Skip to content

Comparative-Ethology Age-Related Sociality — Literature Scoping for Spectral-Framework Integration

Date: 2026-05-13 Branch: research/comparative-ethology-age-sociality Author: Subagent literature scoping for srmech Status: Research notes only (no PR)

1. Summary

Comparative ethology has thoroughly documented a juvenile-permeability / adult-exclusion transition across mammals and birds — Goodall's chimpanzees, Holekamp's hyenas, McComb's elephants, and the cooperative-breeding bird tradition all carry the pattern. The field's methodological vocabulary is graph-theoretic: animal social network analysis (ASNA) is mature, and community detection via Newman's leading-eigenvector modularity is routine (Sosa et al. 2021; Farine 2015). Hierarchical decomposition exists (Hill–Bentley–Dunbar 2008 fractal-scaling). Higher-order structure has entered the field only in the last 3 years (Iacopini et al. 2023; Curley & Ophir 2023 conceptual review), and Hodge-Laplacian / persistent-homology / representation-theoretic decomposition has not — TDA on biological aggregations exists for swarms (Topaz et al. 2015) but not for age-stratified social fibres. The maturity bucket is partial: pairwise-eigenvector spectral methods are standard; bundle-style / Casimir-style decomposition is the open frontier the user is pointing toward.

2. The cross-species pattern (juvenile cooperation → adult competition)

The user's framing maps onto the field's well-known ontogenetic transition in network position. Anchored exemplars:

  • Chimpanzees (Pan troglodytes). Juveniles play cross-kin within community; adult males form cooperative patrols that conduct lethal intergroup raids (Wrangham 1999; Watts et al. 2006; Wilson & Wrangham 2003; Mitani et al. 2010). Wrangham & Glowacki (2012) review the "intergroup hostility / intragroup cooperation" complex.
  • Spotted hyenas (Crocuta crocuta). Smith et al. (2017, Behav Ecol Sociobiol) used ASNA across three ontogenetic stages (communal den → den-independent pre-reproductive → early adulthood). Cubs of both sexes are well-connected; males then disengage from natal clan, females are recruited into matrilineal rank structure. Network position metrics decrease from infancy to adulthood except for "found-alone" rate.
  • African elephants (Loxodonta africana). Matriarchal cores remain stable into old age (McComb, Moss, Durant, Baker & Sayialel 2001, Science 292:491). Bulls leave natal unit at ~14 yr; musth-period adults become exclusionary. Goldenberg et al. (2024, Phil Trans R Soc B) reviews knowledge-transmission consequences of social disruption.
  • Lions (Panthera leo). Female prides are kin-cohesive across life; sub-adult males are ejected at sexual maturity (Hanby & Bygott; reviewed in Pusey & Packer 1987 Behaviour).
  • Wolves (Canis lupus). Pup cohorts play across litters within natal pack; adult breeding-pair excludes unrelated adults; dispersers face high mortality (Mech & Boitani 2003).
  • Cooperative-breeding birds (~9% of species). Family-living with delayed dispersal of helpers, followed by adult-territory competition (Koenig & Dickinson, eds., Cooperative Breeding in Vertebrates, Cambridge 2016; Drobniak et al. 2016).
  • Yellow-bellied marmots (Marmota flaviventris). Wey & Blumstein (2010, Animal Behaviour) showed younger animals as net affiliation-receivers (cohesion-builders), adults as net agonism-initiators (cohesion-reducers) — a direct ASNA confirmation of the user's framing.
  • Spotted dolphins (Stenella longirostris) and bottlenose dolphins (Tursiops). Juvenile crèches are permeable across alliances; adult male alliances are exclusionary (Connor et al. 2000).

3. Bonobos and other counter-examples

The pattern is not universal.

  • Bonobos (Pan paniscus). Adults maintain inter-community tolerance, food-sharing across communities, and female-bonded coalitions — despite female-dispersing societies that should weaken adult female bonds (Hare et al. 2007, Current Biology 17:619; Hare et al. 2012 self-domestication review; Sakamaki et al. 2018 inter-community encounters; lethal events are rare but exist — Mouginot et al. 2025 Sci Rep).
  • Eusocial mammals (naked mole-rat Heterocephalus glaber, Damaraland mole-rat). Adult-status is castelike, not age-graded the same way. Useful boundary case for caste-segregation vs age-segregation.
  • Killer whales (Orcinus orca, resident pods). Both sexes show lifelong natal philopatry; matriarchal post-reproductive females are central network hubs (Brent et al. 2015). No "adult-exclusion" phase.

These exceptions suggest the juvenile-to-adult exclusion gradient is a dimension of variation, not a universal, which is exactly the kind of phenomenon a structural-decomposition framework should resolve into components.

4. Existing structural methods in the domain

What ASNA already does well:

  1. Graph-theoretic node/edge metrics. Degree, strength, betweenness, eigenvector centrality, clustering coefficient. Standard since Croft, James & Krause (2008) Exploring Animal Social Networks, Princeton UP.
  2. Community detection. Newman (2006) leading-eigenvector modularity is the field default; Louvain and edge-betweenness also widely used (Sosa et al. 2021, Methods Ecol Evol 12:10–21).
  3. Multilevel / fractal hierarchical analysis. Hill, Bentley & Dunbar (2008, Biol Lett 4:748–751) used Horton–Strahler scaling on elephants, geladas, hamadryas baboons, orcas; consistent branching ratio ≈ 3 across mammals.
  4. Temporal/dynamic networks. Pinter-Wollman et al. (2014, Behavioral Ecology 25:242); Fisher, Ilany, Silk & Tregenza (2017, J Anim Ecol 86:202) — stochastic actor-oriented models (SAOMs / Snijders).
  5. Higher-order networks (recent). Iacopini, Foote, Fefferman, Derryberry & Silk (2023, Phil Trans R Soc B, arXiv:2309.03783) advocate hypergraphs and simplicial complexes for animal communication. Curley & Ophir (2023, Animal Behaviour — "Conceptual representations of animal social networks") survey the conceptual landscape.
  6. Multilevel toolkits. Sosa, Sueur, Puga-Gonzalez et al. (2020, Sci Rep — ANTs R package) provides global/dyadic/nodal measures and permutation tests.
  7. Network–demography integration. Woodman, Gokcekus, Beck, Green, Nussey & Firth (2024, Phil Trans R Soc B 379:20220464) — "The ecology of ageing in wild societies" — explicitly synthesizes age-structure × social-network research. This is the sharpest single review to read.

What the field uses rarely or never (verified via three targeted searches returning no animal-ethology results): - Hodge Laplacian on age-stratified animal social complexes. Hodge-theoretic decomposition is used in biomolecular data (Wee & Xia 2022, Sci Rep) and biological-aggregation TDA (Topaz, Ziegelmeier & Halverson 2015, PLOS One) but not on social ontogeny networks. - Persistent homology of animal social networks. No matches for age-stratified ontogenetic applications. - Representation theory / Lie-group decomposition / fibre-bundle structure on animal social organisation. Zero results — the conceptual translation has not happened.

5. Distance to the "stream of spectral analysis"

Maturity bucket: PARTIAL.

Method tier Field status
Pairwise graph metrics ✅ mature (standard since 2008)
Eigenvector modularity / spectral community detection ✅ mature (Newman 2006 routine)
Multilevel fractal hierarchical decomposition ✅ exists (Hill–Bentley–Dunbar 2008)
Temporal-dynamic ASNA ✅ mature (Pinter-Wollman 2014, Fisher 2017)
Higher-order (hypergraph / simplicial) 🟡 emerging (2023–2026, Iacopini et al.)
Hodge-Laplacian on ASNA ❌ absent
Persistent homology on age-stratified social networks ❌ absent
Representation-theoretic / Casimir / fibre-bundle decomposition ❌ absent

The gap is precisely the bundle-style decomposition. The field has graphs and is beginning to have simplicial complexes; it does not yet have base-space × fibre decomposition, nor does it have Casimir-like quadratic invariants that would isolate a "juvenile-cooperation fibre" from an "adult-competition fibre" over a common base-space (age or developmental-stage manifold). The Hill–Bentley–Dunbar fractal-scaling finding (consistent branching ratio ≈ 3) is the closest existing result to a Lie-algebraic / structure-constant invariant; it is descriptive rather than representation-theoretic.

6. Specific paper anchors

  1. Woodman, Gokcekus, Beck, Green, Nussey & Firth (2024). The ecology of ageing in wild societies: linking age structure and social behaviour. Phil Trans R Soc B 379(1916):20220464. DOI: 10.1098/rstb.2022.0464. Top-priority review.
  2. Hill, Bentley & Dunbar (2008). Network scaling reveals consistent fractal pattern in hierarchical mammalian societies. Biol Lett 4(6):748–751. DOI: 10.1098/rsbl.2008.0393. Foundational multilevel paper.
  3. Smith, Memenis & Holekamp, with Turner & Watts (2017). Ontogenetic change in determinants of social network position in the spotted hyena. Behav Ecol Sociobiol 72(1):11. DOI: 10.1007/s00265-017-2426-x.
  4. Croft, James & Krause (2008). Exploring Animal Social Networks. Princeton UP. ISBN 978-0-691-12752-1. ASNA textbook.
  5. Sosa, Puga-Gonzalez, Hu, Pansanel, Xie & Sueur (2021). Network measures in animal social network analysis: their strengths, limits, interpretations and uses. Methods Ecol Evol 12:10–21. DOI: 10.1111/2041-210X.13366.
  6. Pinter-Wollman, Hobson, Smith et al. (2014). The dynamics of animal social networks: analytical, conceptual, and theoretical advances. Behavioral Ecology 25(2):242–255. DOI: 10.1093/beheco/art047.
  7. Fisher, Ilany, Silk & Tregenza (2017). Analysing animal social network dynamics: the potential of stochastic actor-oriented models. J Anim Ecol 86(2):202–212. DOI: 10.1111/1365-2656.12630.
  8. Iacopini, Foote, Fefferman, Derryberry & Silk (2023). Not your private tête-à-tête: leveraging the power of higher-order networks to study animal communication. Phil Trans R Soc B (also arXiv:2309.03783). Emerging-frontier paper — hypergraphs + simplicial complexes.
  9. McComb, Moss, Durant, Baker & Sayialel (2001). Matriarchs as repositories of social knowledge in African elephants. Science 292:491–494. DOI: 10.1126/science.1057895.
  10. Wrangham & Glowacki (2012). Intergroup aggression in chimpanzees and war in nomadic hunter-gatherers. Human Nature 23:5–29. DOI: 10.1007/s12110-012-9132-1.
  11. Hare, Melis, Woods, Hastings & Wrangham (2007). Tolerance allows bonobos to outperform chimpanzees on a cooperative task. Current Biology 17(7):619–623. DOI: 10.1016/j.cub.2007.02.040. Counter-example anchor.
  12. Snyder-Mackler, Burger, Gaydosh et al. (2020). Social determinants of health and survival in humans and other animals. Science 368:eaax9553. DOI: 10.1126/science.aax9553. Cross-species ageing-and-sociality synthesis.
  13. Koenig & Dickinson, eds. (2016). Cooperative Breeding in Vertebrates: Studies of Ecology, Evolution, and Behavior. Cambridge UP.
  14. Wey & Blumstein (2010). Social cohesion in yellow-bellied marmots is established through age and kin structuring. Animal Behaviour 79(6):1343–1352. DOI: 10.1016/j.anbehav.2010.03.008.
  15. Topaz, Ziegelmeier & Halverson (2015). Topological data analysis of biological aggregation models. PLOS One 10(5):e0126383. Bridge paper — TDA in animal contexts, but on flocks/swarms not age-stratified social networks.

7. Open research gaps for spectral-framework integration

Three concrete gaps the user's framework could plausibly fill:

  1. Hodge-Laplacian decomposition of age-stratified social complexes. Build a simplicial complex per ontogenetic stage (cub den, juvenile, sub-adult, adult, post-reproductive) from existing ASNA datasets (Holekamp hyena clan, Amboseli elephants, Wytham great tits — Firth/Sheldon — all have multi-year longitudinal data). Compute Betti numbers per stage; track changes in cycle structure across ontogeny. Nobody has done this.

  2. Base-space × fibre decomposition: developmental stage × interaction-type. The user's framing translates to: a base-space B of age/developmental stage, and a fibre F of interaction-type (play, agonism, grooming, alliance, mating). Existing multilayer-network animal work (e.g., Silk, Finn, Porter & Pinter-Wollman 2018, Trends Ecol Evol) is the natural lift-off point but is not formulated as a fibre bundle. Casimir-like invariants would be the obvious next test.

  3. Falsifiable comparison of "human exceptionalism" claim. The user's framing implicitly conjectures humans show the same juvenile-cooperation / adult-exclusion structure as comparative species. A representation-theoretic invariant computed identically on chimpanzee, bonobo, hyena, marmot, and human social datasets (where ethically and methodologically possible) would test this empirically rather than rhetorically. Bonobos as the obvious "tolerant-adult" outlier provide a built-in control.

Note on Hill–Bentley–Dunbar consistent branching ratio ≈ 3. This is suspiciously close to the kind of structure-constant invariant that representation theory would produce for a small Lie algebra. Whether it reflects underlying group structure (e.g., D₃ or related) or is just a cognitive-bandwidth artefact is, as far as I can determine, unsettled. This is the single most spectrally suggestive existing result in the literature.