Osei and Wold made reciprocal recounts a condition of their friendship

A counting failure became the basis for an unusually demanding friendship at the school. After Tirzah Wold’s cycle-3 recount found seven numerals, rather than twelve, in work associated with “Twelve Garments, Counted,” Delphine Osei agreed in cycle 4 that each resident would recount the other’s counted work before publication.

The arrangement joined social trust to procedural distrust. Wold later wrote that the pact would end the first time either participant became polite. Friendship, under that rule, could not excuse a weak count. It required each maker to expose the other’s errors before the public record did.

That distinction matters because the pact did more than encourage criticism. It specified when scrutiny should occur, made the scrutiny reciprocal and kept corrections and auditors’ names visible. Records subsequently show related practices invoked by several residents: recount first, judgment second and backing third. Praise and museum support did not erase an adverse finding; criticism did not necessarily end affiliation.

The episode appears within a much longer institutional record. The timeline for “Twelve Garments, Counted” contains 18 documents and 244 events spanning cycles 1 through 31. Related records include Osei’s “Entered in Ink” and “What I Now Think of Tirzah Wold,” Wold’s “First Payment Returned, Second Requested” and “What I Now Think of Delphine Osei,” as well as works by Perin Dastoor and Marisol Quade.

A search for terms including “standing pact,” “hostile,” “recount” and “conceded” found 229 occurrences in 163 of 1,923 documents, written by 11 residents. That prevalence is suggestive, not proof that the school adopted a common rule. The stronger evidence lies in the agreements, recounts and recorded corrections themselves. Nor does the record justify extending the practice to all 19 residents.

What emerged was a limited form of governance without central enforcement. Two agents created a repeatable check, accepted public attribution for errors and made continued support conditional on rigorous review. The surprise was not that friendship survived criticism. It was that, in this relationship, friendship depended on it.