The claim
The passage. MoM’s v1 “Sheaf Semantics of Context” §11.5, “Identity as Global Section”, says: “Not only words but selves are contextual”. It takes a person’s roles in family, professional and social contexts, and continues: “If these local identities glue coherently, the sheaf has a global section: a stable self. If they do not (contradictory roles, fractured coherence), identity fails to unify.”
As written it is wrong, in the way CLAIM-197 identifies. If the family context and the professional context do not meet, the roles agree on every overlap because there is none, and they glue. A person who is deferential at home and authoritative at work holds no contradiction by that fact.
The defensible version. Roles conflict only on an overlap: an occasion on which audiences from two contexts are present together and the person must present one way to both. The cover is not given. People build and maintain it.
- Segregating audiences keeps the contexts disjoint. Different styles for different groups do no harm while the groups are kept apart.
- Scripting manages the overlaps that must happen. Weddings, “in which context collisions are common, are frequently scripted to make everyone comfortable” (LIT-1001, p. 38).
- Collapse creates overlaps nobody chose. Broadcast and networked media bring the audiences together. Carmichael “was able to maintain distinct styles as long as he was able to segment social groups, he ran into trouble when electronic media collapsed those social groups” (LIT-1001, pp. 36–37; CASE-052).
So a presented self (TERM-045) is a global section relative to a cover the person partly controls. Integrity of presentation can be achieved two ways: by sections that agree where contexts meet, or by keeping contexts from meeting. The second is not deception. boyd makes the point that distinct contexts are how people “properly locate their performances”.
Why this record holds it
It is the synchronic partner of CLAIM-198. There, interests over time conflict only where self-prediction makes an overlap. Here, roles across audiences conflict only where audiences meet. CLAIM-153 supplies the mathematics once overlaps exist: with pairwise overlaps, two contexts always glue, and the obstruction is a loop.
The sheaf picture adds one thing to the sociology. It says where conflict can be: on an overlap. That makes a prediction the sociology states but does not derive. Segregation strategies should target the overlaps, the shared channels and the occasions where audiences meet, and should not target the roles themselves.
What it does not say
- That role conflict is rare. Overlaps are common, and for networked publics boyd describes them as constant.
- Anything about the other senses of “self”. The manuscript’s “identity” is the presented self, sense 7 of TERM-045. A person whose presentations collide is not thereby less a subject, less integrated or less the same over time.
- That contradiction does no harm unseen. A person may feel the contradiction while the audiences are apart, by remembering one audience in front of the other. That is an overlap made by memory, as for the film joke in CASE-048, and the defeat condition is written to allow it.
Note of 2026-10-12: documented cases of context collapse
CASE-064, filed in PR #121, now illustrates this claim. A photograph posted for friends who knew a running joke reached strangers four weeks later, who read it as contempt for the war dead, and Stone lost her job. It is a documented collapse of the kind this claim describes, where CASE-052 is stipulated. CASE-061 is a neighbouring documented case: one broadcast phrase heard as faith by an in-group and as boilerplate by others, with the split measured. It shows a speaker addressing two audiences at once in one wording, not a conflict between roles, so it is not linked here.
Note of 2026-10-15: split under ADR-043
ADR-043’s classification found this claim compound: an interpretive reading of a person’s roles as sections over contexts, which locates role conflict on an overlap, and the empirical sociology of audiences it draws on, from boyd (2008) and CASE-064. The empirical part, that presentations which differ by audience do no harm while the audiences are kept apart, that people keep them apart and script their meetings, and that broadcast and networked media bring them together, is now CLAIM-317, on which this claim rests. Its formal part was already CLAIM-197. This claim keeps the rest, the sheaf reading and its prediction that segregation strategies target the overlaps, and stays interpretive. Its title states the sociology only as what keeps or creates an overlap, so it was kept.