The second half of the book claims that breathing can run the other way: not only turn fear down, but bring something warm up. The book labels that claim craft, because there is no trial of the Trap Door. The closest thing in the literature to a test of the principle is an older experiment in emotion psychology.
In the first study, people were asked to generate joy, anger, fear or sadness and then describe the breathing that went with it. Their descriptions matched what laboratories had recorded during real emotion, were consistent across people, and differed clearly between the four emotions. In the second study, new participants were given breathing instructions built from those descriptions, under a cover story so they could not guess the point. The breathing alone produced the matching feeling states.
That is the mechanism the book's practice leans on, demonstrated once in a lab with self-report as the outcome. It does not show that any particular practice produces a particular emotion reliably over time, which is why the Trap Door stays labeled craft.
The studies
Philippot P, Chapelle G, Blairy S. Respiratory feedback in the generation of emotion. Cognition & Emotion 16(5), 2002, 605–627. Read the abstract
- Sample
- Two studies
- Design
- Study 1: self-described breathing patterns for four emotions. Study 2: breathing instructions derived from Study 1 under a cover story.
- Found
- Patterns were consistent across people and distinct by emotion; following them produced significantly differentiated emotional feeling states.
- Limits
- Sample sizes are not given in the abstract; outcomes are self-reported feelings; one lab.
Where this lands in the book
Techniques
Rooms
Two rules, no exceptions
Never practice any technique on this site while you are driving, operating machinery, or doing anything that needs your full attention. The slow ones can make you drowsy, faster than you’d expect. And never practice a breath-hold, bellows breath, or any vigorous breathing in or near water. Read Before you begin for the doctor’s list and the rest.
Verdict scale: Science means trials or meta-analyses with consistent findings; Promising means real trials that are small, few, or unreplicated; Mixed means the studies point both ways; Mechanism means it shows how something works in the body, not that it helps in life. Checked against the published abstracts in October 2026. Not medical advice.