Towards a More Burkean Approach to Computational Social Choice

Omer LevAAAI 2024 · DOI

TL;DR

This paper argues that the computational social choice (COMSOC) community has focused too much on radical, novel decision-making mechanisms, neglecting centuries of practical experience with existing systems. It advocates for a "Burkean" approach, emphasizing understanding and improving established methods like parties and executive capacity, rather than wholesale replacement.

What Problem Does This Paper Address?

The field of computational social choice (COMSOC) has seen a surge in research on novel mechanisms for large-scale decision-making. However, the authors contend that many of these new approaches are not practical for real-world implementation and overlook the valuable lessons learned from existing, time-tested systems. This focus on radical innovation, while scientifically interesting, risks creating mechanisms that are too detached from human experience and thus unlikely to be adopted. The paper aims to shift the community's focus towards a more grounded approach that builds upon and refines existing structures, drawing inspiration from Edmund Burke's philosophy of gradual, incremental change.

Methods

The paper doesn't propose a new algorithm or formal model in the traditional sense. Instead, it advocates for a philosophical shift in research direction. The "method" is to adopt a "Burkean approach," which involves:

  • Understanding Existing Mechanisms: Deeply analyzing how large-scale human societies have historically made decisions, focusing on common features.
  • Improving, Not Replacing: Suggesting optimizations and tweaks to current systems rather than proposing entirely new ones.
  • Focusing on Under-Investigated Areas: Highlighting specific components of existing systems that have received relatively little attention in COMSOC research, such as the role of parties and executive capacity.
  • Drawing on Historical and Political Science Insights: Integrating knowledge from political science and historical examples to inform computational models.

The paper uses Edmund Burke's philosophy as a guiding principle, emphasizing caution, gradualism, and respect for established institutions. It also introduces a formal definition for executive capacity as a measure of a winner's stability in their victory, even when voters' preferences shift slightly.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • The "Radicalism" Trap: The COMSOC community's fascination with novel, often radical, mechanisms might be hindering practical impact and adoption.
  • Value of "Old" Systems: Centuries of human experience with decision-making systems offer invaluable lessons that should be leveraged, not discarded.
  • Burkean Approach: A philosophy of incremental improvement and understanding existing structures is proposed as a more fruitful path for COMSOC research.
  • Parties as a Key Component: The role of parties and sub-groups of voters and candidates is a fundamental aspect of human decision-making that is under-explored in COMSOC.
  • Executive Capacity Matters: A winner's ability to implement their agenda (executive capacity) is crucial and can be formally defined and studied.
  • Trade-offs Exist: There are likely inherent trade-offs between representability and executive capacity, which warrant further investigation.

Key Results

The paper is primarily a conceptual argument and a call for a new research agenda. It does not present quantitative results from experiments or simulations. Its "results" are in the form of:

  • A critique of current COMSOC research trends.
  • A proposed philosophical framework (the Burkean approach).
  • Identification of under-explored research areas: parties, sub-groups, and executive capacity.
  • A formal definition of executive capacity.

Limitations and Future Work

The authors acknowledge that their approach might be seen as less "ground-breaking" than proposing entirely new systems. They also note that while their critique is not of individual researchers, the community's collective focus on radical change is problematic.

Future work suggested includes:

  • Deeper investigation into parties: Exploring axiomatic approaches and mechanism design for district and intra-party elections, and understanding voter behavior under various informational settings.
  • Formalizing and studying executive capacity: Developing richer frameworks to measure and understand winner robustness and its trade-offs with representability.
  • Exploring other under-investigated components of existing decision-making systems.
  • Developing better metrics and axiomatic frameworks for partisan mechanisms.

Summary updated

BibTeX
@inproceedings{lev2024towards,
  author = {Omer Lev},
  title = {{Towards a More Burkean Approach to Computational Social Choice}},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year = {2024},
  doi = {10.1609/aaai.v38i20.30270},
}
Figure