
$o\left(1\right)$
depends on
${X}^{\prime}$
in only one way: the constant
${C}_{o}$
in Lemma 3.4 may approach
$0$
as
${X}^{\prime}$
approaches
${X}_{o}$
. But even if we knew that
${C}_{o}$
was bounded away from
$0$
, we would still need to use this formulation, simply because the integral on the right hand side of 8
$\left(\right)$
may diverge as
${X}^{\prime}$
approaches
${X}_{o}$
.
Passing to degenerate Kähler forms
Our next goal is to control the constant
${C}_{o}$
. Of course, if
$\sqrt{1}\partial \overline{\partial}\phi $
is degenerate then
${C}_{o}$
must approach zero as
$y$
approaches the complement of
${X}_{o}$
. However, we shall now let
$\varepsilon \to 0$
. It is then clear that the following holds.
Lemma 3.6.
Fix
${X}^{\prime}\subset \subset X$
. Then
$${lim}_{\varepsilon \to 0}{\mathfrak{D}}_{\varepsilon}\left(y\right)=1$$
uniformly on
${X}^{\prime}$
. Moreover, we may take
${C}_{o}=1/2$
for all
$y\in {X}^{\prime}$
as soon as
$\varepsilon $
is small enough.
We now summarize all of our results in the limit as
$\varepsilon \to 0$
.
Proposition 3.7.
Let
$\varepsilon \to 0$
.

(A)
For every
$\delta >0$
there exists
$N$
such that for all
$m\ge N$
,
$${\int}_{X}\frac{{m}^{n}{K}_{\varepsilon}^{\left(m\right)}(x,x)}{r(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(x\right)\le {\int}_{X}{\Omega}_{0}+\delta ,$$
and

(B)
As
$m\to \infty $
,
$${\int}_{X}{\int}_{X}\frac{{m}^{n}\leftr\right(x,y\left){}^{2m}\rightp(x,y){}^{2}}{r(x,x{)}^{m}p(x,x\left)r\right(y,y{)}^{m}p(y,y)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)={\int}_{X}{\Omega}_{0}+o\left(1\right),$$
Proof. Statement (A) is clear from Corollary 3.3 and the uniform boundedness of
${K}_{\varepsilon}^{\left(m\right)}$
. It remains to show statement (B). To this end, fix a compact set
${X}^{\prime}\subset \subset X$
. Since
$\varepsilon \to 0$
, we see from Lemma 3.6 that
${\mathfrak{D}}_{0}\equiv 1$
on
${X}^{\prime}$
and we can take a uniform
${C}_{o}$
, namely
$1/2$
. Both of these functions are clearly independent of
${X}^{\prime}$
.
In view of the remark following Corollary
3.5 , since we have let
$\varepsilon \to 0$
, the right hand side of equation ( 8 ) is now uniform in
${X}^{\prime}$
and we may let
${X}^{\prime}\to X$
. The result is that (B) holds. □
Proof of Theorem 3 .
Fix an orthonormal basis
$\{{s}_{m}^{1},...,{s}_{m}^{{N}_{m}}\}$
of
${H}^{0}(X,{E}^{m}\otimes F)$
. Let
$s={a}^{1}{s}_{m}^{1}+\cdots +{a}^{{N}_{m}}{s}_{m}^{{N}_{m}}$
be a unit vector in
${H}^{0}(X,{E}^{m}\otimes F)$
. Then
$$\begin{array}{ccc}& & {\left{\int}_{X}{\int}_{X}\frac{\left({m}^{n}r(x,y{)}^{m}p(x,y){K}_{\varepsilon}^{\left(m\right)}(x,y)\right)s\left(y\right)\overline{s\left(x\right)}}{r(y,y{)}^{m}p(y,y\left)r\right(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)\right}^{2}\end{array}$$  
$$\begin{array}{ccc}& \le & {\int}_{X}{\int}_{X}\frac{{\left{m}^{n}r(x,y{)}^{m}p(x,y){K}_{\varepsilon}^{\left(m\right)}(x,y)\right}^{2}}{r(y,y{)}^{m}p(y,y\left)r\right(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)\end{array}$$  
$$\begin{array}{ccc}& =& {\int}_{X}{\int}_{X}\frac{{\left{m}^{n}r(x,y{)}^{m}p(x,y)\right}^{2}}{r(y,y{)}^{m}p(y,y\left)r\right(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)\end{array}$$  
$$\begin{array}{ccc}& & +{\int}_{X}{\int}_{X}\frac{{K}_{\varepsilon}^{\left(m\right)}(x,y){K}_{\varepsilon}^{\left(m\right)}(y,x)}{r(y,y{)}^{m}p(y,y\left)r\right(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)\end{array}$$  
$$\begin{array}{ccc}& & 2Re{\int}_{X}{\int}_{X}\frac{{m}^{n}{K}_{\varepsilon}^{\left(m\right)}(x,y)r(y,x{)}^{m}p(y,x)}{r(y,y{)}^{m}p(y,y\left)r\right(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)\end{array}$$  
$$\begin{array}{ccc}& =& {m}^{n}{\int}_{X}{\int}_{X}\frac{{m}^{n}{\leftr(x,y{)}^{m}p(x,y)\right}^{2}}{r(y,y{)}^{m}p(y,y\left)r\right(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)\end{array}$$  
$$\begin{array}{ccc}& & +{m}^{n}{\int}_{X}\frac{{m}^{n}{K}_{\varepsilon}^{\left(m\right)}(x,x)}{r(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(x\right)2Re{\int}_{X}\frac{{m}^{n}r(x,x{)}^{m}p(x,x)}{r(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(x\right)\end{array}$$  
$$\begin{array}{ccc}& =& {m}^{n}{\int}_{X}\Omega \left(x\right)+o\left({m}^{n}\right)+{m}^{n}{\int}_{X}\frac{{m}^{n}{K}_{\varepsilon}^{\left(m\right)}(x,x)}{r(x,x{)}^{m}p(x,x)}\Omega \left(x\right)2{m}^{n}{\int}_{X}\Omega .\end{array}$$  
(In the second to last equation, we used the reproducing property of the Bergman kernel in the second and third integrals.) Thus, in view of Proposition 3.7 with
$\delta =1/2$
in part (A), we have that
$$\begin{array}{c}{\left{\int}_{X}{\int}_{X}\frac{\left({m}^{n}r(x,y{)}^{m}p(x,y){K}_{\varepsilon}^{\left(m\right)}(x,y)\right)s\left(y\right)\overline{s\left(x\right)}}{r(y,y{)}^{m}p(y,y\left)r\right(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)\right}^{2}\le \frac{{m}^{n}}{2}+o\left({m}^{n}\right)\end{array}$$ 
(9)

as
$m\to \infty $
.
Next, let
${\mathsf{\u212d}}^{\left(m\right)}=\left({c}_{\alpha \overline{\beta}}^{\left(m\right)}\right)$
be the Hermitian matrix such that with respect to the basis
$\left\{{s}^{\alpha}\right\}$
of
${H}^{0}(X,{E}^{m}\otimes F)$
,
${R}^{m}\otimes P={c}_{\alpha \overline{\beta}}^{\left(m\right)}{s}^{\alpha}{\overline{s}}^{\beta}$
. Observe that
$${c}_{\alpha \overline{\beta}}^{\left(m\right)}={\int}_{X}{\int}_{X}\frac{r(x,y{)}^{m}p(x,y\left){s}^{\alpha}\right(y)\overline{{s}^{\beta}\left(x\right)}}{r(y,y{)}^{m}p(y,y\left)r\right(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)$$
and thus by ( 9 ) we have
$$\begin{array}{ccc}& & {\left\left({m}^{n}{c}_{\alpha \overline{\beta}}^{\left(m\right)}{\delta}_{\alpha \overline{\beta}}\right){a}^{\alpha}{\overline{a}}^{\beta}\right}^{2}\end{array}$$  
$$\begin{array}{ccc}& =& {\left{\int}_{X}{\int}_{X}\frac{\left({m}^{n}r(x,y{)}^{m}p(x,y){K}_{\varepsilon}^{\left(m\right)}(x,y)\right)s\left(y\right)\overline{s\left(x\right)}}{r(y,y{)}^{m}p(y,y\left)r\right(x,x{)}^{m}p(x,x)}{\Omega}_{0}\left(y\right){\Omega}_{0}\left(x\right)\right}^{2}\end{array}$$  
$$\begin{array}{ccc}& \le & \frac{{m}^{n}}{2}+o\left({m}^{n}\right).\end{array}$$  
But this means there are constants
${B}_{m}\to 0$
(as
$m\to \infty $
) such that for all vectors
$a$
,
$${a}^{\u2020}{\mathsf{\u212d}}^{\left(m\right)}a\ge \frac{\frac{1}{2}{B}_{m}}{{m}^{n}}\left\righta{}^{2}.$$
That is to say, there is an integer
$N>>0$
such that
${\mathsf{\u212d}}^{\left(m\right)}$
is positive definite for any
$m\ge N$
. The proof is complete. □
4 Quotients of squared norms
The Riemann surface case
Let
$M$
be a compact Riemann surface and
$\Lambda \to M$
a holomorphic line bundle.
Theorem 4.1.
A Gmetric
$P\in {\mathfrak{P}}_{1}(M,\Lambda )$
is a quotient of squared norms if and only if there exist holomorphic line bundles
${E}_{1},{E}_{2}\to M$
, a holomorphic section
$\Phi \in {H}^{0}(M,{E}_{1})$
, and a strict Gmetric
$\stackrel{~}{P}$
on
$M$
such that
$\Lambda ={E}_{1}\otimes {E}_{2}$
and
$$\begin{array}{ccc}P=\Phi {}^{2}\stackrel{~}{P},& i.e.,& R(v,\overline{w})=\Phi \left(v\right)\overline{\Phi \left(w\right)}\stackrel{~}{R}(v,\overline{w}).\end{array}$$ 
(10)

Remark.
This theorem was proved by D'Angelo in the case
$M={\mathbb{P}}_{1}$
[
D3]
. Our proof is similar, but it is crucial for this paper that the surface
$M$
is general.
Proof of Theorem 4.1 : The sufficiency of the expression ( 10 ) is a trivial consequence of Theorem 2.6 .
To see the necessity, suppose
$P$
is a quotient of squared norms. Then
$$\begin{array}{ccc}P\cdot {\sum}_{\alpha}{\left{s}^{\alpha}\right}^{2}={\sum}_{\beta}{\left{T}^{\beta}\right}^{2}& & \end{array}$$ 
(11)

for some holomorphic sections
$\left\{{s}^{\alpha}\right\}$
and
$\left\{{T}^{\beta}\right\}$
of (different) line bundles.
Suppose now that
$x\in M$
. Choose a local coordinate function
$z$
near
$x$
with
$z\left(x\right)=0$
, and let
$\xi $
be a nowhere zero section of
${\Lambda}^{*}$
near
$x$
. Then by formula ( 11 ) the function
$p:=P(\xi ,\overline{\xi})$
has the property that for some nonnegative integer
$m$
,
$p/z{}^{2m}$
is bounded and nonzero near
$x$
. (If there exists
$v\in {\Lambda}_{x}^{*}{o}_{{\Lambda}^{*}}\left(x\right)$
such that
$P(v,\overline{v})=0$
, then
$m>0$
.) Clearly the integer
$m$
is unique. We write
$$m=:{D}_{P}\left(x\right).$$
We associate to
$P$
the divisor
$${D}_{P}={\sum}_{x\in M}{D}_{P}\left(x\right)\cdot x.$$
Evidently
${D}_{P}$
is independent of
$\xi $
and
$z$
. Let
${L}_{{D}_{P}}$
be the line bundle associated to
${D}_{P}$
and let
${s}_{{D}_{P}}$
be the canonical section of
${L}_{{D}_{P}}$
whose divisor is
${D}_{P}$
. Since
${D}_{P}$
is effective, i.e.,
${D}_{P}\left(x\right)\ge 0$
for all
$x\in M$
,
${s}_{{D}_{P}}$
is holomorphic.
By construction,
$\stackrel{~}{P}={s}_{{D}_{P}}{}^{2}P$
is a strict Gmetric. An application of Theorem 2.6 completes the proof. □ For the purpose of illustration, we prove Theorem 1 for Riemann surfaces.
Necessity of basic zeros: Let
$P\in {\mathfrak{P}}_{1}(M,\Lambda )$
. First, suppose
$P$
is a quotient of squared norms. In view of Theorem 4.1 ,
$P={s}_{D}{}^{2}\otimes \stackrel{~}{P}$
for some strict Gmetric
$\stackrel{~}{P}$
. It follows that
${s}_{D}$
divides the support space
${V}_{P}$
of
$P$
. Thus for any
${s}_{D}\otimes f\in {V}_{P}$
we have that
$$\frac{{s}_{D}\otimes f{}^{2}}{P}=\frac{f{}^{2}}{\stackrel{~}{P}}$$
is bounded. That is to say, condition ( 4 ) holds.
Sufficiency of basic zeros: Conversely, suppose condition ( 4 ) holds for
$P$
. If
${V}_{P}$
is free, then
$P$
has no zeros and by Theorem 4.1 we are done. Therefore, assume
${V}_{P}$
has a nontrivial base locus.
Then there is a divisor
$D$
on
$M$
whose canonical holomorphic section
${s}_{D}$
divides every element of
${V}_{P}$
. Take
$D$
to be the largest such divisor. Thus
${s}_{D}{}^{2}$
divides
$P$
, and the metric
$\stackrel{~}{P}=P/{s}_{D}{}^{2}$
is a Gmetric for the line bundle
$E=\Lambda \otimes {L}_{D}^{*}$
. Evidently
${V}_{\stackrel{~}{P}}$
is free. Moreover,
$$\frac{s{}^{2}}{P}=\frac{s/{s}_{D}{}^{2}}{\stackrel{~}{P}},$$
And thus by condition ( 4 )
$\stackrel{~}{P}$
has no zeros. By Theorem 4.1 we are done. □
Corollary 4.2.
If
$R$
is a Gmetric on a Riemann surface and
${V}_{R}$
is free, then
$R$
is a quotient of squared norms if and only if
$R$
is a strict Gmetric, i.e.,
$R$
has no nontrivial zeros.
The free case
Corollary 4.2 generalizes to higher dimensions.
Theorem 4.3.
Let
$P\in {\mathfrak{P}}_{1}(X,F)$
be a Gmetric such that
${V}_{P}$
is free. Then
$P$
is a quotient of squared norms if and only if
$P$
is a strict Gmetric.
Proof. The “if ” direction follows from Theorem 2.8 or its generalization, Theorem 3 . We shall now prove the converse. Suppose
$P$
is a quotient of squared norms and
$P(v,\overline{v})=0$
for some
$v\in {F}_{x}^{*}{o}_{{F}^{*}}\left(x\right)$
. Since
$X$
is projective algebraic, there is a Riemann surface
$M\subset X$
passing through
$x$
. Let
$\iota :M\hookrightarrow X$
denote the natural inclusion map. Then the following facts are evident.

(1)
${\iota}^{*}{V}_{P}={V}_{{\iota}^{*}P}$
, and thus
${\iota}^{*}{V}_{P}$
is free.

(2)
${\iota}^{*}P$
is a quotient of squared norms.

(3)
${\iota}^{*}P$
vanishes along
$({\iota}^{*}F{)}_{x}$
.
But these three facts contradict Theorem 4.1 . The proof is complete. □
An illustrative example
D'Angelo showed that the Gmetric
$P\in {\mathfrak{P}}_{1}({\mathbb{P}}_{2},{\mathbb{H}}^{4})$
given in homogeneous coordinates
$[{z}_{0},{z}_{1},{z}_{2}]$
by
$$P(z,\overline{z})={z}_{0}{}^{8}+({z}_{0}{z}_{2}{}^{2}{z}_{1}{}^{4}{)}^{2}$$
is not a quotient of squared norms. The interesting aspect of this Gmetric is that its zero locus
${\mathcal{Z}}_{P}$
is the single point
$[0,0,1]$
. In particular,
${\mathcal{Z}}_{P}$
is an analytic set.
D'Angelo showed that
$P$
is not a quotient of squared norms by showing that
$P$
does not satisfy a certain necessary condition for being a quotient of squared norms: the jet pullback property.
We will recall the jet pullback property in Section
5 below. In the language of the present paper, D'Angelo passes the (rational) curve
$\gamma :{\mathbb{P}}_{1}\to {\mathbb{P}}_{2}$
given by
$$\gamma [x,y]=[{x}^{2},xy,xy+{y}^{2}]$$
through
$[0,0,1]$
and examines the restriction of
$P$
to this curve. He then deduces from the Taylor expansion of the curve near
$[0,0,1]$
that
$P$
cannot be a quotient of squared norms.
The information about the influence of the Taylor expansion of the curve on
$P$
is already contained in Theorem 4.1 . To rephrase D'Angelo's proof in the language of the present paper, note that
${\gamma}^{*}P\in {\mathfrak{P}}_{1}({\mathbb{P}}_{1},{\mathbb{H}}^{8})$
and
$$\begin{array}{ccc}{\gamma}^{*}P\left(\right(x,y),(\overline{x},\overline{y}\left)\right)& =& x{}^{16}+({x}^{3}y+{x}^{2}{y}^{2}{}^{2}xy{}^{4}{)}^{2}\end{array}$$  
$$\begin{array}{ccc}& =& x{}^{8}\left(x{}^{8}+(xy+{y}^{2}{}^{2}y{}^{4}{)}^{2}\right)\end{array}$$  
$$\begin{array}{ccc}& =& \leftx{}^{8}Q\right((x,y),(\overline{x},\overline{y})).\end{array}$$  
One then sees that
$Q$
is a free Gmetric on
${\mathbb{P}}_{1}$
with a nontrivial zero at the point
$[0,1]\in {\mathbb{P}}_{1}$
. By Corollary 4.2
$Q$
is not a quotient of squared norms. Thus
${\gamma}^{*}P$
is not a quotient of squared norms, and hence
$P$
is not a quotient of squared norms.
In order to demonstrate the role of the resolution of singularities in the proof of Theorem
1 , we shall use blowups to show that the Gmetric
$P$
is not a quotient of squared norms. Since the zero locus of
$P$
consists only of the point
$[0,0,1]\in {\mathbb{P}}_{2}$
, it is easiest to work in the affine chart
$\{{z}^{2}\ne 0\}$
.
Let
${x}_{1}={z}_{0}/{z}_{2}$
and
${x}_{2}={z}_{1}/{z}_{2}$
. Then
$$P=\left{z}_{2}{}^{8}\right({x}_{1}{}^{8}+({x}_{1}{}^{2}{x}_{2}{}^{4}{)}^{2}).$$
It suffices to work with the inhomogeneous polynomial
$$p={x}_{1}{}^{8}+({x}_{1}{}^{2}{x}_{2}{}^{4}{)}^{2},$$
since one can always recover the original Gmetric by homogenization. (For more on this point as well as other aspects of homogenization of polynomials with regard to Gmetrics, see [
D3]
.) Now blow up the origin in
${\mathbb{C}}^{2}$
. That is to say, let
$${x}_{1}={y}_{1}{y}_{2}\text{and}{x}_{2}={y}_{2}.$$
Then with
${p}_{1}$
denoting the blowup of
$p$
, one has
$${p}_{1}={y}_{2}{}^{4}\left({y}_{1}^{2}{y}_{2}{}^{4}+({y}_{1}{}^{2}{y}_{2}{}^{2}{)}^{2}\right).$$
Now take one more blowup, namely
$${y}_{1}={t}_{1}\text{and}{y}_{2}={t}_{1}{t}_{2}.$$
Then with
${p}_{2}$
denoting the blowup of
${p}_{1}$
, one has
$${p}_{2}={t}_{1}^{2}{t}_{2}{}^{4}\left({t}_{1}^{2}{t}_{2}{}^{4}+(1{t}_{2}{}^{2}{)}^{2}\right).$$
We have shown that after two blowups,
$p$
is transformed into a Gmetric that is a product of a square (the term
$\left\right({t}_{1}^{2}{t}_{2}{)}^{2}{}^{2}$
) and a free Gmetric with a nontrivial zero, namely the point
$({t}_{1},{t}_{2})=(0,1)$
in the particular affine chart we are working with. It follows from Theorem 4.3 that the Gmetric
$P$
could not have been a quotient of squared norms.
The general case
We remind the reader that we have assumed the absence of zeros in codimension 1.
Let
$\mu :\stackrel{~}{X}\to X$
be a log resolution of
${V}_{P}$
as in Theorem 1.2 . Then there is an effective divisor
$D$
in
$\stackrel{~}{X}$
, a free linear system
$W\subset {H}^{0}(\stackrel{~}{X},F\otimes {L}_{D}^{*})$
and a section
${s}_{D}$
of the line bundle associated to
$D$
such that for all
$s\in {\mu}^{*}{V}_{P}$
,
$$s={s}_{D}\otimes t\text{for some}t\in W.$$
It follows that
${\mu}^{*}P={s}_{D}{}^{2}\otimes \stackrel{~}{P}$
for some Gmetric
$\stackrel{~}{P}\in {\mathfrak{P}}_{1}(X,{\mu}^{*}F\otimes {L}_{D}^{*})$
such that
${V}_{\stackrel{~}{P}}$
is free.
We now show the necessity of condition (
4 ). To this end, if
$P$
is a quotient of squared norms, then so is
$\stackrel{~}{P}$
. Since
${V}_{\stackrel{~}{P}}$
is free, by Theorem 4.3 we see that
$\stackrel{~}{P}$
has no zeros. It follows that if
$s\in {V}_{P}$
and
${\mu}^{*}s={s}_{D}\otimes t$
for some
$t\in {V}_{\stackrel{~}{P}}$
, then the quotient
$t{}^{2}/\stackrel{~}{P}$
is bounded. But then the quotient
$$\frac{{\mu}^{*}s{}^{2}}{{\mu}^{*}P}=\frac{\left{s}_{D}{}^{2}\rightt{}^{2}}{{s}_{D}{}^{2}\stackrel{~}{P}}$$
is bounded on
$\stackrel{~}{X}$
. Since
$\mu $
is an isomorphism on a dense subset, we see that condition ( 4 ) must hold.
Next we turn to the sufficiency of condition (
4 ). To this end, choose a nontrivial line bundle
$E\to X$
and a subspace
$M\subset {H}^{0}(X,E)$
with the following properties:

(1)
The base locus
$G$
of
${\mu}^{*}M$
is a divisor that divides
$D$
.

(2)
There is a metric
$R\in {\mathfrak{P}}_{2}^{\Sigma}(X,E)$
such that
${V}_{R}=M$
.
For example,
$E$
could be very ample and
$M={H}^{0}(X,E)$
. In this case
$G=\varnothing $
.
Let
$U={\mu}^{*}M/{s}_{G}$
. Then
$U$
is free, and we have
${\mu}^{*}R=\stackrel{~}{R}\otimes {s}_{G}{}^{2}$
where
$\stackrel{~}{R}$
is free. It follows from Theorem 3 that there is an integer
$N\ge 0$
such that for all
$m\ge N$
,
$$\begin{array}{ccc}{\stackrel{~}{R}}^{m}\otimes \stackrel{~}{P}={\sum}_{\ell =1}^{r}{\xi}^{\ell}{}^{2}\in {\mathfrak{P}}_{\infty}(\stackrel{~}{X},({\mu}^{*}E\otimes {L}_{G}^{*}{)}^{m}\otimes ({\mu}^{*}F\otimes {L}_{D}^{*}\left)\right).& & \end{array}$$ 
(12)

Tensoring both sides of ( 12 ) by
${s}_{G}{}^{2m}\otimes {s}_{D}{}^{2}$
, we see that
$${\mu}^{*}\left({R}^{m}\otimes P\right)={\sum}_{j=1}^{r}{\kappa}^{j}{}^{2},$$
where
$${\kappa}^{j}:={s}_{G}^{m}\otimes {s}_{F}\otimes {\xi}^{\ell}.$$
It follows from Hartogs' Theorem and the absence of codimension 1 basic zeros for
$P$
that
$${R}^{m}\otimes P={\sum}_{\ell =1}^{r}{\eta}^{\ell}{}^{2},$$
for some sections
${\eta}^{\ell}$
of
${E}^{m}\otimes F$
. By Theorem 3 we see that for all
$m>>0$
,
${R}^{m}\in {\mathfrak{P}}_{\infty}(X,{E}^{m})$
, and thus
$P$
is a quotient of squared norms. The proof is complete. □
5 The jet pullback property
Let us recall the definition of D'Angelo's jet pullback property.
Definition 5.1.
A Gmetric
$P\in {\mathfrak{P}}_{1}(X,F)$
is said to have the jet pullback property if for every compact Riemann surface
$M$
and holomorphic map
$h:M\to X$
the following holds. If
$z$
is a local coordinate in a neighborhood of
$x\in M$
and
$\xi $
is a nowhere zero holomorphic section of
$F$
in a neighborhood of
$h\left(x\right)$
, then the function
${h}^{*}\left(P\right(\xi ,\xi \left)\right)$
has Taylor expansion
$${h}^{*}\left(P\right(\xi ,\xi \left)\right)\left(z\right)=cz{}^{2d}+\text{higher order terms}$$
for some
$c>0$
.
Remark.
Strictly speaking, in D'Angelo's definition of the jet pullback property one is only allowed to use rational curves, i.e.,
$M={\mathbb{P}}^{1}$
. We shall take this stronger definition, since rational curves might not be so plentiful on
$X$
. Indeed, it may happen that
$X$
has no rational curves at all.
It is easy to see that if
$P$
is a quotient of squared norms then
$P$
has the jet pullback property.
In [
D1]
it is asked whether the converse is true. In this section we answer this question in the affirmative.
Proposition 5.2.
If
$P\in {\mathfrak{P}}_{1}(X,F)$
has the jet pullback property then
$P$
is a quotient of squared norms.

Proof.
Suppose
$P$
has the jet pullback property. We shall show that
$P$
has only basic zeros, and then use Theorem 1 .
(i) As in the previous section, let
$\mu :\stackrel{~}{X}\to X$
be a resolution of singularities such that
${\mu}^{*}P=s{}^{2}\stackrel{~}{P}$
for some holomorphic section
$s$
and free Gmetric
$\stackrel{~}{P}$
.
(ii) Observe that
${\mu}^{*}P$
has the jet pullback property. Indeed, if
$h:M\to \stackrel{~}{X}$
is a curve, then
$${h}^{*}\left({\mu}^{*}P\right)=(\mu \circ h{)}^{*}P,$$
and thus it is clear from the jet pullback property for
$P$
that the lowest order term of the Taylor expansion for
${h}^{*}\left({\mu}^{*}P\right)$
at any point of
$M$
has the right form.
(iii) Notice that since
$s{}^{2}\stackrel{~}{P}$
has the jet pullback property, so does
$\stackrel{~}{P}$
.
(iv) If the free Gmetric
$\stackrel{~}{P}$
has the jet pullback property, then
$\stackrel{~}{P}$
must be a strict Gmetric. Indeed, since
$\stackrel{~}{P}$
is free, the pullback
${h}^{*}\stackrel{~}{P}$
of
$\stackrel{~}{P}$
to any curve
$h:M\to \stackrel{~}{X}$
is free. Moreover, it is immediate that
${h}^{*}\stackrel{~}{P}$
has the jet pullback property. But since
$M$
is a curve, the jet pullback property for
${h}^{*}\stackrel{~}{P}$
means that the zero locus of
${h}^{*}\stackrel{~}{P}$
is a divisor. If
$\stackrel{~}{P}$
is not a strict Gmetric, then by choosing the curve
$h:M\to X$
such that
$h\left(M\right)$
passes through a zero of
$\stackrel{~}{P}$
, we see that the zero divisor of
${h}^{*}\stackrel{~}{P}$
is nonempty. It follows as in the proof of Theorem 4.1 that every section in the support space of
${h}^{*}\stackrel{~}{P}$
is divisible by a global section on
$M$
, and this contradicts that
${h}^{*}\stackrel{~}{P}$
is free.
(v) If
$\stackrel{~}{P}$
is a strict Gmetric, then clearly
$${\mu}^{*}\left(\frac{\mathit{\forall}P\mathit{\forall}}{P}\right)=\frac{\mathit{\forall}\stackrel{~}{P}\mathit{\forall}}{\stackrel{~}{P}}$$
is uniformly bounded on
$\stackrel{~}{X}$
. Since
$\mu $
is an isomorphism, we see that
$\frac{\mathit{\forall}P\mathit{\forall}}{P}$
is uniformly bounded on the complement of a set of measure zero in
$X$
. But then
$\frac{\mathit{\forall}P\mathit{\forall}}{P}$
is uniformly bounded on
$X$
, i.e.,
$P$
has only basic zeros. By Theorem 1 we are done. □
6
$\Sigma $
GCS metrics
In the course of the proof of Theorem 1 , we established the following generalization of Theorem 3 .
Theorem 6.1.
Suppose
$P\in {\mathfrak{P}}_{1}(X,F)$
has only basic zeros, i.e.,
$P$
satisfies condition ( 4 ) above. Then for any holomorphic line bundle
$E\to X$
and
$R\in {\mathfrak{P}}_{2}^{\Psi}(X,E)$
there exists
$N>0$
such that for all
$m\ge N$
,
$${R}^{m}\otimes P\in {\mathfrak{P}}_{\infty}(X,{E}^{m}\otimes F).$$
Thus we have weakened the conditions required from
$P$
in Theorem 3 , allowing it to have basic zeros. We also pointed out that
$R$
itself may have zeros, so long as its zero locus is not larger than that of
$P$
.
This discussion suggests that perhaps the requirements we place on GCS metrics in the notion of
$\Psi $
GCS are too stringent. In fact, this is the case. The methods we have developed thus far make for a generalization of Theorem 6.1 . To state this generalization, it is convenient to introduce another class of GCS metrics that contains the previous classes.
Definition 6.2.
A Gmetric
$R\in {\mathfrak{P}}_{1}(X,E)$
is said to be a
$\Sigma $
GCS metric, written
$$R\in {\mathfrak{P}}_{2}^{\Sigma}(X,E),$$
if the following conditions hold.

(
$\Sigma $
1)
$R\in {\mathfrak{P}}_{1}^{\u266e}(X,E)\cap {\mathfrak{P}}_{2}(X,E)$
.

(
$\Sigma $
2)
For any
$v,w\in E$
,
$\leftR\right(v,\overline{w}){}^{2}<R(v,\overline{v}\left)R\right(w,\overline{w})$
unless

(a)
$v$
and
$w$
lie in the same fiber of
$E$
, or

(b)
$R(v,\overline{v})R(w,\overline{w})=0$
.

(
$\Sigma $
3)
$\sqrt{1}\partial \overline{\partial}logr(x,x)$
is strictly positive, in the sense of currents, away from an analytic subvariety.
With this definition, we have the following theorem.
Theorem 4.
If
$P\in {\mathfrak{P}}_{1}^{\u266e}(X,F)$
and
$R\in {\mathfrak{P}}_{2}^{\Sigma}(X,E)$
then there is an integer
$N>0$
such that for all
$m\ge N$
,
$${R}^{m}\otimes P\in {\mathfrak{P}}_{\infty}(X,{E}^{m}\otimes F).$$
Proof. Recall (see Remark 2 ) that we have assumed the base locus of
${V}_{R}$
contains no divisors. Let
$\mu :\stackrel{~}{X}\to X$
be a log resolution, i.e., a birational map such that
${\mu}^{*}{V}_{R}=W\otimes {s}_{D}$
, where
$D$
is an effective divisor and
$W\subset {H}^{0}(\stackrel{~}{X},{\mu}^{*}E\otimes {L}_{D}^{*})$
is a free linear system. Let
$$\stackrel{~}{R}:={\mu}^{*}R/{s}_{D}{}^{2}.$$
Evidently
$\stackrel{~}{R}\in {\mathfrak{P}}_{2}^{\Psi}(\stackrel{~}{X},{\mu}^{*}E\otimes {L}_{D}^{*})$
.
Let
$\stackrel{~}{P}:={\mu}^{*}P$
. Then
$\stackrel{~}{P}\in {\mathfrak{P}}_{1}^{\u266e}(\stackrel{~}{X},{\mu}^{*}F)$
. It follows from Theorem 6.1 that there is an integer
$N$
such that for all
$m\ge N$
,
$${\stackrel{~}{R}}^{m}\otimes \stackrel{~}{P}\in {\mathfrak{P}}_{\infty}(\stackrel{~}{X},({\mu}^{*}E\otimes {L}_{D}^{*}{)}^{m}\otimes {\mu}^{*}F).$$
Thus
$${\mu}^{*}({R}^{m}\otimes P)={s}_{D}{}^{2m}\otimes {\stackrel{~}{R}}^{m}\otimes \stackrel{~}{P}\in {\mathfrak{P}}_{\infty}(\stackrel{~}{X},{\mu}^{*}({E}^{m}\otimes F)).$$
Since
$\mu $
is an isomorphism outside the exceptional divisor, the result follows once more by an application of Hartogs' Theorem. □
Remark.
As will be shown in the next section, any Gmetric having the property that some power of it is a sum of squared norms can have only basic zeros.
7 Projective isometric maps
Weak
$\Sigma $
GCS metrics
If we take
$F$
to be the trivial bundle
$\mathcal{O}$
and
$P\equiv 1\in {H}^{0}(X,\mathcal{O})$
, then Theorem 4 says that all sufficiently large powers of a
$\Sigma $
GCS metric are sums of squared norms. Following [
CD2]
, we can rephrase this corollary in the language of projective isometric maps.
Corollary 7.1.
Let
$R$
be a
$\Sigma $
GCS metric and set
${N}_{m}:=dim\left({V}_{{R}^{m}}\right)$
. Let
${\omega}_{m}$
the Gmetric on
$\mathbb{H}\to {\mathbb{P}}_{{N}^{m}1}$
associated to the identity Hermitian form. Then there are holomorphic maps
${T}_{m}:X\to {\mathbb{P}}_{{N}_{m}}$
such that
$${R}^{m}={T}_{m}^{*}{\omega}_{m}.$$
Thus Corollary 7.1 gives a Gmetric analogue of Kodaira's celebrated Embedding Theorem. In fact, the conditions of
$\Sigma $
GCS are not necessary for obtaining Corollary 7.1 . To illustrate, consider the following simple example.
Example 7.2.
Let
$X={\mathbb{P}}_{2}$
with homogeneous coordinates
$[{z}^{0},{z}^{1},{z}^{2}]$
, and consider any Gmetric
$R\in {\mathfrak{P}}_{2}^{\Sigma}({\mathbb{P}}_{1},{\mathbb{H}}^{2d})$
that is not a quotient of squared norms. For example, fix an integer
$d$
and a real number
$\lambda \in \left(\left(\genfrac{}{}{0ex}{}{2d}{d}\right),{2}^{2d1}\right)$
, and take
$$R\left(\right({z}^{1},{z}^{2}),({\overline{z}}^{1},{\overline{z}}^{2}\left)\right)=\left(\right{z}^{1}{}^{2}+{z}^{2}{}^{2}{)}^{2d}\lambda {z}^{1}{z}^{2}{}^{2d}.$$
(To see that
$R\in {\mathfrak{P}}_{2}({\mathbb{P}}_{1},{\mathbb{H}}^{2d})$
, see [
DV]
.) Then
$R$
can be thought of as a Gmetric in
${\mathfrak{P}}_{2}({\mathbb{P}}_{2},{\mathbb{H}}^{2d})$
that does not depend on
${z}^{0}$
, and as such it does not satisfy the hypotheses of Definition 2.9 . In particular, property
$(\Sigma 3)$
fails. Nevertheless, by thinking of
$R$
as a Gmetric over
${\mathbb{P}}_{1}=\left\{\right[{z}^{1},{z}^{2}\left]\right\}$
we see
$${R}^{m}\in {\mathfrak{P}}_{\infty}({\mathbb{P}}_{2},{\mathbb{H}}^{2md}).$$
Said another way, Example 7.2 shows that sometimes a Gmetric
$R$
may be defined on a space
$X$
that has more than the necessary degrees of freedom required by
$R$
.
In this section, we will attempt to deal with this phenomenon of excessive spatial degrees of freedom. We begin with the classical Veronesetype construction from algebraic geometry. Fix a free Gmetric
$R$
with support space
${V}_{R}$
. Let
$\phi :X\to \mathbb{P}\left({V}_{R}^{*}\right)$
be the map sending
$x\in X$
to the hyperplane
$$\phi \left(x\right)=\{s\in {V}_{R};s(x)=0\}.$$
Since
$R$
is free,
$\phi $
is a well defined map. Moreover, by definition we have that
$${\phi}^{*}\left({\mathcal{O}}_{\phi \left(X\right)}\left(1\right)\right)={\mathcal{O}}_{X}\left(E\right).$$
Indeed, a linear functional on
${V}_{R}^{*}$
is simply an element of
${V}_{R}$
, and so
$\phi $
pulls back global sections of
$\mathbb{H}$
(over
$\phi \left(X\right)$
) to elements of
${V}_{R}$
, which are global sections of
$E$
.
Definition 7.3.
Let
${K}_{\phi}$
denote the kernel of the map
$D\phi :{T}_{X}\to {T}_{\mathbb{P}\left({V}_{R}^{*}\right)}$
.
In homogeneous coordinates, the mapping
$\phi $
can be given as follows: if
$\{{s}^{1},...,{s}^{N}\}$
is a basis for
${H}^{0}(X,E)$
, then up to a choice of homogeneous coordinates,
$\phi =[{s}^{0},...,{s}^{N}]$
. Thus the fiber of
${K}_{\phi}$
over
$x$
consists of all those vectors
$\xi $
such that
$$d{f}^{j}\left(x\right)\xi =0,1\le j\le N,$$
where
${f}^{j}={s}^{j}/{s}^{0}$
and the basis is chosen so that
${s}^{0}\left(x\right)\ne 0$
. This is equivalent to the slightly more geometric statement that, with
$s=({s}^{1},...,{s}^{N})$
, there exists
$\lambda \in \mathbb{C}$
such that
$$\partial s\left(x\right)\xi =\lambda s\left(x\right).$$
Recall that we write
${e}^{\phi}$
for the Hermitian metric on
$E$
induced by
$R$
.
Definition 7.4.
The free Gmetric
$R\in {\mathfrak{P}}_{2}(X,E)$
is said to be critically convex if outside an analytic subvariety of
$X$
,
$$ker\left(\sqrt{1}\partial \overline{\partial}\phi \right)\subset {K}_{\phi}.$$
That is to say, if
$\sqrt{1}\partial \overline{\partial}\phi (\xi ,\xi )=0$
then
$\left(D\phi \right)\xi =0$
.
Remark.
Note that the freeness hypothesis can be dropped for the purpose of this definition. Indeed, if
$R$
is not free then the map
$\phi $
is only defined off an analytic subvariety of codimension 2 in
$X$
.
Definition 7.5.
A WGCS metric is a Gmetric
$R\in {\mathfrak{P}}_{1}(X,E)$
that satisfies (
$\Sigma $
1) and (
$\Sigma $
2) of Definition 6.2 , and also has the following property.

(K)
$R$
is critically convex.
We write
$R\in {\mathfrak{P}}_{2}^{\u266d}(X,E)$
.
With this definition, we have the following result.
Theorem 5.
If
$R\in {\mathfrak{P}}_{2}^{\u266d}(X,E)$
then there is an integer
$N\ge 0$
such that for all
$m\ge N$
,
$${R}^{m}\in {\mathfrak{P}}_{\infty}(X,{E}^{m}).$$
The necessity of basic zeros
The methods already introduced allow us to show that in fact, any Gmetric that is a root of a sum of squared norms can have only basic zeros.
Proposition 7.6.
If
$R$
is a Gmetric such that
${R}^{m}$
is a sum of squared norms for some nonnegative integer
$m$
then
$R$
has only basic zeros.
Proof. Without loss of generality, we may assume that the metric
$R$
in question is free. Indeed, as we have now shown several times, after resolution of singularities we can peel off the part of the resolved Gmetric that vanishes on a divisor.
Suppose now that
$R\in {\mathfrak{P}}_{1}(X,E)$
is a free metric such that some power of
$R$
is a sum of squared norms. To arrive at a contradiction, suppose
$R$
has a zero
$x\in {\mathcal{Z}}_{R}$
. Let
$$M\stackrel{\iota}{\hookrightarrow}X$$
be a Riemann surface in
$X$
passing through
$x$
. Write
$S={\iota}^{*}R$
. Then as before
$$x\in {\mathcal{Z}}_{S},{\iota}^{*}{V}_{R}={V}_{S}and{V}_{S}\text{is free.}$$
Moreover, there is an integer
$m\ge 0$
such that
$${S}^{m}\in {\mathfrak{P}}_{\infty}(M,m({\iota}^{*}E\left)\right).$$
But it follows that the zero locus of
${S}^{m}$
is a divisor. By taking roots, we see that the zero locus
$D$
of
$S$
is a
$\mathbb{Q}$
divisor.
We claim that
$D$
is a
$\mathbb{Z}$
divisor. Indeed, suppose some point
$p$
is in the support of
$D$
. Then
$S\left(p\right)=0$
. Choose a local coordinate
$z$
on
$M$
vanishing at
$p$
. We claim that since
$S$
is nonnegative, we must have
$$S\left(z\right)=a\leftz{}^{2k}\right(P(z,\overline{z}))$$
for some realanalytic function
$P(z,\overline{z})$
that does not vanish at the origin. Indeed, if the lowest order terms contain a pure factor of
$z$
or
$\overline{z}$
, then
$S$
must have both signs. Thus the multiplicity of
$D$
at
$p$
is an integer, as claimed.
Since
$S$
has an analytic zero locus
$D$
, we can write
$$S=\stackrel{~}{S}\otimes {s}_{D}{}^{2},$$
so
${V}_{S}$
is not free. This is the desired contradiction. □
Critical metrics
As we will show momentarily, the map
$\phi $
defined above naturally pushes the metric
$R$
onto a Gmetric on
$\phi \left(X\right)$
. The trouble is that
$\phi \left(X\right)$
may be singular, and we do not have a notion of Gmetrics for this setting. Even though such a definition can be given using sections of sheaves instead of line bundles, our methods of Section 3 for Gmetrics on line bundles do not extend to Gmetrics on sheaves.
The fortunate fact is that our objects are smooth on a Zariskiopen set. We will apply resolution of singularities to obtain a Gmetric on a smooth space. The construction is noncanonical, but nevertheless we will refer to the metric obtained after such a resolution as a
critical metric induced by
$R$
. Let us be more specific.
First note that the Gmetric
$R\in {\mathfrak{P}}_{1}(X,E)$
naturally induces a Gmetric in
${\mathfrak{P}}_{1}(\mathbb{P}({V}_{R}^{*}),\mathbb{H})$
, which we denote
${R}_{\phi}$
. To be explicit, if we fix a basis
$\left\{{s}^{\alpha}\right\}$
of
${H}^{0}(X,E)$
, then the elements of this basis are linear functionals on
${V}_{R}^{*}$
, and thus
$${R}_{\phi}={c}_{\alpha \overline{\beta}}{s}^{\alpha}{\overline{s}}^{\beta}$$
defines a Gmetric in
${\mathfrak{P}}_{1}(\mathbb{P}({V}_{R}^{*}),\mathbb{H})$
.
Let
${\mathcal{O}}_{\phi \left(X\right)}\left(1\right)$
be the restriction of
${\mathcal{O}}_{\mathbb{P}\left({V}_{R}^{*}\right)}(\mathbb{H})$
to
$\phi \left(X\right)$
. Consider a resolution of singularities (as in Theorem 1.1 )
$\lambda :Y\to \phi \left(X\right)$
. Let
$${\mathcal{O}}_{Y}(\Lambda ):={\lambda}^{*}{\mathcal{O}}_{X}\left(1\right)Y.$$
Then
${\lambda}^{*}{R}_{\phi}\in {\mathfrak{P}}_{1}(Y,\Lambda ).$
Another way to construct the line bundle
$\Lambda $
is as follows.
$\mathbb{H}$
is the line bundle corresponding to a hyperplane. One intersects
$\phi \left(X\right)$
with this hyperplane to obtain a divisor
$D$
on
$\phi \left(X\right)$
, and takes
$\Lambda $
to be the line bundle associated to the divisor
${\lambda}^{*}D$
in
$Y$
.
We will write
$$\hat{R}:={\lambda}^{*}{R}_{\phi},$$
even though the metric
$\hat{R}$
depends on the resolution. Every such metric
$\hat{R}$
will be called a critical metric induced by
$R$
, and we use the notation
$$Y=\hat{X}\text{and}\Lambda =\hat{E}.$$
Lemma 7.7.
If
$R\in {\mathfrak{P}}_{2}^{\u266d}(X,E)$
then
$\hat{R}\in {\mathfrak{P}}_{2}^{\Sigma}(\hat{X},\hat{E}).$
Proof. Since
$R$
has only basic zeros, the same is true for any resolution of
$R$
, and thus property
$(\Sigma 1)$
holds. Property
$(\Sigma 2)$
holds trivially. It remains only to show that property
$(\Sigma 3)$
holds.
By the critical convexity hypothesis, we see that when restricted to
$\phi \left(X\right)$
, the
$(1,1)$
form
$\sqrt{1}\partial \overline{\partial}logr(x,x)$
is nondegenerate on the complement of an analytic subvariety of
$\phi \left(X\right)$
. Thus the same is true of
${\mu}^{*}R$
on
$Y$
. □ Although
$\phi :X\to \phi \left(X\right)$
is in general not an isomorphism, the Gmetric
$R$
is completely determined by its image under
$\phi $
. It is in this sense that there are additional degrees of freedom that are not needed in the definition of
$R$
. Simply put, we have
$${\phi}^{*}{R}_{\phi}=R.$$
Now,
${R}_{\phi}$
is isomorphic to
$\hat{R}$
away from the singular set of
$\phi \left(X\right)$
, and thus if we show that
$\hat{R}$
is a root of a sum of squared norms, then the same is true of
$R$
. Indeed, suppose there are holomorphic sections
${f}_{1},...,{f}_{N}$
of the appropriate line bundle such that
$\hat{R}=\sum {f}_{j}{}^{2}$
. Then
$$R=\sum \left{\phi}^{*}\right({\lambda}^{1}{)}^{*}{f}_{j}{}^{2}$$
holds away from an analytic subset. But as the sections are bounded (since this is so on
$\hat{X}$
), they extend across the analytic subset by Riemann's Removable Singularities Theorem. Thus
$R$
can be written as a sum of squared norms. Of course, a similar argument holds if we replace
$R$
by some power of
$R$
.
Proof of Theorem 5 : Let
$\hat{R}\in {\mathfrak{P}}_{2}^{\Sigma}(\hat{X},\hat{E})$
be a critical metric induced by
$R$
. (The fact that
$\hat{R}$
is a
$\Sigma $
GCS metric is part of the hypothesis.) Now apply Theorem 4 to
$\hat{R}$
and the Gmetric
$P\equiv 1$
in
${\mathfrak{P}}_{1}(\hat{X},\mathcal{O})$
to see that
$${\hat{R}}^{m}\in {\mathfrak{P}}_{\infty}(\hat{X},{\hat{E}}^{m}).$$
As we just discussed, this means that
${R}^{m}\in {\mathfrak{P}}_{\infty}(X,{E}^{m})$
. □
8 Proof of Theorem 2
Suppose given a Gmetric
$P\in {\mathfrak{P}}_{1}(X,E)$
with only basic zeros. We begin by reducing to the case where
$P$
is a free metric. This reduction is, by this point in the paper, standard: we find a resolution of singularities
$\mu :\stackrel{~}{X}\to X$
such that
${\mu}^{*}{V}_{P}={V}_{\stackrel{~}{P}}\otimes {s}_{D}$
for some free metric
$\stackrel{~}{P}$
. Evidently
$${\mu}^{*}\mathit{\forall}P\mathit{\forall}=\mathit{\forall}\stackrel{~}{P}\mathit{\forall}\otimes {s}_{D}{}^{2},\frac{{\mu}^{*}\mathit{\forall}P\mathit{\forall}}{{\mu}^{*}P}=\frac{\mathit{\forall}\stackrel{~}{P}\mathit{\forall}}{\stackrel{~}{P}},$$
and
$\mathit{\forall}\stackrel{~}{P}{\mathit{\forall}}^{m}\stackrel{~}{P}$
is a sum of squared norms if and only if this is the case for
$\mathit{\forall}P{\mathit{\forall}}^{m}P$
.
From here on we assume, without loss of generality, that
$P$
is a free Gmetric, i.e.,
$P\in {\mathfrak{P}}_{1}^{\u266f}(X,E)$
.
We begin with the following proposition.
Proposition 8.1.
If
$P\in {\mathfrak{P}}_{1}^{\u266f}(X,E)$
then
$\mathit{\forall}P\mathit{\forall}$
is critically convex.
Proof. Let
$\mathcal{A}$
be the set of points
$x\in X$
such that for some
$\xi \in {T}_{X,x}$
,
$$(\sqrt{1}\partial \overline{\partial}log\mathit{\forall}P\mathit{\forall})(\xi ,\xi )=0.$$
Let us choose a distinguished basis
${s}^{1},...,{s}^{k}$
for
${V}_{P}$
, i.e., a basis such that, with
$s=({s}^{1},...,{s}^{k})$
denoting the ktuple of this basis of global sections,
$$\mathit{\forall}P\mathit{\forall}=s\cdot \overline{s}.$$
Then, up to projective isomorphism,
$\phi =\left[s\right]$
.
We compute that
$$\sqrt{1}\partial \overline{\partial}log\mathit{\forall}P\mathit{\forall}=\frac{\sqrt{1}\partial s\wedge \overline{\partial s}}{\mathit{\forall}P\mathit{\forall}}\frac{\sqrt{1}(\overline{s}\cdot \partial s)\wedge (s\cdot \overline{\partial s})}{\mathit{\forall}P{\mathit{\forall}}^{2}}$$
The semipositivity of this
$(1,1)$
form follows from the CauchySchwartz inequality, and in particular we see that
$x\in \mathcal{A}$
if and only if there is a
$\xi \in {T}_{X,x}$
such that for some
$\lambda \in {\mathbb{C}}^{*}$
,
$$\partial s\left(x\right)\xi =\lambda s\left(x\right).$$
But this condition means exactly that we are in the critical locus of
$\phi $
, which is what we were trying to show. The proof is complete. □ Proof of Theorem 2 : By Lemma 7.7 the critical convexity of
$\mathit{\forall}P\mathit{\forall}$
implies that
$$\mathit{\forall}\hat{P}\mathit{\forall}\in {\mathfrak{P}}_{2}^{\Sigma}(\hat{X},\hat{E}).$$
The proof of Theorem 2 now follows easily from Proposition 8.1 . Indeed, Theorem 4 applies with the metrics
$\hat{P}\in {\mathfrak{P}}_{1}(\hat{X},\hat{E})$
and
$\mathit{\forall}\hat{P}\mathit{\forall}\in {\mathfrak{P}}_{2}^{\Sigma}(\hat{X},,\hat{E})$
, and then we may invoke the discussion preceding the proof of Theorem 5 . □ References

Berman, R., Bergman kernels and local holomorphic Morse inequalities. Math. Z. 248 (2004), no. 2, 325–344.

Abramowitz, D., De Jong, A. J., Smoothness, semistability, and toroidal geometry. J. Algebraic Geom. 6 (1997), no. 4, 789–801.

Bogomolov, F., Pantev, T., Weak Hironaka theorem. Math. Res. Lett. 3 (1996), no. 3, 299–307.

Boutet de Monvel, Sjöstrand, J., Sur la singularité des noyaux de Bergman et de Szegö. (French) Journées: Équations aux Dérivées Partielles de Rennes (1975), pp. 123–164. Asterisque, No. 3435, Soc. Math. France, Paris, 1976.

Catlin, D., The Bergman kernel and a theorem of Tian. Analysis and geometry in several complex variables (Katata, 1997), 1–23, Trends Math., Birkhäuser Boston, Boston, MA, 1999.

Catlin, D., D'Angelo, J., A stabilization theorem for Hermitian forms and applications to holomorphic mappings. Math. Res. Lett. 3 (1996), no. 2, 149–166.

Catlin, D., D'Angelo, J., An isometric embedding theorem for holomorphic bundles. Math. Res. Lett. 6 (1999) no. 1, 43–60.

D'Angelo, J., Inequalities from complex analysis. Carus Mathematical Monographs, 28. Mathematical Association of America, Washington, DC, 2002.

D'Angelo, J., Bordered complex Hessians. J. Geom. Anal. 11 (2001), no. 4, 561–571.

D'Angelo, J., Complex Variables Analogues of Hilbert's Seventeenth Problem To appear in Internat. J. Math.

D'Angelo, J., Varolin, D., Positivity conditions for Hermitian symmetric functions. Asian J. Math, 7, No. 4 (2003) 118.

Fefferman, C., The Bergman kernel and biholomorphic mappings of pseudoconvex domains. Invent. Math. 26 (1974), 1–65.

Hironaka, H., Resolution of singularities of an algebraic variety over a field of characteristic zero. I, II. Ann. of Math. (2) 79 (1964), 109203, 205326.

Lazarsfeld, R., Positivity in Algebraic Geometry I and II Springer 2004.

Paranjape, K., The Bogomolov–Pantev resolution, an expository account. New Trends in Algebraic Geometry (ed. K. Hulek, F. Catanese, C. Peters, M. Reid), pages 347358, Cambridge University Press 1999. (See also math.AG/9806084 in the mathematics ArXiV at http://front.math.ucdavis.edu/)

Quillen, D., On the representation of Hermitian forms as sums of squares. Invent. Math. 5 (1968) 237–242.

Zelditch, S., Szegö kernels and a theorem of Tian. Internat. Math. Res. Notices 1998, no. 6, 317–331.