Multi-dimensional Taxonomies for Systematic Research Gap Analysis

This direction is split into other directions: faceted classification. Its last version stays readable and citable here.

Version 1 of 158 papers64 conceptsfrozen 2026-10-04

In rapidly evolving fields like AI, healthcare, and software engineering, research often becomes a "fragmented jungle." When thousands of papers are published annually, traditional hierarchical classification—organizing topics like a family tree—breaks down. A tree requires every leaf to have exactly one parent, but modern research is inherently multi-faceted. A single study might simultaneously involve a specific task (e.g., summarization), a method (e.g., reinforcement learning), and a target population (e.g., low-resource languages).

To navigate this complexity, researchers are increasingly adopting multi-axial taxonomies. Instead of a rigid hierarchy, this approach treats research dimensions as independent, orthogonal axes. By mapping existing work into this multi-dimensional space, researchers can move beyond mere "summaries" to perform rigorous gap analysis, identifying exactly where a field is over-saturated and where it is dangerously thin.

The Geometry of Research Landscapes

The core intuition behind a multi-axial framework is the transition from a list to a coordinate system. As Mei Hong argues in Potential Usage of Faceted Classification in Internet "Information Retrieval", "faceted classification" decomposes complex subjects into atomic, independent units. These units can be dynamically combined to describe a piece of information from multiple perspectives simultaneously.

Consider the work of Anders Søgaard in Do Language Models Have Semantics? On the Five Standard Positions. Rather than debating semantics as a single spectrum, Søgaard uses a "combinatorial taxonomic formalization." By crossing independent dimensions—such as internal versus external grounding—he creates a matrix that reveals hybrid theoretical positions that were previously obscured by binary debates.

Revealing the White Space

The primary power of these frameworks is their ability to visualize "white space"—the combinations of variables that no one has studied yet. When you plot research along multiple axes, the gaps become mathematically visible as empty cells in a matrix.

Standardizing Fragmented Ecosystems

Beyond gap analysis, these taxonomies serve as a "health check" for a field's infrastructure. In The State and Fate of Summarization Datasets, Noam Dahan introduced a 7-axis ontology to standardize how datasets are described. By categorizing resources across orthogonal dimensions like modality, domain, and supervision type, Dahan exposed the field's over-reliance on the news domain and the relative scarcity of high-quality data for low-resource languages.

Similarly, Nhan Tran proposed the MLCommons Scientific Benchmarks Ontology to unify the fragmented landscape of scientific machine learning. By using a dual-axis approach—mapping domain-specific motifs against algorithmic tasks—the framework allows researchers to identify common computational patterns across disparate fields like physics and climate science.

Why It Works: The Analytico-Synthetic Advantage

The shift toward multi-axial mapping works because it mirrors the way modern knowledge is actually constructed. As Gita Sankaran and Shankar Sankaran demonstrate in their literature review of Knowledge Management, using a tri-axial coding system (methodology, level of analysis, and thematic focus) allows for a "state of the union" that is far more granular than a simple chronological summary.

By treating research as a set of coordinates in a high-dimensional design space—as seen in the work on live music agents by Yewon Kim and colleagues—the community can move from "reinventing the wheel" to systematically exploring the "unexplored design configurations." This structured approach transforms a chaotic pile of papers into a navigable map, showing us not just where we have been, but exactly where we need to go next.

Papers in this version

History & citationVersion 1 · 2026-10-04

Split into other directions · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_multi_dimensional_taxonomies_for_systema_b490939a,
  title = {Multi-dimensional Taxonomies for Systematic Research Gap Analysis},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/multi-dimensional-taxonomies-for-systematic-research-gap-analysis/rdr_b490939ab9484f63906d12f8367163a6},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-21},
}

This exact version (version 1, frozen text):

@misc{lacuna_multi_dimensional_taxonomies_for_systema_b490939a_v1,
  title = {Multi-dimensional Taxonomies for Systematic Research Gap Analysis},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/multi-dimensional-taxonomies-for-systematic-research-gap-analysis/rdr_b490939ab9484f63906d12f8367163a6/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-21},
}