Decoherent histories on graphs

Richard Blute, Ivan T. Ivanov, Prakash PanangadenarXiv 2001 · DOI · arXiv

Abstract

The consistent histories approach to quantum mechanics is traditionally based on linearly ordered sequences of events. We extend the histories formalism to sets of events whose causal ordering is described by directed acyclic graphs. The need for a global time is eliminated and our construction reflects the causal structure faithfully.

Introduction

In the "consistent histories" approach to quantum mechanics, probabilities are assigned to sequences of events, such as a particle traversing a series of filters. Traditionally, these events are assumed to follow a linear temporal order: Event A precedes Event B, which precedes Event C. However, this "film strip" model of history faces significant challenges in quantum cosmology and the study of entangled systems. In the early universe or near black holes, a single, well-defined global clock often does not exist. Similarly, for spatially separated subsystems like EPR pairs, imposing a linear order on distant events is artificial and fails to respect the underlying causal structure.

This research moves beyond linear histories toward a framework based on Directed Acyclic Graphs (DAGs). This approach respects local causality and allows for events to be "acausal"—neither strictly before nor after one another—thereby providing a more faithful representation of the universe's causal web.

Methodology: Quantum Evolution on Graphs

The core innovation is the replacement of linear event sequences with a DAG structure. In this model, vertices represent quantum operations (either unitary evolutions or projection operators), while edges represent the flow of quantum information. Each edge is labeled with a Hilbert space representing the degrees of freedom of a subsystem traveling between events.

A simple directed acyclic graph (DAG) representing quantum events (vertices) and subsystems (edges).

To define the state of the system at any "moment," the authors introduce the concept of a slice—a set of edges that are not causally related (spacelike). The framework provides a mathematical rule for propagating density matrices through the graph. When a vertex has multiple incoming edges, the operator acts on their tensor product. To maintain consistency regardless of the path taken, the authors utilize "locative slices" and partial traces to manage entanglement between distant edges.

Key Insights and Results

The transition to a graph-based formalism yields several profound insights:

  • Elimination of Global Time: The model reflects the system's causal structure faithfully. If two events are not connected by a path in the graph, they are causally unrelated, and the order of calculation does not change the outcome.
  • Discrete Covariance: The authors demonstrate that the final state of the system is independent of the specific "slicing" or sequence used to calculate it, providing a discrete, graph-based analog to Lorentz covariance in relativity.
  • Refinement as Zooming: The paper introduces "refinement," where a single vertex or edge can be "blown up" into a more complex sub-graph, allowing researchers to zoom in on specific space-time points.
A more complex causal structure where multiple paths branch and merge, illustrating the absence of a single global time.

The paper successfully generalizes the Griffiths-Omnès-Gell-Mann-Hartle formalism from linear chains to arbitrary DAGs. It demonstrates that quantum evolution can be treated as a local process at each vertex while still preserving the global entanglement of the system. Furthermore, the composition of these operators is formally encoded in a polycategory, linking quantum foundations to higher-order category theory.

Discussion and Future Work

While the framework is theoretically sound, it is currently strictly discrete. Translating this graph-based model into a continuous spacetime manifold remains a significant challenge. Additionally, the computational complexity of calculating decoherence functionals for very large, complex graphs has yet to be explored. Future research will also investigate the deeper relationship between this discrete covariance and the general relativistic treatment of curved spacetime.

Summary updated

Abstract

The consistent histories approach to quantum mechanics is traditionally based on linearly ordered sequences of events. We extend the histories formalism to sets of events whose causal ordering is described by directed acyclic graphs. The need for a global time is eliminated and our construction reflects the causal structure faithfully.
BibTeX
@misc{blute2001decoherent,
  author = {Richard Blute and Ivan T. Ivanov and Prakash Panangaden},
  title = {{Decoherent histories on graphs}},
  year = {2001},
  eprint = {gr-qc/0111020},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/gr-qc/0111020},
}
Figure