Kobayashi-Royden pseudometric versus Lempert function (Q415431)

For a domain \(D\subset\mathbb C^n\) let \[ \ell_D^\ast(z,w):=\{|\alpha|: \exists {\varphi\in\mathcal O(\mathbb D, D)}: \varphi(0)=z, \varphi(\alpha)=w\},\quad z,w\in D, \] be the Lempert function, and let \[ \varkappa_D(z;X):=\inf\{|\alpha|: \exists {\varphi\in\mathcal O(\mathbb D, D)}: \varphi(0)=z,\;\alpha\varphi'(0)=X\},\qquad z\in D,\quad X\in\mathbb C^n, \] be the Kobayashi--Royden pseudometric. It is known that if \(D\) is taut, then \[ \varkappa_D(z;X)=\lim\limits_{_{\substack{ z'\to z,\; X'\to X \\ \mathbb C\setminus\{0\}\ni t\to0}}}\frac1{|t|}\ell_D^\ast(z',z'+tX')=:(\mathcal D\ell_D^\ast)(z;X),\quad z\in D,\;X\in\mathbb C^n \] (cf. [\textit{M.-Y.~Pang}, Pac. J. Math. 162, No. 1, 121--141 (1944; Zbl 0788.32020)]). In fact, it suffices to assume that \(D\) is Kobayashi hyperbolic and \(\varkappa_D\) is continuous (cf. [\textit{N.~Nikolov} and \textit{P.~Pflug}, Ann. Mat. Pura Appl. (4) 187, No. 3, 547--553 (2008; Zbl 1150.32004)]). It is an open question to find for each \(n\in\mathbb N\) a domain \(D\subset \mathbb C^n\), a point \(z\in D\), and a vector \(X\in\mathbb C^n\) such that \(\varkappa_D(z;X)>(\mathcal D\ell_D^\ast)(z;X)\). The paper under review gives a partial answer to this question for \(n\geq8\). Let \(\varOmega_3\) be the spectral unit ball consisting of those \(3\times 3\) complex matrices whose all eigenvalues are in \(\mathbb D\). Let \(\widetilde\varOmega_3:=\big\{A\in\varOmega_3: \text{trace}A=0\big\}\). We may consider \(\widetilde\varOmega_3\) as a pseudoconvex domain in \(\mathbb C^8\). Put \[ A:=\left[\begin{matrix} 0&0&0 \\ 0&0&1 \\ 0&0&0\end{matrix}\right],\qquad B:=\left[\begin{matrix} 1&0&0 \\ 0&\omega&0 \\ 0&0&\omega^2\end{matrix}\right]\in\widetilde\varOmega_3, \] where \(\omega:=e^{2\pi i/3}\). Then \[ \varkappa_{\widetilde\varOmega_3}(A;B)>0=\lim_{\mathbb C\setminus\{0\}\ni t\to0}\frac{1}{|t|}\ell_{\widetilde\varOmega_3}^\ast(A,A+tB). \]