Mapping foundational contributions in complex systems and network science
MINERVA is a community-driven initiative to map the intellectual foundations of complex systems and network science. We are collecting foundational papers from researchers across the field: not to rank papers or authors, but to build a structured, paper-centered map of the ideas, methods, models, and applications that shaped how we think about complexity.
Why MINERVA?
Contributions from different generations, disciplines, and perspectives are essential. The obvious papers matter, but so do overlooked works and ideas imported from neighboring fields. The goal is to build a coherent, field-wide reference that reflects how the discipline is actually used and understood across different subdomains.
In recent years, large-scale models and automated systems have made it possible to synthesize vast amounts of scientific information. However, identifying what is foundational — what truly shapes the conceptual and methodological backbone of a field — still requires distributed expert judgment. This effort is designed to complement algorithmic approaches by leveraging collective intelligence: many independent perspectives, aggregated into a structured view.
Not a ranking
A structured map of foundational contributions, not a hierarchy of papers, authors, or citation scores.
Expert-informed, community-refined
Expert curation provides an initial scaffold; community recommendations help identify missing or underrepresented work.
Conceptual breadth
The scope includes adjacent domains that shaped key ideas such as information, dynamics, computation, organization, and adaptation.
From expert core to community-driven expansion
MINERVA currently combines an expert-group core selection, an expert augmentation layer, and a community-driven candidate layer under evaluation.
Expert-group core selection
Initial expert-defined core corpus. Most entries are confirmed; a small subset is still under independent expert revision.
Bars are scaled relative to the largest current count: 270 community recommendations. Unique IDs refer to distinct candidate papers and may partially overlap with the expert-defined corpus.
Community candidate layer
The community-driven layer currently contains 270 recommendations, corresponding to 45 unique candidate paper IDs. These are being evaluated against the expert-group core and augmentation layers.
Snapshot terminology: “confirmed” refers only to the expert-group core selection. Community-driven recommendations are candidate entries under evaluation, not yet validated additions.
Scope of the initiative
An initial corpus has been assembled drawing from expert input and curated sources. The current phase focuses on expanding and refining this corpus through community contributions, with particular attention to identifying missing or underrepresented work.
Importantly, the scope is not limited to papers explicitly labeled as “complex systems” or “network science.” It also includes foundational contributions from adjacent domains that shaped key concepts used in the field — such as information, dynamics, evolution, computation, organization, and adaptation. The objective is to capture the conceptual foundations of the field, not only its explicitly defined boundaries.
How to contribute
If you work in complex systems, network science, or closely related areas, your input would be valuable. The survey asks you to submit a small set of key papers using DOI references only, including both widely recognized contributions and work you consider essential but potentially overlooked.
Contributions will be aggregated, deduplicated, and analyzed across subfields and types of contribution — for example theoretical frameworks, methods, applications, interdisciplinary bridges, and foundational imports from adjacent areas.
Relation to previous efforts
This initiative complements computational and bibliometric mappings of complex systems research, including approaches based on keyword extraction, large bibliographic databases, and automated topic discovery. Such methods are valuable for identifying explicit topical structure, but they may miss conceptual ancestors or foundational imports that do not use the modern vocabulary of the field.
This effort also draws on experience from community-driven initiatives, such as Complexity Explained (now available in 15 languages), and divulgative initiatives, such as Complexity Thoughts, aimed at structuring and communicating research in complex systems. The present project extends that work toward a more systematic, community-informed synthesis.