Demonstration of dependencies of references. Within b.c we find the reference to x; we resolve that to a vertex which contains conjuncts from line 1 and and line 4. We spot that line 4 is in the same (or ancestor) lexical scope as b.c, and so we ensure that if line 5 survives, then line 4 must too, otherwise c: x will reference an unknown field. See also 53, 54, and 55. -- a.cue -- a: x: 5 b: { a x: int c: x } -- out/trim -- == a.cue a: x: 5 b: { a x: int c: x }