Stanton Samenow, the criminologist who spent decades interviewing incarcerated offenders, compresses the framework's foundational premise into one image:1 [POPULAR SOURCE]
It is impossible to commit a crime that is out of character. It would be like asking a building to fly; it is not within the building's nature to do so.1
The building cannot fly. The flying is not just unlikely — it is not within the nature of the thing. Samenow's claim, applied to violent action, is structurally identical: the crime of passion that appears to come from nowhere does not actually come from nowhere. The actor's history contains the seeds. The seeds may have been hushed up, disregarded, or unrecognized at the time, but the seeds are there.
Lieberman extends Samenow's compression with the elaboration:1
Blustery, inflexible, and impatient, each demand that others do what he wants. They may flare up at even minor slights. Instead of coping constructively with unpleasant situations, they compound their problems. When frustrated or disappointed they blame others... The "out of character" crime may be preceded by a long series of threats or assaults that were hushed up or disregarded by the family. Despite appearances, when the homicide is finally committed, it is by a person to whom violence is no stranger.1
The implication is operationally important: violent action is predictable in retrospect. The retrospective predictability implies prospective observability — the same patterns that explain the violence after it has occurred would have been observable before, if the right framework had been applied. Lieberman's compressed framework provides the prospective observation instrument.
People don't just snap. There are almost always identifiable behaviors that will allow you to know when violence may be looming.1
Lieberman compresses the entry-level threat-assessment inventory into six observable patterns:1
— When angered, does he lash out at inanimate objects — punching walls, throwing objects — or engage in symbolic destruction, such as ripping photos, destroying documents, or throwing his wedding ring?
The inanimate-object register is the most reliable early-warning sign. The actor is producing physical violence in a contained register that has not yet escalated to person-directed violence. The escalation to person-directed violence is statistically more likely from this baseline than from baselines that do not contain inanimate-object violence.
— Does she tend to make threats or use violence in an attempt to resolve conflict or to get her way?
The threats-as-tool register. The actor uses threat or violence not as last resort but as habitual conflict-resolution tool. The habituation is the diagnostic — once threats are in the actor's standard toolkit, the threshold for escalation is lower than for actors who have never deployed threats.
— Does he overreact to little things and assume others have a personal motive for crossing him? For instance, if the secretary gives him the wrong information or someone relays poor instructions, does he become enraged, believing their motivation was intentional and personal?
The personalization-of-randomness register. The actor processes random events (the secretary's mistake, the colleague's misdirection) as deliberately personal attacks. The personalization implies a perceptual orientation in which the actor is the target of others' intentions rather than an incidental subject of others' randomness.
— Is she cruel to animals or, for that matter, to people? Does she say hurtful things or seek to embarrass or humiliate others, particularly those who cannot easily defend themselves?
The asymmetric-cruelty register. The actor produces cruelty toward those least able to retaliate — animals, children, junior colleagues, social subordinates. The asymmetry is the diagnostic — cruelty toward equals can be conflict; cruelty toward those who cannot defend themselves reveals enjoyment of the power-differential, which is structurally different.
— Has he not moved up the corporate ladder, and does he show frustration with his lack of progress? Does he feel that no one appreciates his contributions or feel that others take credit for his work and that everyone is out to get him?
The persistent-grievance register. The actor maintains a stable narrative of being wronged, undervalued, and undermined. The grievance does not resolve through specific events; it is maintained across years and contexts. The maintenance is the diagnostic — durable grievance is not the response to specific wrongs but a perceptual orientation that produces wrong-detection across many contexts.
— Has there been a sudden decline in her attitude, performance, or behavior? Does she suddenly seem disinterested and unaffected by the goings-on at work or at home?
The acute-disengagement register. The actor's previous patterns have shifted suddenly — withdrawal from previous engagements, loss of interest in previous priorities, behavioral flattening. The shift is the diagnostic — sudden change in baseline pattern is statistically more concerning than long-stable patterns even when the long-stable pattern is itself worrying.
Lieberman's quantitative finding for the substance-abuse-and-mental-illness intersection:1
Research finds that 31 percent of people who had both a substance abuse disorder and a psychiatric disorder committed at least one act of violence in a year, compared with 18 percent of people with a psychiatric disorder alone.1
The 31% figure is not the population-baseline violence rate; it is the rate within the specific intersection of substance-abuse-disorder-plus-psychiatric-disorder. The 18% comparison is for psychiatric-disorder alone. The framework's operational implication: substance abuse roughly doubles the violence rate within already-elevated populations.
The framework's broader risk-factor catalog:1
Being a young male or a substance abuser puts you at a greater risk for violent behavior than being mentally ill — and combining the risk factors points to an even greater statistical propensity.1
The combination of risk factors is multiplicative rather than additive. The framework directs attention to specific intersections (substance abuse + psychiatric disorder + young male + access to means) where the cumulative statistical likelihood is highest.
Lieberman flags two specific verbal patterns as additional warning signs:1
If someone talks at all about being "fed up" or "sick and tired" of "everyone and everything" or generally about a plan to get even or solve his problems, be alert.1
The fed up and sick and tired registers compress two specific signals: (a) the speaker's frustration has accumulated past a threshold they themselves recognize, and (b) the speaker is articulating the world as the problem rather than specific situations as problems. The total-exhaustion register is structurally different from situation-specific frustration — it implies a perceptual orientation in which the entire context is intolerable, which removes the fine-grained discrimination that allows for non-violent resolution.
The framework's full alarm list:1
You should be extra vigilant if he has detailed plans to commit acts of violence, speaks about settling debts or getting respect, and has easy access to a weapon. Other troubling indicators include if he even jokes or comments about weapons or settling the score, displays pervasive anger and frustration or utters statements of hopelessness, or has a litany of endless grievances.1
Each item in the catalog independently elevates concern. Multiple items together produce convergent signal. The framework's discipline is to attend to combinations of indicators rather than to any single utterance.
The colleague pattern audit. A colleague has been increasingly difficult over six months. Run the six diagnostic questions. Has there been inanimate-object violence (slammed doors, thrown items, broken equipment)? Has the colleague used threat as a conflict-resolution tool (this had better get fixed or there will be consequences)? Have they personalized random events as personal attacks? Have they shown asymmetric cruelty to those who cannot defend themselves (junior colleagues, contractors, support staff)? Do they maintain persistent grievance about lack of recognition or advancement? Has there been sudden disengagement or behavioral flattening? Multiple yes-answers on the inventory is convergent signal that warrants escalation to formal HR or threat-assessment processes. Single yes-answers are concerning but not definitive.
The household-violence triage. A friend confides that their partner has shown some concerning patterns. Run the inventory while listening. Does the partner lash out at inanimate objects? Use threats? Personalize randomness? Show cruelty to pets or to the friend's children? Maintain persistent grievance? Show sudden behavioral changes? The framework's value in this context is to provide structure to a situation where the friend's own perceptual filtering may be downplaying signals. The friend may not have categorized the inanimate-object violence as a precursor to person-directed violence; the framework names it as such. The triage outcome is not to make the leave-or-stay decision for the friend but to ensure they are processing the situation against the right framework.
The own-pattern audit. Apply the inventory reflexively during difficult periods. Have you been engaging in inanimate-object violence (kicking things, slamming doors, breaking objects in frustration)? Have you been using threats in conflicts? Have you been personalizing random events? Have you been cruel to people who cannot easily defend themselves? Have you maintained persistent grievance about being undervalued or wronged? Have you shown sudden disengagement from previous priorities? The reflexive audit catches early-stage cognitive-condition assembly before it has converged into action. The intervention is to recognize the inventory's hits on yourself and pursue support before the assembly completes.
Evidence:
[POPULAR SOURCE]Tensions:
The inventory produces many false positives. Many people who engage in inanimate-object violence, use threats, or personalize random events never act violently. The framework has high sensitivity (catches most actual cases) but lower specificity (also catches many non-cases). Resource-allocation decisions based on the inventory should be calibrated for the false-positive rate.
Samenow's character-consistency claim is empirically contested. The impossible to commit a crime that is out of character framing is rhetorically strong but does not capture the empirical reality that some violent acts are committed by individuals with no clear pre-existing pattern. The framework's directional advice (look for the pattern) is operationally useful but should not be misread as an absolute claim about the universality of pre-existing patterns.
The substance-abuse-multiplier finding has interpretive complexity. The 31% vs 18% finding compares within already-elevated populations (people with psychiatric disorder). Generalizing the finding to broader populations requires controlling for the base rate, which the casual presentation does not do.
Cultural-context confounds the cruelty-to-animals indicator. What counts as cruelty to animals varies across cultural contexts. The framework's diagnostic is calibrated to American urban-suburban norms; cross-cultural application would require recalibration.
Open Questions:
Stanton Samenow built the inside the criminal mind framework through extensive interviewing of incarcerated offenders, beginning at St. Elizabeths Hospital under Samuel Yochelson in the 1960s and 1970s. Samenow's central claim — that violent action is not out of character for the actor — is grounded in offender-interview data showing that violent actors typically have long pre-existing patterns consistent with the violence that ultimately occurred. The methodology has been criticized for non-representative sampling (incarcerated offenders are not representative of all violent actors) but the central pattern has held across multiple subsequent studies.
The MacArthur Violence Risk Assessment Study (Appelbaum, Robbins, Monahan and colleagues) brought systematic empirical rigor to violence prediction in psychiatric populations, producing the substance-abuse-plus-psychiatric-disorder finding (31% vs 18% violence rate within one year) that Lieberman cites. The MacArthur methodology has been one of the most extensively replicated frameworks in violence-risk research.
Lieberman's contribution is the integrated framework that takes Samenow's character-consistency claim and the MacArthur empirical findings and produces a deployable inventory for everyday observation. The integration moves the framework from clinical and forensic settings (where the original research was conducted) into general-population threat-assessment territory.
The genuine convergence: both Samenow's offender-interview research and the MacArthur empirical research agree that violent action has identifiable precursors that are observable to trained attention. The convergence — across qualitative offender-interview methodology and quantitative empirical methodology — is the framework's strongest empirical anchor.
The genuine tension: Samenow's framework is largely categorical (the offender either has the character pattern or does not). The MacArthur framework is probabilistic (specific risk factors elevate violence likelihood by specific amounts within specific populations). Lieberman's framework presents the diagnostic in a register that mixes the two — the inventory feels categorical (does the person show inanimate-object violence yes or no?) but operates probabilistically (multiple yes-answers elevate risk; single yes-answers are concerning but not definitive). The mixed register is operationally useful but can be misread as more categorical than the underlying empirical research supports.
Plain version: violent action almost always has identifiable precursors in the actor's behavior, language, and life patterns — if you know what to look for. Two adjacent vault frameworks structurally illuminate why this works.
Behavioral Mechanics — JACA Threat Assessment Framework: JACA Threat Assessment Framework documents the structured four-axis evaluation (Justification, Alternatives, Consequences, Ability) that operates after the inventory has flagged concern. The Pre-Violence Diagnostic Inventory provides the triage before JACA is deployed. Read together, the two pages produce the operational chain: the inventory catches general behavioral patterns of concern, JACA evaluates whether the cognitive conditions for violence have been assembled, and the linguistic-signature component (qualifier-without-retractor) provides a third independent data source. The structural insight neither page generates alone: violence-prediction operates as layered triage rather than as single-instrument assessment. The inventory is broad and catches many patterns of concern; JACA narrows the assessment to whether the cognitive conditions are assembled; the linguistic signature provides additional independent evidence. Single-layer assessment produces both false positives and false negatives at unacceptable rates; layered assessment reduces both. The combined deployment is operationally robust in ways that any single layer alone is not.
Psychology — Anger as Fear Compensation: Five Modus Operandi Types: Anger as Fear Compensation: Five Modus Operandi Types documents the framework for anger as defensive response to underlying fear. The Pre-Violence Diagnostic Inventory's six questions catch behavioral surfaces that the anger-as-fear-compensation framework explains the underlying mechanism for. The inanimate-object violence register, the personalization-of-randomness register, the persistent-grievance register — each is an audible expression of the fear-compensation machinery operating at high intensity. Read together, the two pages produce the integrated chain: underlying fear → anger compensation → behavioral expression (inventory items) → cognitive condition assembly (JACA) → potential action. The structural insight neither page generates alone: violent-action prediction is partly fear-state inference, because the behavioral inventory items are downstream expressions of the fear-state that produces them. Reading the inventory items as data about the underlying fear-state lets you understand why the patterns predict violence — the same fear-state that produces the inventory patterns is the engine that ultimately produces violent action when the cognitive conditions converge.
Behavioral Mechanics — Sociopath Diagnostic Architecture: Sociopath Diagnostic Architecture documents the five-axis sociopath read. The Pre-Violence Diagnostic Inventory operates at a complementary level — the inventory catches general violence-precursor patterns across the broader population, while the sociopath architecture catches the specific clinical extreme where violence is more likely to be cold and calculated rather than fear-driven. Read together, the two pages produce a fuller picture of violence-risk landscape. The inventory is the wider net; the sociopath axis is the narrower one. Most people who score positive on the inventory are not sociopaths and operate from fear-compensation logic; sociopaths produce a different signature where fear is largely absent and the cognitive conditions can assemble in colder ways. The structural insight neither page generates alone: violence-prediction frameworks must distinguish between fear-driven and predator violence, because the diagnostic surfaces and the intervention requirements differ substantially across the two cases. The inventory is calibrated more for the fear-driven case; the sociopath architecture is calibrated for the predator case. Both are needed for comprehensive assessment.
The Sharpest Implication
The framework's most uncomfortable consequence: most violence prevention happens through dismissal of warning signs that were observable to multiple people in advance. Post-event reviews of violent incidents (workplace shootings, domestic violence escalations, school shootings) regularly identify multiple individuals who had observed warning patterns — inanimate-object violence, threats used as tools, personalization of random events, persistent grievance, sudden behavioral changes — but had dismissed the patterns as not-quite-actionable. The dismissal is the cumulative failure mode. Each individual observer's dismissal seems reasonable in isolation; the aggregate effect across multiple observers is that no one acts and the warning signs go unaddressed until the action occurs.
This implies that organizational and community threat-prevention requires not only training in the diagnostic frameworks but also training in cross-observer aggregation. The single observer who has noticed one inventory item is uncertain whether to act; the same observer who knows that three other observers have each noticed one inventory item should be much more certain. The aggregation does not happen automatically — it requires institutional structures that enable observers to share concerns without bearing full personal responsibility for triggering action. Most organizational structures fail this requirement, which is why the post-event reviews repeatedly find that the warning signs had been observed but never aggregated.
The corollary the inventory's reflexive application forces: your own engagement with any of the inventory items is data about your current emotional load. Catching yourself producing inanimate-object violence (slamming doors, breaking things in frustration), using threats in conflict, personalizing random events, or maintaining persistent grievance is data that your fear-compensation machinery is operating at high intensity. The inventory is not only a tool for assessing others; it is a low-cost monitor for assessing yourself. The reflexive audit is harder to perform than the audit of others because the defensive machinery is fully active when you are the target. The audit's value lives in catching specific moments — that was inventory-item-positive behavior, what's actually going on? — rather than in attempting comprehensive self-assessment.
Generative Questions