Most decision advice begins after the candidate moves already exist: compare the options, score the tradeoffs, choose the strongest one.
Perspective Expansion begins earlier. Its question is whether the current representation of the problem is hiding actions before evaluation even starts.
The method separates the desired state from the assumed method. Wanting evidence that work is reaching people is not identical to needing more streams. Wanting income is not identical to needing one specific job. Wanting proximity to an industry is not identical to living in one particular city.
The sequence
The published method uses: Desired State → Assumed Method → Representation Check → Reframe → Expanded Action Set → Evaluate → Choose.
The point of the reframe is not to generate random alternatives. New actions still have to be materially relevant to the desired state. The move is to remove unnecessary constraints that entered through interpretation rather than reality.
Useful reframing operations include changing scale, changing time horizon, separating outcome from method, examining a constraint as a variable, changing observer perspective, or asking what becomes possible if one assumed mechanism disappears.
Why it belongs upstream
If a person can materially take ten useful actions but recognizes only three, a perfect ranking system applied to those three is still solving the wrong problem. Perspective Expansion treats candidate discovery as part of reasoning rather than a prelude to it.
That makes it naturally complementary to Path Expansion. First expand the represented action set. Then ask which visible action creates the strongest future possibility space.
The compact instruction survives the longer framework: change the model before choosing the move.