C.28.CM:11 - SoTA-Echoing

The comparison concerns usable construction and criticism for a bounded causal question. It does not rank causal-discovery algorithms or claim a universally best empirical workflow.

Practice question and selected move	Alternative, trade-off and concrete use	Source role, limits and reopening
How can a working account become explicit causal assumptions? Adapt extraction, translation and critical integration of subject claims.	Directly drawing the preferred story is cheaper but can hide missing or conflicting premises. :4.2–4.3 recover meanings and alternatives; a full evidence synthesis is reserved for questions that need it.	Ferguson et al., 2020, Table 1 and Discussion supplies a developed construction comparator. Its integration heuristics do not establish causal truth. Reopen when this extraction loses a subject mechanism or a simpler adequate construction is available.
Which conditioning changes the causal question or opens a misleading path? Adopt whole-graph path inspection; reject deciding from a variable’s name or one three-node picture.	Classifying a few familiar motifs is quicker but can miss another open path or conditioning through selection. :4.4 works the entire small graph and :5.3 preserves assignment and selection.	Geiger, Verma and Pearl, 1990 gives the formal separation basis; Cinelli, Forney and Pearl, 2022 supplies the current practice comparison for controls. These are conditional graphical results, not tests of omitted subject premises. Change the rule when the graph class or estimand changes.
What if the graph itself is uncertain? Adapt comparison of the causal query over admissible structures.	Selecting one convenient graph simplifies calculation but can suppress a result-changing rival. :4.5 retains the remaining family and states its shared or differing consequences.	Padh et al., 2025, §§2 and 6 supplies a computational line with explicit structural uncertainty; its procedure assumes no hidden confounding and has optimization limitations. Peters et al., 2014 shows how additional model assumptions can identify structure. Neither licenses assumption-free discovery. Reopen when a qualified result excludes a live rival or reveals another.
How should feedback affect construction? Adopt temporal distinction where it answers the query; otherwise request a suitable cyclic model.	Forcing an equilibrium loop into a DAG can delete the operative mechanism. :4.4 and :5.2 retain the horizon and the controller dependency.	Bongers et al., 2021 supplies existence and interpretation conditions for cyclic structural models. Its theory makes a cyclic alternative available under conditions, not automatically solvable. Reopen when the chosen temporal resolution or equilibrium premise changes.
Is specialized construction needed at all? Retain direct use of sufficient general modeling or a supplied mechanism model.	B.5.FM can construct the thermal limiting argument without a causal-model family; C.28.MR can transform supplied equations directly. This pattern adds work only when the causal account or a material alternative is missing.	These internal alternatives determine :1 and :4.7. In :5.2 the additional work distinguishes supply, display and feedback before choosing a transformation. The examples establish conditional reasoning, not superior field performance. Prefer the simpler route when it supplies that distinction already.