New characterizations for Fock spaces Thanks: Ma is supported by NNSF of China (Grant numbers 11801572 and 12171484), the Natural Science Foundation of Hunan Province (Grant number 2023JJ20056), the Science and Technology Innovation Program of Hunan Province (Grant number 2023RC3028), and Central South University Innovation-Driven Research Programme (Grant number 2023CXQD032). Bao is partially supported by NNSF of China (Grant number 12271328). Thanks: Pan Ma is the corresponding author. Guanlong Bao, Pan Ma, Kehe Zhu Abstract . We show that the maximal Fock space F α ∞ F^{\infty}_{\alpha} on ℂ n {{\mathbb{C}}^{n}} is a Lipschitz space, that is, there exists a distance d α d_{\alpha} on ℂ n {{\mathbb{C}}^{n}} such that an entire function f f on ℂ n {{\mathbb{C}}^{n}} belongs to F α ∞ F^{\infty}_{\alpha} if and only if | f ( z ) − f ( w ) | ≤ C d α ( z , w ) |f(z)-f(w)|\leq Cd_{\alpha}(z,w) for some constant C C and all z , w ∈ ℂ n z,w\in{{\mathbb{C}}^{n}} . This can be considered the Fock space version of the following classical result in complex analysis: a holomorphic function f f on the unit ball 𝔹 n {\mathbb{B}^{n}} in ℂ n {{\mathbb{C}}^{n}} belongs to the Bloch space if and only if there exists a positive constant C C such that | f ( z ) − f ( w ) | ≤ C β ( z , w ) |f(z)-f(w)|\leq C\beta(z,w) for all z , w ∈ 𝔹 n z,w\in{\mathbb{B}^{n}} , where β ( z , w ) \beta(z,w) is the distance on 𝔹 n {\mathbb{B}^{n}} in the Bergman metric. We also present a new approach to Hardy-Littlewood type characterizations for F α p F^{p}_{\alpha} . 2020 Mathematics Subject Classification 30H20, 46E15, 46E22. 1. Introduction For \alpha>0 and 0
0, are usually called Fock spaces. Each F^{p}_{\alpha} is closed in the Lebesgue space L^{p}_{\alpha} . In particular, F^{p}_{\alpha} is a Banach space when 1\leq p\leq\infty . It is clear that L^{2}_{\alpha}=L^{2}({{\mathbb{C}}^{n}},d\lambda_{\alpha}) , where d\lambda_{\alpha}(z)=\left(\frac{\alpha}{\pi}\right)^{n}e^{-\alpha|z|^{2}}\,dv(z) is the Gaussian measure. The orthogonal projection P_{\alpha}:L^{2}({{\mathbb{C}}^{n}},d\lambda_{\alpha})\to F^{2}_{\alpha} is an integral operator, namely, P_{\alpha}f(z)=\int_{{\mathbb{C}}^{n}}e^{\alpha z\overline{w}}f(w)\,d\lambda_{\alpha}(w), (1) where z\overline{w}=z_{1}\overline{w}_{1}+\cdots+z_{n}\overline{w}_{n} . It is well known that F^{p}_{\alpha}=P_{\alpha}L^{p}_{\alpha} for all 1\leq p\leq\infty . The function K_{w}(z)=K(z,w)=e^{\alpha z\overline{w}} is the reproducing kernel of the Hilbert space F^{2}_{\alpha} . We will need to use the normalized reproducing kernels k_{w}=K_{w}/\|K_{w}\|_{2,\alpha} , which are unit vectors in F^{2}_{\alpha} . It is clear that k_{w}(z)=\frac{K(z,w)}{\sqrt{K(w,w)}}=e^{\alpha z\overline{w}-(\alpha|w|^{2}/2)}. See [ 9 ] for an introduction to Fock spaces. The main result of the paper is Theorem A below. Theorem A . Suppose \alpha>0 , f\in H({{\mathbb{C}}^{n}}) , and d is the distance function on {{\mathbb{C}}^{n}} defined by d(z,w)=\int_{{\mathbb{C}}^{n}}\left|e^{\alpha z\overline{u}}-e^{\alpha w\overline{u}}\right|\,d\lambda_{\alpha/2}(u). Then the following conditions are equivalent. (a) f\in F^{\infty}_{\alpha} . (b) There exists a positive constant C such that |f(z)-f(w)|\leq Cd(z,w),\qquad z,w\in{{\mathbb{C}}^{n}}. (c) The function Rf(z)/(1+|z|^{2}) belongs to L^{\infty}_{\alpha} , where Rf(z)=z_{1}\partial_{1}f(z)+\cdots+z_{n}\partial_{n}f(z) is the radial derivative of f with \partial_{k}f=\partial f/\partial z_{k} for 1\leq k\leq n . This result has a well-known analogue in the more classical theory of Bergman spaces. Recall that the Bergman space A^{p} , 0
0 , 0
0 there exists a constant C>0 such that C^{-1}e^{\beta^{2}|z|^{2}/(4\alpha)}\leq d_{\alpha,\beta}(z,0)\leq Ce^{\beta^{2}|z|^{2}/(4\alpha)},\qquad|z|\geq r. This clearly implies that there is another positive constant C such that C^{-1}\sqrt{e^{\beta^{2}|z|^{2}/(2\alpha)}-1}\leq d_{\alpha,\beta}(z,0)\leq C\sqrt{e^{\beta^{2}|z|^{2}/(2\alpha)}-1},\qquad|z|\geq r. On the other hand, it is clear from the unitary invariance of the Gaussian measure (or the distance d_{\alpha,\beta} ) that we can use the special points z=(z_{1},0,\cdots,0) below to obtain \lim_{|z|\to 0}\frac{d_{\alpha,\beta}(z,0)}{|z|}=\lim_{|z|\to 0}\int_{{\mathbb{C}}^{n}}\frac{|e^{\beta z\overline{u}}-1|}{|z|}\,d\lambda_{\alpha}(u)=\beta\int_{{\mathbb{C}}^{n}}|u_{1}|\,d\lambda_{\alpha}(u). (2) It follows that d_{\alpha,\beta}(z,0)\sim\sqrt{e^{\beta^{2}|z|^{2}/(2\alpha)}-1},\qquad|z|\to 0. The proof of the proposition is complete when we combine this with the estimate at the end of the previous paragraph. ∎ To simplify notation, we will write d_{\alpha}(z,w)=d_{\alpha/2,\alpha}(z,w) from this point on. This distance function arises naturally in the study of the spaces F^{p}_{\alpha} . For example, we have the following result. Lemma 3 . For any \alpha>0 we have d_{\alpha}(z,w)\sim\sup\left\{|f(z)-f(w)|:f\in F^{\infty}_{\alpha},\|f\|_{\infty,\alpha}\leq 1\right\} for z,w\in{{\mathbb{C}}^{n}} . Proof. It is well known that the integral operator P_{\alpha} defined in ( 1 ) maps L^{\infty}_{\alpha}({{\mathbb{C}}^{n}}) boundedly onto F^{\infty}_{\alpha} . It follows easily that for f\in F^{\infty}_{\alpha} we have \|f\|_{\infty,\alpha}\sim\inf\left\{\|g\|_{\infty,\alpha}:f=P_{\alpha}g,\ g\in L^{\infty}_{\alpha}({{\mathbb{C}}^{n}})\right\}. Let d^{\prime}_{\alpha}(z,w) denote the supremum in the lemma, which is also a distance on {{\mathbb{C}}^{n}} (see [ 10 ] for many other examples of distance functions induced by spaces of analytic functions). Then \displaystyle d^{\prime}_{\alpha}(z,w) \displaystyle\sim\sup_{\|g\|_{\infty,\alpha}\leq 1}|P_{\alpha}g(z)-P_{\alpha}g(w)| \displaystyle=\sup_{\|g\|_{\infty,\alpha}\leq 1}\left|\int_{{\mathbb{C}}^{n}}(e^{\alpha z\overline{u}}-e^{\alpha w\overline{u}})g(u)\,d\lambda_{\alpha}(u)\right| \displaystyle=\int_{{\mathbb{C}}^{n}}|e^{\alpha z\overline{u}}-e^{\alpha w\overline{u}}|e^{\alpha|z|^{2}/2}\,d\lambda_{\alpha}(u) \displaystyle=2^{n}\int_{{\mathbb{C}}^{n}}\left|e^{\alpha z\overline{u}}-e^{\alpha w\overline{u}}\right|\,d\lambda_{\alpha/2}(u). This proves the desired estimates. ∎ It is natural to wonder if the limit in ( 2 ) can be computed at points away from the origin. When n=1 , it is clear that \lim_{w\to z}\frac{d_{\alpha}(z,w)}{|z-w|}=\lim_{w\to z}\int_{{\mathbb{C}}}\frac{|e^{\alpha z\overline{u}}-e^{\alpha w\overline{u}}|}{|z-w|}\,d\lambda_{\alpha/2}(u)=\alpha\int_{{\mathbb{C}}}|u|\left|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u). However, the limit above does NOT exist when n>1 and z\not=0 . In fact, if we choose w=z+(t,0,\cdots,0) , then \lim_{t\to 0}\int_{{\mathbb{C}}^{n}}\frac{|1-e^{\alpha(w-z)\overline{u}}|}{|w-z|}|e^{\alpha z\overline{u}}|\,d\lambda_{\alpha/2}(u)=\alpha\int_{{\mathbb{C}}^{n}}|u_{1}e^{\alpha z\overline{u}}|\,d\lambda_{\alpha/2}(u). Other similar “partial derivatives” will yield the following sub-limits: \alpha\int_{{\mathbb{C}}^{n}}|u_{k}e^{\alpha z\overline{u}}|\,d\lambda_{\alpha/2}(u),\qquad 1\leq k\leq n, which clearly depend on k . More specifically, it follows from polar coordinates and the one-dimensional case that \int_{{\mathbb{C}}^{n}}|u_{k}e^{\alpha z\overline{u}}|\,d\lambda_{\alpha/2}(u)\sim(1+|z_{k}|)e^{\alpha|z|^{2}/2}. On the other hand, we can write \frac{d_{\alpha}(z,w)}{|z-w|}=\int_{{\mathbb{C}}^{n}}\frac{|1-e^{\alpha(w-z)\overline{u}}|}{|(w-z)\overline{u}|}\,\frac{|(w-z)\overline{u}|}{|z-w|}\,|e^{\alpha z\overline{u}}|\,d\lambda_{\alpha/2}(u). It follows from the Cauchy-Schwarz inequality for vectors in {{\mathbb{C}}^{n}} that \limsup_{w\to z}\frac{d_{\alpha}(z,w)}{|z-w|}\leq\alpha\int_{{\mathbb{C}}^{n}}|u||e^{\alpha z\overline{u}}|\,d\lambda_{\alpha/2}(u). (3) Thus we want to determine the growth rate of the integral E(z)=\int_{{\mathbb{C}}^{n}}|u|\,|e^{\alpha z\overline{u}}|\,d\lambda_{\alpha/2}(u) as |z|\to\infty , which will be used several times later on. Lemma 4 . We have \frac{\alpha}{2}\int_{{\mathbb{C}}^{n}}\left|u|^{2}|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u)=\left(n+\frac{\alpha}{2}|z|^{2}\right)e^{\alpha|z|^{2}/2} for all z\in{{\mathbb{C}}^{n}} . Proof. It follows from the reproducing property in F^{2}_{\alpha/2} that \int_{{\mathbb{C}}^{n}}\left|e^{\frac{\alpha}{2}z\overline{u}}\right|^{2}\,d\lambda_{\alpha/2}(u)=e^{\frac{\alpha}{2}|z|^{2}} for all z\in{{\mathbb{C}}^{n}} . Apply \partial^{2}/\partial z_{k}\partial\overline{z}_{k} to both sides of the above identity. The result is \left(\frac{\alpha}{2}\right)^{2}\int_{{\mathbb{C}}^{n}}|u_{k}|^{2}\left|e^{\frac{\alpha}{2}z\overline{u}}\right|^{2}\,d\lambda_{\alpha/2}(u)=\frac{\alpha}{2}\left(1+\frac{\alpha}{2}|z_{k}|^{2}\right)e^{\frac{\alpha}{2}|z|^{2}}. Summing over k , we obtain \frac{\alpha}{2}\int_{{\mathbb{C}}^{n}}|u|^{2}\left|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u)=\left(n+\frac{\alpha}{2}|z|^{2}\right)e^{\frac{\alpha}{2}|z|^{2}}, completing the proof of the lemma. ∎ Lemma 5 . For any \alpha>0 we have E(z)\sim(1+|z|)e^{\alpha|z|^{2}/2} on {{\mathbb{C}}^{n}} . Proof. We write z=rw , where r=|z| (so |w|=1 ). Then \int_{{\mathbb{C}}^{n}}e^{\alpha rw\overline{u}/2}e^{\alpha r\overline{w}u/2}\,d\lambda_{\alpha/2}(u)=\int_{{\mathbb{C}}^{n}}\left|e^{\alpha z\overline{u}/2}\right|^{2}\,d\lambda_{\alpha/2}(u)=e^{\alpha|z|^{2}/2}=e^{\alpha r^{2}/2}. Take the derivative with respect to r on both sides. We obtain \frac{\alpha}{2}\int_{{\mathbb{C}}^{n}}(\overline{w}u+w\overline{u})\left|e^{\alpha r\overline{u}}\right|\,d\lambda_{\alpha/2}(u)=\alpha re^{\alpha r^{2}/2}, or \int_{{\mathbb{C}}^{n}}\re(w\overline{u})\left|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u)=|z|e^{\alpha|z|^{2}/2}. Therefore, |z|e^{\alpha|z|^{2}/2}\leq\int_{{\mathbb{C}}^{n}}|u|\left|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u). It is clear that the integral on the right-hand side above is a strictly positive continuous function of z for |z|\leq 1 , so there is a positive constant c such that c\,(1+|z|)e^{\alpha|z|^{2}/2}\leq\int_{{\mathbb{C}}^{n}}|u|\left|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u) for all z\in{{\mathbb{C}}^{n}} . On the other hand, it follows from Hölder’s inequality and Lemma 4 that \displaystyle\left[\int_{{\mathbb{C}}^{n}}|u|\left|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u)\right]^{2} \displaystyle\leq\int_{{\mathbb{C}}^{n}}|u|^{2}\left|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u)\int_{{\mathbb{C}}^{n}}\left|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u) \displaystyle=\left(\frac{2n}{\alpha}+|z|^{2}\right)e^{\frac{\alpha}{2}|z|^{2}}e^{\frac{\alpha}{2}|z|^{2}} \displaystyle=\left(\frac{2n}{\alpha}+|z|^{2}\right)e^{\alpha|z|^{2}}. Thus there exists a positive constant C=C_{\alpha} such that \int_{{\mathbb{C}}^{n}}|u|\left|e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u)\leq C(1+|z|)\,e^{\frac{\alpha}{2}|z|^{2}} for all z\in{\mathbb{C}} . Combining this with the estimate in the previous paragraph, we complete the proof of the lemma. ∎ Finally in this section we note that for any \alpha>0 , \beta>0 , and p>1 , the function d(z,w)=\left[\int_{{\mathbb{C}}^{n}}|e^{\beta z\overline{u}}-e^{\beta w\overline{u}}|^{p}\,d\lambda_{\alpha}(u)\right]^{\frac{1}{p}} is also a distance on {{\mathbb{C}}^{n}} . For 0
0 and f\in H({{\mathbb{C}}^{n}}) . Then f\in F^{\infty}_{\alpha} if and only if there exists a positive constant C such that |f(z)-f(w)|\leq Cd_{\alpha}(z,w) for all z,w\in{{\mathbb{C}}^{n}} . Proof. The “only if” direction is a direct consequence of Lemma 3 . Alternatively, for any f\in F^{\infty}_{\alpha} we have \displaystyle|f(z)-f(w)| \displaystyle=\left|\int_{{\mathbb{C}}^{n}}(e^{\alpha z\overline{u}}-e^{\alpha w\overline{u}})f(u)\,d\lambda_{\alpha}(u)\right| \displaystyle\leq\left(\frac{\alpha}{\pi}\right)^{n}\int_{{\mathbb{C}}^{n}}|e^{\alpha z\overline{u}}-e^{\alpha w\overline{u}}||f(u)e^{-\alpha|u|^{2}/2}|e^{-\alpha|u|^{2}/2}\,dv(u) \displaystyle\leq 2^{n}\|f\|_{\infty,\alpha}\int_{{\mathbb{C}}^{n}}|e^{\alpha z\overline{u}}-e^{\alpha w\overline{u}}|\,d\lambda_{\alpha/2}(u) \displaystyle=2^{n}\|f\|_{\infty,\alpha}d_{\alpha}(z,w) for all f\in F^{\infty}_{\alpha} and z,w\in{{\mathbb{C}}^{n}} . On the other hand, if f satisfies the Lipschitz condition, then in particular, |f(z)-f(0)|\leq Cd_{\alpha}(z,0) for all z\in{{\mathbb{C}}^{n}} . By Proposition 2 , there is another positive constant C such that |f(z)-f(0)|\leq C\sqrt{e^{\alpha|z|^{2}}-1} for all z\in{{\mathbb{C}}^{n}} , which clearly implies that the function f(z)e^{-\alpha|z|^{2}/2} is bounded on {{\mathbb{C}}^{n}} . This completes the proof of the theorem. ∎ Again, to put the result above in proper perspective, we should think of it as the Fock space version of the following well-known result for holomorphic functions f on the unit ball {\mathbb{B}^{n}} : f belongs to the Bloch space {\mathcal{B}} if and only if there exists a constant C such that |f(z)-f(w)|\leq C\beta(z,w) for all z,w\in{\mathbb{B}^{n}} , where \beta(z,w) is the distance between z and w in the Bergman metric. See [ 8 ] . For a function f\in H({{\mathbb{C}}^{n}}) we will write \nabla f=(\partial_{1}f,\cdots,\partial_{n}f) for the holomorphic gradient of f . The equivalence of conditions (a), (b), and (c) below is known, and we include a simple proof here. But condition (d) appears to be new, interesting, and non-trivial. Theorem 7 . Suppose \alpha>0 and f is an entire function on {{\mathbb{C}}^{n}} . Then the following conditions are equivalent: (a) f\in F^{\infty}_{\alpha} . (b) There exists a positive constant C such that |\partial_{k}f(z)|\leq C(1+|z|)e^{\alpha|z|^{2}/2} for all z\in{{\mathbb{C}}^{n}} and 1\leq k\leq n . (c) There exists a positive constant C such that |\nabla f(z)|\leq C(1+|z|)e^{\alpha|z|^{2}/2} for all z\in{{\mathbb{C}}^{n}} . (d) There exists a positive constant C such that |Rf(z)|\leq C(1+|z|^{2})e^{\alpha|z|^{2}/2} for all z\in{{\mathbb{C}}^{n}} . Proof. It is obvious that conditions (b) and (c) are equivalent. If f\in F^{\infty}_{\alpha} , then by Theorem 6 , there exists a positive constant C such that |f(z)-f(w)|\leq C\int_{{\mathbb{C}}^{n}}\left|e^{\alpha z\overline{u}}-e^{\alpha w\overline{u}}\right|\,d\lambda_{\alpha/2}(u) for all z and w in {\mathbb{C}} . Let w=z+(h,0,\cdots,0) , divide both sides of the inequality above by |h| , and then let h\to 0 . The result is |\partial_{1}f(z)|\leq C\alpha\int_{{\mathbb{C}}^{n}}\left|u_{1}e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u). It is clear that we also have |\partial_{k}f(z)|\leq C\alpha\int_{{\mathbb{C}}^{n}}\left|u_{k}e^{\alpha z\overline{u}}\right|\,d\lambda_{\alpha/2}(u) for all 1\leq k\leq n . By Lemma 5 , there exists another constant C such that |\partial_{k}f(z)|\leq C(1+|z|)e^{\alpha|z|^{2}/2} for all 1\leq k\leq n and z\in{\mathbb{C}} . Thus condition (a) implies (b). By the Cauchy-Schwarz inequality, we have |Rf(z)|\leq|z||\nabla f(z)| , which yields \frac{|Rf(z)|}{1+|z|^{2}}\leq\frac{|z||\nabla f(z)|}{1+|z|^{2}}\leq\frac{2|\nabla f(z)|}{1+|z|} (4) for all z\in{{\mathbb{C}}^{n}} . This shows that condition (c) implies (d). Finally, we assume that condition (d) holds. For any z\in{{\mathbb{C}}^{n}} we can write \displaystyle f(z)-f(0) \displaystyle=\sum_{k=1}^{n}z_{k}\int_{0}^{1}\partial_{k}f(tz)\,dt=\int_{0}^{1}Rf(tz)\,\frac{dt}{t} (5) \displaystyle=\int_{0}^{1/2}Rf(tz)\,\frac{dt}{t}+\int_{1/2}^{1}Rf(tz)\,\frac{dt}{t}. Let I_{1}(z) and I_{2}(z) denote the two integrals above, respectively. Then \displaystyle|I_{2}(z)| \displaystyle\leq\int_{1/2}^{1}|Rf(tz)|\,\frac{dt}{t}\leq 2\int_{1/2}^{1}|Rf(tz)|\,dt \displaystyle\leq 2C\int_{0}^{1}(1+t^{2}|z|^{2})e^{\alpha t^{2}|z|^{2}/2}\,dt \displaystyle\leq 2C\left[e^{\alpha|z|^{2}/2}+\int_{0}^{|z|}se^{\alpha s^{2}/2}\,ds\right] \displaystyle=2C\left[e^{\alpha|z|^{2}/2}+\frac{1}{\alpha}\left(e^{\alpha|z|^{2}/2}-1\right)\right] \displaystyle\sim e^{\alpha|z|^{2}/2},\qquad z\in{{\mathbb{C}}^{n}}. To estimate I_{1}(z) , note that the assumption |Rf(z)|\leq C(1+|z|^{2})e^{\alpha|z|^{2}/2} is clearly equivalent to |Rf(z)|\leq C|z|(1+|z|)e^{\alpha|z|^{2}/2} (with a possibly different constant). Thus \displaystyle|I_{1}(z)| \displaystyle\leq\int_{0}^{1/2}|Rf(tz)|\,\frac{dt}{t} \displaystyle\leq C\int_{0}^{1/2}|z|(1+t|z|)e^{\alpha t^{2}|z|^{2}/2}\,dt \displaystyle\leq\frac{C}{2}|z|\left(1+\frac{|z|}{2}\right)e^{\alpha|z|^{2}/8} \displaystyle\leq C^{\prime}e^{\alpha|z|^{2}/2},\qquad z\in{{\mathbb{C}}^{n}} for another positive constant C^{\prime} . Combining the estimates for I_{1}(z) and I_{2}(z) , we find another positive constant C such that |f(z)-f(0)|\leq Ce^{\alpha|z|^{2}/2} for all z\in{{\mathbb{C}}^{n}} . This shows f\in F^{\infty}_{\alpha} and completes the proof of the theorem. ∎ We warn any inexperienced reader that the factor 1-|z|^{2} on the unit ball {\mathbb{B}^{n}} can be replaced by 1-|z| , while 1+|z|^{2} is critically different from 1+|z| in Theorem 7 above! 4. Hardy-Littlewood type theorems for Fock spaces It is known that Theorem 7 can be extended to all Fock spaces F^{p}_{\alpha} in terms of higher order derivatives. Theorem 8 . Suppose 0
0 , f\in H({{\mathbb{C}}^{n}}) , and N is a positive integer. Then f\in F^{p}_{\alpha} if and only if the functions \partial^{m}f(z)/(1+|z|)^{m} , where |m|=N , all belong to L^{p}_{\alpha} . Proof. See [ 1 , 2 , 3 , 5 , 6 ] . ∎ In this section, we provide a new approach to Theorem 8 in the one-dimensional case. This approach is based on the main theorem in [ 4 ] and is much easier than the arguments used in other papers in the literature. We begin with the case of first order derivatives. Theorem 9 . Suppose f is an entire function on {\mathbb{C}} , \alpha>0 , and 0
1/2 . Thus f^{\prime}(z)/z\in F^{p}_{\alpha} implies f^{\prime}(z)/z-f(z)/z^{2}\in F^{p}_{\alpha} , which yields f\in F^{p}_{\alpha} for p>1/2 . For 0
1/2
as well), we write
\frac{f^{\prime}(z)}{z}=\left[\frac{f^{\prime}(z)}{z}-\frac{f(z)}{z^{2}}\right]+\frac{f(z)}{z^{2}}.
By the triangle inequality for the distance
d(f_{1},f_{2})=\int_{{\mathbb{C}}}|f_{1}(z)-f_{2}(z)|^{p}\,d\lambda_{\alpha}(z)
in
F^{p}_{\alpha}
(technically, we should first work with
f_{r}(z)=f(rz)
,
0 0
.
(a)
If
0 0
, and
f\in H({{\mathbb{C}}^{n}})
. Then
f\in F^{p}_{\alpha}
if and only
if the function
Rf(z)/(1+|z|^{2})
belongs to
L^{p}_{\alpha}
. More generally, if
N
is any
positive integer, then
f\in F^{p}_{\alpha}
if and only if the function
R^{N}f(z)/(1+|z|^{2})^{N}
belongs to
L^{p}_{\alpha}
.
The “only if” parts above follows from Theorem
8
and the expression of
R^{N}f
in terms of partial derivatives. For example, if
N=2
, we have
\displaystyle R^{2}f
\displaystyle=R(z_{1}\partial_{1}f+\cdots+z_{n}\partial_{n}f)
\displaystyle=\sum_{k=1}^{n}R(z_{k}\partial_{k}f)=\sum_{k=1}^{n}\sum_{j=1}^{n}z_{j}\partial_{j}(z_{k}\partial_{k}f)
\displaystyle=\sum_{k=1}^{n}\left[z_{k}\partial_{k}f+\sum_{j=1}^{n}z_{j}z_{k}\frac{\partial^{2}f}{\partial z_{j}\partial z_{k}}\right]
\displaystyle=Rf+\sum_{j,k=1}^{n}z_{j}z_{k}\frac{\partial^{2}f}{\partial z_{j}\partial z_{k}}.
Similar formulas can be obtained for
R^{N}f
when
N
is any positive integer.
It is clear from the previous sections that, for each
\alpha>0
, the distance function
d_{\alpha}(z,w)
plays a signficant role in the study of the Fock spaces
F^{p}_{\alpha}
.
However, we have very limited information about these distance functions.
Proposition
2
gives a good estimate for
d_{\alpha}(0,z)
. A natural question is
whether or not we can use the estimate in Proposition
2
together with Weyl
unitary operators (see
[
9
]
) to obtain optimal estimates for the distance
function
d_{\alpha}(z,w)
. Our attempts so far have been unsuccessful.
Recall that Bergman spaces
A^{p}
can be characterized by Lipschitz type conditions
|f(z)-f(w)|\leq\beta(z,w)\left[g(z)+g(w)\right],
where
g\in L^{p}({\mathbb{B}^{n}},dv)
and
\beta(z,w)
is the Bergman distance between
z
and
w
. See
[
7
]
. It is natural to ask whether or not something similar is also
true for Fock spaces. We make the following conjecture here.
Conjecture 14
.
Suppose
\alpha>0
,
0