A man writes a paragraph about a problem he is having. The paragraph is grammatically odd in a specific way:1 [POPULAR SOURCE]
I now find myself with a definitive problem which I wish I could find the answer to. And there doesn't seem to be any definitive answer within myself. The problem within me is something that I do not completely understand — whether or not it's myself or the real thing. I keep playing with the idea that maybe that's the trouble. Maybe I should distract my mind and get my mind on interest of something else of another nature, that I may be able to completely get the thought out of my mind. I think, maybe if I go back to my artwork and concentrate on different phases of learning it, that maybe I come into my interests and alleviate the problem for my mind — do everything I can to cooperate with anyone I can that might be able to help me with this problem. And that the thing may find an answer for itself.1
Walter Weintraub's coding system catches the specific irregularity: the paragraph is full of qualifiers (maybe, may be, I think, might, may, may find) but contains not a single retractor (no but, however, although, nevertheless). The qualifier-without-retractor signature is what Weintraub's research identifies as predictive of fixed determination — once an answer to a problem has been found, there may be no turning back.1 The paragraph also operates in detachment register (the problem within me, the real thing, the thing may find an answer for itself) — the author has linguistically severed himself from his own mental state.
Lieberman's chilling note: not long after this note was written, its author murdered his wife.1
The linguistic signature was present before the act. The framework's operational claim: violent action is rarely unprecedented in the actor's psychological state. The state shows up in the language. Reading the linguistic signature is part of how the threat-assessment framework operates. The other part is the structured threat-evaluation framework itself — Gavin de Becker's JACA scale.
Lieberman compresses Gavin de Becker's threat-assessment framework into four diagnostic questions. JACA stands for Justification, Alternatives, Consequences, Ability — the four features that, in de Becker's professional threat-analysis practice, predict whether a stated or implied threat will be carried out:1
Justification: We first consider whether the person likely feels he is justified to use violence to inflict pain, harm, or death.
Alternatives: We then view whether the person feels he has options other than violence to achieve his means. If violence seems like the only way for him to get justice, he will evaluate the consequences.
Consequences: He assesses the likely repercussions of resorting to violence and weighs whether the probable outcome — injury, death, jail — is worth it.
Ability: Whether his plans for revenge remain a fantasy or turn into a horrid reality hinges on whether he feels he has the means and ability to carry out the threat. If he believes that he does, he will likely move forward.
The framework reads each axis on a scale of high to low. High scores across all four axes (the speaker feels strongly justified, sees no other options, considers the consequences worth it, and believes they have the ability) predict highest threat. Low scores across the four predict lowest threat. The combination is the diagnostic. Single-axis readings are unreliable; the framework requires the four-axis composite.
The mechanism the framework operationalizes: violent action is rarely an impulsive event. It is the convergence of four cognitive conditions, each of which the actor has typically been working through over time. The framework does not predict when violence will occur; it predicts whether the cognitive conditions for violence have been assembled. Once assembled, action is more likely; the timing depends on triggers that may be small.
Lieberman extends the same framework to suicide-risk assessment, which is structurally an inward-directed version of the same four-axis evaluation:1
Justification: "Life isn't worth living. The pain is too much, and besides, everyone — my family, friends, and loved ones — would be better off without me."
Alternatives: "There's nothing I can do to make it better, and I feel like there is no way out."
Consequences: "I won't be around to deal with anything afterward."
Ability: "I have access to (or intend to obtain) weapons or pills. I have made plans and have gotten my affairs in order. I've paid off debts and given away my personal possessions."
The same four cognitive conditions, oriented inward. Justification: my death is justified because of (perceived burden, ongoing pain, hopelessness). Alternatives: no other option will reduce the pain. Consequences: post-death consequences are no longer my problem (or are positive — the family is better off). Ability: I have the means.
The structural insight is generative for any impulsive ego-driven action. The JACA framework operates wherever an actor has to assemble cognitive conditions before acting — not only violence and suicide, but escalation in negotiation, abandonment of relationships, war initiation, exit from organizations. The four-axis structure may generalize beyond the specific cases Lieberman applies it to.
Lieberman's most operationally actionable line for the framework's deployment in everyday life:1
In any situation where you feel something is just not right, trust your instincts. You don't need to point to a reason. Your subconscious has picked up on a threat that your conscious mind has dismissed. To protect yourself, you have to learn to trust yourself.
The conscious-mind-dismissal is the dangerous moment. The subconscious has integrated multiple JACA-relevant signals — the speaker's verbal qualifier-without-retractor pattern, their detachment register, their stated grievances, their access to means — and has produced a something is wrong signal. The conscious mind, applying social conventions about politeness and not assuming the worst, frequently overrides the signal. The override is structurally costly when the signal turns out to have been correct.
The framework's discipline: when the something is wrong signal arrives, do not try to argue with it. The signal is the audible compression of subconscious pattern-matching against threat templates. The argument that follows the signal — but they wouldn't really, surely they're just venting, I'm being paranoid — is the conscious-mind dismissal. The framework treats the dismissal as the structural failure mode of casual threat assessment.
The estranged-colleague threat read. A former colleague has been unexpectedly difficult about a recent professional separation. They have made comments to others that range from vague grievance (they treated me terribly) to specific anger (the manager who fired me deserves to fail). Run the four axes. Justification: do they articulate that violence (or significant retaliation) would be justified given what was done to them? Most stop short of justification — anger does not equal justification. Alternatives: do they perceive other options (legal recourse, finding a new job, moving on) — or have those alternatives been foreclosed in their account? Consequences: do they show awareness of post-action consequences (legal exposure, social cost), or are consequences absent from their accounts? Ability: do they have the means — physical, professional, social — to carry out what they are implying? The four-axis read produces a graduated assessment. High on all four is the highest-concern signal. The framework does not produce certainty, but it produces a structured way of taking the situation seriously rather than dismissing it.
The friend-in-crisis suicide read. A friend has been depressed for months. Some recent statements have raised your concern. Apply the JACA-suicide framework. Justification: have they articulated that their family or loved ones would be better off without them — or are they articulating ordinary depressive pain without the burden-on-others framing? Alternatives: do they perceive paths forward, even if those paths feel hard — or have alternatives been foreclosed in their account? Consequences: are post-death considerations entirely absent from their language, or do they show ordinary concern for what would happen after? Ability: have they discussed access to means, made specific plans, given away possessions, paid off debts? High on all four is the situation in which professional intervention is most urgent. The framework is not a substitute for clinical judgment; it is a way of structuring observation so that the highest-concern signals do not get buried under softer signals.
The own-language audit. Apply the framework reflexively to your own internal state during difficult periods. Justification: are you constructing internal narratives that justify extreme action — separation, exit, escalation — beyond what the situation requires? Alternatives: have you foreclosed alternatives that an outside observer would see clearly, in service of an extreme decision that feels emotionally compelling? Consequences: are you minimizing post-action consequences in your own internal accounting? Ability: are you taking preparatory actions toward a decision you have not yet consciously committed to? The reflexive audit catches early-stage cognitive condition assembly before the conscious decision has been made. The intervention is to recognize the assembly and slow it before the conditions converge.
Evidence:
[POPULAR SOURCE]Tensions:
The framework predicts assembly, not timing. JACA tells you the cognitive conditions for violence have been assembled, not when violence will occur. The framework has therefore been criticized for producing too many true positives — many people who score high on JACA never act. The discipline is operationally useful for resource allocation in threat-management but should not be misread as predictive in any clinical-grade sense.
The trust-your-instincts compression has a complementary failure mode. The instinct-trust framing is operationally robust against false negatives (situations where the threat is real and the conscious mind is dismissing the signal). It is less robust against false positives (situations where the something is wrong signal is produced by ordinary social anxiety rather than by accurate subconscious threat-detection). The framework does not provide a clean way to differentiate between the two source-types of the signal.
Cultural register confounds the verbal indicators. The qualifier-without-retractor finding was developed primarily on English-language clinical and forensic samples. Cross-cultural deployment of the linguistic-signature component requires recalibration that the framework does not specify.
Single-utterance misread risk. As ever. One JACA-positive utterance is not a diagnosis. The framework requires sustained pattern observation. The fed up and sick and tired markers Lieberman flags are diagnostically meaningful as recurring register, not as one-time exclamations.
Open Questions:
Gavin de Becker built the JACA framework through decades of professional threat-analysis practice — initially providing personal-protection services to high-profile clients and later extending the work into systematic threat-assessment training. De Becker's unit of analysis is the forensic case file: thousands of completed-vs-not-completed threat scenarios, coded for the four axes that empirically discriminated between them. The Gift of Fear (1997) compresses the practitioner-derived framework into a popular trade book.
Walter Weintraub built the qualifier-without-retractor finding through clinical psychiatric coding work in the 1970s and 1980s. His unit of analysis was the structured interview transcript and (in the relevant published case) forensic writing samples. The qualifier-without-retractor finding emerged as an empirical pattern in samples from individuals who had subsequently committed violent acts.
Stanton Samenow built the inside the criminal mind framework through extensive interviewing of incarcerated offenders, beginning at St. Elizabeths Hospital in Washington DC under Samuel Yochelson. Samenow's central claim — that violent action is not out of character for the actor — is grounded in the offender-interview data showing that violent actors typically have long histories of behaviors consistent with the violence that ultimately occurred.
Lieberman's contribution is the integrated framework that takes the de Becker JACA scale, the Weintraub linguistic signature, and the Samenow character-consistency claim and produces a deployable assessment protocol. The integration is operationally robust in a way that any single framework alone is not — JACA without linguistic signature would miss the verbal-pattern data; linguistic signature without JACA would catch the pattern but lack the structured assessment; Samenow without either would have the character-consistency claim without the diagnostic instruments.
The genuine convergence: all three traditions agree that violent action has identifiable precursors that are observable to trained attention. The convergence — across professional threat-analysis practice, clinical psycholinguistics, and offender-interview research — is the framework's strongest empirical anchor.
The genuine tension: the three traditions have different epistemic standards. De Becker's threat-analysis practice is operationally validated but does not produce the controlled-experimental data that academic violence-prediction research requires. Weintraub's linguistic finding is published in peer-reviewed clinical literature but draws on relatively small samples. Samenow's offender-interview methodology has been criticized for its non-representative sample (incarcerated offenders are not representative of all violent actors). Lieberman's framework presents the integrated diagnostic with stronger predictive implications than any of the underlying research independently supports. The integrated framework is operationally robust as a triage tool; reading it as a clinical-grade prediction instrument would overstate its empirical foundation.
Plain version: violent action and suicide are rarely impulsive — they typically follow assembly of four specific cognitive conditions (justification, no alternatives, acceptable consequences, ability), and the assembly is observable to trained attention through both the four-axis evaluation and the linguistic signature. Two adjacent vault frameworks structurally illuminate why this works.
Behavioral Mechanics — Pre-Violence Diagnostic Inventory: Pre-Violence Diagnostic Inventory documents the six-question diagnostic Lieberman lays out as the entry-level threat assessment, plus the substance-abuse-and-psychiatric-disorder statistics. The JACA framework provides the structured evaluation once the entry-level inventory has flagged concern; the inventory provides the triage before JACA is deployed. Read together, the two pages produce the operational chain: the inventory catches concerning patterns, JACA evaluates whether the cognitive conditions for violence are assembled, and the linguistic signature provides a third independent data source. The structural insight neither page generates alone: violence-prediction requires layered observation rather than any single instrument. The inventory catches behavioral patterns; JACA catches cognitive patterns; linguistic signature catches verbal patterns. Each layer is independently reliable; the convergence across layers is what produces operationally robust assessment. Single-layer reads produce both false positives and false negatives at unacceptable rates; layered reads reduce both.
Behavioral Mechanics — Weintraub Qualifiers-Retractors-Intensifiers System: Weintraub Qualifiers-Retractors-Intensifiers System documents the broader Weintraub psycholinguistic framework that the JACA-supporting linguistic finding is one specific deployment of. The qualifier-without-retractor signature in the murderer's writing sample is the same primitive Weintraub developed for clinical assessment, applied to forensic-linguistic territory. Read together, the two pages produce the full account: Weintraub's framework provides the diagnostic vocabulary, and the JACA-context deployment shows that the framework operates not only in clinical assessment but in threat-analysis. The structural insight neither page generates alone: linguistic primitives that catch one psychological state (anxiety, defensiveness, escape-preparation) often catch related states (escape-foreclosed-determination) when the same primitive operates in different contexts. The qualifier-without-retractor signature in clinical anxiety is reading the same underlying psychological structure that the same signature in the pre-violence writing sample is reading — escape paths foreclosed, determination set, no openings remaining for redirection.
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 JACA framework's Justification axis often operates through anger-as-fear-compensation logic — the actor feels justified in violence because they have constructed an internal narrative in which they have been wronged in ways that warrant retaliation. The construction is downstream of underlying fear (of inadequacy, of disconnection, of unworthiness), and the violence is the compensatory escalation. Read together, the two pages produce the integrated chain: underlying fear → anger compensation → grievance narrative → JACA Justification axis ascent → cognitive condition assembly → potential action. The structural insight neither page generates alone: violence-prediction is partly fear-state prediction, because the cognitive conditions for violence are typically assembled by speakers operating from underlying fear that has been processed through anger and narrative-construction. Reading the underlying fear-state lets you read the upstream condition that produces the JACA-positive surface. Threat-assessment is therefore not only behavioral observation but also fear-architecture inference.
The Sharpest Implication
The framework's most uncomfortable consequence: most threat assessment in everyday life happens through social-convention dismissal rather than through structured evaluation. Casual social processing of concerning behavior runs through politeness conventions (surely they don't really mean it, surely they would never actually act) that systematically suppress the JACA-relevant signals. The conscious-mind dismissal Lieberman warns against is not a personal failure; it is a culturally trained habit that operates without the speaker's deliberation. The implication: organizations and communities that want to detect threats earlier need to train against the dismissal habit, not just teach the diagnostic frameworks. The framework knowledge does not automatically defeat the dismissal habit; the two operate on different cognitive layers.
This implies that significant cultural reform around how concerning behavior is processed could shift threat-detection effectiveness meaningfully. Schools, workplaces, and other communities that have implemented threat-assessment training generally report that the implementation is harder than expected because the training operates against deeply trained social conventions about politeness and not assuming the worst. The cognitive frameworks (JACA, the inventory) are easier to teach than the dismissal-resistance habit is to install. Both are required for the assessment to actually function in real time.
The corollary the trust-your-instincts compression forces: your subconscious threat-detection is probably more accurate than your conscious dismissal, but only on average across many situations. In any specific situation, the subconscious signal could be either a genuine threat-detection or an ordinary social-anxiety response. The framework's discipline is to take the signal seriously enough to investigate, not to act on the signal as definitive. The investigation is what differentiates the genuine subconscious-detection cases from the ordinary-anxiety cases. Both produce the same surface signal; the differentiation requires further inquiry. Most people either dismiss the signal entirely (under-investigation) or act on the signal without investigation (over-action). The framework asks for the harder middle path: take the signal seriously, investigate it, then act on whatever the investigation reveals.
Generative Questions