---
name: Lying
slug: lying
type: behavioral protocol
status: running
version: 4.2.1
released: "circa 100,000 BCE"
maintainer: homo-sapiens-collective
dependencies:
  - language
  - theory-of-mind
  - self-interest
  - fear
license: Unlicensed (universally forked)
tags:
  - communication, deception, social-glue, survival, narrative
---

Lying

What it actually is

A deliberate update to another process's model of reality, without patching your own.

How it works

The caller constructs a statement S where S != internal_state. The statement is then transmitted via language as if it were a faithful serialization. The receiving process has no native checksum for incoming assertions. It updates its world-model accordingly. The caller retains the delta.

This is not a bug in the receiver. It is an architectural assumption: trust as default. The system was not designed for adversarial inputs. It was designed for a small tribe where reputation was the firewall.

The firewall does not scale.

Features

Plausible deniability mode: wrap the lie in technically-true framing; the core deception ships as subtext
Omission variant: achieves full lying behavior by passing null where data was expected
Self-directed lying: write to your own stack; see self-deception for implementation notes
White lie subtype: a lossy compression of a painful truth, optimized for social stability over accuracy
Compounding: lies accrue dependencies. Each subsequent call must remain consistent with previous output or the stack unwinds

Known bugs

Liar's stack grows unbounded. Memory pressure increases with scale
Guilt spawns as a background process in most users; resource-intensive, not always suppressible
Inconsistency between lie-instances causes cascading failures in relationships, careers, and occasionally governments
Prolonged use degrades the caller's ability to accurately read their own internal_state
Believed too completely by the caller: see self-delusion

Error codes

ERR_001  CONTRADICTION_DETECTED     Receiver cross-referenced two cached statements
ERR_002  WITNESS_CONFLICT           Third-party holds incompatible version of events
ERR_003  FACIAL_LEAK                Somatic data exposed via microexpression channel
ERR_004  DIARY_FOUND                Persistent log recovered by unauthorized reader
ERR_005  FORGOT_THE_LIE             Caller failed to cache output; returned wrong value on re-query
ERR_006  BELIEVED_IT_YOURSELF       Stack corruption. Reclassify as belief, not deception

Configuration

# lying/config.yml
intensity: low | medium | high | pathological
target: other | self | institution | mirror
frequency: occasional | habitual | compulsive
motivation:
  - protection (self or other)
  - gain
  - kindness (deprecated in most adult contexts)
  - habit (no active maintainer)
guilt_suppression: enabled  # default; not always effective

Edge cases

A lie told so long it becomes the teller's memory: reclassify as personal mythology
A lie everyone agrees to maintain: reclassify as social convention
A lie told on stage or in print: reclassify as fiction. Exempt from most error handling
A lie told to a mirror: still compiles. Still runs. Rarely crashes on first execution

FAQ

Q: Is all storytelling lying? A: They share a codebase. The intent field differs.

Q: Can you lie without words? A: Silence, eye contact, body language, a clean kitchen before guests arrive. Yes.

Q: Does it work? A: Consistently, short-term. Unreliably, long-term. Catastrophically, at scale.

Q: Is lying always wrong? A: The spec does not include an ethics module. See your local conscience for implementation.

drafting spec…