04
Lateral Channel, Deep Network, Three Defence Layers
The previous two chapters covered what is protected and which patterns trigger the protecting. This chapter covers the route: which signals run through which channel, why only one channel is readable, and where the pressure goes when you block a layer.
Two processing channels
You see a nail on the floor. Your foot avoids it before you can think “that is a nail”. That is one channel.
You sit down to solve a maths problem. You have time, you try several routes, you correct yourself. That is the other channel.
| Lateral processing | Deep-network processing | |
|---|---|---|
| Mechanism | Matching against stored patterns | Analysis by intelligence |
| Time | Close to zero — reflex | Unlimited |
| Can it be faked? | No, or only with great difficulty | Yes |
| Requires | The experience network | Knowledge plus time |
| Example | Avoiding a nail, startling, recoiling, a frown lasting a quarter of a second | Solving a maths problem, preparing a statement, constructing a plausible story |
The foundational theorem of the whole system
You can read only what runs through the lateral channel. Everything that goes through the deep-network channel can be constructed — and with enough preparation time, constructed very well.
This explains three things at once, and all three are pillars of this book:
- The true leakage zone is the fast-reflex zone, far from consciousness.
- The false leakage zone is speech, deliberate facial expression, rehearsed posture.
- The counter-strategy of the modern approach is to increase cognitive load — make the deep-network channel do too many things at once, until it no longer has the resources to manage the other one. Chapter 25 is devoted to this.
The leakage ladder — further from the brain, truer
In the body, awareness of each part falls off with distance from the head. This is not a metaphor; it is a consistently replicated observation.
The further a body part is from the brain, the less aware we are of what it is doing. After the face, we are aware of what the arms and hands are doing, then the chest and abdomen. We are least aware of our legs, and we all but forget our feet.
An experiment that settled the argument
In a series of staged interviews, managers were asked to lie convincingly. The result, regardless of sex: every one of them significantly increased the number of unconscious foot movements — while still keeping good control of the face and hands.
A related result: observers detect lies markedly better when they can see the whole body than when they can see only the upper half.
Consequences for spatial design — three independent sources pointing the same way
“If you are not sure whether someone is deceiving you, look under the table.”
“The ideal interview setting places the interviewee in a position with no obstruction — such as a table or a desk — blocking the view of the whole body.”
And a classic practical recommendation: a chair without arms, placed in an open area.
In other words: the table is the most expensive thing in your room, not because of what it cost but because it hides the most honest region of the person opposite you. Chapter 23 goes into removing it in detail — and into why ninety-five per cent of refusals are delivered while people are sitting behind a desk.
Three defence layers
This is the most important part of the chapter, and the organising frame for both Book Two and Book Three. People defend in three concentric layers, activated from the outside in, always using the cheapest layer first.
The layer-shift rule, running in real life
In a crowded lift
Layer 1 blocked — no retreat possible. The result: people fold their arms, turn to face the floor indicator, hold a briefcase to the chest. The pressure drops to layer 2.
In the interview room
A fixed chair with no arms, a glass table, a closed door. Layers 1 and 2 are both blocked. The result: all the pressure is forced down into layer 3 — fidgeting rises sharply, ankles lock, constant pacifying.
This is the mechanism behind every interview-room design you have ever seen. It works. The problem is that it works in two directions at once, and few people mention the second.
Two faces of the same design
The good face. Block layers 1 and 2 and you force the subject to leak at layer 3 — the layer they control least. You see more.
The bad face. It also produces inhibition, and inhibition lowers the quality of the information you get. You see more of a person who is saying less, and saying it worse.
The right balance
There is a fairly sharp line between two kinds of “blocking”, and it lets you take the good face without the bad:
Should be blocked — physical barriers
- The table between the two of you
- Chair arms that make folding your arms comfortable
- Enough objects on the table to build a wall
These block only your view. Removing them takes nothing away from the other person.
Must not be blocked — autonomy
- The right to leave
- The right to choose where to sit
- The right not to answer a question
- The right to change the subject
Blocking these trades the quality of your information for compliance. That is a bad bargain.
Fitting three chapters together
By this point Book One has built three quarters of the foundation, and the pieces fit together like this:
- There is one objective function
The body protects a ranked list of vulnerabilities. Chapter 02.
- The protecting is triggered by pattern matching
Patterns come from two different libraries, and one of them produces a great many false alarms. Chapter 03.
- Reactions run through two channels, and only one is readable
And it is unevenly distributed over the body: the further from the brain, the truer. This chapter.
- The protecting escalates through three layers, and blocking one drives it inward
Giving you both a reading map and a set of controls. This chapter.
- Still missing: when everything above stops working
That is chapter 05, and it decides whether you read correctly or read at random.
Chapter exercises
One. Over one day, note three situations where you see someone in a “layer 1 blocked” state — a lift, a queue, a bus, a cramped meeting room. For each one, find the matching layer 2 behaviour.
Two. Look again at the space where you usually meet people. How many free barriers are already sitting there? How many of them can you move?
Three. Observe yourself: when you are tense in a meeting you cannot leave, what does your layer 3 do? There is almost certainly a repeated behaviour you have never noticed. Find it.
Chapter summary
- Two processing channels. Lateral: matching against stored patterns, time close to zero, cannot be faked or very nearly so. Deep network: analysis by intelligence, unlimited time, can be faked.
- The foundational theorem of the whole system: you can read only what runs through the lateral channel. Everything that goes through the deep-network channel can be constructed — and with enough preparation time, constructed very well.
- The false leakage zone is speech, deliberate facial expression, rehearsed posture. The counter-strategy of the modern approach is to increase cognitive load: make the deep-network channel do too many things at once until it no longer has the resources to manage the other one.
- The leakage ladder — the further from the brain, the truer. Awareness falls off from the face, to the arms and hands, to the chest and abdomen, to the legs, and is lowest at the feet, which are all but forgotten. The face is the most managed and therefore the best liar.
- The experiment that settled the argument: managers were asked to lie convincingly. Regardless of sex, every one of them significantly increased the number of unconscious foot movements — while still keeping good control of the face and hands.
- Observers detect lies markedly better when they can see the whole body than when they can see only the upper half. If you are not sure, look under the table.
- The table is the most expensive thing in your room. It hides the most honest region of the person opposite, gives them a free barrier, and signals hierarchy — ninety-five per cent of refusals are delivered while people are sitting behind a desk.
- Three concentric defence layers, always using the cheapest first. Layer 1, distance, fails when retreat is impossible · layer 2, the body axis, fails under observation or when it becomes impolite · layer 3, facial features and micro-movements, the layer that does NOT fail but only escalates, and therefore the most reliable layer.
- Block a layer and the pressure does not vanish — it drops to the layer inside. A crowded lift blocks layer 1: people fold their arms, turn to face the floor indicator, hold a briefcase to the chest. An interview room with a fixed armless chair, a glass table and a closed door blocks both outer layers: fidgeting rises sharply, ankles lock, constant pacifying.
- That design works in two directions at once, and few people mention the second: you see more of a person who is saying less, and saying it worse.
- There is a fairly sharp line between the two kinds of “blocking”. Physical barriers should be blocked — the table, chair arms, objects enough to build a wall — because they block only your view, and removing them takes nothing away from the other person. Autonomy must not be blocked: the right to leave, the right to choose where to sit, the right not to answer a question, the right to change the subject. Blocking those trades the quality of your information for compliance.