RQ4 · structural recurrence

Cross-saga comparison

Only Njáls saga has a hand-annotated evaluation; the other three run the same bootstrap entity table and rule pipeline without curator review, so treat their absolute numbers as provisional (see Methods). What is worth comparing is whether the same structural shape recurs.

Basic network statistics

SagaNodesEdgesDensityAvg. clusteringCommunitiesModularityAssortativityBalance ratio
Njáls saga (Brennu-Njáls saga)1603820.0300.235100.553-0.15271.4%
Laxdæla saga1212270.0310.21190.495-0.15993.8%
Grettis saga Ásmundarsonar1192130.0300.154140.466-0.16575.0%
Egils saga Skallagrímssonar951650.0370.148110.524-0.12790.0%

Does the shape recur?

Degree assortativity (all four negative → hub-and-spoke, not clique-of-hubs)

njala-0.15laxdaela-0.16grettla-0.16egla-0.13

Structural balance ratio

njala0.71laxdaela0.94grettla0.75egla0.90

RQ5 · centrality vs. narrative prominence, by saga

Sagamentions ~ degreementions ~ betweennessmentions ~ pagerankdialogue ~ pagerank
Njáls saga (Brennu-Njáls saga)0.7870.5880.7930.62
Laxdæla saga0.7860.6780.6750.721
Grettis saga Ásmundarsonar0.6470.5730.560.504
Egils saga Skallagrímssonar0.6740.5890.64

Spearman ρ between raw mention/dialogue counts and network centrality. Consistently strong positive correlation across all four sagas is itself informative: it suggests structural centrality and textual prominence are not independent measurements of “importance” here, largely because both are downstream of how often a character is written about at all.