1308 0663

A FREE PRODUCT FORMULA FOR THE SOFIC DIMENSION
ROBERT GRAHAM AND MIKAEL PICHOT

arXiv:1308.0663v1 [math.DS] 3 Aug 2013
Abstract. It is proved that if G = G1 G3 G2 is free product of probability

measure preserving s-regular ergodic discrete groupoids amalgamated over an
amenable subgroupoid G3 , then the sofic dimension s(G) satisfies the equality
s(G) = h(G01 )s(G1 ) + h(G02 )s(G2 ) h(G03 )s(G3 )
where h is the normalized Haar measure on G.
Let G be a group. The sofic dimension of G is an asymptotic invariant that

accounts for the number of unital maps
Fn Sym(d)
from the Cayley ball Fn of radius n in G into the symmetric group Sym(d), where
F G is a finite set, n is an integer, d is a very large integer and the maps are
multiplicative and free up to an error > 0 relative to the normalized Hamming
distance on Sym(d) (see 1 below). If SA(F, n, , d) is the (finite) set of all such
maps, and NSA = {F , SA}, the sofic dimension of F is:
s(F ) = inf inf lim sup

log NSA(F, n, , d)
nN >0 d d log d
(so the limit is on d first, and then on and n). This definition was considered in
[DKP1] and [DKP2]. It is a combinatorial version of Voiculescus (microstate) free
entropy dimension (F ), which can be defined by a similar formula involving maps
Fn U (d)
into the unitary group U (d) (see [Voi96, Jun1]). It can be shown that the value of
s(F ) doesnt dependent on the finite generating set F of G and is therefore denoted
s(G). A limiting process allows to define of s(G) for an arbitrary group G.
The definition of the sofic dimension can be extended to probability measure
preserving (pmp) actions of countable groups, their orbit equivalence relations, and
more generally to discrete pmp groupoids. We refer to [DKP2, Definition 2.3] for
the general groupoid definition. An interesting feature of s is to provide combina-
torial proofs of statements in orbit equivalence theory (for example, Corollary 7.5
in [DKP1] reproves Gaboriaus theorem that the free groups on p generators are
pairwise non orbit equivalent using the counting method).
Let G be a pmp groupoid and assume that G = G1 G3 G2 is an amalgamated
free product over a subgroupoid G3 . A free product formula of the form
s(G) = s(G1 ) + s(G2 ) s(G3 )
is known to hold in the following cases (under some technical assumptions, for
example finitely generated and/or s-regularity):
(1) G, G1 , G2 , and G3 are a pmp equivalence relations on (X, ) and G3 is
amenable as an equivalence relation: see [DKP1, Theorem 1.2].
1
2 ROBERT GRAHAM AND MIKAEL PICHOT
(2) G, G1 , G2 , and G3 are countable groups and G3 is an amenable group: see

[DKP2, Theorem 4.10].
(3) G is the crossed product groupoid G = G1 G2 X of a pmp action (G1
G2 ) (X, ), where G1 and G2 are countable groups and X is a standard
probability space. Here G3 is assumed to be the trivial group but the action
of G1 G2 is not necessarily free (if free this is covered by (1)): see [DKP2,
Theorem 6.4].
The general strategy to establish this sort of formula was devised by Voiculescu for
the free entropy dimension: see in particular [Voi91, Voi96, Voi98].
The proofs of the above results in [DKP1, DKP2] apply distinct tools to handle
the amenable amalgamated part, namely the ConnesFeldmann-Weiss theorem in
(1), and the OrnsteinWeiss quasi-tiling theorem in (2). This was a reason why
it was hardly conceivable to incorporate an amenable amalgamated subgroup G3
in (3); in fact, the technical details would presumably (to quote [DKP2, 6]) be
formidable even if the action of G3 X is essentially free.
We follow a different approach here, based on the use of Bernoulli shifts as a
correspondence principle
groupoids equivalence relations
by which we mean that proving a result for (pmp) equivalence relations automat-
ically implies an a priori more general statement for (pmp) groupoids in a variety
of situation, and in particular for the computation of s (see 9 for more details).
The exact assumptions that we need for the free product formula are described
in the following statement, which is the main result of this paper:
Theorem 1. Let G be a discrete pmp groupoid of the form G = G1 G3 G2 , where
G1 , G2 are s-regular ergodic subgroupoids of G and G3 is an amenable groupoid,
then
s(G) = h(G01 )s(G1 ) + h(G02 )s(G2 ) h(G03 )s(G3 )
where h is the normalized Haar measure on G and G0 is the object space of G.
The more technical assumptions in this result can be weakened slightly. For
example, one way to remove the s-regularity assumption, following Voiculescus
idea, see e.g. [Voi98, Remark 4.8], is to replace the lim sup in the definition of
s(F ) by a limit along a fixed ultrafilter . What is rather unclear is the extend
to which the assumption that G3 is amenable is essential. Cohomological tools
can be used to prove a similar formula for the first L2 Betti number 1 under the
much weaker assumption that 1 (G3 ) = 0. (This is a result of Lck see [BDJ,
Theorem A.1] for the group case.) Furthermore Mineyev and Shlyakhtenko [MS]
have shown that Voiculescus non-microstate free entropy dimension satisfies
(G) = 1 (G) 0 (G) + 1 for any finitely generated group G, and therefore we
have the formula
(G1 G3 G2 ) = (G1 ) + (G2 ) (G3 )
where G1 and G2 are finitely generated groups and G3 is a group such that
1 (G3 ) = 0. A fundamental relation between the microstate and the non-microstate
approach to free entropy is provided by the BianeCapitaineGuionnet inequality
[BCG]. A free product formula for 0 has been established in [BDJ] for
amalgamation of (0 -regular) groups over an amenable subgroup (where 0 is
a technical modification of not depending on the generating set of the group,
A FREE PRODUCT FORMULA FOR THE SOFIC DIMENSION 3
see [Voi96, Section 6] and [Voi98]). We also note that the above correspondence
principle for 0 is probably less useful as the amenable part can always be handled
uniformly using the hyperfiniteness of von Neumann algebra LG3 (see in partic-
ular [Jun2]; for example, the proof in [BDJ] in the group case doesnt rely on
quasi-tilings). Concerning pmp equivalence relations, a free product formula has
been established by Gaboriau [Gab] for the cost, allowing for amalgamations over
amenable subrelations, and by Shlyakhtenko [Shl] for 0 , for free product with
trivial amalgamation.
Question 2. Can the assumption that G3 is amenable in Theorem 1 be weakened
(for example to 1 (G3 ) = 1)?
The paper is organized as follows: 1 and 2 establish basic facts about pmp
groupoids and their actions. In the case of s(G), the correspondence groupoids
equivalence relations is achieved by using the formula s(G) = s(G X0G ), where
G X0G is a Bernoulli shift, see Theorem 28 in 7 (other applications of the
correspondence principle are given in 9). The proof of this formula uses the idea
in a result of L. Bowen [Bow, Theorem 8.1] for the sofic entropy, as explained in
4. Other difficulties inherent to the groupoid setting are dealt with in 3, 5, 6
(these difficulties were avoided in [DKP2] by working with groups and their actions
rather than with general groupoids). The proof of Theorem 28 is given in Section
7. In 8 we prove a scaling formula for s(G). In 9 we prove Theorem 1 by putting
together these ingredients.
1. Review of s(G)
Recall that a discrete standard Borel groupoid G with base (=set of objects) G0 ,
source map s G G0 and range map r G G0 , is said to be probability measure
preserving (pmp) with respect to a Borel probability measure on G0 if the left
and right Haar measures h and h1 on G coincide:
h = h1
where
h(A) = Ae d(e) and h1 (A) = Ae d(e)
G0 G0
and A G is a Borel set with Ae = r1 (e) A and Ae = s1 (e) A for e G0 . Then
hG0 = h1 G0 = so we simply denote by h the measure on G .
0
A bisection is a Borel subset s G such that the restrictions of s and r to s

are Borel isomorphisms onto G0 . The set of bisections form a group called the full
group of G and are denoted [G]. A partial bisection is a Borel subset s G such
that s and r are injective in restriction to s. The set of partial bisection form a
Polish inverse monoid called the full inverse semigroup (or the full pseudogroup)
of G and denoted JGK. For s JGK let dom(s) = s1 s G0 and ran(s) = ss1 G0 .
For example, if G = {1, . . . , d}2 is the transitive equivalence relation on the set
{1, . . . d} with d Z1 elements, then [G] = Sym(d) is the symmetric group on d
letters and JGK is the inverse semigroup of partial permutations. We denote the
latter by JdK.
The semigroup JGK (and [G] JGK) is Polish with respect to the uniform distance
s t = h{e G0 se te}.
If G = {1, . . . , d}2 then the uniform distance is the normalized Hamming distance
on JdK.
The (von Neumann) trace on JGK is given by
(s) = h(s Ge ) = h{e G0 se = e}.
It is the restriction to JGK LG of the finite trace on the von Neumann algebra LG
of G.
We have
s t = (s1 s) + (t1 t) (s1 st1 t) (st1 ).
We will write tr for the trace on JdK. So
tr() = d1 number of fixed points of JdK.
If F JGK is a finite subset and n Z1 then Fn denotes the set of all products
of at most n elements of F = F F 1 {Id}. For F JGK we let F JGK denote
the set of sums of elements of F with pairwise orthogonal domains and pairwise
orthogonal ranges.
By definition, G is sofic if its full inverse semigroup JGK is sofic:
Definition 3. A pmp groupoid G is called sofic if for every finite set F JGK, > 0
and n Z1 there exist d Z1 and a map
Fn JdK
such that
(i) (st) (s)(t) <
for every s, t Fn such that st Fn ( is -multiplicative) and
(ii) tr (s) (s) <
for every s Fn ( is -trace-preserving).
Remark 4. If G is a group, one can replace JdK by Sym(d) as is easily seen.
Remark 5. The notion of sofic pmp equivalence relations was introduced by Elek
and Lippner in [EL] in terms of graph approximation and in [Oz] by requiring
that JGK is sofic as in the definition above (compare [DKP1, DKP2]). The sofic
property was first considered for groups by Gromov and Weiss in terms of (Cayley)
graph approximation and was studied by Elek and Szabo. It is a simultaneous
generalization of amenability and the LEF property of Vershik and Gordon (see
[Pe] for more details).
Let
SA(F, n, , d) = { Fn JdK satisfying (i), (ii)}
and define
SA(F, n, , d)E = {E SA(F, n, , d)}
for E F .
Definition 6. For E F JGK finite, n Z1 and > 0 define successively
log SA(F, n, , d)E
sE (F, n, ) = lim sup
d d log d
sE (F, n) = inf sE (F, n, )
>0
sE (F ) = inf sE (F, n)
nZ1
If K JGK is an arbitrary subset the sofic dimension of K is

s(K) = sup inf sE (F )
E F
where E F K are finite subsets. The sofic dimension of G is s(G) = s(JGK).

One defines similarly the lower sofic dimension s and the sofic dimension s for a
ultrafilter on Z1 by replacing lim supd by lim inf d and limd respectively.
Voiculescus regularity condition reads:
Definition 7. A pmp groupoid G is s-regular if s(G) = s(G).
Finally we recall
Definition 8. A subset K JGK is transversally generating if for any t JGK and
> 0 there exist n Z1 and s Kn such that t s .
This definition appears in [DKP1, Definition 2.4] where it is called dynam-
ically generating. The more classical notion of generating set for pmp equiva-
lence relations (and groupoids) (as in [DKP1, Definition 2.2]) is that of Connes
FeldmannWeiss. While being distinct notions, a groupoid is finitely generated in
the ConnesFeldmannWeiss sense if and only if it is transversally finitely generated
(by an argument similar to that in [DKP1, Proposition 2.6]) so finitely generated
is unambiguous for groupoids (and coincide with the usual notion in the group
case).
The following result is proved in [DKP1, Theorem 4.1].
Theorem 9 (Invariant of s under orbit equivalence). Let R be pmp equivalence
relation and K, L be transversally generating sets. Then s(K) = s(L), s(K) = s(L)
and s (K) = s (L).
The result in [DKP1] is stated for finitely generated equivalence relations but
the same proof works in the general case (as does the proof of [DKP1, Theorem
1.2]). The proof in full generality for groupoids is given in [DKP2, Theorem 2.11].
We wont use this more general result here but will rather deduce it from Theorem
9 as an illustration of the correspondence principle.
Remark 10. It is sometimes convenient to use the 2-norm on LG and its restriction
to JGK:
s t22 = ((s t)(s t)1 ) = ((s t)1 (s t))
Observe
s t22 s t
(with equality on [G] JGK) as
s t22 = (s1 s) + (t1 t) 2 (st1 )
and (st1 ) (s1 st1 t).
2. Actions of groupoids
Let X be a standard Borel space endowed with a Borel fibration p X G0 where
G is the base of G. If (e )eG0 is a Borel field of probability measures on X we
0
define a probability measure on X by = G0 e dh(e) where h is the invariant

Haar measure on G0 . Recall that a pmp action of G on the fibered space (X, ) is
a measurable map
G G0 X (g, x) gx X
(where G fibers via the source map s G XG ) satisfying the usual axioms of an
action, and such that
g s(g) = r(g)
for ae g G. Groupoid actions are denoted G X. The crossed product groupoid is
the fiber bundle GG0 X endowed with groupoid law defined by (s, x)(t, y) = (st, y)
whenever t(y) = x.
Example 11 (Bernoulli shifts). Given a pmp groupoid G with invariant Haar
measure h and a probability space (X0 , 0 ), consider the probability space
(X0G , ) = (X0G , )dh(e),

e e
G
0
G0
(where X0G = Ge X0 is the infinite Cartesian product over Ge = r1 (e)), endowed

e
with the fibration X0G G0 and field of measures (0 G )eG0 . Every element
e
x X0G can be viewed as a sequence x = (xt )tGe of elements of X0 . The Bernoulli

action G X0G is given by
s((xt )tGr(s) ) = ((xs1 t )tGr(s) ).
Remark 12. The notion of groupoid action has long been used in ergodic group
theory (see for example [Ram]). They can equivalently be described as actions on
bundles (cocycles) as above, or as groupoid extensions which are fiber bijective. In
[Bow2] L. Bowen discusses Bernoulli shifts using the latter description.
We first prove a few lemmas that will be used in the proof of Theorem 1.
Lemma 13. Let G X0G be a Bernoulli action and H G be an ergodic sub-
groupoid with G0 = H 0 . Then the action H X0G is isomorphic to a Bernoulli
shift over H.
Proof. By the von Neumann selection theorem, we can find a measurable section
s RG G of (s, r) G RG (=the pmp equivalence relation associated with G)
such that that s(RG0 ) = G0 and s(RH) H and, since the set G(e)/H(e) is
countable for e G0 , measurable sections (gj Dj G0 SG)jJ of SG G0 such
that for ae e G0 , G(e) = jJ H(e)gj (e). By ergodicity we may assume h(Dj ) = 1.
Let (i )iI be a sequence in Aut(G0 ) such that {RH[i e]}iI form a partition of
RG[e] for ae e G0 (see [FSZ]). Then
Ge = i f)
H(e)gj (e)s(e, 1
iI, jJ (e,f )RH
since any g G can be written uniquely in the form h0 gj (e)s(e, 1

e
i f ) for i
and f such that (e, f ) RH and f = i s(g) so gs(e, 1 f )1
G(e) and h0 =
gs(e, i1 f )1 gj (e)1 H(e).
i
Consider the measurable field of maps e X0G (X0IJ )H defined by sending

e e
x X0G to
e
((xh0 gj (e)s(e,1 f ) )(i,j)IJ ) .

i h0 H(e),(e,f )RH
These maps are measure preserving and if we consider the Bernoulli action of H
with base X0IJ then we see it is H-equivariant: for h H e say h = h1
1 s(d, e)
1
where h1 H(e)
d (h(x)) = (h(xh0 gj (e)s(e,1 f ) )(i,j)IJ,h0 H(e),(e,f )RH )
f ) )(i,j)IJ,h0 H(e),(e,f )RH )

i
= ((xs(d,e)h1 h0 gj (e)s(e,1
i
= ((xs(d,e)h1 h0 gj (e)s(e,1
i f)
)(i,j)IJ )
h0 H(e),(e,f )RH
= h ((xh0 gj (e)s(e,1 f ) )(i,j)IJ )

i h0 H(e),(e,f )RH
= h(e (x))

We say a groupoid action G X is essentially free if for ae s G G 0
s(s) (Fix(s)) = 0
where
Fix(s) = {x X s(s) sx = x}.
Lemma 14. If the pmp groupoid action G X is essentially free then G X is a
pmp equivalence relation.
Proof. Since G X is essentially free and Ge is countable the set
X0e = {x X e sx x, s Ge , s e}
has measure 1 in X e for every e A G0 a measurable subset with h(A) = 1. Let
(X0A , ) (A, h) be the measure fibration corresponding to (X0e )eA . Since G is
isomorphic to GA X0A , we may assume that sx x for all s G G0 , x X s(s).
Then for (s, x) G X
r(s, x) = s(s, x) x Fix(s)
so r(s, x) s(s, x) for every s G G0 and x X. This shows that G X is an
equivalence relation. It is an easy exercise to check that it is pmp (more generally
if G is pmp and G X is a pmp action, then G X is a pmp groupoid).
Lemma 15. If G is transversally finitely generated then so is GX for any ergodic
pmp action G X.
Proof. Let R be the orbit equivalence relation of G X. Since R is ergodic, there
exists an ergodic automorphism JRK, which is orbit equivalent to a Bernoulli shift
Z {0, 1}Z ([Dye]). For i {0, 1} let Bi be the cylinder set {x {0, 1}Z x0 = i}.
Let p1 , p2 be the projection in L (X) corresponding to B0 and B1 in Z {0, 1}Z.
If F JGK be a finite transversally generating set for G, then F {, p1 , p2 } is a
finite generating set for G X.
Lemma 16. The Bernoulli action G X0G is essentially free if G has infinite
fibers (i.e. Ge = for ae e G0 ), and the support of 0 contains at least two
points. If in addition 0 is diffuse then the action is essentially free.
Proof. Let s Ge , s e, such that Ge = . We show that e (Fix(s)) = 0. If
r(s) s(s) this is clear so we assume s G(e). Note that (xt )tGe is s-invariant
if and only if xsn t = xt for all n Z, t Ge . Thus we can find an infinite family of
pairwise disjoint pairs {si , ti }iI , si ti Ge , such that xsi = xti for every x Fix(s)
and i I. Since I = , the set of x X0G such that xsi = xti is negligible, so
e
e (Fix(s)) = 0. If in addition 0 is diffuse then the set of x X0G such that xs = xe

e
is negligible so e (Fix(s)) = 0 in this case.

Lemma 17. If G is ergodic with infinite fibers then the Bernoulli action G X0G
is ergodic for any base space (X0 , 0 ).
Proof. If A X0G is a nonzero G-invariant subset and p X0G G0 , then p(A) is
G-invariant and nonzero therefore = e (Ae ) 0 and is almost surely constant by
the ergodicity of G. Let > 0 be arbitrary. Since Ge = almost surely we can find
by the von Neumann selection theorem a nonzero measure field (F e )eB of finite
subsets of (Ge )eB such that e (Ae AF e ) > and a section s of s such that
se F e F r(se ) = for ae e B. By invariance r(e) (Ar(se ) Ase F e ) > so
r(e) (AF r(se ) Ase F e ) > 2
and r(e) (AF r(se ) Ase F e ) = r(e) (AF r(se ) )r(e) (Ase F e ) also
r(e) (AF r(se ) Ase F e ) < 2 + 6 + 22 .
Letting 0, we get 2 so 1 and (A) = 1.
Lemma 18. Let G X be a groupoid action and suppose that G = G1 G3 G2 .
Then
G X (G1 X1 ) (G3 X3 ) (G2 X2 )
where Xi = p1 (G0i ).
Proof. First note that Gi acts on Xi and that Gi Xi is naturally a subgroupoid of
G X. Given an arbitrary groupoid H and groupoid morphisms f1 G1 X1 H,
f2 G2 X2 H with f1 G X = f2 G X we want to show there is a unique
3 3 3 3
morphism k G X H such that the following diagram commutes (where the
unlabeled edges are the inclusion map).
G3 X3
G1 X1 GX G2 X2
f1 k f2
Since kG X = f1 G X , kG X = f2 G X and the values of k are determined

1 1 1 1 2 2 2 2
on G1 X1 and G2 X2 which generate G X, this gives uniqueness. To show it
is well defined note that if gi G1 X G1 G2 X G2 and g1 ...gn = Ids(g1 ) and if gi
are the corresponding elements of G1 and G2 then g1 ...gn = e, so using the fact that
G1 and G2 are in free product over G3 this shows that k(g1 )...k(gn ) = Ids(f1 (g1 )) or
Ids(f2 (g1 )) as appropriate.
3. Overlapping generators
Let F JGK be a finite subset, be a partition of F , and F JdK be a map.
Denote F the set of e G0 such that
e ss1
sF
and
s, t F, es = et (s) = (t).
Here and below we will view a partition of F as a map from F to {1, ..., } = ran
for some Z0 .
Similarly let F be the set of e {1, . . . d} (identified with the base space of the
transitive relation on {1, . . . , d}) such that
e (s)(s)1
sF
and
s, t F, e(s) = e(t) (s) = (t).
The following lemma will be useful in the proof of Theorem 28. We denote a
projection onto a set A by pA .
Lemma 19. Given F JGK, n 1, let
Fn = F {pF0 F0 Fn , partition of F0 }
If SA(Fn , 4nFn + 1, , d), F0 Fn , and is a partition of F0 then
F0

h(F0 ) < c1 (F, n).
d
where c1 (F, n) = 176(32n)F 2n .
Proof. Fix some F0 Fn and let 1 , 2 , . . . , m be the family of partitions of F0
(so m = Be(F0 ) is the Bell number of F0 ). For convenience let pj = pF0j and
pj = pF0
.
If j (s) = j (t), pj = st1 pj so
j
(s)(t)1 (pj ) (pj ) = (s)(t)1 (pj ) (st1 pj )

(s)(t)1 (pj ) (s)(t1 )(pj ) + (s)(t1 )(pj ) (st1 pj )
4 + 3 = 7
by [DKP1, Lemma 3.1(8)].
Similarly if j (s) j (t) then
(s)(t)1 (pj ) (st1 pj ) 7.
As (st1 pj ) = 0 we have
tr((s)(t)1 (pj )) = tr((s)(t)1 (pj )) (st1 pj )
tr((s)(t)1 (pj )) tr (st1 pj ) + tr (st1 pj ) (st1 pj )
7 + = 8
by [DKP1, Lemma 3.1 (4)]
Moreover
(pj ) (pj )(pj ) = (p2j ) (pj )(pj ) < .
m
and since pj = ss1 using [DKP1, Lemma 3.5]
j=1 sF0
RRR m RRR RRR RRR
RR R R R
RRR (pj ) (s)(s)1 RRRR < RRRR ss1 (s)(s)1 RRRR + 150(2F + 1)2n
RRRj=1 sF0 R R
RR RR sF0 sF0 RRR
F0 4 + F0 2 + + 150(2F + 1)2n
157(32n)F 2n .
If V denotes the set of integers 1 k d such that for all 1 j m and s, t F0 :
(s)(t)1 (pj )k = (pj )k
(s)(t)1 (pj )k k
(pj )k = (pj )(pj )k
m
( (pj ))k = ( (s)(s)1 )k
j=1 sF0
then
V (1 7Fn 7Fn Fn Fn 157(32n)F 2n ) d
(1 175(32n)F 2n ) d
So, for all j we have that (pj )V is a projection and
j (s) = j (t) (s)(t)1 (pj )V = (pj )V and tr ((s)(t)1 (pj )V ) = 0
Hence ran((pj )pV ) ran(pj pV ), but since we also have that
m m
( (pj ))pV = ( (s)(s)1 )pV = ( pj )pV
j=1 sF0 j=1
we conclude (pj )pV = pj pV which implies tr (pj ) F0

j
/d < 175(32n )F 2n
therefore
h(F0 j ) F0 /d = tr(pj ) F0 /d

j j
tr(pj ) tr (pj ) + tr (pj ) F0

j
/d
+ 175(32n)F 2n 176(32n )F 2n .

4. Bernoulli shifts and random partitions

For 1 i q let
Bi = {x {1, . . . , q}G xe = i}
where X = {1, . . . , q}G is endowed with the Bernoulli action G X, and for d Z1
let A1 , . . . , Aq be a partition of {1, . . . , d}. Given a set F JGK and function
F {1, . . . , q} we define
B = sB(s)
sF
and
A = (s)A(s)
sF
where F JdK.
The next lemma adapts the proof of Theorem 8.1 from L. Bowens paper [Bow].
Lemma 20. If d is large enough (depending on F , n and ) then there is a partition
{A1 , A2 , . . . , Aq } of {1 . . . d} such that if SA(Fn , 4nFn + 1, , d) then for every
subset F Fn and every function F {1, . . . , q}
A
(B ) < c2 (F, n)
d
where c2 (F, n) = 2c1 (F, n) Be(F n ) and Be is the Bell number.
Proof. Fix F Fn and F {1, . . . , q}. Create a random partition
{A1 , A2 , . . . , Aq }
of {1 . . . d} using the following scheme: for each k {1, . . . , d} place k in Ai with
probability 0 (i).
We will find the probability that (B ) A /d < c2 (F, n).
Let 1 , 2 , . . . , m be the family of partitions of F . Let sj ran j F be an
arbitrary section of j . If

0 if s, t F such that j (s) = j (t) and (s) (t)
j =

1 otherwise
then

(B ) = sB(s)
sF
m
= j h(Fj ) 0 ((sj (r)))
j=1 rran j
Since for SA(Fn , 4nFn + 1, , d), we have

A m A Fj

=
d j=1 d
it will suffice to show by the triangle inequality that for all j
RRR A F RRR
RRR j R RRR < 2c (F, n).

RRRR j h(F ) ((s (k))) RRR

RR RR
j 0 j 1
kran j d
First suppose j = 0, so for some s, t F , j (s) = j (t) and (s) (t). Then
A F
= , for if k A F
then k (s)A(s) (t)A(t) F
which
implies (s) k = (t) k A(s) A(t) , a contradiction.
j j j
1 1
Next suppose j = 1. For 1 k d let

1 if k F
A
Zk =

j

0 otherwise
We wish to compute E(Zk ).
For k F
we have E(Zk ) = 0. Otherwise, k F

A if and only if
sj (r) k A(sj (r)) for all r ran j and since
j j
1
s1
j (r)
k s1 k, r r ran j
j (r )

we obtain
E(Zk ) = 0 ((sj (r)))
rran j
Thus if we let Z = dk=1 Zk = F

j
A then
E(Z) = F

j
0 ((sj (r))).
rran j
Now lets bound var(Z):

var(Z) = E(Z 2 ) E(Z)2 = E(Zk Zl ) E(Z)2
k,l{1...,d}
For k, l we have E(Zk Zl ) = 0 = E(Zk )E(Zl ). On the other hand if k, l

F
k = s1
j

F then Z k and Zl are not independent if s1
j (r)

l for some r, r . Thus
j (r )
are at most Fn 2 F non independent pairs (k, l). Now clearly for these pairs
j

E(Zk Zl ) E(Zk )E(Zl ) + 1. So returning to our equation above

j
E(Zk Zl ) E(Z) E(Zk )E(Zl ) + F Fj E(Z )

2 n2 2
1k,ld 1k,ld
= E(Z 2 ) + Fn 2 F

j
E(Z 2 ) = Fn 2 F

j

Now we can apply Chebyshevs inequality to Z
d
for a > 0
var( Zd ) Fn 2 F

Fn 2
Pr ( a)
Z E(Z)

j
d d a2 a2 d2 a2 d
Since Z = A F
j
and E(Z) = F
j
rran j 0 ((sj (r))), and, by Lemma
19,
RR F RRR
RRR j R

RRRh(Fj ) RR < c1 (F, n)
RRR d RRRR
R
we have
R RRR
RRR RRR Fn 2
Pr RRRRh(Fj ) 0 ((sj (k))) F
/d R
RRR (F, 2
RRR
+

A a c n)
R
1
R R
kran j
j
a d
Let a = c1 (F, n) so then
R RR
RRR RRR Fn 2
Pr RRRRh(Fj ) 0 ((sj (k))) F
/d RRR 2c1 (F, n)
RRR RRRR

(c1 (F, n))2 d
A
R
j
kran j
So for d large enough there is some partition (A1 , . . . , Aq ) such that

RRR RRR
RRR RRR
RRR h(F ) ((s (k))) F
A /d RRR < 2c1 (F, n)
RR RR
j 0 j j
R kran j
R
Thus, for d large enough there will be a partition so that this true for all j so that
A
(B ) < c2 (F, n).
d
Indeed for large enough d nearly all partitions will satisfy this and hence we will
have non zero probability that for all F Fn and F {1, . . . , q} the inequality
(B ) A /d < c2 (F, n) holds.
5. A lemma on approximate equivariance

Given a groupoid G acting on a set X for any finite set of projections P
L (X, ) and finite set F JGK then PF denotes the set of all projections of the
form
sps
sF
where ps P and F F .

Let P = {pBi } where

Bi = {x {1, . . . , q}G xe = i}
and X = {1, . . . , q}G is endowed with the Bernoulli action G X. Fix F JGK and
a basis {pB1 , pB2 ...pB } for span(PFn ) in L (X, ) associated with i Fi
{1, . . . , q}, 1 i .
We set
= max ai 1.
,1 ,..., ,a1 ,...a pB =i=1 ai pB
i
Both and depend on F and n only.

Lemma 21. Let SA(Fn , 4nFn + 1, , d) and take a partition {A1 , . . . , Aq }
satisfying the conclusion of Lemma 20.
If for some F {1, . . . , q}

pB = ai pBi
i=1
then

pA ai pAi < c3 (F, n)
i=1 2
where c3 (F, n) = ((1 + (3 + ) Be()N ))c2 (F, n) and N = 2 .
Proof. For I {1, . . . } define
DI = {x X x Bi i I}
and similarly
DI = {1 k d k Ai i I}.
We want to bound (DI ) DI /d.
Note that for any I {1, . . . }
{x X i I, x Bi } = ( Bi ) .
iI
Now either for some s Fi1 Fi2 we have i1 (s) i2 (s) and so iI Bi = or
otherwise iI Bi = B for some F {1, . . . , q}. In the first case iI Ai =
as well and in the second case by Lemma 20
(B ) A /d < c2 (F, n)
and
A /d = Ai /d = {1 k d i I, k Ai }/d
iI
So in general we conclude that {x X i I, x Bi } is within c1 (F, n) of

{1 k d i I, k Ai }/d. By inclusion-exclusion

(DI ) = (1)kI ({x X i I , x Bi })
k=I I I
I =k
and similarly
DI
= (1)kI {1 k d i I , k Ai }
1
d k=I I I
d
I =k
so by applying the triangle inequality at most N = 2 times we have
(DI ) DI /d c2 (F, n)N .
Now
(B ) = (DI ).
I{1,...};iI ai =1
So if iI ai = 1 then (iI Bi ) = (iI Bi B ) so by a similar argument

{x B i I, x Bi } is within c2 (F, n) of {k A i I, k Ai }/d So if
we let
I = {k A x Ai i I}
D
then
(DI ) D
I /d c2 (F, n)N
so
DI D

I 2c2 (F, n)N .
d d
Let a(I) = iI ai .

pA ai pAi
X X
i=1 2
X
X
X X
X
X
=XX
X ( )p X
X
X
X
X
p a I X
X
X
X X
X
A i D
X I{1,...} iI
X2
X
X
X X
X
X X
X
X X
X
X
X
X X
X X
X X
X
XX
X X
X X
X ( )p X
X
X I XX X
X I XX
p p + a
X X X X

X X X X
A D i D
X X2 X X2

a(I)=1 a(I){0,1} iI
X
X
X X
X
X X
X
X X
X
X
X X X X
XX
X
X D I X X
X
X
X
X
X
X D I DI X X
X
X + DI /d
X X X X
p p + p p
X X
X2 X X

X X X X
A
X Xa(I)=1 X2

a(I)=1 a(I)=1 a(I){0,1}

RRR A D RRRR D
RRR I RR I DI + Be()c (F, n)N

RR

RR +
RRRR d a(I)=1 d RRR a(I)=1
R 2
d d

A RRR RRRR
D
R
(B ) + RRRR (DI ) I RR
RR
RRRa(I)=1 RRR
R R
d d

a(I)=1
+ Be()2c2 (F, n)N + Be()c2 (F, n)N

D

c2 (F, n) + (DI ) I + Be()2c2 (F, n)N + Be()c2 (F, n)N
d

a(I)=1
c2 (F, n) + (3 + )Be()Nc2 (F, n)

= c3 (F, n) .
6. Sofic dimension and groupoid actions

We now briefly recall the group action formulation of s(G X) given in [DKP2,
Section 5] rephrasing it here in the framework of groupoid actions.
Let G X be a pmp action of a pmp groupoid. Let 1 F JGK and P be
a partition of X. We write HA(F, P, n, , d) for the set of all pairs (, ) where
SA(F, n, , d) and is a map
PFn JdK
defined on PFn JGK and satisfying
(i) tr (p) (p) < for all p PFn
(ii) s(p) (s) (p) < for all p P and s Fn
(iii) (p1 p2 ) (p1 )(p2 ) < for all p1 , p2 , p1 p2 PFn .
(iv) (pX ) pd < , where pd = p{1,...,d} .
(Note that since P partitions X and F contains the identity we have pX span PFn .)
We observe that if satisfies these conditions then it is automatically approxi-
mately linear:
Lemma 22. If (, ) HA(F, 4n, , d) then

(p1 + p2 ) ((p1 ) + (p2 )) < 146
for all p1 , p2 , p1 +p2 PFn with p1 p2 = 0 and (p1 )+(p2 ) is defined as in [DKP1,
Def. 3.3].
Proof. Using (p1 + p2 )pi = pi we have
(pi ) (p1 + p2 )(pi ) <
for i = 1, 2. Let i ((p1 ), (p2 )) be defined as in [DKP1, Def. 3.3] so
(p1 ) + (p2 ) = (p1 )1 ((p1 ), (p2 )) + (p2 )2 ((p1 ), (p2 )).
By [DKP1, Lemma 3.4] using the approximate homomorphism property of we
obtain
i (p1 , p2 ) (pi )(pi )1 < 40
for i = 1, 2. Since (pi ) (pi )(pi )1 < 8 we obtain
(p1 + p2 )i (p1 , p2 ) (pi ) < 48
and so
(p1 + p2 )(p1 , p2 ) ((p1 ) + (p2 )) < 48
with (p1 , p2 ) = 1 ((p1 ), (p2 )) + 2 ((p1 ), (p2 )).
Then
tr((p1 + p2 )(p1 , p2 )) = tr((p1 + p2 )1 (p1 , p2 ))) + tr((p1 + p2 )2 (p1 , p2 ))
> tr((p1 )) + tr((p2 )) 96
> (p1 ) + (p2 ) 98
= (p1 + p2 ) 98.
Therefore
(p1 + p2 ) ((p1 ) + (p2 )) < 146.
(Similarly one can also show that (p) is approximately a projection for every
p PFn .)
Definition 23. Given E, Q, F , P , n and define successively
sE,Q (F, P, n, ) = lim sup log( HA(F, P, n, , d)E,Q )

1
d d log(d)
sE,Q (F, P, n) = inf sE,Q (F, P, n, )
>0
sE,Q (F, P ) = inf sE,Q (F, P, n)
nN
If K JGK is a transversally generating set and R is a dynamically generating
family of projections of L (X, ) define
s(K, R) = sup sup inf inf sE,Q (F, P )
E Q F P
where E and F range over finite subsets of K and P , Q range over finite subparti-
tions of R. Set s(G, X) = s(JGK, L (X, {0, 1})).
Since JGK L (X, {0, 1}) is a transversally generating set of the crossed product
groupoid we obtain by Theorem 9 (if the action G X is essentially free) and
[DKP2, Theorem 2.11] in general (compare [DKP2, Proposition 5.2]):
Proposition 24. s(G X) = s(G, X).
Moreover if F is a finite transversally generating subset of JGK and P is a finite

and dynamically generating partition of unity, then s(G X) = sF,P (F, P ).
Proposition 25. s(G X) s(G).
Proof. For every (, ) HA(F, P, n, , d) there are at most Qd restrictions Q
and hence
HA(F, P, n, , d)E,Q Qd SA(F, n, , d)E
so we have sE,Q (F, P ) sE (F ) and the result follows directly by the proposition
above.
Remark 26. The set HA differs from the set HA introduced in [DKP2] in that we
do not assume to be a strict homomorphism. Furthermore the maps in [DKP2]
are defined on L (X) with values in Md using the 2-norm. It is often convenient to
adopt the latter point of view for computational purposes and we will do so below.
The definition of HA given above has the advantage of being purely finitary, in the
spirit of the sofic property. We have maintained the notation HA in view of (iii)
and (iv).
7. The computation of s(G (X0 , 0 )G )

Lemma 27. If G {1, . . . , q}G , then s(G ({1, . . . , q}, 0 )G ) = s(G) for any
probability measure 0 on {1, . . . , q} and pmp groupoid G. The same holds true for
s and s .
Proof. Let X = {1, . . . , q}G . By Proposition 25 we have to prove s(G X) s(G).
Let P = {pBi } where Bi = {x {1, . . . , q}G xe = i}. Let {pB1 , pB2 ...pB }
be a basis for span(PFn ) L (X, ), where i Fj {1, . . . , q}. Let
SA(Fn , 4nFn + 1, , d). Let be as defined before Lemma 21. Let

1 = min { ai , 1 , .... , a1 ..., a , T {1, . . . , } and pB = ai pBi } /{0}
iT i=1

2 = min { ai 1 , 1 , .... , a1 ..., a , T {1, . . . , } and pB = ai pBi } /{0}

i=1
R RR
iT
RR
R
2 = min RRRR ai bj RRRR , 1 , .... , a1 ..., a , T, T {1, . . . , } and

RR RRR
RRiT,jT R

pB = ai pBi , pB = bi pBi } /{0}
i=1 i=1
= min{1 , 2 , 3 }
so that depends only on F and n and 1. We want to find such that

(, ) HA(F, P, n, 9PFn 2 2 5 2 qc3 (F, n)2 , d)
1

(for sufficiently large d). Using Proposition 24 this will complete the proof.
Take a partition {A1 , . . . , Aq } such that the conclusion of Lemma 20 holds
(namely a random partition for d large). For each pBi , i = 1 . . . , let
0 (pBi ) = pAi
and extend 0 linearly to span(PFn ) L (X, ) with values in Md . We will check

that 0 satisfies the following properties, where 0 = 3 1 2 2 qc3 (F, n)2 (compare
Remark 26):
(1) tr 0 (p) (p) < 0 for all p PFn
(2) 0 s(p) (s) 0 (p)2 < 0 for all p P and s Fn
(3) 0 (p1 p2 ) 0 (p1 )0 (p2 )2 < 0 for all p1 , p2 span PFn
(4) 0 (pX ) pd 2 < 0 , where pd = p{1,...,d} .
One can see (as shown below) that properties (1)-(4) are closely related to the
properties (i)-(iv) defined above.
(1) Note for any x {1 . . . , d}, and a1 , ...an if

( ai pAi pA ) x 0
i=1
then

( ai pAi pA ) x min{1 , 2 } .
i=1
Let pB PFn say pB = i=1 ai pBi then

0 (pB ) = ai pAi
i=1
so
A
tr 0 (pB ) (pB ) tr( ai pAi ) tr pA + (B )
i=1 d
2
1
ai pAi pA + c2 (F, n)
i=1

2
c3 (F, n) + c2 (F, n) c3 (F, n)2 , for < 1

1 2 2 1

where we use both Lemma 20 and 21
(2) Here we use Lemma 21. Let s Fn and suppose

spBi = ai pBi .
i=1
Then

0 (spBi ) = ai pAi
i=1
and

ai pAi (s)pAi < c3 (F, n) .
i=1 2
On the other hand

pBi = bi pBi 0 (pBi ) = bi pAi
i=1 i=1
and

bi pAi (1)pAi < c3 (F, n) .
i=1 2
Thus
0 (spBi ) (s)0 (pBi )2
0 (spBi ) (s)pAi 2 + (s)pAi (s)(1)pAi 2
+ (s)(1)pAi (s)0 (pBi )2

< 2c3 (F, n) + = 3c3 (F, n)
(3) Let p1 = i=1 ai pBi and p2 = i=1 bi pBi then

0 (p1 p2 ) 0 (p1 )0 (p2 )2 = ai bi (0 (pBi pBj ) pAi pAj )2
i,j=1

2 0 (pBi pBj ) pAi pAj 2
i,j=1
So it is enough to show that for all i, j

0 (pBi pBj ) pAi pAj 2 < c3 (F, n) .
But as we have seen in the proof of Lemma 21, either pBi pBj and
pAi pAj are both 0 in which case we are done or pBi pBj = pB and
pAi pAj = pA for some in which case if pB = i=1 ci pBi then by
Lemma 21

0 (pBi pBj ) pAi pAj 2 = ci pAi pA 2 < c3 (F, n)
i=1
(4) pX = i=1
q
pBi and Id = i=1 pAi
q
so by the triangle inequality it suffices to
show
0 (pBi ) pAi 2 < 2c3 (F, n)
By Lemma 21

0 (pBi ) (1)pAi 2 < c3 (F, n)
so
0 (pBi ) pAi 2 0 (pBi ) (1)pAi 2 + (1)pAi pAi 2

< 2c3 (F, n) + < 2c3 (F, n) .
We now show how to define using 0 . For each pB

0 (pB ) pA 2 c3 (F, n)
by Lemma 21. Let M be the number of x {1 . . . , d} such that 0 (pB )x = pA x
then
0 (pB ) pA 2
dM 2

d
hence
M d(1 c3 (F, n)2 2 )
1

Similarly for pB , pB we have

0 (pB )0 (pB ) (pB )(pB )2 22 c3 (F, n)
by (3) so
{x [d] 0 (pB )0 (pB )x = 0} d(1 c3 (F, n)2 4 )

1
2
Let V be the set of all x {1, ...d} such that for all pB (arbitrarily represented)
we have
0 (pB )x = pA x
and for all pB , pB with pB pB = 0 we have
0 (pB )0 (pB )x = 0
then
V d(1PFn c3 (F, n)2 PFn 2 c3 (F, n)2 4 2 ) d(12PFn 2 c3 (F, n)2 4 2 )
1 1 1
2
and for any p PFn 0 (p)V JdK. We finally define
PFn JdK
p 0 (p)V
Let us check that satisfies (i)-(iv):
(i)
tr (pB ) (B ) tr (pB ) tr 0 (pB ) + 0
2PFn 2 c3 (F, n)2 4 2 + 0

1

(ii)
(spBi ) (1)(pBi ) (spBi ) (1)(pBi )2
(spBi ) 0 (spBi )2
+ (1)(pBi ) (1)0 (pBi )2 + 0
1
4PFn c3 (F, n)2 + 0

(iii)
(p1 )(p2 ) (p1 p2 ) (p1 )(p2 ) (p1 p2 )2
(p1 )(p2 ) 0 (p1 )0 (p2 )2
+ (p1 p2 ) 0 (p1 p2 )2 + 0
1
2 2 q4PFn c3 (F, n)2 + 0

(iv)
1
(pX ) pd 2PFn c3 (F, n)2 + 0

Thus for every SA(Fn , 4nFn + 1, , d) withd sufficiently large we found so
that (, ) HA(F, P, n, 9PFn 2 12 5 2 qc3 (F, n)2 , d).
We now check that
s(G) = sup inf inf inf lim sup SA(Fn , 4nFn + 1, , d)E
1
E F nN >0 d d log d
Recall that Fn = F {pF0 F0 Fn , partition of F0 }.

The right hand side equals

sup inf inf sE (Fn , 4nFn + 1).
E F nN
Let > 0 and choose n0 so that

sE (Fn0 , 4n0 Fn0 + 1) inf sE (Fn , 4nFn + 1) <
n
and
SE (F, 4n0 Fn0 + 1) inf sE (F, 4nFn + 1) < .
n
Clearly sE (Fn0 , 4n0 Fn0 + 1) sE (F, 4n0 Fn0 + 1) but on the other hand if
SA(Fn0 , 4n0 (Fn0 )n 0 +1, , d) then SA(Fn0 , 4n0 Fn0 +1, , d) since 4n0 (Fn0 )n 0 +
1 4n0 Fn0 + 1. So
inf sE (F, 4n0 Fn0 + 1) = inf sE (Fn0 , 4n0 Fn0 + 1)
F F
Since was arbitrary
inf inf sE (Fn , 4nFn + 1) = inf inf sE (F, 4nFn + 1) = sE (G)
F n F n
since
inf sE (F, n) = inf sE (F, 4nFn + 1).
n n
However

SA(Fn , 4nFn + 1, , d)E HA(F, P, n, 9PFn 2 (F,
1 5 2 2
qc 3 n) , d)E
2
so s(G) s(GX) by Proposition 24. Replacing lim sup by lim inf or limd above,
we get a similar inequality for s and s .
Theorem 28. Let G be a pmp groupoid, (X0 , 0 ) be a standard probability space,
and G X0G be the corresponding a Bernoulli action. Then s(G X0G ) = s(G).
The same holds true for s and s .
Proof. Again we only need to prove s(G) s(G X0G ). Let U = {U1 , U2 , . . . , Uq }
be any finite partition of X0 into measurable sets. Let
BU = {BU1 , . . . , BUq }
where
BUi = {x X0G e G0 , xe Ui }
Let M = {pUi } for all possible choices of U and all i, so that M is dynamically
generating *-subalgebra of L (X, ). Now if PU = {pU1 , . . . , pUq } then
s(G X) = sup sup inf inf sE,Q (F, PU ) = sup sup inf inf sE,Q (F, PU )
E Q F PU E Q PU F
Since the map X0 {1 . . . , q} defined by Ui x i extends G-equivariantly to

X0G {1, . . . , q}G we have
inf s(F ) inf sE,Q (F, PU ).
F F
Hence s(G) s(G X). The same holds for s and s .

Corollary 29. A pmp groupoid G is s-regular if and only if the crossed product
groupoid G X0G associated with the Bernoulli action G X0G is s-regular for any
base space (X0 , 0 ).
Proof. Since
s(G) = s(G X0G ) s(G X0G ) = s(G)
s(G) = s(G) if and only if s(G X0G ) = s(G X0G ).
8. The scaling formula

Proposition 30 (Scaling formula). Let G be an ergodic pmp groupoid, 0 p
L (G0 ). Then
s(G) 1 = h(p)(s(pGp) 1)
hpGp
where pGp is endowed with the normalized Haar measure h(p)
. Furthermore the
same equality holds for s and s therefore:
G is s-regular pGp is s-regular.
We will start by showing in the rational case, which is easier to handle:
Lemma 31. If h(p) Q then
s(pGp)
1 1
s(G) + 1 .
h(p) h(p)
Furthermore the same inequality holds for s and s .
Proof. Write h(p) = N k
N
and choose s1 , s2 ...sk JGK with
i si p, h(si si ) =
1
s1 1
N
and
k
si si = 1 p.
1
i=1
Let S = {s1 , ...sk } so pGp S is generating G and let E, F JpGpK be finite subsets.
We may assume that p E F and s1 i si E F for all i. Let SApGp (F, 4n +
5, , d). In light of [DKP2, Lemma 2.13] we may also assume that N kd. Let
d = NNk d and partition {1, . . . , d } into sets A0 ...Ak with A0 = {1, ...d} and Ai = N
d
.
Since
i si )
h(s1
Fix ((s1 i si )) d d
h(p)
we may choose a subset Bi A0 with exactly d
N
elements such that
i si )) Bi
Fix ((s1 < d .
For each i = 1 . . . k let (si ) be a bijection Bi Ai (we have ( N

k
d
!) choices).
As
j si = si sj = 0 i, j 1
s1 1
i sj = 0 i j 1
s1
ssi = s1
i s = 0 i 1, s pGp
i si F
s1
each element of n (JpGpK S)n can be written (not necessarily uniquely) as si f s1
for i, j 0 where s0 = p and f JpGpK. Let (s0 ) = pA0 We define a map
j
(F S)2n+5
Jd K
by
(si f s1
j ) = (si ) (si si f sj sj )(sj ) .
1 1 1
where we denote (f ) the permutation ( (f ) 0

0 0
) acting in A0 {1, . . . , d }. Note
that is well defined: if si1 f1 sj1 = si2 f2 sj2 then since the si have disjoint ranges
1 1
we have i1 = i2 j1 = j2 so si1 f1 s1
j1 = si1 f2 sj1 hence si1 si1 f1 sj1 sj1 = si1 si1 f2 sj1 sj1
1 1 1 1 1
so that
i1 si1 f1 sj1 sj1 )(sj1 )

(si1 ) (s1 i1 si1 f2 sj1 sj1 )(sj1 ) .
= (si1 ) (s1
1 1 1 1
then extend linearly to (F S)2n+5 .

Let = 5N 2 + 150N 2 (2F S + 1)2(2n+5) . We claim that:
SAG (F S, n, , d).
Let us first see what this implies. Note if 1 (f ) 2 (f ) then 1 (f ) = pA0 1 (f )pA0 =
1 (f )
2 (f ). Next note if 1 (si ) 2 (si ) Bi Ai are two bijections then if we
have
1 (si ) = 1 (si ) (p) = 2 (si ) (p) = 2 (si )
we must have 1 (si )f ix((p)) = 2 (si )f ix((p)) . But since Fix((p)) {1, .., .d}
d we have at most JdK (d + 1)d choices for 2 such that 2 (si ) 1 (si )
and 2 (si ) = 1 (si ). Therefore there are at most (d+1)dk choices for 2 such
k
that 1 ES 2 ES Thus each E SApGp (F, n, , d) gives at least

(d
N
!)
(d+1)dk
distinct elements ES SAG (F S, n, , d ) so we obtain
(N
k
d
!)
SApGp (F, n, , d)E SAG (F S, 4n + 5, , d )ES /
(d + 1)dk
So let us show that SAG (F S, n, , d ). We first show that is approxi-
mately linear. Suppose in the span we have
M
si f sj = si f sj ,
=1
then if we let L(u, v) be the set of all with i = u j = v then si f sj = L(i,j) si f s1

j
and su 0sv = 0 = L(u,v) sp f s1
q for (u, v) (i, j). So we may assume i = i, j = j
for all then
(si f sj ) = (si ) (s1

i si f sj sj )(sj )
1 1
M
i ( si f sj ) sj ) (sj )
= (si ) (s1 1 1
=1
M
= (si ) ( s1
i si f sj sj ) (sj )
1 1
=1
so by [DKP1, Lemma 3.5]

M M
(si f sj ) (si f sj ) = (si f sj ) (si ) (s1
i si f sj sj )(sj )
1 1
=1 =1
< 150(2F S + 1)2(n+5) .

Thus if f = M
=1 f , then
RRR RRR
R R
(f ) (f ) = RRRR ( si f sj ) ( si f sj )RRRR
M N M N
=1 R
RR i,j=1 =1 i,j=1 RRR
N M
(si f sj ) (si f sj )
i,j=1 =1
< 150N (2F S + 1)2(n+5) .

2
Thus to show that is -multiplicative it will suffice to show that for a, b

(F S)n such that ab (F S)n we have
(ab) (a) (b) < 5.
Indeed then for i ai , i bi (F S)n with (i ai ) (j bj ) (F S)n
RR R
RRR RR
RR ( ai ) bj ( ai ) bj RRRR
RRR i j j RRR
R i R
RR R
RRR RRR
RRR ai bj ( (ai )) (bj ) RRRR + (150N 2(2F S + 1)2(2n+5) )
2
RR i,j j RR
R i R
RRR RRR
R R
RRRR (ai bj ) (ai ) (bj )RRRR + 2 (150N 2 (2F S + 1)2(2n+5) )
2
RRR i,j i,j R

RR
(assuming < 1)
5N 2 + 2 (150N 2(2F S + 1)2(2n+5) )
2
So say we have a = si1 f1 s1

j1 and b = si2 f2 sj2 where f1 , f2 F . If j1 i2 then
1 n
s1
j1 si2 = (sj1 ) (si2 ) = 0 so we are done. Suppose otherwise, then
1
i1 si1 f1 (sj1 sj1 ) f2 sj2 sj2 )(sj2 )

(ab) (a) (b) = (si1 ) (s1 1 3 1 1
i1 si1 f1 sj1 sj1 )(sj1 ) (sj1 ) (sj1 sj1 f2 sj2 sj2 )(sj2 )
(si1 ) (s1 1 1 1 1 1
Note first that for any i > 0
(si )1 (si ) (s1

i si ) Fix((s1
i si )) Bi + (si si ) pFix((s1
1 1
d i si ))
2
and for g1 , g2 F2n+5 such that g1 g2 F2n+5
(g1 g2 ) (g1 ) (g2 ) =

d(g1 g2 ) (g1 )(g2 )
1
d
so by repeated applications of the triangle inequality
(ab) (a) (b) 2 + 3 = 5.
Let us now show that is -trace-preserving. Let si f s1 j (F S) . If i j
n
then
j ) = tr((si ) (si si f sj sj )(sj ) ) = 0,
tr(si f s1 1 1 1
suppose otherwise, then

tr( (si f s1
i )) (si f si ) = tr ((si ) (si si f si si )(si ) ) (si f si )
1 1 1 1 1
= tr ( (s1
i si f si si )(si ) (si )) (f si si )
1 1 1
tr ( (s1
i si f si si )(si ) (si ))) tr ( (si si f si si ))
1 1 1 1
(N k) (f s1
i si )
+ tr ( (s1
i si f si si ))
1

N h(p)
(2 + ) + <
where we used the computation above.
So we have proved
(N
k
d
!)
SApGp (F, 4n + 5, , d)E SAG (F S, n, , d )ES /

(d + 1)dk
So
k
log ( SAG (F S, n, , d )ES / (d+1) dk )

N
( d !)
log SApGp (F, 4n + 5, , d)E

d log d d log d
N log SAG (F S, n, , d )ES
=
N k d log d
)!

d
log(d + 1)
+ dk
log( N
k
d log d d log d
Now for > 0 arbitrary and d sufficiently large we have 1
log d
1
log d
+ log d so the
left hand side is at most
log SAG (F S, n, , d )ES ) )!

d
log(d + 1)
(1 + ) + dk
N log( N
k
N k
d log d d log d d log d
By Stirlings approximation
d d d d d
log( )! = log + + O(log )
N N N N N
therefore
spGp,E (F, 2n + 5, ) (1 + ) lim sup sG,ES (F S, n, , d ) +
N k
+ k
N k d N
which gives
spGp,E (F, 4n + 5) (1 + )sG,ES (F S, n) +

1 k
h(p) N
= (1 + )sG,ES (F S, n) + 1
1 1
h(p) h(p)
From here it is clear that
s(pGp)
1 1
s(G) + 1 .
h(p) h(p)
The same proof works with lim inf d and limd instead of lim supd .
Next we prove the other direction in the general case:
Lemma 32. s(G) h(p)s(pGp) + 1 h(p) and similarly for s and s .

Proof. Let S = {s1 , s2 , . . . , sk } JGK be such that
k
i si p and si si = 1 p.
s1 1
i=1
and let E, F JpGpK be finite subsets. We may assume that p E F and

i , si si E F for all i.
si s1 1
Let SAG (F S, n+4, , d) and assume that d is large enough so d = h(p)d >
. Let B0 = Fix (p) so
1
B0
h(p) < .
d
Thus we may arbitrarily extend or shrink B0 to a subset B such that either B B0
or B B0 and B = d where
(p) pB (p) pFix (p) + pFix (p) pB
= {x {1 . . . d} (p)2 x (p)} + B0 h(p) + B h(p)

1 1 1
d d d
< 3.
Let = 20
h(p)
. We will show that
= B SApGp (F, n, , d ).
Cleary is well defined, let us show it is -multiplicative that is for a, b Fn
with ab Fn that
(ab) (a) (b) < .
So
(ab) (a) (b) = {x B (ab)x (a) (b)x}
1
d xB
= pB (ab)pB pB (a)pB (b)pB

d
d
20 2
(15 + 5) = < , for < 1

h(p) h(p)
We now show is -trace-preserving. Let a Fn then
Fix (a) G (a) 1 Fix (a)
= G (a)
d h(p) h(p) d /h(p)
1 Fix (a)
+ G (a)
1 h(p) 1
h(p) d d h(p) d
Fix pB (a)pB Fix (a)
( + tr (a) G (a) + 2 )
1

h(p) d d
(pB (a)pB (a) + + 2 )

1
h(p)
9 + 2
((6 + 2) + + 2 ) = < , for < 1
1
h(p) h(p)
So SApGp (F, n, , d ) as claimed.
Next we study the map

SAG (F S, n + 4, , d) SApGp (F, n, , d )

Note that
Fix (si )(si )1 Fix (si )(si )1 Fix (si s1
i )
h(si s1
i )
d d d
+ tr (si s1
i ) h(si si )
1
(si )(si )1 (si s1

i ) +
5 + < 6
and for i j
ran((si )) ran((sj )) = {x {1, . . . , d} (sj )(sj )1 (si )(si )1 x 0}
d (sj )(sj )1 (si )(si )1 < 9d
so
ran((si )) B = {x {1, . . . , d} pB (si )(si )1 x 0}
d pB (si )(si )1 (3 + 5)d 8d
So for all d sufficiently large
ran((si )) ( ran((sj )) B)
ji
contains at least di = h(si s1

i )d (14 + 9k)d elements so we can find
Ai ran((si )) ( ran((sj )) B)
ji
be a subset with size Ai = di and Ai Aj = for i j and Ai B = . Similarly

we have
Fix (si )1 (si )
i si ) < 6
h(s1
d
and
dom (si ) Fix (s1
i si ) B {x {1 . . . , d}PB PFix (s1
i
si ) x (si ) (si )x}
1
d PB PFix (s1
i
si ) (si ) (si ) < d6
1
So we can find subsets Bi dom((si )) Fix (s1 i si ) B with Bi = di . Let

dk+1 = d d ki=1 di . Also recall that Fix (si s1
i ) d(h(si si ) + ).
1
Let SAG (F S, n + 4, , d)ES be the number of elements of SAG (F S, n +

4, , d) where we distinguish elements by their values on {pB (f )pB f E} and

{pAi (si )pBi }. We have shown with the above computations that
SAG (F S, n + 4, , d)E,S

( ) SApGp (F, n, , d )E
d
d , d1 , . . . , dk+1
i si ) + )
d(h(s1
k k
( ) di !
i=1 di i=1
where the former term accounts for the choice of the Ai and B and the latter terms
for the choice of the Bi and (si ) respectively.
Now note that for f E
pB (f )pB (f ) (6 + 2) = 8
and similarly for i j
pAi (si )pBi (si ) pAi (si )1 (si ) + pBi (si )(si )1 + 3
(14 + 9k + 6) + (14 + 9k + 6) + + 3, for d sufficiently large
= (44 + 18k)
so we can find > 0 with 0 as 0 so that for a given choice of pB (f )pB
there are dd choices for (f ) and similarly for (si ) (see [DKP2, Lemma 2.5]) so
SAG (F S, n + 4, , d)E,S
( ) SApGp (F, n, , d )E
d
d , d1 , . . . , dk+1
i si ) + )
d(h(s1
k k
( ) (di !)ddES .
i=1 di i=1
Therefore
log SAG (F S, n + 4, , d)E,S log + log SApGp (F, n, , d )E
d!
d !dk+1 !
i si ) + )!
d(h(s1
k
+ log + dE S log d
i=1 i si ) + ) di )!
di !(d(h(s1
hence
log SAG (F S, n + 4, , d)E,S + log d ! log d! + log SApGp (F, n, , d )E + (, d).
with
i si ) + )!
d(h(s1
k
(, d) = log + dE S log d
i=1 di !(d(h(si si ) + ) di )!
1
so since limd log d !

d log d
= h(p) and inf limd (, d) = 0 we obtain finally
SApGp (F, n, , d )E
sG,ES (F S, n + 4) + h(p) 1 + inf lim sup .
>0 d d log d
Using [DKP2, Lemma 2.13]
sG,ES (F S, n + 4) + h(p) 1 + h(p)spGp,E (F, n)
and we deduce
s(G) h(p)s(pGp) + 1 h(p).
The same proof works with lim inf d and limd instead of lim supd .
Finally we deduce the scaling formula:
Proof of Proposition 30. Let pn p qn be projections with h(pn ), h(qn ) Q and
pn qn 0.
h(p)s(pGp) h(pn )s(pn Gpn ) + h(p) h(pn )
by the previous lemma so lim inf n s(pn Gpn ) s(pGp). Similarly
h(qn )s(qn Gqn ) h(p)s(pGp) + h(qn ) h(p)
by the previous lemma so lim supn s(qn Gqn ) s(pGp). However the two lemmas
combined imply
s(G) = h(pn )s(pn Gpn ) + 1 h(pn ).
so
lim sup s(pn Gpn ) = (s(g) 1) + 1
1
n h(p)
and
s(G) = h(qn )s(qn Gqn ) + 1 h(qn )
so
lim inf s(qn Gqn ) = (s(G) 1) + 1.
1
n h(p)
Therfore
s(pGp) = (s(G) 1) + 1
1
h(p)
and the same argument works for s and s .
So
s(G) = s(G) s(pGp) = s(pGp)
so G is s-regular if and only if pGp is s-regular.
9. Applications of the correspondence principle and the proof of

Theorem 1
Since the work of [Dye], orbit equivalence relations have been used to prove
results in group theory on several occasions, with the Bernoulli action X0
serving as a link. A recent example is the use of the GaboriauLyons theorem [GL]
(that the orbit relation of [0, 1] of a non amenable group contains a nona-
menable subtreeing) as a way to replace the assumption contains a nonamenable
free group on by is non amenable. The correspondence principle discussed
here is a straightforward but useful extension of this wellknown idea from groups
to groupoids.
For example let us prove the following result using the principle:
Proposition 33. Let G X be a pmp action of an amenable pmp groupoid G.
Then s(G) = s(G X). In particular s(G X) doesnt depend on the action.
Let us first observe that the measure of the set of finite orbits of an action G X
doesnt depend on the action:
Lemma 34. If G X is an essentially free pmp action of a pmp groupoid G and
D X is a fundamental domain for the set of finite orbits of G in X, then
1
(D) =
G Ge
dh(e)
0
so (D) depends on G only.

Proof. Let
Xf = {x X Gx < }.
Since G X is free Gx = Gr(x) almost surely so

0 if Ge =
e (Xfe ) =

1 else
for ae e G0 and if D Xf is a fundamental domain then if e1 , e2 ..., en are the units

isomorphic to e we have
n
i (D i ) =
e e n
i=1 Ge
therefore
(D) =
1
G Ge
dh(e).
0
using invariance of .
Remark 35. An analog of the so-called fixed price problem [Gab] for s is the
question of whether s(G X) depends on G only and not on the essentially free
pmp action G X. This fixed sofic dimension problem holds for example for
groupoids with fixed price 1 (=all their free pmp actions have cost 1) provided all
their free pmp actions are sofic, but it is open in general. Also open is whether a
groupoid is sofic if and only if all its free pmp actions are sofic.
Proof of Proposition 33. Since the Bernoulli action G [0, 1]G is essentially free
G [0, 1]G is a pmp equivalence relation so by Theorem 28 and [DKP1, Corollary
5.2]
s(G) = s(G [0, 1]G ) = 1 (D) = 1
1
G Ge
dh(e)
0
where D is the fundamental domain of the set of finite classes of the amenable pmp
equivalence relation G [0, 1]G .
If G X is a pmp action then the action GX [0, 1]GX is pmp and essentially
free so by Theorem 28
s(G X) = s ((G X) [0, 1]GX ) .
Now the measure isomorphism
(G X) [0, 1]GX G (X [0, 1]GX )
((s, x), y) (s, (x, y))
is an isomorphism of pmp equivalence relations, where the action G X [0, 1]GX
is diagonal and G [0, 1]GX on the first coordinate.
So
s(G X) = s (G (X [0, 1]GX ))
= 1 (D )
=1
1
G0 Ge
dh(e).
where D is the is the fundamental domain of the set of finite classes of the amenable
pmp equivalence relation G X [0, 1]GX . So
s(G) = s(G X).

Another example is the proof that the invariance of s under orbit equivalence
(Theorem 9) established in [DKP1, Theorem 4.1] is equivalent to the statement for
groupoids in [DKP2, Theorem 2.1].
Theorem 36. Let G be a pmp groupoid, E, F be transversally generating sets, then
s(E) = s(F ), s(E) = s(F ) and s (E) = s (F ).
Proof. If F is a transversally generating set of G (which we assume to have infinite

fibers) then F {pB0 , pB1 } (as defined in Lemma 15) is a transversally generating
set of the pmp equivalence relation G {0, 1}G . By Theorem 28
s(F ) = s(F {pB0 , pB1 })
so by Theorem 9 we have
s(E) = s(E {pB0 , pB1 }) = s(F {pB0 , pB1 }) = s(F ).
The same applies to s and s .
We conclude with the proof Theorem 1 which is our main illustration of the
correspondence principle.
Proof of Theorem 1. Let G i = Gi G0 and G = GG0 = G 2 then for every
1 G3 G
3 3

essentially free pmp action G X of G we have
s(G) 1 = h(G0 ) (s(G
3
2 ) 1) by Proposition 30
1 G3 G
h(G03 ) (s(G 1 G3 G 2 X) 1) by Proposition 25
= h(G03 ) (s ((G 1 X1 ) (G X ) (G
3 3
2 X2 )) 1) by Lemma 18
= h(G03 ) (s(R1 R3 R1 ) 1) where Ri = G

i Xi are pmp
equivalence relations by Lemma 14
= h(G03 ) (s(R1 ) + s(R2 ) s(R3 ) 1) by [DKP1, Theorem 1.2]
= h(G03 ) (s(R1 ) + s(R2 ) s(G3 ) 1) by Proposition 33
X is Bernoulli then
If G
s(G 2 ) = s(G
1 G3 G 2 X)
1 G3 G
by Theorem 28 and G i Xi for i = 1, 2 are (isomorphic to) Bernoulli actions by
Lemma 13. Therefore s(Ri ) = s(G i ), i = 1, 2 by Theorem 28. Then
s(G) 1 = h(G03 ) (s(R1 ) + s(R2 ) s(G3 ) 1)
= h(G03 ) (s(G
1 ) + s(G 2 ) s(G3 ) 1) by Lemma 13 and
Theorem 28
= h(G ) (s(G
0
3
1 ) 1) + h(G ) (s(G
3
0 2 ) 1) h(G0 ) (s(G3 ) 1)
3
= h(G01 ) (s(G1 ) 1) + h(G02 ) (s(G2 ) 1) h(G03 ) (s(G3 ) 1)

by Proposition 30
= h(G01 )s(G1 ) + h(G02 )s(G2 ) h(G03 )s(G3 ) h(G01 ) h(G02 ) + h(G03 )
= h(G01 )s(G1 ) + h(G02 )s(G2 ) h(G03 )s(G3 ) 1
(We note that the inequality s(G1 G3 G2 ) s(G1 ) + s(G2 ) s(G3 ) could also be
proved by a direct argument.)
References
[BCG] Biane, P., Capitaine, M., Guionnet, A. Large deviation bounds for matrix Brownian mo-
tion. Invent. Math., 152(2), 433459. (2003)
[Bow] L. Bowen. Measure conjugacy invariants for actions of countable sofic groups. J. Amer.
Math. Soc. 23 (2010), 217245.
[Bow2] , Entropy theory for sofic groupoids I: the foundations arXiv:1210.1992
[BDJ] N.P. Brown, K. Dykema, and K. Jung, Free entropy dimension in amalgamated free prod-
ucts, Proc. London Math. Soc. 97 (2008), no. 2, 339367.
[Dye] H.A. Dye, On groups of measure preserving transformations. I, Amer. J. Math. 81 (1959),
no. 1, 119159.
[DKP1] K. Dykema, D. Kerr, M. Pichot, Orbit equivalence and sofic approximation,
arXiv:1102.2556
[DKP2] , Sofic dimension for discrete measure groupoids, arXiv:1111.2842
[EL] G. Elek and G. Lippner, Sofic equivalence relations, J. Funct. Anal. 258 (2010), no. 5,
16921708.
[FSZ] Feldman, J., Sutherland, C. E., Zimmer, R. J. Subrelations of ergodic equivalence relations.
Erg. Th. Dyn. Sys., 1988, 9(02), 239269.
[Gab] D. Gaboriau, Cot des relations dquivalence et des groupes, Invent. Math. 139 (2000),
no. 1, 4198.
[GL] Gaboriau, D., Lyons, R. A measurable-group-theoretic solution to von Neumanns problem.
Invent. Math., 177(3), (2009) 533540.
[Jun1] Jung, K. . A free entropy dimension lemma. Pacific J. Math. (2003), 177, 265271.
[Jun2] , The free entropy dimension of hyperfinite von Neumann algebras. Trans. AMS,
34. (2003).
[MS] Mineyev, I., Shlyakhtenko, D. Non-microstates free entropy dimension for groups, GAFA
(December 2005), 113.
[OW] D. S. Ornstein and B. Weiss. Entropy and isomorphism theorems for actions of amenable
groups. J. Analyse Math. 48 (1987), 1141.
[Oz] N. Ozawa, Hyperlinearity, sofic groups and applications to group theory, Notes from a 2009
talk available at http://people.math.jussieu.fr/pisier/taka.talk.pdf
[Pe] V. Pestov, Hyperlinear and sofic groups: a brief guide. Bull. Symbolic Logic 14 (2008), no.
4, 449480.
[Ram] Ramsay, A. Virtual groups and group actions. Advances in Mathematics, 322 (1971), 253
322.
[Shl] D. Shlyakhtenko, Microstates free entropy and cost of equivalence relations, Duke Math. J.
118 (2003), no. 3, 375426.
[Voi91] D. Voiculescu, Limit laws for random matrices and free products, Invent. Math. 104
(1991), no. 1, 201220.
[Voi96] , The analogues of entropy and of Fishers information measure in free probability
theory III: The absence of Cartan subalgebras, Geom. Funct. Anal. 6 (1996), no. 1, 172199.
[Voi98] , A strengthened asymptotic freeness result for random matrices with applications
to free entropy, Int. Math. Res. Not. 1998 (1998), no. 1, 4163.

1308 0663

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1308 0663

Uploaded by

Copyright:

Available Formats

A FREE PRODUCT FORMULA FOR THE SOFIC DIMENSION

ROBERT GRAHAM AND MIKAEL PICHOT

Abstract. It is proved that if G = G1 G3 G2 is free product of probability

Let G be a group. The sofic dimension of G is an asymptotic invariant that

s(F ) = inf inf lim sup

(2) G, G1 , G2 , and G3 are countable groups and G3 is an amenable group: see

A bisection is a Borel subset s G such that the restrictions of s and r to s

If K JGK is an arbitrary subset the sofic dimension of K is

where E F K are finite subsets. The sofic dimension of G is s(G) = s(JGK).

define a probability measure on X by = G0 e dh(e) where h is the invariant

(X0G , ) = (X0G , )dh(e),

(where X0G = Ge X0 is the infinite Cartesian product over Ge = r1 (e)), endowed

x X0G can be viewed as a sequence x = (xt )tGe of elements of X0 . The Bernoulli

since any g G can be written uniquely in the form h0 gj (e)s(e, 1

Consider the measurable field of maps e X0G (X0IJ )H defined by sending

((xh0 gj (e)s(e,1 f ) )(i,j)IJ ) .

f ) )(i,j)IJ,h0 H(e),(e,f )RH )

= h ((xh0 gj (e)s(e,1 f ) )(i,j)IJ )

e (Fix(s)) = 0. If in addition 0 is diffuse then the set of x X0G such that xs = xe

is negligible so e (Fix(s)) = 0 in this case.

Since kG X = f1 G X , kG X = f2 G X and the values of k are determined

(s)(t)1 (pj ) (pj ) = (s)(t)1 (pj ) (st1 pj )

we conclude (pj )pV = pj pV which implies tr (pj ) F0

tr(pj ) tr (pj ) + tr (pj ) F0

4. Bernoulli shifts and random partitions

Since for SA(Fn , 4nFn + 1, , d), we have

RRRR j h(F ) ((s (k))) RRR

Next suppose j = 1. For 1 k d let

Thus if we let Z = dk=1 Zk = F

Now lets bound var(Z):

For k, l we have E(Zk Zl ) = 0 = E(Zk )E(Zl ). On the other hand if k, l

E(Zk Zl ) E(Zk )E(Zl ) + 1. So returning to our equation above

E(Zk Zl ) E(Z) E(Zk )E(Zl ) + F Fj E(Z )

So for d large enough there is some partition (A1 , . . . , Aq ) such that

5. A lemma on approximate equivariance

Let P = {pBi } where

Both and depend on F and n only.

So in general we conclude that {x X i I, x Bi } is within c1 (F, n) of

so by applying the triangle inequality at most N = 2 times we have

(DI ) DI /d c2 (F, n)N .

So if iI ai = 1 then (iI Bi ) = (iI Bi B ) so by a similar argument

+ Be()2c2 (F, n)N + Be()c2 (F, n)N

c2 (F, n) + (3 + )Be()Nc2 (F, n)

6. Sofic dimension and groupoid actions

Lemma 22. If (, ) HA(F, 4n, , d) then

sE,Q (F, P, n, ) = lim sup log( HA(F, P, n, , d)E,Q )

Moreover if F is a finite transversally generating subset of JGK and P is a finite

7. The computation of s(G (X0 , 0 )G )

and extend 0 linearly to span(PFn ) L (X, ) with values in Md . We will check

Let pB PFn say pB = i=1 ai pBi then

c3 (F, n) + c2 (F, n) c3 (F, n)2 , for < 1

So it is enough to show that for all i, j

{x [d] 0 (pB )0 (pB )x = 0} d(1 c3 (F, n)2 4 )

2PFn 2 c3 (F, n)2 4 2 + 0

Recall that Fn = F {pF0 F0 Fn , partition of F0 }.

The right hand side equals

Let > 0 and choose n0 so that

Since the map X0 {1 . . . , q} defined by Ui x i extends G-equivariantly to

Hence s(G) s(G X). The same holds for s and s .

8. The scaling formula

For each i = 1 . . . k let (si ) be a bijection Bi Ai (we have ( N

where we denote (f ) the permutation ( (f ) 0

i1 si1 f1 sj1 sj1 )(sj1 )