Canonical: https://knowlpedia.com/noncommutative-geometry/dimension-spectrum/

# Dimension spectrum of a spectral triple

The discrete set of possible poles of weighted spectral zeta functions associated with a regular spectral triple.

Let \((\mathcal A,H,D)\) be a [regular spectral triple](https://knowlpedia.com/noncommutative-geometry/regular-spectral-triple/), and let \(\mathcal B\) be the algebra generated by all \(\delta^k(a)\) and \(\delta^k([D,a])\), where \(a\in\mathcal A\), \(k\geq0\), and \(\delta(T)=[|D|,T]\). A discrete set \(\Sigma\subset\mathbb C\) is a **dimension spectrum** if, for every \(b\in\mathcal B\), the initially defined
\[
\zeta_b(s)=\operatorname{Tr}(b|D|^{-s})
\]
extends meromorphically to \(\mathbb C\) with all poles contained in \(\Sigma\), and \(\Sigma\) is minimal with this property. The triple has **simple dimension spectrum** when all those poles are simple. Thus the definition concerns a family of [weighted spectral zeta functions](https://knowlpedia.com/noncommutative-geometry/spectral-zeta-function/), not only \(\zeta_D\).

## Geometric meaning

Pole locations encode the exponents in high-energy or small-time asymptotic
expansions. For the canonical Dirac [spectral triple](https://knowlpedia.com/noncommutative-geometry/spectral-triple/) on a closed
\(d\)-dimensional manifold, the dimension spectrum is contained in an
arithmetic sequence of integers [bounded above](https://knowlpedia.com/real-analysis/bounded-above/) by \(d\); parity and vanishing
of heat coefficients can remove some candidate poles. The largest actual pole
often recovers the metric dimension, but the full set contains finer
information.

## Role in the local index formula

The local index formula uses residues of the functions \(\zeta_b\) to build
cyclic cocycles. Regularity alone supplies the algebra \(\mathcal B\), but it
does not imply the required meromorphic continuation. Discreteness and bounds
on pole order are separate analytic hypotheses.

## Conventions and scope

**Warning.** Some authors specify only a pole-containing set rather than the
minimal set of actual poles, and some enlarge \(\mathcal B\) to an abstract
pseudodifferential algebra. Either choice can enlarge the reported spectrum.
The chosen coefficient algebra, kernel convention for \(|D|^{-s}\), and
allowed pole orders should therefore accompany any assertion about a
dimension spectrum.

## References

1. A. Connes and H. Moscovici, “The Local Index Formula in Noncommutative Geometry,” *Geometric and Functional Analysis* 5 (1995), 174–243. [DOI record](https://doi.org/10.1007/BF01895667). Relevant: §§II–III, dimension spectrum and residue cocycles.
2. N. Higson, “The Local Index Formula in Noncommutative Geometry,” in *Contemporary Developments in Algebraic K-Theory*, ICTP Lecture Notes 15, 2004. [Author-hosted manuscript](https://nigel.higson.ca/uploads/1/2/1/4/121496570/higson_-_2004_-_the_local_index_formula_in_noncommutative_geometry.pdf). Relevant: zeta-function asymptotics and dimension-spectrum hypotheses.
