nucleation

LIT-716: Network Epistemology: Communication in Epistemic Communities

2 min read

Properties4
StatusActive — read 2026-10-05 (NOTE-567): read in full from the author's posted manuscript. Worth reading as the short survey of network epistemology up to 2012. Its main lesson is that the best communication structure depends on the learning problem. Pooling estimates favours dense, regular networks. Costly transmission favours sparse ones, often a star. Active learning in bandit problems favours minimally connected ones. Intuitions about "too little" or "too centralized" communication are often wrong. A survey, weighed as one.
Tagsepistemology, network-science, philosophy-of-science
Cited as “paper” byNOTE-567 — Network Epistemology: Communication in Epistemic Communities
Cited as “source” byTHEORY-422 — In the standard formal models of group learning, no communication network is best for every epistemic problem: costly transmission favours minimally connected or star networks, pooling of estimates by averaging favours regular networks, and learning by trying in bandit problems can favour sparse ones, so a group's network cannot be judged until its learning problem is known

Graph View

Kevin J. S. Zollman (2013), Philosophy Compass 8(1):15–27 — DOI-10.1111/j.1747-9991.2012.00534.x

Key takeaways

  • Two lessons (§1): “the best way for group members to communicate with each other depends critically on the epistemic problem”, and “our intuitions can often be poor guides”.
  • Information transmission (§2). When information passes second-hand without loss and links are costly, minimally connected networks are optimal: the directed cycle in one model, various minimal trees in others. With even slight loss over distance, stars become optimal. Unequal centrality is therefore no evidence of distortion. Optimal networks are often stable, but the empty network can be too, so groups need not find the optimum on their own.
  • Pooling (§3.1). When agents repeatedly average with neighbours (DeGroot), regular networks are best in the long run, and denser ones get there faster. In the bounded-confidence version the complete graph remains optimal.
  • Active learning (§4). With a “truth pull” on at least one agent and a path to it, any connected network converges. In bandit problems, where evidence depends on what agents choose, minimally connected networks were best in small simulated groups “even when communication is free” (LIT-724, LIT-735).

Standing in the record

Filed on 2026-10-05 at the owner’s request. Of the three Zollman papers it is the one that best backs the essay’s sentence as written: research on epistemic networks shows the effect of communication depends on the problem, not that less is better. It predates LIT-733, which narrows the bandit result further.

Backlinks