Coarse quotient mappings between metric spaces (Q499007)

Understanding the geometry of Banach spaces in the uniform category is a well-developed topic in nonlinear Banach space geometry that can be traced back at least to the mid 1930's. The relevance of understanding the geometry of Banach spaces in the coarse category was brought to light at the turn of the 21st century, when deep connections with noncommutative geometry, topology and geometric group theory were uncovered. While in the uniform theory one deals with the small scale structure, the coarse theory is related to the large scale structure. Quite surprisingly, in the Banach space setting, the uniform theory and the coarse theory are not as remote as one would expect. Indeed, a uniform equivalence is a coarse equivalence, but only a recent and delicate example of \textit{N. J. Kalton} [Math. Ann. 354, No. 4, 1247--1288 (2012; Zbl 1268.46018)] shows that there are separable Banach spaces that are coarsely equivalent but not uniformly equivalent. As far as embeddings are concerned, many results that hold for uniform embeddings are also known to hold for their coarse counterparts. It is worth noting that it is still open at the time this review is written, whether a Banach space \(X\) uniformly embeds into a Banach space \(Y\) if and only if \(X\) coarsely embeds into \(Y\). Uniform quotients have been introduced by Bates, Johnson, Lindenstrauss, Preiss and Schechtman [\textit{S. Bates} et al., Geom. Funct. Anal. 9, No. 6, 1092--1127 (1999; Zbl 0954.46014)] in the late 1990's. A viable definition of a coarse quotient has been missing until the author of the article under review filled this gap. Recall that a map \(f: X\to Y\) between two metric spaces \((X,d_X)\) and \((Y,d_Y)\) is said to be coarsely continuous if \(\omega_f(t)<\infty\) for all \(t>0\), where \(\omega_f\) is the expansion modulus of \(f\) defined by \[ \omega_f(t):=\sup\{d_Y(f(x),f(y)): d_X(x,y)\leq t\}. \] A ``co'' version of coarse continuity is defined by the author as follows. {Definition.} A map \(f: X\to Y\) is said to be co-coarsely continuous with constant \(K\in[0,\infty)\) if for every \(\varepsilon>0\) there exists \(\delta:=\delta(\varepsilon)>0\) so that for every \(x\in X\), \[ B_Y(f(x),\varepsilon)\subset f(B_X(x,\delta))^K, \] where for a subset \(Z\) of a metric space \(Y\) the notation \(Z^K\) means the \(K\)-neighborhood of \(Z\), i.e., \(Z^K:=\{y\in Y: d_Y(y,z)\leq K\text{ for some } z\in Z\}\). A map \(f: X \to Y\) is then said to be a coarse quotient map with constant \(K\) if \(f\) is coarsely continuous and co-coarsely continuous with constant \(K\). Finally, \(Y\) is a coarse quotient of \(X\) if for some \(K\in[0,\infty)\) there exists a coarse quotient map \(f: X\to Y\) with constant \(K\). Note that if \(f: X\to Y \) is a coarse quotient map with constant \(K\), then \(f\) is not necessarily surjective but \(f(X)^K=Y\). This notion of coarse quotient map fulfills all the requirements that one would expect from such a map (the composition of coarse quotient maps is a coarse quotient map, a coarse equivalence is a coarse quotient map). A key classical ingredient in the study of uniform quotients is the passage to Lipschitz quotients when working at the ultrapower level. The author proves a similar theorem for coarse quotients. {Theorem 1.} Let \(X\) and \(Y\) be Banach spaces and \(\mathcal{U}\) be a free ultrafilter on the natural numbers \(\mathbb{N}\). If \(Y\) is a coarse quotient of \(X\), then the ultrapower \(Y_{\mathcal{U}}\) is a Lipschitz quotient of the ultrapower \(X_{\mathcal{U}}\). The proof of Theorem 1 is similar to, but more involved than, its uniform counterpart. It relies in part on a delicate property specific to the Banach space framework, namely, whenever a Banach space \(Y\) is a coarse quotient of a Banach space \(X\), there exists a coarse quotient map with constant \(0\) from \(X\) onto \(Y\). This crucial observation follows from a clever transfinite argument. With Theorem 1 at hand, coarse analogues of profound results in the uniform setting follow, e.g., a Banach space that is a coarse quotient of \(L_p\), \(p\in(1,\infty)\), (resp. a Hilbert space) must be isomorphic to a linear quotient of \(L_p\) (resp. must be isomorphic to a Hilbert space). The last part of the article is devoted to the ``coarsification'' of a result of \textit{V. Lima} and \textit{N. L. Randrianarivony} [Isr. J. Math. 192, Part A, 311--323 (2012; Zbl 1271.46026)]. This result, which is based on the technical and qualitative ``fork argument'', states that \(\ell_q\) is not a uniform quotient of \(\ell_p\) when \(1<p<q<\infty\) (the critical case here is when \(q=2\) since in the other cases one can apply the results from Bates, Johnson, Lindenstrauss, Preiss and Schechtman [loc. cit.]). The author shows that the ``fork argument'' can be implemented in the coarse framework as well, and he obtains the following theorem (and some other interesting variations of it). {Theorem 2.} \(\ell_q\) is not a coarse quotient of \(\ell_p\) when \(1<p<q<\infty\). A consequence of Theorem 2 (and some deep results from the linear theory) is that a coarse quotient of \(\ell_p\), \(p\in (1,2)\), must be isomorphic to a linear quotient of \(\ell_p\). The proof of Theorem 2 was significantly simplified in [\textit{F. P. Baudier} and \textit{S. Zhang}, J. Lond. Math. Soc., II. Ser. 93, No. 2, 481--501 (2016; Zbl 1348.46025)] where the technical fork argument was circumvented by splitting its proof mechanism into two distinct quantitative problems, interesting in their own right, that can be treated by rather elementary techniques.