Here are “multi-polar traps” you keep running into when a lineage tries to become a Darwinian individual (an organism) while the old units (cells, nuclei, lineages) still have their own fitness gradients. “Multi-polar” here means there are multiple selection targets with partially incompatible objective functions: cells vs collective, different cell lineages vs each other, nuclei within a syncytium, and sometimes “fast” within-lifetime selection vs “slow” between-generation selection.

1. Relatedness collapse: aggregation and chimeras (the classic “different genotypes in one body” trap) If your multicellular stage is built by aggregation or frequent fusion, you get genetically mixed groups. Then within-group selection becomes loud: selfish lineages can treat the collective as a public-goods buffet. The mechanistic problem is simple: selection can “see” cell-level replication differences inside the group. The reason clonal development matters is it throttles that signal. This is why the aggregative mode gets framed as intrinsically cheater-vulnerable: “cheating within non-clonal groups has the potential to undermine multicellular adaptation.” (Springer) Names to anchor this: Pedro Márquez-Zacarías, William C. Ratcliff (reviewing why aggregative lineages “stayed simple”). (Springer)
2. The kin-recognition specificity vs speed trade-off (aggregation wants to be fast; policing wants to be picky) Even if you “solve” Trap #1 by evolving kin discrimination, there’s a second pole: selection for rapid aggregation can oppose selection for high-specificity kin recognition. The Márquez-Zacarías review states it bluntly: “selection for the ability to aggregate quickly may constrain the evolution of highly specific kin recognition, sowing the seeds for persistent evolutionary conflict.” (Springer) This is a good example of multi-polarity that isn’t just “cheaters exist”: two collective-level desiderata can conflict (ecological timing vs genetic homogeneity).
3. “Renegade soma” / reversion: cells that abandon maintenance and claw back reproduction Once you have division of labour, you’ve created distinct fitness landscapes for “maintenance specialists” vs “reproducers.” That invites a specific defection: soma lineages that dedifferentiate, stop paying the maintenance tax, and try to re-enter the germline channel. A review of Leo W. Buss summarizes the core move: “Cells that gave up their maintenance role to reproduce… gained an immediate advantage.” (Academia) This is basically the conceptual ancestor of cancer-as-cheating, but it applies far earlier: any somatic commitment device that’s leaky creates a within-organism arms race.
4. No life cycle, no individuality: the “collective reproduction is not a trivial issue” trap You can get groups, mats, clusters, even some functional differentiation… and still fail to become a proper evolutionary individual because you cannot reproduce as a collective in a heritable way. Paul B. Rainey and Benjamin Kerr put the knife in: “the critical problem is the evolution of a means of collective reproduction.” (icts.res.in) This is the “new pole” problem: selection at the group level needs a heredity channel at the group level, not just transient clustering.
5. The bottleneck trap (and the even nastier twist: bottlenecks may start as exaptations, not ‘policing adaptations’) Orthodox story: single-cell bottlenecks jack up relatedness and suppress internal conflict. But even sympathetic sources note the historical ambiguity about why bottlenecks appear when they do. A plant-focused review notes: “standard multilevel selection theory… posits that a unicellular ‘bottleneck’ is required to reduce intercellular conflict,” then immediately adds: “unicellular bottlenecks may be exaptations conferring immunity to future cell–cell conflicts rather than being adaptations per se.” (OUP Academic) That matters because it implies an easy failure mode: if your ecology doesn’t hand you a bottleneck (or anything that behaves like one), you may never get conflict under control.
6. “Cheats as propagules”: when cheating creates a primitive germline (useful, but unstable) Rainey & Kerr’s point is that cheats aren’t only destroyers: under some ecological scaffolds, cheating lineages can become the first “reproductive” propagules that instantiate a life cycle. Their abstract summary is explicit: a life cycle can arise “from conflicts among levels of selection and invokes cheats as a primitive germ line.” (icts.res.in) This is a trap because it’s double-edged: you get a reproduction channel, but the reproduction channel is literally powered by defection, so you now need secondary policing to keep it from eating the whole system.
7. Facultative life cycles and pleiotropy: selection keeps snapping back to “unicellular priorities” Even if you could, in principle, build complex multicellular traits, if your lineage mostly lives as single cells, cell-level selection dominates the long-run response to selection. The aggregative-multicellularity review states: “Because they primarily exist in a unicellular state, it may be difficult… to evolve multicellular traits that carry pleiotropic cell-level fitness costs.” (Springer) This is the “time budget” version of multi-polarity: the pole that gets more generations per unit time wins unless you actively firewall it.
8. Syncytia and multi-nuclear bodies: within-cytoplasm conflicts (nuclei as competing lineages) Once you allow multiple nuclei to share a cytoplasm (common in fungi), you’ve created a new competitive arena: nuclei can act like “selfish genotypes” with partial independence. Buss’s review flags this as a distinct problem class: “Fungi with coenocytia face special problems because the different nuclei can come into conflict.” (Academia) This is a literal “multi-polar” trap: the poles aren’t even cells anymore, they’re replicators inside the same cell.
9. Coordination and division-of-labour constraints: incompatible functions force either temporal separation or cell-type specialization A non-cheater trap: you can fail because some functions are physically incompatible at the same time in the same cell. A worked example in the volvocine algae is the “flagellation vs cell division” incompatibility; one summary explains that in Volvox carteri, the conflict is “solved… where these incompatible functions are segregated into two different cell types.” (Normale Sup) This is multi-polarity as constraint satisfaction: you need a developmental program that can allocate tasks without letting reproduction re-invade everything.
10. “Cheats take over and everyone dies”: the dispersal/segregation failure mode One sharp ecological reason early collectives are fragile is that cooperators often can’t just leave when a cheat appears. A heterodox-ish but clear popular writeup says: “They cannot disperse when cheats arrive. Cheats, therefore, take over… leading to the eventual downfall of all.” (Psychology Today) If you want this in a more meme-dense rationalist framing, a LessWrong post compresses the endpoint as: “Multicellular organisms can only exist because they’ve evolved powerful internal mechanisms to outlaw evolution.” (lesswrong.com) Those are rhetorically different, but they’re pointing at the same dynamical trap: if within-group selection is allowed to run unconstrained, the collective is a temporary arrangement.
11. Temporal-scale mismatch: “fast” selection within collectives vs “slow” selection among collectives Even before full individuality, the system is already “set for the evolution of conflict,” and one route out is that “solutions… arise from selection operating at different temporal scales.” (Annual Reviews) This is the deep structure behind a lot of the above: you’re trying to get slow, between-collective selection to install constraints that suppress fast, within-collective defection.

If you want, I can translate these into a compact ABM checklist (“if your simulated lineage has X life cycle, expect Y trap; here are the control knobs”) rather than prose.