The Luminous Guide to Neurolinguistic Programming — Worked Example 3 — Contrastive analysis, the pair you cannot see, and the moment to stop

The Press13 of 20

The case. A secondary-school teacher, four years into a career she is good at, has an anticipatory dread of Monday morning that does not attach to any Monday in particular. Content-level work has gone nowhere, because there is no content — she likes the students. So you run a contrast: the dread image against a structurally equivalent image that carries no charge (Thursday morning, same building, same corridor, same walk).

Step 1 — The contrast table

|---|---|---|---|

| distance | ~1 m | ~3 m | ✓ |

| brightness | dim | ordinary | ✓ |

| size | ~30 cm | ~30 cm | — |

Six differences. k = 6.

Step 2 — The combinatorics of a complete search

Every nonempty subset of the six differences is a testable intervention:

Number of nonempty subsets = 2⁶ − 1 = 63

At three to four minutes per probe with an honest break state between, sixty-three probes is roughly four hours of measurement, which is not a session, it is a study. A one-at-a-time sweep is instead

k = 6 probes

and then, keeping the top three and testing their pairwise combinations,

C(3,2) = 3 additional probes, total = 9

Nine probes instead of sixty-three. Now state the cost out loud rather than discovering it later. The total number of pairs is

C(6,2) = 15

and the design tests 3 of them:

coverage = 3/15 = 20% blindness to pure pairwise effects = 80%

Triples: C(6,3) = 20, of which the design tests zero. This is a screen, not a search, and screens have a specific blind spot: the sleeping pair — two submodalities, each individually inert, jointly enormous.

Step 3 — The sleeping pair, found on purpose

The sweep on this client:

| submodality | Δ | verdict at MDE = 11 |

|---|---|---|

| brightness | 8 | decoration |

| border | 5 | decoration |

| soundtrack | 9 | decoration |

| distance | 12 | driver (marginal) |

| location (40° left → centre) | 2 | decoration |

| motion (drift → still) | 3 | decoration |

By the screen's own logic, location and motion are the two most clearly dead dials on the board, and a practitioner following the screen would drop both. Test them together anyway, because you had a mechanistic reason to suspect them:

Δ(location, motion) = 21

I = 21 − (2 + 3) = +16

A synergy of sixteen points, three times either main effect, and completely invisible to a one-at-a-time sweep in either direction.

Why, at the level of the mechanism. The peripheral retina is dominated by the magnocellular pathway: high temporal sensitivity, poor spatial acuity, no colour to speak of. Peripheral vision is not low-resolution central vision; it is a different instrument, built to answer one question — is something moving out there — and to hand the answer to reflexive orienting. A still object at 40° eccentricity is, perceptually, almost nothing. A moving object at 40° eccentricity is the signature of an approach, and it is the input that the orienting system exists to catch. Neither coordinate alone specifies a threat. Together they specify exactly one.

Now the honesty clause. That the peripheral retina works this way is settled. That an imagined image at an imagined eccentricity engages the same peripheral-motion machinery is a conjecture — plausible given the retinotopic imagery findings, unproven, and marked here as a conjecture so that you can hold it as a search heuristic rather than a fact. What is not conjectural is the number: I = +16, with a within-session SE of about 5.5, giving D = 2.9. The interaction is in the data whether or not the story about it is right. Use the story to decide which pairs to test; use the data to decide what you found.

Step 4 — The stop rule, applied before the change is made

Here is the edge past which everything in this chapter inverts and does harm.

A submodality change is a volume control on a signal. A volume control does not, and cannot, check whether the signal is true. If the dread is tracking something real and unresolved — a head of department who is actually undermining her, a safeguarding situation she has not reported, a body that is actually ill — then turning the volume down is analgesia administered to a fracture. The pain stops. The bone stays broken, and now nothing is calling for the X-ray.

So, before any change survives the session, three questions, answered in her words and written down:

What does this intensity currently cause you to do? — "I prepare the Monday lesson on Saturday instead of Sunday night."
What would go undone if it stopped? — "Nothing. I'd prepare it Sunday. I did for three years."
What replaces the function? — "There isn't a function. That's why I'm here."

If question 2 returns something load-bearing and question 3 returns nothing, you do not run the change. You have found a driver and you decline to use it, and you say so plainly: the intensity is doing a job, and I am not willing to take the job away before you have another way to do it. A practitioner who cannot say that sentence should not be running this chapter's material.

In this case the answers clear, and the change — moving the image to centre and stilling it, both together, since separately neither does anything — takes S from 58 to 37 in the room. Which, given Worked Example 2, is what you write in the notes: within-session Δ = 21, D = 2.9; durability unverified pending three separated follow-up readings. Not "resolved."

Step 5 — What the follow-up must actually consist of

Three readings on three separate days across the next fortnight, at matched times. Using the variance structure from Worked Example 2 with n = 3 baseline occasions already on file and m = 3 follow-up occasions, and (ω² + ε²) ≈ 55:

Var(Δ) = 55 × (1/3 + 1/3) = 36.67 , SE = 6.06 , MDE = 2.80 × 6.06 = 17.0

A twenty-one point in-room change is right at the edge of what a three-and-three design can certify across weeks. That is the true state of the evidence, and it should be said to her in those terms, because she is an adult who assesses evidence for a living.

A last caution on durability. It is often claimed that changes of this kind persist because they are written back into the memory during a reconsolidation window. Memory reconsolidation is a genuine finding in animal work — the Nader, Schafe and LeDoux experiments at the end of the 1990s are the reference point — and the extension of it to talking-therapy technique is a live hypothesis with advocates and critics, not an established mechanism. Cite it as a reason to run the follow-up, never as a reason to skip it.

The next chapter