Independent scholarship · Open accessThe monthly digest

Indexed text attachment

representation-theory-in-higher-categories.tex

Source for Representation Theory in Higher Categories

View paper · Download original file

Lines 1901–2000 of 3021. Source text is displayed literally.

1901 1902\section{Two Kan extensions}1903 1904Let $i:BH\to BG$ come from the inclusion of discrete groups1905$H\subseteq G$. The functor $i^*$ is restriction, with adjunctions1906$i_!\dashv i^*\dashv i_*$. Under the linearization of1907Theorem~\ref{thm:derived-linearization} these are1908\[1909 \begin{aligned}1910 \operatorname{Ind}_H^G W &= k[G]\otimes^L_{k[H]} W,\\1911 \operatorname{Coind}_H^G W1912   &= \operatorname{RHom}_{k[H]}(k[G],W).1913 \end{aligned}1914\]1915The first occurrence of $k[G]$ is a left $k[G]$, right $k[H]$1916bimodule. In the second it is a left $k[H]$, right $k[G]$ bimodule,1917with left $G$ acting on the resulting Hom by $(g f)(a)=f(ag)$.1918The tensor formulas identify the adjoints by the natural tensor--Hom1919bijections on modules. Since $k[G]$ is free over $k[H]$ on either side,1920the displayed functors can be computed degreewise on complexes.1921At the module level, a map out of the induced object is determined1922by its values on $1\otimes W$. A map from a $G$-module $M$ into the1923coinduced object is determined by evaluation at $1$, and has the1924explicit form $m\mapsto(a\mapsto\phi(am))$ for an $H$-map $\phi:M\to W$.1925 1926\begin{proposition}[Finite-index comparison]1927If $[G:H]$ is finite, induction and coinduction are naturally1928isomorphic. More generally there is a natural comparison1929\[1930 g\otimes w\longmapsto1931 \bigl(a\longmapsto \operatorname{pr}_{k[H]}(ag)w\bigr),1932\]1933where $\operatorname{pr}_{k[H]}$ is coefficient projection.1934\end{proposition}1935 1936\begin{proof}1937Projection is $H$-linear on each side. Right linearity gives the1938tensor relation, and left linearity gives $H$-equivariance of the1939resulting map in $a$. The left $G$-action sends the pairing for $g$1940to the pairing for its left translate. For a set $T$ of1941representatives of the cosets $gH$, the inverse elements $t^{-1}$1942represent the cosets $Hg$. Induction has one direct-sum copy of1943$W$ for each $t$, and coinduction has one product copy, recovered1944by evaluating at $t^{-1}$. As $\operatorname{pr}_{k[H]}(t^{-1}s)$1945is $1$ for $s=t$ and $0$ for distinct representatives, the map is1946the inclusion of the direct sum into the product. It is an1947isomorphism for finite $T$, degreewise on complexes as well.1948\end{proof}1949 1950For example, if $H=1$ and $G=\mathbb Z$, induction from $k$ yields1951finitely supported sequences on $\mathbb Z$ and coinduction yields1952all sequences, after indexing by inversion where appropriate.1953The natural comparison has proper image over every nonzero $k$.1954In the continuous compact setting the $L^2$ completion and1955boundedness of intertwining operators enter the story. Finite1956index permits a sum/product comparison in either setting.1957 1958\section{Double cosets and base change}1959 1960Let $K$ be another subgroup of $G$. The homotopy pullback1961$BK\times_{BG}BH$ is a groupoid of pairs of subgroup objects1962together with an arrow between their images. Choosing that arrow1963to go from the $H$ object to the $K$ object identifies the objects1964with $g\in G$, and morphisms with the action1965$(k,h)\cdot g=kgh^{-1}$. Its connected components are indexed by1966$K\backslash G/H$. The stabilizer at $g$ is identified with1967$L_g=K\cap gHg^{-1}$ through projection onto $K$, and the projection1968onto $H$ sends $\ell\in L_g$ to $g^{-1}\ell g$. Its classifying1969diagram is1970\[1971 \begin{tikzcd}[column sep=large,row sep=large]1972  \coprod_{[g]}BL_g \arrow[r] \arrow[d] & BH \arrow[d] \\1973  BK \arrow[r] & BG .1974 \end{tikzcd}1975\]1976For an $H$-object $W$ write $\prescript{g}{}{W}$ for its restriction1977along $L_g\to H$ given by conjugation by $g^{-1}$.1978 1979\begin{theorem}[Mackey decomposition]1980For $W\in D(k[H])$,1981\[1982 \operatorname{Res}_{K}^{G}\operatorname{Ind}_{H}^{G} W1983   \simeq \bigoplus_{[g]\in K\backslash G/H}1984       \operatorname{Ind}_{L_g}^{K}\prescript{g}{}{W}.1985\]1986\end{theorem}1987 1988\begin{proof}1989As a $(k[K],k[H])$-bimodule, $k[G]$ is the direct sum of1990$k[KgH]$. For each component the bimodule map1991\[1992 k[K]\otimes_{k[L_g]} k[H]\longrightarrow k[KgH],1993 \qquad k_0\otimes h_0\longmapsto k_0 g h_01994\]1995uses $\ell\cdot h_0=(g^{-1}\ell g)h_0$ on the second factor. It is1996well defined since $\ell g=g(g^{-1}\ell g)$. Choose representatives1997$t$ of the cosets $k_0L_g$. Every element of $KgH$ has the form1998$tgh_0$ and this form is unique. Indeed, $tgh_0=t'gh'_0$ would force1999$t'^{-1}t\in L_g$ and hence $t=t'$ and $h_0=h'_0$. This gives the2000bimodule isomorphism. Tensoring over $k[H]$ gives the stated