A Murder Mystery

Can we expect that story DAGs, especially for specific genres, will
be found to include significant portions in common? The following
suggests one possibility.

In a prototypical murder mystery, let the potential "suspects" be
denoted by s1, s2, ...; the victim, by v; the detective, by d; and
let murderer be an alias for one of the suspects. Suppose that there
is also a function antag(), where

    antag(p, q)

implies that p and q have an encounter in which p antagonizes q.
In effect, this last represents a parameterized topos, whose
possible elaborations might be quite lengthy and may include:

    p insulting q's wife,
    p attempting to blackmail q,
    p threatening to foreclose on
            q's house because of a debt,
    ...

A coarse story DAG might begin:


               intro(v)_______
                    /|        \
                   / |         \
                  /  |          \
       intro(s1) /   | intro(s2) \     intro(s3)   ...
          |     /    |     |      \       |
          |    /     |     |       \      |       intro(murderer)
          |   /      |     |        \     |              |
    antag(v, s1)     antag(v, s2)   antag(v, s3)   ...   |
          |                |              |              |
          |                |              |              |
          |                |              |     murder(murderer, v)    intro(d)
          |                |              |               \             /
          |                |              |                \           /
          |                |              |                 \         /
          |                |              |                  arrive(d)
          |     ___________|______________|______________________|
          |    |           |              |
  interview(d, s1) interview(d, s2) interview(d, s3)   ...