The case
Two teams have been rivals for generations (A2’s example: the Yankees and the Red Sox). Players, coaches, owners and supporters on both sides have all been replaced. New supporters still learn that this game matters more than others, learn the old incidents as the rivalry’s history, and reproduce its chants and its antagonism. A2: “A young fan might learn that a particular opponent matters before understanding why.”
What is held fixed: the two franchises and the competitive stakes. What varies: every person who carries the antagonism.
What it can show
That a relation between two collectives can be continuous while every carrier is replaced, and that present antagonism can be transmitted rather than generated by present conflicts. A2 separated interpersonal antagonism, collective antagonism (membership disposes one to oppose) and institutionalized rivalry (an enduring relation that structures both sides’ practices). Only the third is the case’s point.
A5 §5 set it beside the dancing couple (CASE-046, CLAIM-175) as a contrast, not a further instance. The couple coordinates continuously and may carry a collective variable (CLAIM-172). A rivalry has no obvious contemporaneous dynamical variable. It is held by archives, scheduled encounters and transmitted expectations, which is the time-lagged coupling CLAIM-157 says its block test cannot see (QUESTION-116).
Given its standing, it cannot show that any actual rivalry survived in this way, that every supporter takes part in it, or that it is a system. Berendt and Uhrich’s survey of Bundesliga supporters (LIT-1147, unread) ties rivalry to in-group distinctiveness and collective self-esteem, which is evidence about identity, not about the rivalry as a unit. A5’s discriminating variation: hold the teams and stakes fixed and interrupt transmission of the rivalry’s history; does the antagonism persist?
Note of 2026-10-29: Berendt and Uhrich, read
LIT-1147 has now been read (NOTE-1303). The sentence above calls it unread and judges it “evidence about identity, not about the rivalry as a unit”. That still holds, with one addition. In the survey of 571 Bundesliga supporters, rivalry’s gains for fans (perceived distinctiveness, cohesion, regard from other clubs’ fans) run through perceived reciprocity, the belief that the rival returns the antagonism. They do not run through fans’ own disidentification from the rival (THEORY-1113).
So what individual fans take from a rivalry depends on how they believe the other side sees the relation. That is a perceived two-sidedness, the individual-level shadow of the institutionalized rivalry this case is about, and still not evidence that the rivalry is a unit.
The paper also reports that 164 of 748 intercepted supporters, mostly at clubs without a famous rivalry, named no archrival at all. Not every supporter carries a rivalry.
A5’s discriminating variation is untouched: the study does not interrupt transmission. Talarico and Moore (LIT-1125), on memory of a rivalry game, is now Unavailable. Qi, Giulianotti and Wiltshire (LIT-1131) is Unfetched: a CC BY-NC copy exists at SAGE and this client was refused it.
Also filed unread on 2026-10-29:
- Botzung et al. 2010 (LIT-1734), an open study of rival fans’ memories of one game, in place of Talarico and Moore;
- Kilduff, Elfenbein and Staw (LIT-1755), on rivalry as a relationship rather than a matter of stakes;
- Tyler and Cobbs (LIT-1682), on rivalry as identity threat.
QUESTION-298 asks whether perceived reciprocity is accurate. That is the paired-survey design this case would need to treat the rivalry as a two-sided relation.
Note of 2026-10-29: Kilduff, Elfenbein and Staw, read
Kilduff et al. (LIT-1755) has been read (NOTE-1292). Of the works the record holds, it comes closest to treating a rivalry as a unit. Sportswriters rated how much their team saw each conference opponent as a rival, with players’ and coaches’ ratings agreeing (r = .89). Across 71 NCAA basketball teams, half the variance in those ratings (50.4%) belonged to the particular pair, not to either team. The rivalry was largely returned from the other side (r = .85 net of team effects).
It also bears on inheritance. Rivalry followed the all-time head-to-head record more than recent form, and the all-time record of even results outweighed similarity in status. Current players inherit a relation made by contests they never played in, which is the case’s claim, measured in the team rather than the fans.
It does not settle what the case leaves open. The ratings are current perceptions, so they show a relation that is reproduced, not who reproduces it. A5’s variation, interrupting transmission, is still untried.
Note of 2026-10-29: Botzung et al., read
Botzung et al. (NOTE-1313) is about memory of one game, not about transmission, but its design is the case in miniature. Twenty-three Duke and UNC fans learned a 2001 rivalry game none of them knew, from archive footage and among fellow fans. A day later their side set the valence of each remembered shot (88.9% of ratings), and they recalled their own side’s successes more accurately and with more confidence. As an inference, a rivalry learned second-hand is stored as one record with opposite signs, each side’s version tilted to its own successes. A5’s discriminating variation, interrupting transmission, is still untried.