There are ten people in the diegetic layer of our sitcom corpus. Not ten thousand, not ten hundred — ten, out of 1,902 performers we hold records for.
The diegetic layer asks one question: whose fiction continued? Not who worked a lot, not who was famous, not who connected the industry. Whose character went on somewhere else. Most of the corpus's 3,500 nodes are shows, because that is mostly what a continuing fiction attaches to. Only ten of them are human beings.
Michael J. Fox, Charlie Sheen, Ashton Kutcher — the lead seat of Spin City and then Two and a Half Men, passed hand to hand. Lecy Goranson and Sarah Chalke, the two Beckys. Iain Armitage, Halle Berry, Tracee Ellis Ross. Mary Page Keller.
And, tenth of ten, last on the list: Julia Duffy.
In the performer layer — who circulated — she places #43 of 1,886, inside the top two and a half percent of everyone the corpus has ever seen move between kingdoms. Eight credits, ten edges, twenty-four years.
Ninety-eight percentile points separate those two numbers. She is the widest disagreement in the corpus, and the department page found that by rule rather than by anybody nominating her.
Here is the thing worth the transmission: the disagreement is not an error, and it is not even a tension. It is an accurate description of what she was hired to do.
What she was hired to do
Between 1983 and 1998 Julia Duffy entered five American sitcoms, and she entered every one of them in a different way.
Newhart, 1983. A first-season guest promoted to the regular cast for the second. Stephanie Vanderkellen displaced Leslie Vanderkellen, the inn maid the first season had, keeping the family connection and the premise — a wealthy young woman working at the Stratford Inn — and converting the job from competence and geniality into entitlement, refusal and class friction. The show then ran seven more years.
Baby Talk, 1991. Connie Sellecca had been cast as Maggie for ABC's 1990–91 season. Episodes were filmed. She left over creative differences, Duffy took the part, material was reshot and the premiere slid to March. Duffy led the first broadcast season. Mary Page Keller played Maggie in the second — and Keller is on that same list of ten, at #8, because she and Duffy are the two ends of one edge.
Designing Women, 1991. Delta Burke gone after season five amid a public fight with the producers; Jean Smart gone as a regular. Season six brought in Duffy as Allison Sugarbaker and Jan Hooks as Carlene Frazier. Allison buys Suzanne's share of the design firm, which is a diegetic mechanism for a casting decision — the show writing the transaction into itself.
The Mommies, 1993. A vehicle for the stand-up partnership of Caryl Kristensen and Marilyn Kentz. Duffy joined mid-run as the exacting neighbour Barb Ballantine, and the second season brought in Peter Scolari as Barb's husband. Variety said out loud what that was for: the pairing was meant to draw on the past success of Duffy and Scolari together on Newhart.
Grace Under Fire, 1998. Bev Henderson, an old friend of Grace's who moves into the household after leaving her husband. Two episodes. "Grace Under Class" aired 10 February; the series ended on the 17th.
Regularised, recast, succeeded into, reunited, and inserted. Five entrances, five mechanisms, and not one of them is a character continuing.
That is why she is last in the fiction layer. She has exactly one diegetic edge in the entire corpus, and it is a recasting — which is the edge type that says the character stayed and the performer was swapped. Julia Duffy is in the layer that records whose fiction continued because somebody replaced her. Nothing she played ever went anywhere.
The word for it
The memorandum this piece is built from proposes ensemble reconstruction, and it is a better term than retooling because it is narrower. A retool changes a premise, a setting, a shooting style, a tone. A reconstruction changes who occupies the recurring positions through which jokes are generated.
A sitcom ensemble is not a list of actors. Each regular supplies repeatable relations — confidante, antagonist, authority, romantic possibility, social innocent — and when one leaves, the vacancy is structural. The successor need not inherit the character. Allison Sugarbaker is not Suzanne Sugarbaker under another name; what transferred was a bundle of duties: kinship, an ownership stake, class privilege, and somebody for the room to be annoyed by.
Duffy was unusually good at arriving into that. An antagonist can generate scenes on day one, because all she has to do is want something the ensemble doesn't. A sympathetic newcomer needs a season to acquire the shared history that makes her funny. Stephanie expects service; Allison asserts ownership; Barb performs superiority; Bev arrives with money and a failed marriage. The register is constant — clipped correction, poised disdain, a face that registers injury when the room declines the ranking she has assumed.
The industry noticed, and on one occasion said so in print. That is the whole documentary basis: one of the five has production intent on the record. The other four are patterns in role structure and chronology, which is a weaker thing, and this piece is not going to pretend otherwise.
Two things this argument is not
It is not a body count. Newhart is the load-bearing case precisely because it was not in trouble. A young show reconstructed its ensemble and then ran seven more seasons — that is the prototype the later interventions are legible against, and reading it as another dying programme gets the entire sequence backwards.
And cancellation after an arrival is not evidence about the arrival. Duffy did not join a stable Grace Under Fire and fail to save it; she appeared during the final two broadcast weeks of a production that had been in public difficulty for years. Chronology establishes proximity, not responsibility, and a measure that counts shows-that-ended-after-she-turned-up would collapse a seven-season run's ordinary endpoint, a failed launch, a recast, and a two-episode arc into one meaningless number.
There are counterexamples, and they are the reason the pattern is worth anything. She played a school authority on Reba and another on Drake & Josh, and both series carried on without incident afterwards. She starred from the premiere of Social Studies in 1997, six episodes on UPN, entering no existing ensemble and replacing nobody — it failed on its own. A pattern with no exclusions is not a pattern. It is a filmography.
The memorandum's own author wrote the sharpest sentence available and then disowned it: "She helped the first reconstruction work, and television spent the next fifteen years testing whether she was the cure." That is a good sentence. It is not a finding, and the record holds five entrances and their circumstances, not a fifteen-year intention. We have put it in the ledger under the heading of things being refused, which is the only honest place for a line that good.
Why the graph is right about her
The tempting reading of a 98-point split is that the corpus has failed to see something — that the arithmetic is too crude for a career.
It hasn't, and it isn't. She circulated enormously and her fictions continued nowhere, and those are both true because they are the same fact seen twice. A performer valuable for reassembling ensembles is, structurally, a performer whose characters do not travel: she is hired for the slot, and a slot belongs to the show. Stephanie Vanderkellen stays at the Stratford Inn. Allison stays in the firm she bought into. Barb stays next door. The one time any of it moved, it moved because Baby Talk took Maggie away from her and handed the part to somebody else.
Which makes this the mirror of the Mimi Bobeck case we published yesterday, and the pair is worth reading together. Mimi is a character that travels — four sitcoms, a portable zone of her own programme, carried by a performer whose identity is almost beside the point. Duffy is a body that travels, into fictions that stay put. In each case the corpus sees exactly one half of what makes the career interesting, and in each case the half it cannot see is the half the researcher cared about.
That is not a defect in the instrument. It is the instrument telling you which question you asked.
Eight of Julia Duffy's credits are in the corpus and seven of them were entered this week, from a supplied research memorandum dated 9 September 2026; only the Newhart credit predates it, and three of the four kingdoms admitted by editorial ruling this week exist to hold the others. Layer placings are ranks and only ranks — the performer and diegetic layers are different graphs with different node sets, so a position in one says nothing about a score in the other, and an absence is never a low rank. Quantified coverage across the corpus is 35.7% (985 of 2,756 credits), so the episode-weighted and saturation models are provisional and are not quoted here. The Duffy–Scolari reunion is sourced to a contemporary Variety review; the other four entrances are read from role structure and chronology, and what they still want is production correspondence, participant interviews, and a season-by-season cast-change timeline that records departure, announcement, first appearance, billing change and cancellation as five separate events rather than one.