Independent scholarship · Open accessThe monthly digest

Indexed text attachment

galois-theory-in-higher-categories.tex

Source for Galois Theory in Higher Categories

View paper · Download original file

Lines 1821–1920 of 4071. Source text is displayed literally.

1821Its fibers are \(K(\mathbb Z,2)\) and the homotopy sequence1822shows that \(Z\) is a connected \(2\)-type with1823\(\tau_{\leq1}Z\simeq Y\).1824The identity of \(Y\) cannot lift. The map1825\(\mathbb T^3\to\mathbb T^3\) that sends the third coordinate1826to its identity and leaves the first two coordinates fixed1827does lift, because it factors through the two-torus.1828The lift components in the second case form a torsor for1829\(H^2(\mathbb T^3;\mathbb Z)=\mathbb Z^3\). At a chosen lift1830the degree-one and degree-two homotopy groups of the lift1831space are \(\mathbb Z^3\) and \(\mathbb Z\), respectively.1832All these claims follow from the exterior-algebra cohomology1833of the torus and Theorem~\ref{thm:ob-postnikov}. They distinguish1834the vanishing test, the moduli of solutions, and the group of1835automorphisms of a solution.1836\end{example}1837 1838\section{Continuous lifting and convergence}1839 1840Let \(P=(P_i)\) be a pro-space as in Chapter~\ref{ch:profinite}.1841Suppose that a partial monodromy map to a space \(Y\) is1842represented by \(f_i:P_i\to Y\). For \(j\) over \(i\), write1843\(f_j:P_j\to Y\) for its pullback. The lifting problem for a1844target map \(Z\to Y\) has a useful stage description.1845 1846\begin{proposition}\label{prop:ob-pro-lifts}1847There is a natural equivalence1848\[1849\begin{aligned}1850 &\operatorname{fib}_{[f_i]}1851     \bigl(\Map_{\Pro(\SSet)}(P,Z)\longrightarrow\Map_{\Pro(\SSet)}(P,Y)\bigr)\\1852 &\qquad\simeq1853 \colim_{j\to i}1854   \operatorname{fib}_{f_j}\bigl(\Map(P_j,Z)\longrightarrow\Map(P_j,Y)\bigr).1855\end{aligned}1856\]1857For a Postnikov step as in Theorem~\ref{thm:ob-postnikov}, a1858lift exists exactly when the pulled-back obstruction vanishes1859at some refining stage.1860\end{proposition}1861 1862\begin{proof}1863Restricting to refinements of the chosen stage is cofinal in the1864filtered direction. The map in the proposition then compares1865the filtered colimits of the indicated diagrams of mapping1866spaces. The compatible basepoint diagram is constant at a point1867and induces \([f_i]\). Filtered colimits of spaces commute with1868the pullback defining its fiber. This proves the equivalence.1869A filtered colimit of spaces has a component exactly when some1870stage does, since \(\pi_0\) preserves colimits. The result now1871follows from the vanishing criterion above.1872\end{proof}1873 1874A space \(Z\) is the limit of its Postnikov tower. There is1875therefore a comparison of spaces1876\begin{equation}\label{eq:ob-pro-tower}1877 \colim_i\lim_n \Map(P_i,\tau_{\leq n} Z)1878 \longrightarrow1879 \lim_n\colim_i \Map(P_i,\tau_{\leq n} Z).1880\end{equation}1881The left side classifies maps from \(P\) to the constant pro-space1882\(Z\). The right side classifies maps from \(P\) to the pro-object1883consisting of the truncations of \(Z\). A lift for each separate1884obstruction need not use one fixed stage \(i\). A bounded target,1885or a constant source, gives an equivalence in1886\eqref{eq:ob-pro-tower}. The following computation shows how the1887two sides differ without either condition.1888 1889\begin{example}[A compatible tower with no common stage]1890Fix an abelian group \(A\neq 0\), for instance \(\mathbb Z\).1891Let1892\(P_i=\prod_{k\geq i}K(A,k)\) for \(i\geq2\), with maps1893\(P_{i+1}\to P_i\) given by inserting the basepoint in the \(i\)th1894coordinate. Let \(Z=P_2\). Every \(P_i\) is \((i-1)\)-connected.1895For fixed \(n\), the map1896\(\Map(*,\tau_{\leq n} Z)\to\Map(P_i,\tau_{\leq n}Z)\)1897is an equivalence for \(i>n\), since the target is \(n\)-truncated.1898The right side of \eqref{eq:ob-pro-tower} is consequently \(Z\)1899and is connected.1900 1901On the left side, inclusions \(P_i\to Z\) represent one compatible1902element of \(\colim_i\pi_0\Map(P_i,Z)\). This element differs1903from the component of the constant map. Indeed, for every1904refining stage \(P_j\), restrict the inclusion to1905\(K(A,j)\) using the \(j\)th factor and project \(Z\) to1906that factor. The composite is the identity and induces1907\(\id_A\) on \(\pi_j\), whereas the constant map induces zero.1908Thus every bounded observation becomes trivial after refinement,1909while the unbounded monodromy test detects a nontrivial component.1910\end{example}1911 1912There is a second issue even when the source is an ordinary1913space. The inverse limit of the spaces of lifts encodes1914compatibility among choices. Passing to their component sets1915before forming the limit can lose information. For a countable1916tower of pointed spaces, the homotopy limit is the homotopy1917equalizer of the identity and shift on the product. An explicit1918point of the homotopy equalizer consists of a point in each1919space and a path from that point to the image of the next1920point. This model yields a direct example.