This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

ubsection]section 0

The spectrum of simplicial volume
with fixed fundamental group

Clara Löh
(00footnotetext: © C. Löh 2022. This work was supported by the CRC 1085 Higher Invariants (Universität Regensburg, funded by the DFG).
MSC 2010 classification: 57N65, 53C23, 20F65
)
Abstract

We study the spectrum of simplicial volume for closed manifolds with fixed fundamental group and relate the gap problem to rationality questions in bounded (co)homology. In particular, we show that in many cases this spectrum has a gap at zero. For such groups, this leads to corresponding gap results for the minimal volume entropy semi-norm and for the minimal volume entropy in dimension 44.

1 Introduction

The simplicial volume of an oriented closed connected manifold is the 1\ell^{1}-semi-norm of its \mathbb{R}-fundamental class [Gro82] (Section 2.1). The simplicial volume is connected to amenability, negative curvature, and Riemannian volume estimates [Gro82].

Definition 1.1 (spectrum of simplicial volume).

Let dd\in\mathbb{N} and let Mfdd\operatorname{Mfd}_{d} denote the class of all oriented closed connected dd-manifolds. The spectrum of simplicial volume in dimension dd is the set

SV(d):={M|MMfdd}0.\operatorname{SV}(d):=\bigl{\{}\|M\|\bigm{|}M\in\operatorname{Mfd}_{d}\bigr{\}}\subset\mathbb{R}_{\geq 0}.

Given a group Γ\Gamma, we write

SVΓ(d):={M|MMfdd,π1(M)Γ}0.\operatorname{SV}_{\Gamma}(d):=\bigl{\{}\|M\|\bigm{|}M\in\operatorname{Mfd}_{d},\ \pi_{1}(M)\cong\Gamma\bigr{\}}\subset\mathbb{R}_{\geq 0}.

A subset V0V\subset\mathbb{R}_{\geq 0} has a gap at 0 if there exists a c>0c\in\mathbb{R}_{>0} with V(0,c)=V\cap(0,c)=\emptyset. The sets SV(d)\operatorname{SV}(d) are known not to have a gap at zero whenever d4d\geq 4 (Section 1.1). However, the problem is open for the spectrum with fixed fundamental group:

Question 1.2 (gap problem with fixed fundamental group).

Let dd\in\mathbb{N} and let Γ\Gamma be a finitely presented group with dimHd(Γ;)<\dim_{\mathbb{R}}H_{d}(\Gamma;\mathbb{R})<\infty. Does the set SVΓ(d)\operatorname{SV}_{\Gamma}(d) have a gap at zero?

Fundamental groups of closed manifolds are finitely presented. In the context of homological properties of groups, it is reasonable to further restrict the class of groups: We say that a group Γ\Gamma has type FHd\operatorname{FH}_{d} if it is finitely presented and satisfies dimHd(Γ;)<\dim_{\mathbb{R}}H_{d}(\Gamma;\mathbb{R})<\infty.

In the present article, we give partial positive answers to Question 1.2 and put this problem into a geometric context.

1.1 The spectrum of simplicial volume

We first recall known results on the spectrum of simplicial volume. On the one hand, we have generic structural results:

Theorem 1.3 (general structure [HL21a, Remark 2.3][HL22, Theorem B/E]).

Let dd\in\mathbb{N}.

  1. 1.

    The set SV(d)\operatorname{SV}(d) is countable and closed under addition.

  2. 2.

    The set SV(d)\operatorname{SV}(d) is contained in the set of right-computable real numbers; in particular, if AA\subset\mathbb{N} is a subset that is recursively enumerable but not recursive, then nA2n\sum_{n\in A}2^{-n} is not in SV(d)\operatorname{SV}(d).

On the other hand, classification results in low dimensions and stable commutator length, respectively, can be used to exhibit concrete real numbers as simplicial volumes:

Theorem 1.4 ((no) gap [HL21a, Example 2.4/2.5, Theorem A]).
  1. 1.

    The sets SV(0),,SV(3)\operatorname{SV}(0),\dots,\operatorname{SV}(3) have a gap at zero.

  2. 2.

    If d4d\in\mathbb{N}_{\geq 4}, then SV(d)\operatorname{SV}(d) is dense in 0\mathbb{R}_{\geq 0}.

The most specific information is available in dimension 44:

Theorem 1.5 (dimension 44).

The set SV(4)\operatorname{SV}(4) contains

  • all non-negative rationals [HL21a, Theorem B];

  • a dense set of transcendental numbers that is linearly independent over the field of algebraic numbers [HL22, Theorem A, Theorem C];

  • certain irrational algebraic numbers [FFL21, Theorem 1.10].

The constructions from Theorem 1.5 can be performed with fundamental groups with a bounded number of generators and relations [HL21a, Section 8.4], but it is not clear from the constructions whether it is possible to fix the group.

In contrast to the closed case, the spectrum of the (locally finite) simplicial volume of oriented connected not necessarily compact manifolds without boundary in dimensions 4\geq 4 coincides with 0{}\mathbb{R}_{\geq 0}\cup\{\infty\} [HL21b].

1.2 Gaps and rationality

We show that the gap behaviour of a given fundamental group is driven by the rationality properties of the zero-norm subspace of singular homology.

Definition 1.6.

Let dd\in\mathbb{N} and let XX be a topological space or a group.

  • Then we write

    Nd(X;)\displaystyle N_{d}(X;\mathbb{R}) :={αHd(X;)|α1=0}Hd(X;),\displaystyle:=\bigl{\{}\alpha\in H_{d}(X;\mathbb{R})\bigm{|}\|\alpha\|_{1}=0\bigr{\}}\subset H_{d}(X;\mathbb{R}),
    Bd(X;)\displaystyle B^{d}(X;\mathbb{R}) :={φHd(X;)|φ is bounded}Hd(X;).\displaystyle:=\bigl{\{}\varphi\in H^{d}(X;\mathbb{R})\bigm{|}\text{$\varphi$ is bounded}\bigr{\}}\subset H^{d}(X;\mathbb{R}).
  • A subspace VHd(X;)V\subset H_{d}(X;\mathbb{R}) is rational if VHd(X;)V\cap H_{d}(X;\mathbb{Q}) generates VV over \mathbb{R}. A subspace VHd(X;)V\subset H^{d}(X;\mathbb{R}) is rational if VHd(X;)V\cap H^{d}(X;\mathbb{Q}) generates VV over \mathbb{R}.

Theorem 1.7 (Section 3).

Let d4d\in\mathbb{N}_{\geq 4} and let Γ\Gamma be a group of type FHd\operatorname{FH}_{d}. Then the following are equivalent:

  1. 1.

    The set SVΓ(d)\operatorname{SV}_{\Gamma}(d) has a gap at zero.

  2. 2.

    The set {α1|αHd(Γ;) is integral}\bigl{\{}\|\alpha\|_{1}\bigm{|}\text{$\alpha\in H_{d}(\Gamma;\mathbb{R})$ is integral}\bigr{\}} has a gap at zero.

  3. 3.

    The subspace Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) is rational in Hd(Γ;)H_{d}(\Gamma;\mathbb{R}).

  4. 4.

    The subspace Bd(Γ;)B^{d}(\Gamma;\mathbb{R}) is rational in Hd(Γ;)H^{d}(\Gamma;\mathbb{R}).

The proof in Section 3 shows that the implication 2. \Longrightarrow 1. as well as the equivalence of the properties 2., 3., and 4. also hold for d{0,1,2,3}d\in\{0,1,2,3\}.

The rationality property 4. is related to a problem of Frigerio and Sisto in the context of quasi-isometrically trivial extensions [FS20, Question 16].

1.3 Examples

The characterisation in Theorem 1.7 allows us to establish that many groups admit a positive answer to Question 1.2. Let dd\in\mathbb{N}. We write Gap(d)\operatorname{Gap}(d) for the class of all groups Γ\Gamma of type FHd\operatorname{FH}_{d} such that SVΓ(d)\operatorname{SV}_{\Gamma}(d) has a gap at zero.

If Nd(;)N_{d}(\;\cdot\;;\mathbb{R}) is trivial or the full homology, then Nd(;)N_{d}(\;\cdot\;;\mathbb{R}) is rational in Hd(;)H_{d}(\;\cdot\;;\mathbb{R}) (and similarly for cohomology). Therefore, we obtain:

Example 1.8 (base cases).

Let d4d\in\mathbb{N}_{\geq 4}. The class Gap(d)\operatorname{Gap}(d) contains the following groups:

  • all amenable groups of type FHd\operatorname{FH}_{d} because they have trivial bounded cohomology [Gro82, Iva85] (and thus SVΓ(d)={0}\operatorname{SV}_{\Gamma}(d)=\{0\} [Gro82]);

  • more generally, all boundedly acyclic groups of type FHd\operatorname{FH}_{d}; this includes the Thompson group FF [Mon22];

  • all hyperbolic groups because they are of finite type and the 1\ell^{1}-semi-norm is a norm by the duality principle and Mineyev’s results [Min01];

  • all finitely presented groups with dimHd(Γ;)1\dim_{\mathbb{R}}H_{d}(\Gamma;\mathbb{R})\leq 1;

  • all groups Γ\Gamma of type FHd\operatorname{FH}_{d} whose comparison map Hbd(Γ;)Hd(Γ;)H_{b}^{d}(\Gamma;\mathbb{R})\longrightarrow H^{d}(\Gamma;\mathbb{R}) is trivial; this includes all groups of type FHd\operatorname{FH}_{d} whose classifying space admits an amenable open cover of multiplicity at most dd [Gro82, Iva85, LS20]. Good bounds for such amenable multiplicities are, e.g., known for right-angled Artin groups [Li22]. More generally, one can also consider multiplicities of (uniformly) boundedly acyclic open covers [Iva20, LLM22].

Example 1.9 (Thompson group TT).

The Thompson group TT lies in Gap(d)\operatorname{Gap}(d) for all d4d\in\mathbb{N}_{\geq 4}: It is well-known that TT is finitely presented and has finite-dimensional cohomology in every degree [GS87]. Moreover, B(T;)B^{*}(T;\mathbb{R}) is generated by the cup-powers of the Euler class [FFLM22, MN21]. Because the Euler class is rational, we see that B(T;)B^{*}(T;\mathbb{R}) is rational. We can thus apply Theorem 1.7 to conclude.

We have the following inheritance properties (proofs are given in Section 4):

Example 1.10 (inheritance properties).

For d4d\in\mathbb{N}_{\geq 4}, we have:

  • The class Gap(d)\operatorname{Gap}(d) is closed under taking (finite) free products.

    More generally, there is an inheritance principle for graphs of groups with amenable edge groups and vertex groups in Gap(d)\operatorname{Gap}(d) (Lemma 4.2).

  • Let Γk{2,,d}Gap(k)\Gamma\in\bigcap_{k\in\{2,\dots,d\}}\operatorname{Gap}(k) and Λk{2,,d}Gap(k)\Lambda\in\bigcap_{k\in\{2,\dots,d\}}\operatorname{Gap}(k). Then

    Γ×ΛGap(d).\Gamma\times\Lambda\in\operatorname{Gap}(d).
  • If Γ\Gamma is a group that contains a finite index subgroup in Gap(d)\operatorname{Gap}(d), then also ΓGap(d)\Gamma\in\operatorname{Gap}(d).

  • Let 1AΓΛ11\longrightarrow A\longrightarrow\Gamma\longrightarrow\Lambda\longrightarrow 1 be an extension of groups with boundedly acyclic (e.g., amenable) kernel AA. If ΛGap(d)\Lambda\in\operatorname{Gap}(d) and Γ\Gamma is of type FHd\operatorname{FH}_{d}, then ΓGap(d)\Gamma\in\operatorname{Gap}(d).

  • More generally: Let f:ΓΛf\colon\Gamma\longrightarrow\Lambda be a group homomorphism that induces a surjection Hbd(f;):Hbd(Λ;)Hbd(Γ;)H^{d}_{b}(f;\mathbb{R})\colon H^{d}_{b}(\Lambda;\mathbb{R})\longrightarrow H^{d}_{b}(\Gamma;\mathbb{R}). If ΛGap(d)\Lambda\in\operatorname{Gap}(d) and Γ\Gamma is of type FHd\operatorname{FH}_{d}, then also ΓGap(d)\Gamma\in\operatorname{Gap}(d).

However, it remains an open problem whether for all groups Γ\Gamma of type FHd\operatorname{FH}_{d} the space Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) is rational or not.

If we drop the finiteness conditions, then, in general, we cannot expect a gap on integral classes:

Example 1.11.

There exists a countable group Γ\Gamma such that {α1αH2(Γ;) is integral}\{\|\alpha\|_{1}\mid\alpha\in H_{2}(\Gamma;\mathbb{R})\text{ is integral}\} has no gap at zero: For each n>0n\in\mathbb{N}_{>0}, there exists a finitely presented group Γn\Gamma_{n} with an integral class αnH2(Γn;)\alpha_{n}\in H_{2}(\Gamma_{n};\mathbb{R}) satisfying

0<αn1<1n;0<\|\alpha_{n}\|_{1}<\frac{1}{n};

for example, such groups and elements can be constructed via stable commutator length [HL21a, Theorem C]. Then the infinite free product Γ\Gamma of the (Γn)n(\Gamma_{n})_{n\in\mathbb{N}} has the claimed property. Clearly, this example Γ\Gamma is not finitely generated and dimH2(Γ;)=\dim_{\mathbb{R}}H_{2}(\Gamma;\mathbb{R})=\infty.

Taking products with fundamental groups of oriented closed connected hyperbolic manifolds and the standard cross-product estimates for 1\|\cdot\|_{1} [HL21a, Proposition 2.9] show that such examples also exist in all degrees 4\geq 4.

1.4 Gap phenomena for geometric volumes

In dimensions d4d\geq 4, it does not seem to be known whether the set of minimal volumes of all oriented closed connected smooth dd-manifolds has a gap at 0 or not. For a smooth manifold MM, the minimal volume is defined by

minvol(M):=inf{vol(M,g)|gRiem1(M)},\operatorname{minvol}(M):=\inf\bigl{\{}\operatorname{vol}(M,g)\bigm{|}g\in\operatorname{Riem_{1}}(M)\bigr{\}},

where Riem1(M)\operatorname{Riem_{1}}(M) denotes the set of all complete Riemannian metrics on MM whose sectional curvature lies everywhere in [1,1][-1,1]. The following connections with the simplicial volume are classical [Gro82, Section 0.5]:

  • Main inequality. For all oriented closed connected smooth dd-manifolds MM, we have

    M(d1)dd!minvol(M).\|M\|\leq(d-1)^{d}\cdot d!\cdot\operatorname{minvol}(M).
  • Isolation theorem. For each dd\in\mathbb{N}, there exists a constant εd>0\varepsilon_{d}\in\mathbb{R}_{>0} with the following isolation property: If MM is an oriented closed connected smooth dd-manifold with minvol(M)<εd\operatorname{minvol}(M)<\varepsilon_{d}, then M=0\|M\|=0.

It is not known whether the vanishing of simplicial volume implies the vanishing of the minimal volume. Therefore, the gap results from Section 1.3 do not directly give gap results for the minimal volume with fixed fundamental group.

Similarly, the corresponding gap problem for the minimal volume entropy is open. For dd\in\mathbb{N}, we write Gapminvolent(d)\operatorname{\operatorname{Gap}_{\operatorname{minvolent}}}(d) for the class of all groups Γ\Gamma of type FHd\operatorname{FH}_{d} such that the set of minimal volume entropies minvolent(M)\operatorname{minvolent}(M) of oriented closed connected smooth dd-manifolds MM with fundamental group isomorphic to Γ\Gamma has a gap at 0. In dimension 44, gaps for simplicial volume lead to gaps for minimal volume entropy:

Corollary 1.12 (minimal volume entropy gaps in dimension 44).
  1. 1.

    We have Gap(4)Gapminvolent(4)\operatorname{Gap}(4)\subset\operatorname{\operatorname{Gap}_{\operatorname{minvolent}}}(4).

  2. 2.

    In particular, all the examples of groups in Gap(4)\operatorname{Gap}(4) listed in Section 1.3 lie in Gapminvolent(4)\operatorname{\operatorname{Gap}_{\operatorname{minvolent}}}(4).

Proof.

The second part is clear. For the first part, on the one hand, we use that the minimal volume entropy is a linear upper bound for the simplicial volume [BCG91]; on the other hand, in dimension 44, the vanishing of simplicial volume implies the vanishing of the minimal entropy [SS09, Theorem A] and whence of the minimal volume entropy [BCG91]. ∎

The volume entropy semi-norm E\|\cdot\|_{E} is equivalent to the 1\ell^{1}-semi-norm on singular homology [BS19]. Let GapE(d)\operatorname{\operatorname{Gap}_{E}}(d) be the class of all groups Γ\Gamma of type FHd\operatorname{FH}_{d} such that the set of volume entropy semi-norms [M]E\|[M]_{\mathbb{R}}\|_{E} of oriented closed connected smooth dd-manifolds MM with fundamental group isomorphic to Γ\Gamma has a gap at 0.

Corollary 1.13 (volume entropy semi-norm gaps).

Let dd\in\mathbb{N}.

  1. 1.

    We have Gap(d)GapE(d)\operatorname{Gap}(d)\subset\operatorname{\operatorname{Gap}_{E}}(d).

  2. 2.

    In particular, all the examples of groups in Gap(d)\operatorname{Gap}(d) listed in Section 1.3 lie in GapE(d)\operatorname{\operatorname{Gap}_{E}}(d).

Proof.

The first part follows from the fact that E\|\cdot\|_{E} and 1\|\cdot\|_{1} are equivalent on singular homology [BS19, Theorem 1.3], whence on fundamental classes of smooth manifolds. The second part is clear. ∎

The smooth Yamabe invariant can be viewed as a curvature integral sibling of the minimal volume, defined in terms of scalar curvature instead of sectional/Riemannian curvature. If d5d\in\mathbb{N}_{\geq 5} and Γ\Gamma is of type FHd\operatorname{FH}_{d}, then it is known that the truncated smooth Yamabe invariant on oriented closed connected smooth spin dd-manifolds with fundamental group isomorphic to Γ\Gamma has a gap at 0; this is implicitly contained in the surgery inheritance results for this version of the Yamabe invariant [ADH13, Section 1.4].

Organisation of this article

Basic notions are recalled in Section 2. In Section 3, we prove Theorem 1.7. Finally, Section 4 treats the inheritance properties listed in Section 1.3.

Acknowledgements

I would like to thank Bernd Ammann and Francesco Fournier-Facio for interesting discussions on related topics and the anonymous referee for carefully reading the manuscript.

2 Preliminaries

We collect basic terminology and properties on simplicial volume and bounded cohomology [Gro82].

2.1 The 1\ell^{1}-semi-norm and simplicial volume

Definition 2.1 (1\ell^{1}-semi-norm).

Let XX be a space or a group and let dd\in\mathbb{N}. For αHd(X;)\alpha\in H_{d}(X;\mathbb{R}), we set

α1:=inf{|c|1|cCd(X;)c=0[c]=α}0.\|\alpha\|_{1}:=\inf\bigl{\{}|c|_{1}\bigm{|}\text{$c\in C_{d}(X;\mathbb{R})$, $\partial c=0$, $[c]=\alpha$}\bigr{\}}\in\mathbb{R}_{\geq 0}.

Here, C(X;)C_{*}(X;\mathbb{R}) denotes the singular chain complex if XX is a space; if XX is a group, C(X;)C_{*}(X;\mathbb{R}) can be taken to be the chain complex of the simplicial resolution or the singular chain complex of a classifying space BΓB\Gamma (these chain complexes are boundedly chain homotopy equivalent with respect to ||1|\cdot|_{1}). Moreover, ||1|\cdot|_{1} denotes the 1\ell^{1}-norm on C(X;)C_{*}(X;\mathbb{R}) with respect to the basis given by all singular simplices (or all simplicial tuples, respectively).

The 1\ell^{1}-semi-norm on H(;)H_{*}(\;\cdot\;;\mathbb{R}) is functorial in the following sense: If f:XYf\colon X\longrightarrow Y is a continuous map (or group homomorphism, respectively) and αHd(X;)\alpha\in H_{d}(X;\mathbb{R}), then

Hd(f;)(α)1α1.\bigl{\|}H_{d}(f;\mathbb{R})(\alpha)\bigr{\|}_{1}\leq\|\alpha\|_{1}.
Definition 2.2 (simplicial volume [Mun80, Gro82]).

The simplicial volume of an oriented closed connected dd-manifold MM is defined as

M:=[M]1,\|M\|:=\bigl{\|}[M]_{\mathbb{R}}\bigr{\|}_{1},

where [M]Hd(M;)[M]_{\mathbb{R}}\in H_{d}(M;\mathbb{R}) denotes the \mathbb{R}-fundamental class of MM.

2.2 Bounded cohomology and duality

The bounded cohomology of groups or spaces is Hb(;):=H(C(;)#),H_{b}^{*}(\;\cdot\;;\mathbb{R}):=H^{*}\bigl{(}C_{*}(\;\cdot\;;\mathbb{R})^{\#}\bigr{)}, where C(;)#C_{*}(\;\cdot\;;\mathbb{R})^{\#} denotes the topological dual with respect to ||1|\cdot|_{1} (the latter is introduced in Definition 2.1). Forgetting boundedness induces a natural transformation comp:Hb(;)H(;)\operatorname{comp}^{*}\colon H^{*}_{b}(\;\cdot\;;\mathbb{R})\Longrightarrow H^{*}(\;\cdot\;;\mathbb{R}), the comparison map. Classes in the image of the comparison map are called bounded. Evaluating cocycles on cycles induces a Kronecker product ,\langle\;\cdot\;,\!\;\cdot\;\rangle, which is compatible with the comparison map.

Proposition 2.3 (duality principle [Gro82, p. 16]).

Let dd\in\mathbb{N}, let XX be a space/group, and let αHd(X;)\alpha\in H_{d}(X;\mathbb{R}). Then

α1=sup{1φ|φHbd(X;),φ,α=1}.\|\alpha\|_{1}=\sup\Bigl{\{}\frac{1}{\|\varphi\|_{\infty}}\Bigm{|}\varphi\in H^{d}_{b}(X;\mathbb{R}),\ \langle\varphi,\alpha\rangle=1\Bigr{\}}.

We will also use the following version of the duality principle:

Corollary 2.4.

Let dd\in\mathbb{N}, let XX be a space/group with dimHd(X;)<\dim_{\mathbb{R}}H_{d}(X;\mathbb{R})<\infty. Then

Bd(X;)={φHd(X;)|αNd(X;)φ,α=0}.B^{d}(X;\mathbb{R})=\bigl{\{}\varphi\in H^{d}(X;\mathbb{R})\bigm{|}\forall_{\alpha\in N_{d}(X;\mathbb{R})}\;\;\;\langle\varphi,\alpha\rangle=0\bigr{\}}.
Proof.

By the duality principle (Proposition 2.3), we have

Nd(X;)\displaystyle N_{d}(X;\mathbb{R}) ={αHd(X;)|φHbd(X;)φ,α=0}\displaystyle=\bigl{\{}\alpha\in H_{d}(X;\mathbb{R})\bigm{|}\forall_{\varphi\in H_{b}^{d}(X;\mathbb{R})}\;\;\;\langle\varphi,\alpha\rangle=0\bigr{\}}
={αHd(X;)|φBd(X;)φ,α=0}.\displaystyle=\bigl{\{}\alpha\in H_{d}(X;\mathbb{R})\bigm{|}\forall_{\varphi\in B^{d}(X;\mathbb{R})}\;\;\;\langle\varphi,\alpha\rangle=0\bigr{\}}.

Because Hd(X;)H_{d}(X;\mathbb{R}) is finite-dimensional and Hd(X;)Hom(Hd(X;),)H^{d}(X;\mathbb{R})\cong_{\mathbb{R}}\operatorname{Hom}_{\mathbb{R}}(H_{d}(X;\mathbb{R}),\mathbb{R}) via the evaluation map, the annihilator

{φHd(X;)|αNd(X;)φ,α=0}\bigl{\{}\varphi\in H^{d}(X;\mathbb{R})\bigm{|}\forall_{\alpha\in N_{d}(X;\mathbb{R})}\;\;\;\langle\varphi,\alpha\rangle=0\bigr{\}}

of this null space coincides with Bd(X;)B^{d}(X;\mathbb{R}). ∎

2.3 Normed Thom realisation

Classical Thom realisation and surgery allow us to construct manifolds from group homology classes with controlled simplicial volume:

Theorem 2.5 ([HL21a, (proof of) Theorem 8.1]).

Let d4d\in\mathbb{N}_{\geq 4}. Then, there exists a constant Kd>0K_{d}\in\mathbb{N}_{>0} with the following property: If Γ\Gamma is a finitely presented group and αHd(Γ;)\alpha\in H_{d}(\Gamma;\mathbb{R}) is an integral class, then there exists an oriented closed connected dd-manifold MM with π1(M)Γ\pi_{1}(M)\cong\Gamma and a K{1,,Kd}K\in\{1,\dots,K_{d}\} such that

M=Kα1.\|M\|=K\cdot\|\alpha\|_{1}.

3 Gaps via rationality

In this section, we prove Theorem 1.7. More precisely, we show:

  • the equivalence 1. \Longleftrightarrow 2. in Section 3.1 via the mapping theorem and normed Thom realisation;

  • the equivalence 2. \Longleftrightarrow 3. in Section 3.2 through basic properties of integer lattices in vector spaces;

  • the equivalence 3. \Longleftrightarrow 4. in Section 3.3 by the duality principle;

3.1 The integral lattice

Let XX be a space or a group. A class in Hd(X;)H_{d}(X;\mathbb{R}) is called integral if it is in the image of the change of coefficients map Hd(X;)Hd(X;)H_{d}(X;\mathbb{Z})\longrightarrow H_{d}(X;\mathbb{R}). We write

Zd(X):={αHd(X;)|α is integral}Z_{d}(X):=\bigl{\{}\alpha\in H_{d}(X;\mathbb{R})\bigm{|}\text{$\alpha$ is integral}\bigr{\}}

for the \mathbb{Z}-submodule of Hd(X;)H_{d}(X;\mathbb{R}) of integral classes. Normed Thom realisation shows that SVΓ(d)\operatorname{SV}_{\Gamma}(d) is roughly the same as {α1αZd(Γ)}\{\|\alpha\|_{1}\mid\alpha\in Z_{d}(\Gamma)\}:

Proof of Theorem 1.7, 2. \Longrightarrow 1.

Let MMfddM\in\operatorname{Mfd}_{d} satisfying π1(M)Γ\pi_{1}(M)\cong\Gamma and let f:MBΓf\colon M\longrightarrow B\Gamma be the classifying map. As ff induces an isomorphism on the level of fundamental groups, we obtain from the mapping theorem [Gro82, Section 3.1] and the duality principle (Proposition 2.3) that

M=[M]1=Hd(f;)([M])1.\|M\|=\bigl{\|}[M]_{\mathbb{R}}\bigr{\|}_{1}=\bigl{\|}H_{d}(f;\mathbb{R})([M]_{\mathbb{R}})\bigr{\|}_{1}.

Moreover, [M]Hd(M;)[M]_{\mathbb{R}}\in H_{d}(M;\mathbb{R}) is an integral class and so Hd(f;)([M])Zd(Γ)H_{d}(f;\mathbb{R})([M]_{\mathbb{R}})\in Z_{d}(\Gamma).

Hence, if 1\|\cdot\|_{1} has a gap at zero on Zd(Γ)Z_{d}(\Gamma), then also SVΓ(d)\operatorname{SV}_{\Gamma}(d) has a gap at zero. ∎

Proof of Theorem 1.7, 1. \Longrightarrow 2.

Let SVΓ(d)\operatorname{SV}_{\Gamma}(d) have a gap cc at zero and let Kd>0K_{d}\in\mathbb{N}_{>0} be a constant for normed Thom realisation in dimension dd (Theorem 2.5). Then c/Kdc/K_{d} is a gap for 1\|\cdot\|_{1} on Zd(Γ)Z_{d}(\Gamma):

Let αZd(Γ)\alpha\in Z_{d}(\Gamma) with α10\|\alpha\|_{1}\neq 0. Normed Thom realisation shows that there exists an MMfddM\in\operatorname{Mfd}_{d} with π1(M)Γ\pi_{1}(M)\cong\Gamma and M=Kα1\|M\|=K\cdot\|\alpha\|_{1} with K{1,,Kd}K\in\{1,\dots,K_{d}\}. In particular, we obtain α1M/Kc/Kd\|\alpha\|_{1}\geq\|M\|/K\geq c/K_{d}, as claimed. ∎

Remark 3.1 (lattices).

Let VV be a finite-dimensional \mathbb{R}-vector space. Then VV carries a canonical topology (induced by any Euclidean inner product on VV). A lattice in VV is a \mathbb{Z}-submodule that is discrete with respect to the canonical topology. We recall two basic facts on lattices:

  • If \|\cdot\| is a norm on VV and LVL\subset V is a lattice, then {xxL{0}}\{\|x\|\mid x\in L\setminus\{0\}\} has a gap at zero.

    (The corresponding statement for semi-norms is false, in general: The semi-norm x|x12x2|x\mapsto|x_{1}-\sqrt{2}\cdot x_{2}| on 2\mathbb{R}^{2} does not have a gap on the standard lattice 22\mathbb{Z}^{2}\subset\mathbb{R}^{2}. Even worse, this semi-norm is non-degenerate on 2\mathbb{Z}^{2}.)

  • If LVL\subset V is a cocompact lattice, then VV has an \mathbb{R}-basis consisting of elements of LL.

Our main example is: Let dd\in\mathbb{N} and let XX be a space/group satisfying dimHd(X;)<\dim_{\mathbb{R}}H_{d}(X;\mathbb{R})<\infty. Then, by the universal coefficient theorem, Zd(X)Z_{d}(X) is a lattice in Hd(X;)H_{d}(X;\mathbb{R}).

3.2 Rationality of the zero-norm subspace

In the following, we consider the quotient space Qd(Γ;):=Hd(Γ;)/Nd(Γ;)Q_{d}(\Gamma;\mathbb{R}):=H_{d}(\Gamma;\mathbb{R})/N_{d}(\Gamma;\mathbb{R}). By construction, the quotient semi-norm of 1\|\cdot\|_{1} on Qd(Γ;)Q_{d}(\Gamma;\mathbb{R}) is a norm and the canonical projection π:Hd(Γ;)Qd(Γ;)\pi\colon H_{d}(\Gamma;\mathbb{R})\longrightarrow Q_{d}(\Gamma;\mathbb{R}) is isometric. We denote the quotient norm also by 1\|\cdot\|_{1}.

Proof of Theorem 1.7, 3. \Longrightarrow 2.

Let Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) be rational in Hd(Γ;)H_{d}(\Gamma;\mathbb{R}). Because Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) is rational and Zd(Γ)Z_{d}(\Gamma) is a lattice in Hd(Γ;)H_{d}(\Gamma;\mathbb{R}) (Remark 3.1), the image π(Zd(Γ))\pi(Z_{d}(\Gamma)) is a lattice in the finite-dimensional \mathbb{R}-vector space Qd(Γ;)Q_{d}(\Gamma;\mathbb{R}) [Bar08, Corollary 10.3]. In particular, the norm 1\|\cdot\|_{1} has a gap at 0 on π(Zd(Γ))\pi(Z_{d}(\Gamma)) (Remark 3.1). Therefore, also

{α1|αZd(Γ)}={π(α)1|αZd(Γ)}={β1|βπ(Zd(Γ)}\bigl{\{}\|\alpha\|_{1}\bigm{|}\alpha\in Z_{d}(\Gamma)\bigr{\}}=\bigl{\{}\|\pi(\alpha)\|_{1}\bigm{|}\alpha\in Z_{d}(\Gamma)\bigr{\}}=\bigl{\{}\|\beta\|_{1}\bigm{|}\beta\in\pi(Z_{d}(\Gamma)\bigr{\}}

has a gap at zero. ∎

Proof of Theorem 1.7, 2. \Longrightarrow 3.

Let 1\|\cdot\|_{1} have a gap cc at zero on Zd(Γ)Z_{d}(\Gamma). We show that Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) is rational in Hd(Γ;)H_{d}(\Gamma;\mathbb{R}):

Because Zd(Γ)Z_{d}(\Gamma) is a lattice in Hd(Γ;)H_{d}(\Gamma;\mathbb{R}) (Remark 3.1), there exists a tuple (v1,,vn)(v_{1},\dots,v_{n}) of elements of Zd(Γ)Z_{d}(\Gamma) that is an \mathbb{R}-basis for Hd(Γ;)H_{d}(\Gamma;\mathbb{R}) (Remark 3.1). Let αNd(Γ;)\alpha\in N_{d}(\Gamma;\mathbb{R}). We write

α=j=1nλjvj\alpha=\sum_{j=1}^{n}\lambda_{j}\cdot v_{j}

with λ1,,λn\lambda_{1},\dots,\lambda_{n}\in\mathbb{R}. Given N>0N\in\mathbb{N}_{>0}, simultaneous Dirichlet approximation [Sch80, Theorem II.1.A] shows that there exist pN,1,,pN,np_{N,1},\dots,p_{N,n}\in\mathbb{Z} and qN{1,,N}q_{N}\in\{1,\dots,N\} with

j{1,,n}|λjpN,jqN|<1qNN1/n.\forall_{j\in\{1,\dots,n\}}\;\;\;\Bigl{|}\lambda_{j}-\frac{p_{N,j}}{q_{N}}\Bigr{|}<\frac{1}{q_{N}\cdot N^{1/n}}.

Then the class αN:=j=1npN,jvj\alpha_{N}:=\sum_{j=1}^{n}p_{N,j}\cdot v_{j} lies in Zd(Γ)Z_{d}(\Gamma) and

qNααN1j=1n|qNλjpN,j|vj1j=1n1N1/nvj1.\|q_{N}\cdot\alpha-\alpha_{N}\|_{1}\leq\sum_{j=1}^{n}|q_{N}\cdot\lambda_{j}-p_{N,j}|\cdot\|v_{j}\|_{1}\leq\sum_{j=1}^{n}\frac{1}{N^{1/n}}\cdot\|v_{j}\|_{1}.

Because qNαNd(Γ;)q_{N}\cdot\alpha\in N_{d}(\Gamma;\mathbb{R}), we obtain αN1=qNααN1\|\alpha_{N}\|_{1}=\|q_{N}\cdot\alpha-\alpha_{N}\|_{1} and so the previous estimate and the gap cc show that αN1=0\|\alpha_{N}\|_{1}=0 for all large enough NN. Hence, αNNd(Γ;)Zd(Γ)\alpha_{N}\in N_{d}(\Gamma;\mathbb{R})\cap Z_{d}(\Gamma) and 1/qNαNNd(Γ;)Hd(Γ;)1/q_{N}\cdot\alpha_{N}\in N_{d}(\Gamma;\mathbb{R})\cap H_{d}(\Gamma;\mathbb{Q}).

We now consider the standard topology on the finite-dimensional \mathbb{R}-vector space Hd(Γ;)H_{d}(\Gamma;\mathbb{R}). Then the choice of the approximating coefficients shows that limN1/qNαN=α\lim_{N\to\infty}1/q_{N}\cdot\alpha_{N}=\alpha.

In conclusion, α\alpha lies in the closure of Nd(Γ;)Hd(Γ;)N_{d}(\Gamma;\mathbb{R})\cap H_{d}(\Gamma;\mathbb{Q}) with respect to the standard topology. As \mathbb{R}-subspaces of finite-dimensional \mathbb{R}-vector spaces are closed in the standard topology, α\alpha lies in the \mathbb{R}-subspace generated by Nd(Γ;)Hd(Γ;)N_{d}(\Gamma;\mathbb{R})\cap H_{d}(\Gamma;\mathbb{Q}). This shows that Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) indeed is rational. ∎

3.3 Rationality of the bounded subspace

Proof of Theorem 1.7, 3. \Longleftrightarrow 4.

This is a consequence of Corollary 2.4: By linear algebra over \mathbb{Q}, an \mathbb{R}-subspace of Hd(Γ;)H_{d}(\Gamma;\mathbb{R}) is rational if and only if its annihliator is rational in the dual \mathbb{R}-vector space. Thus, Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) is rational in Hd(Γ;)H_{d}(\Gamma;\mathbb{R}) if and only if Bd(Γ;)B^{d}(\Gamma;\mathbb{R}) is rational in the dual Hd(Γ;)H^{d}(\Gamma;\mathbb{R}) of Hd(Γ;)H_{d}(\Gamma;\mathbb{R}). ∎

4 Inheritance properties

We prove the inheritance properties from Section 1.3.

Lemma 4.1 (free products).

Let d4d\in\mathbb{N}_{\geq 4}. Then Gap(d)\operatorname{Gap}(d) is closed under taking (finite) free products.

Proof.

Let Γ,ΛGap(d)\Gamma,\Lambda\in\operatorname{Gap}(d). We show that ΓΛGap(d)\Gamma*\Lambda\in\operatorname{Gap}(d):

With Γ\Gamma and Λ\Lambda also ΓΛ\Gamma*\Lambda is of type FHd\operatorname{FH}_{d} (finitely presented groups are closed under free products and the homology is finite-dimensional by the Mayer–Vietoris sequence). By Theorem 1.7, we know that Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) and Nd(Λ;)N_{d}(\Lambda;\mathbb{R}) are rational and it suffices to show that Nd(ΓΛ;)N_{d}(\Gamma*\Lambda;\mathbb{R}) is rational:

The inclusions/projections i,ji,j and p,qp,q, respectively, of the summands of the free product ΓΛ\Gamma*\Lambda induce the Mayer–Vietoris \mathbb{R}-isomorphism φ:Hd(ΓΛ;)Hd(Γ;)Hd(Λ;)\varphi\colon H_{d}(\Gamma*\Lambda;\mathbb{R})\longrightarrow H_{d}(\Gamma;\mathbb{R})\oplus H_{d}(\Lambda;\mathbb{R}). Under this isomorphism, Nd(ΓΛ;)N_{d}(\Gamma*\Lambda;\mathbb{R}) corresponds to Nd(Γ;)Nd(Λ;)N_{d}(\Gamma;\mathbb{R})\oplus N_{d}(\Lambda;\mathbb{R}): If αNd(ΓΛ;)\alpha\in N_{d}(\Gamma*\Lambda;\mathbb{R}), then

Hd(p;)(α)1α1=0andHd(q;)(α)1α1=0\bigl{\|}H_{d}(p;\mathbb{R})(\alpha)\bigr{\|}_{1}\leq\|\alpha\|_{1}=0\quad\text{and}\quad\bigl{\|}H_{d}(q;\mathbb{R})(\alpha)\bigr{\|}_{1}\leq\|\alpha\|_{1}=0

and so φ(α)Nd(Γ;)Nd(Λ;)\varphi(\alpha)\in N_{d}(\Gamma;\mathbb{R})\oplus N_{d}(\Lambda;\mathbb{R}). Conversely, if (α,β)Nd(Γ;)Nd(Λ;)(\alpha,\beta)\in N_{d}(\Gamma;\mathbb{R})\oplus N_{d}(\Lambda;\mathbb{R}), then

φ1(α,β)1=Hd(i;)(α)+Hd(j;)(β)1α1+β10\bigl{\|}\varphi^{-1}(\alpha,\beta)\bigr{\|}_{1}=\bigl{\|}H_{d}(i;\mathbb{R})(\alpha)+H_{d}(j;\mathbb{R})(\beta)\bigr{\|}_{1}\leq\|\alpha\|_{1}+\|\beta\|_{1}\leq 0

and thus φ1(α,β)Nd(ΓΛ;)\varphi^{-1}(\alpha,\beta)\in N_{d}(\Gamma*\Lambda;\mathbb{R}).

Because φ\varphi maps rational subspaces to rational subspaces, also Nd(ΓΛ;)N_{d}(\Gamma*\Lambda;\mathbb{R}) is rational. ∎

Lemma 4.2 (graphs of groups).

Let d4d\in\mathbb{N}_{\geq 4}, let GG be a graph of groups on a finite graph (V,E)(V,E), whose vertex groups (Gv)vV(G_{v})_{v\in V} lie in Gap(d)\operatorname{Gap}(d) and whose edge groups (Ge)eE(G_{e})_{e\in E} are amenable. Let Γ\Gamma be the fundamental group of GG. If Γ\Gamma is of type FHd\operatorname{FH}_{d}, then also ΓGap(d)\Gamma\in\operatorname{Gap}(d).

Proof.

By Theorem 1.7, Nd(Gv;)N_{d}(G_{v};\mathbb{R}) is rational for all vVv\in V and it suffices to show that Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) is rational.

We consider the following commutative diagram:

Hbd(Γ;)\textstyle{H^{d}_{b}(\Gamma;\mathbb{R})\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}Fb\scriptstyle{F_{b}}compΓd\scriptstyle{\operatorname{comp}^{d}_{\Gamma}}vVHbd(Gv;)\textstyle{\bigoplus_{v\in V}H^{d}_{b}(G_{v};\mathbb{R})\ignorespaces\ignorespaces\ignorespaces\ignorespaces}vVcompGvd\scriptstyle{\bigoplus_{v\in V}\operatorname{comp}^{d}_{G_{v}}}Hd(Γ;)\textstyle{H^{d}(\Gamma;\mathbb{R})\ignorespaces\ignorespaces\ignorespaces\ignorespaces}F\scriptstyle{F}vVHd(Gv;)\textstyle{\bigoplus_{v\in V}H^{d}(G_{v};\mathbb{R})}

Here, FbF_{b} and FF denote the maps induced by the inclusions of the vertex groups on bounded cohomology and cohomology, respectively. The upper horizontal arrow FbF_{b} is surjective [BBF+14]. Hence, the diagram implies that

F(Bd(Γ;))=vVBd(Gv;).F\bigl{(}B^{d}(\Gamma;\mathbb{R})\bigr{)}=\bigoplus_{v\in V}B^{d}(G_{v};\mathbb{R}).

The hypothesis that GvGap(d)G_{v}\in\operatorname{Gap}(d) for all vVv\in V shows that the right-hand side is rational. Moreover, the map FF is rational because it is induced by group homomorphisms; in particular, the kernel of FF is rational. Therefore, also Bd(Γ;)B^{d}(\Gamma;\mathbb{R}) is rational. ∎

The statement of Lemma 4.2 can be generalised to uniformly boundedly acyclic edge groups by using the corresponding result on bounded cohomology of such graphs of groups [LLM22, Theorem 8.11].

In the situation of Lemma 4.2, we have the following sufficient condition for the group Γ\Gamma to be of type FHd\operatorname{FH}_{d}: By hypothesis, all vertex groups are of type FHd\operatorname{FH}_{d}. If all edge groups are of type FHd+1\operatorname{FH}_{d+1}, then the Mayer–Vietoris sequence in the proof of Lemma 4.2 shows that Γ\Gamma is of type FHd\operatorname{FH}_{d}.

Lemma 4.3 (products).

Let d4d\in\mathbb{N}_{\geq 4} and let Γk{2,,d}Gap(k)\Gamma\in\bigcap_{k\in\{2,\dots,d\}}\operatorname{Gap}(k) and Λk{2,,d}Gap(k)\Lambda\in\bigcap_{k\in\{2,\dots,d\}}\operatorname{Gap}(k). Then Γ×ΛGap(d)\Gamma\times\Lambda\in\operatorname{Gap}(d).

Proof.

As Γ\Gamma and Λ\Lambda are of type FHd\operatorname{FH}_{d}, also Γ×Λ\Gamma\times\Lambda is of type FHd\operatorname{FH}_{d} (finitely presented groups are closed under finite products; and the cohomological Künneth theorem).

By Theorem 1.7, we know that Nk(Γ;)N_{k}(\Gamma;\mathbb{R}) and Nk(Λ;)N_{k}(\Lambda;\mathbb{R}) are rational for all k{2,,d}k\in\{2,\dots,d\} and it suffices to show that Nd(Γ×Λ;)N_{d}(\Gamma\times\Lambda;\mathbb{R}) is rational:

More precisely, we show that, under the Künneth isomorphism, Nd(Γ×Λ;)N_{d}(\Gamma\times\Lambda;\mathbb{R}) corresponds to

N:=j=0d(Nj(Γ;)Hdj(Λ;)+Hj(Γ;)Ndj(Λ;)).N:=\sum_{j=0}^{d}\bigl{(}N_{j}(\Gamma;\mathbb{R})\otimes_{\mathbb{R}}H_{d-j}(\Lambda;\mathbb{R})+H_{j}(\Gamma;\mathbb{R})\otimes_{\mathbb{R}}N_{d-j}(\Lambda;\mathbb{R})\bigr{)}.

Because the Künneth isomorphism preserves rational subspaces and because N0(;)=0N_{0}(\;\cdot\;;\mathbb{R})=0 and N1(;)=H1(;)N_{1}(\;\cdot\;;\mathbb{R})=H_{1}(\;\cdot\;;\mathbb{R}) are always rational, this would finish the proof.

The standard estimate for the homological cross-product (via the shuffle description of the Eilenberg–Zilber map) shows that NNd(Γ×Λ;)N\subset N_{d}(\Gamma\times\Lambda;\mathbb{R}). In order to prove the converse inclusion Nd(Γ×Λ;)NN_{d}(\Gamma\times\Lambda;\mathbb{R})\subset N, we proceed as follows:

We consider the bilinear form

,:Bd(Γ;)×Hd(Γ;).\langle\;\cdot\;,\!\;\cdot\;\rangle\colon B^{\leq d}(\Gamma;\mathbb{R})\times H_{\leq d}(\Gamma;\mathbb{R})\longrightarrow\mathbb{R}.

The description of the bounded part from Corollary 2.4 and elementary finite-dimensional linear algebra show that there exist families (φi)iI1(\varphi_{i})_{i\in I_{1}} in Bd(Γ;)B^{\leq d}(\Gamma;\mathbb{R}) and (αi)iI1I0(\alpha_{i})_{i\in I_{1}\sqcup I_{0}} in Hd(Γ;)H_{\leq d}(\Gamma;\mathbb{R}) with the following properties:

  • The family (αi)iI0(\alpha_{i})_{i\in I_{0}} is an \mathbb{R}-basis of Nd(Γ;)N_{\leq d}(\Gamma;\mathbb{R}).

  • The family (αi)iI0I1(\alpha_{i})_{i\in I_{0}\sqcup I_{1}} is an \mathbb{R}-basis of Hd(Γ;)H_{\leq d}(\Gamma;\mathbb{R}).

  • The family (φi)iI1(\varphi_{i})_{i\in I_{1}} is an \mathbb{R}-basis of Bd(Γ;)B^{\leq d}(\Gamma;\mathbb{R}).

  • For all i,jI1i,j\in I_{1}, we have

    φi,αj=δij.\langle\varphi_{i},\alpha_{j}\rangle=\delta_{ij}.

Similarly, we obtain such families (ψj)jJ1(\psi_{j})_{j\in J_{1}} and (βj)jJ1J0(\beta_{j})_{j\in J_{1}\sqcup J_{0}} for Λ\Lambda.

Let αNd(Γ×Λ;)\alpha\in N_{d}(\Gamma\times\Lambda;\mathbb{R}). Using the Künneth isomorphism, we write (where I:=I1I0I:=I_{1}\sqcup I_{0} and J:=J1J0J:=J_{1}\sqcup J_{0})

α=(i,j)I×Jλijαi×βj\alpha=\sum_{(i,j)\in I\times J}\lambda_{ij}\cdot\alpha_{i}\times\beta_{j}

for suitable real coefficients λij\lambda_{ij}. Let (i1,j1)I1×J1(i_{1},j_{1})\in I_{1}\times J_{1}. Then λi1,j1=0\lambda_{i_{1},j_{1}}=0 as the following computation shows:

|λi1,j1|\displaystyle|\lambda_{i_{1},j_{1}}| =|φi1×ψj1,(i,j)I×Jλijαi×βj|\displaystyle=\Bigl{|}\Bigl{\langle}\varphi_{i_{1}}\times\psi_{j_{1}},\sum_{(i,j)\in I\times J}\lambda_{ij}\cdot\alpha_{i}\times\beta_{j}\Bigr{\rangle}\Bigr{|}
=|φi1×ψj1,α|\displaystyle=\bigl{|}\langle\varphi_{i_{1}}\times\psi_{j_{1}},\alpha\rangle\bigr{|}
φi1ψj1α1\displaystyle\leq\|\varphi_{i_{1}}\|_{\infty}\cdot\|\psi_{j_{1}}\|_{\infty}\cdot\|\alpha\|_{1}
=0\displaystyle=0

Therefore, αN\alpha\in N. ∎

Lemma 4.4 (finite index supergroups).

Let d4d\in\mathbb{N}_{\geq 4} and let Γ\Gamma be a group that contains a finite index subgroup Λ\Lambda with ΛGap(d)\Lambda\in\operatorname{Gap}(d). Then ΓGap(d)\Gamma\in\operatorname{Gap}(d).

Proof.

By Theorem 1.7, Nd(Λ;)N_{d}(\Lambda;\mathbb{R}) is rational and it suffices to show that Nd(Γ;)N_{d}(\Gamma;\mathbb{R}) is rational and that Γ\Gamma has type FHd\operatorname{FH}_{d}:

Let i:ΛΓi\colon\Lambda\longrightarrow\Gamma denote the inclusion. Because [Γ:Λ]<[\Gamma:\Lambda]<\infty and [Γ:Λ][\Gamma:\Lambda] is a unit in \mathbb{R}, there is a homological transfer map td:Hd(Γ;)Hd(Λ;)t_{d}\colon H_{d}(\Gamma;\mathbb{R})\longrightarrow H_{d}(\Lambda;\mathbb{R}), which satisfies

Hd(i;)td=[Γ:Λ]idHd(Γ;).H_{d}(i;\mathbb{R})\circ t_{d}=[\Gamma:\Lambda]\cdot\operatorname{id}_{H_{d}(\Gamma;\mathbb{R})}.

In particular, dimHd(Γ;)dimHd(Λ;)<\dim_{\mathbb{R}}H_{d}(\Gamma;\mathbb{R})\leq\dim_{\mathbb{R}}H_{d}(\Lambda;\mathbb{R})<\infty. Moreover, because Γ\Gamma contains a finitely presented subgroup of finite index (namely Λ\Lambda), also Γ\Gamma is finitely presented.

We now show that Nd(Γ;)=Hd(i;)(Nd(Λ;))N_{d}(\Gamma;\mathbb{R})=H_{d}(i;\mathbb{R})\bigl{(}N_{d}(\Lambda;\mathbb{R})\bigr{)}: Clearly, the right-hand side is contained in Nd(Γ;)N_{d}(\Gamma;\mathbb{R}). Conversely, let αNd(Γ;)\alpha\in N_{d}(\Gamma;\mathbb{R}). We consider α~:=1/[Γ:Λ]td(α)Hd(Λ;)\widetilde{\alpha}:=1/[\Gamma:\Lambda]\cdot t_{d}(\alpha)\in H_{d}(\Lambda;\mathbb{R}). The explicit construction of the transfer tdt_{d} through lifts of singular simplices shows that

α~11[Γ:Λ][Γ:Λ]α1=0.\|\widetilde{\alpha}\|_{1}\leq\frac{1}{[\Gamma:\Lambda]}\cdot[\Gamma:\Lambda]\cdot\|\alpha\|_{1}=0.

Hence, α~Nd(Λ;)\widetilde{\alpha}\in N_{d}(\Lambda;\mathbb{R}). By construction,

α=1[Γ:Λ]Hd(i;)(td(α))=Hd(i;)(α~).\alpha=\frac{1}{[\Gamma:\Lambda]}\cdot H_{d}(i;\mathbb{R})\bigl{(}t_{d}(\alpha)\bigr{)}=H_{d}(i;\mathbb{R})(\widetilde{\alpha}).

This proves the claimed description of Nd(Γ;)N_{d}(\Gamma;\mathbb{R}).

Finally, because Hd(i;)H_{d}(i;\mathbb{R}) preserves rational subspaces, the rationality of the subspace Nd(Λ;)N_{d}(\Lambda;\mathbb{R}) implies the rationality of Nd(Γ;)N_{d}(\Gamma;\mathbb{R}).

Alternatively, one could also use the cohomological transfer in (bounded) cohomology. ∎

Lemma 4.5 (epis on bounded cohomology).

Let d4d\in\mathbb{N}_{\geq 4}, let f:ΓΛf\colon\Gamma\longrightarrow\Lambda be a group homomorphism that induces a surjection Hbd(f;):Hbd(Λ;)Hbd(Γ;)H^{d}_{b}(f;\mathbb{R})\colon H^{d}_{b}(\Lambda;\mathbb{R})\longrightarrow H^{d}_{b}(\Gamma;\mathbb{R}), let ΛGap(d)\Lambda\in\operatorname{Gap}(d), and let Γ\Gamma be of type FHd\operatorname{FH}_{d}. Then ΓGap(d)\Gamma\in\operatorname{Gap}(d).

Proof.

By Theorem 1.7, Bd(Λ;)B^{d}(\Lambda;\mathbb{R}) is rational and it suffices to show that Bd(Γ;)B^{d}(\Gamma;\mathbb{R}) is rational. The commutative diagram

Hbd(Λ;)\textstyle{H^{d}_{b}(\Lambda;\mathbb{R})\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}Hbd(f;)\scriptstyle{H^{d}_{b}(f;\mathbb{R})}compΛd\scriptstyle{\operatorname{comp}^{d}_{\Lambda}}Hbd(Γ;)\textstyle{H^{d}_{b}(\Gamma;\mathbb{R})\ignorespaces\ignorespaces\ignorespaces\ignorespaces}compΓd\scriptstyle{\operatorname{comp}^{d}_{\Gamma}}Hd(Λ;)\textstyle{H^{d}(\Lambda;\mathbb{R})\ignorespaces\ignorespaces\ignorespaces\ignorespaces}Hd(f;)\scriptstyle{H^{d}(f;\mathbb{R})}Hd(Γ;)\textstyle{H^{d}(\Gamma;\mathbb{R})}

and the surjectivity of the upper arrow Hbd(f;)H^{d}_{b}(f;\mathbb{R}) imply that

Bd(Γ;)=Hd(f;)(Bd(Λ;)).B^{d}(\Gamma;\mathbb{R})=H^{d}(f;\mathbb{R})\bigl{(}B^{d}(\Lambda;\mathbb{R})\bigr{)}.

As Bd(Λ;)B^{d}(\Lambda;\mathbb{R}) is rational in Hd(Λ;)H^{d}(\Lambda;\mathbb{R}) and as the induced homomorphism Hd(f;)H^{d}(f;\mathbb{R}) preserves rationality, we obtain that also Bd(Γ;)B^{d}(\Gamma;\mathbb{R}) is rational. ∎

Lemma 4.6 (boundedly acyclic extensions).

Let d4d\in\mathbb{N}_{\geq 4}, let 1AΓΛ11\longrightarrow A\longrightarrow\Gamma\longrightarrow\Lambda\longrightarrow 1 be an extension of groups with boundedly acyclic kernel AA, let ΛGap(d)\Lambda\in\operatorname{Gap}(d), and let Γ\Gamma be of type FHd\operatorname{FH}_{d}. Then ΓGap(d)\Gamma\in\operatorname{Gap}(d).

Proof.

Let π:ΓΛ\pi\colon\Gamma\longrightarrow\Lambda be the epimorphism of the given short exact sequence. Because kerπA\ker\pi\cong A is boundedly acyclic, the map Hbd(π;):Hbd(Λ;)Hbd(Γ;)H^{d}_{b}(\pi;\mathbb{R})\colon H^{d}_{b}(\Lambda;\mathbb{R})\longrightarrow H^{d}_{b}(\Gamma;\mathbb{R}) is an isomorphism; this can be seen from the Hochschild–Serre spectral sequence in bounded cohomology [Mon01, Chapter 12]. Therefore, Lemma 4.5 applies. ∎

Statements and declarations

Funding. This work was supported by the CRC 1085 Higher Invariants (Universität Regensburg, funded by the DFG).

Conflicts of interest. There are no conflicts of interest to declare.

Author contribution. This is a single-author paper. The contribution is 100%100\%.

Data availability statement. No external data is processed in this project.

References

  • [ADH13] Bernd Ammann, Mattias Dahl, and Emmanuel Humbert. Smooth Yamabe invariant and surgery. J. Differential Geom., 94(1):1–58, 2013.
  • [Bar08] Alexander Barvinok. Integer points in polyhedra. Zürich Lectures in Advanced Mathematics. European Mathematical Society, 2008.
  • [BBF+14] Marc Bucher, Michelle. Burger, Roberto Frigerio, Alessandra Iozzi, Cristina Pagliantini, and M. Beatrice Pozzetti. Isometric embeddings in bounded cohomology. J. Topol. Anal., 6(1):1–25, 2014.
  • [BCG91] Gérard Besson, Gilles Courtois, and Sylvestre Gallot. Volume et entropie minimale des espaces localement symétriques. Invent. Math., 103(2):417–445, 1991.
  • [BS19] Ivan Babenko and Stéphane Sabourau. Volume entropy semi-norm. 2019. arXiv:1909.10803 [math.GT].
  • [FFL21] Francesco Fournier-Facio and Yash Lodha. Second bounded cohomology of groups acting on 11-manifolds and applications to spectrum problems. 2021. arXiv:2111.07931 [math.GR].
  • [FFLM22] Francesco Fournier-Facio, Clara Löh, and Marco Moraschini. Bounded cohomology and binate groups. J. Aust. Math. Soc, DOI 10.1017/S1446788722000106, 2022.
  • [FS20] Roberto Frigerio and Alessandro Sisto. Central extensions and bounded cohomology. 2020. arXiv:2003.01146 [math.GR].
  • [Gro82] Michael Gromov. Volume and bounded cohomology. Inst. Hautes Études Sci. Publ. Math., (56):5–99 (1983), 1982.
  • [GS87] Étienne Ghys and Vlad Sergiescu. Sur un groupe remarquable de difféomorphismes du cercle. Comment. Math. Helv., 62(2):185–239, 1987.
  • [HL21a] Nicolaus Heuer and Clara Löh. The spectrum of simplicial volume. Invent. Math., 223(1):103–148, 2021.
  • [HL21b] Nicolaus Heuer and Clara Löh. The spectrum of simplicial volume of non-compact manifolds. Geom. Dedicata, 215:243–253, 2021.
  • [HL22] Nicolaus Heuer and Clara Löh. Transcendental simplicial volumes. Annales de l’Institut Fourier, 2022. to appear.
  • [Iva85] Nikolai V. Ivanov. Foundations of the theory of bounded cohomology. Zap. Nauchn. Sem. Leningrad. Otdel. Mat. Inst. Steklov. (LOMI), 143:69–109, 177–178, 1985. Studies in topology, V.
  • [Iva20] Nikolai V. Ivanov. Leray theorems in bounded cohomology theory. 2020. arXiv:2012.08038 [math.AT].
  • [Li22] Kevin Li. Amenable covers of right-angled Artin groups. 2022. arXiv:2204.01162 [math.GR].
  • [LLM22] Kevin Li, Clara Löh, and Marco Moraschini. Bounded acyclicity and relative simplicial volume. 2022. arXiv:2202.05606 [math.AT].
  • [LS20] Clara Löh and Roman Sauer. Bounded cohomology of amenable covers via classifying spaces. Enseign. Math., 66(1-2):151–172, 2020.
  • [Min01] Igor Mineyev. Straightening and bounded cohomology of hyperbolic groups. Geom. Funct. Anal., 11(4):807–839, 2001.
  • [MN21] Nicolas Monod and Sam Nariman. Bounded and unbounded cohomology of homeomorphism and diffeomorphism groups. 2021. arXiv:2111.04365 [math.GT].
  • [Mon01] Nicolas Monod. Continuous bounded cohomology of locally compact groups, volume 1758 of Lecture Notes in Mathematics. Springer, 2001.
  • [Mon22] Nicolas Monod. Lamplighters and the bounded cohomology of Thompson’s group. Geom. Funct. Anal., DOI 10.1007/s00039-022-00604-9, 2022.
  • [Mun80] Hans J. Munkholm. Simplices of maximal volume in hyperbolic space, Gromov’s norm, and Gromov’s proof of Mostow’s rigidity theorem (following Thurston). In Topology Symposium, Siegen 1979 (Proc. Sympos., Univ. Siegen, Siegen, 1979), volume 788 of Lecture Notes in Math., pages 109–124. Springer, 1980.
  • [Sch80] Wolfgang M. Schmidt. Diophantine approximation, volume 785 of Lecture Notes in Mathematics. Springer, 1980.
  • [SS09] Pablo Suárez-Serrato. Minimal entropy and geometric decompositions in dimension four. Algebr. Geom. Topol., 9:365–395, 2009.

Clara Löh
    Fakultät für Mathematik, Universität Regensburg, 93040 Regensburg     [email protected], https://loeh.app.ur.de