The Setting
The question is how kernel energy can grow relative to fibre dimension while a representation retains a prescribed common operator bound. The constructions involve finite families of rank-one weights, with their labels retained when the associated group elements coincide.
Let be a discrete group and a finite set. The fibre is the Euclidean space of complex functions on , of dimension . The space in (1) consists of square-summable fibre-valued functions, and denotes its bounded complex-linear operators. (1)
For , coordinate evaluation sends to , and coordinate inclusion places a vector at and is zero elsewhere.✦ The row, column and block of are the maps in (2). (2)
For , its coefficient matrix is the vector in whose entry is , where is the standard basis vector.✦ Its Euclidean norm is the unnormalized Hilbert-Schmidt norm. The kernel energy of is the sum in (3). (3)✦
For a bijection , the coordinate permutation is the operator on given in (4). The left regular representation of on is the monoid homomorphism given in (4); for , right translation is the operator on given in the same display. (4)
The translation commutant is the set of that commute with every .✦ Conjugation by and the coboundary of an operator are defined in (5). (5)
Let be a finite label set, and for . For , the rank-one operator on and its coefficient matrix are defined in (6). The collision kernel is the -valued function on whose value is the sum of the coefficient matrices of the rank-one operators over the labels with . (6)
A mask on is a function with . For , the weighted rank-one shift , the masked operator , and the unmasked operator are the bounded operators on defined in (7). (7)
For a nonnegative integer , a real and a nonnegative integer , put . Here and is the natural logarithm.
Definition 1. The endpoint capacity is the nonnegative integer in (8). (8)✦ The fibre dimension is the nonnegative integer in (9). (9)✦
The Hilbert direct sum of a family of complex Hilbert spaces is the Hilbert space of families with , with squared norm equal to that sum.✦ The direct sum operator of a uniformly bounded family is the bounded operator on given by .✦
For a dimension sequence , the stage spaces , their stage sum , and the obstruction space are specified in (10). (10)
A representation on a complex Hilbert space is a monoid homomorphism . A representation of on a complex Hilbert space is similar to a unitary representation if there is a bounded invertible complex-linear map satisfying (11), with no restriction on . (11)✦
The Claim
Theorem 1. Let be an integer and a real number. The dimensions satisfy (12) for every . (12)✦ If , the dimension ratio has the limit in (13). (13)✦ For and a real sequence satisfying for every , the energy ratio has the limit in (14). (14)✦
Let be a discrete group, finite sets, and let and . The unmasked operator has the coefficient kernel in (15) for every . (15)✦ When every is a unit vector, the collision kernel and the unmasked operator satisfy (16) and (17), respectively. (16)✦ (17)✦
Let be positive integers and let , with in the translation commutant on . Suppose that a common real satisfies the bounds in (18), together with the energy growth specified there. (18) For every real , there exists a representation on with the properties in (19). (19)✦
In the paper, the result reads: for a countable nonamenable group and the parameters of Lemma 3.1, the quantities , and satisfy ; for unit rank-one weights and arbitrary product collisions, has ; rescaling by , for , preserves divergence of energy divided by fibre dimension and gives common bound in Proposition 2.2, hence a nonunitarizable representation with bound .
What Is Checked
The statements concern arbitrary discrete groups, finite label and fibre sets, and Hilbert direct sums indexed by natural numbers starting at zero; the rescaling statement assumes positive stage dimensions and the commutant condition, bounds and energy growth in (18).
The coordinate permutations in (4) are complex-linear isometries. A family of coordinate operators with a common nonnegative norm bound has a direct sum operator of norm at most . The rank-one operators in (6) have norm , and the fibre maps in (7) have this same common upper bound.