Threat Timing, Action–Reaction, and the Start and Stop of Police Force

Overview

Police use-of-force decisions occur in time. An officer confronted with a developing deadly threat must perceive what is happening, recognize or interpret the threat, decide whether force is necessary, initiate a physical response, execute that response, perceive any change in the threat, decide whether force should stop, and inhibit or terminate the physical response – and none of those stages is instantaneous. This creates two related principles that operate in tension: an officer does not always have to wait until an assailant completes the final movement of a deadly attack before using defensive force, and once the justification for force ends, continued force must also end, though perception, cognition, and motor stopping each require some finite amount of time.

The Fourth Amendment does not require an officer to wait until a gun is fired, a knife makes contact, a sword is swung, a vehicle actually strikes someone, or another person has already been seriously injured before acting against a sufficiently immediate deadly threat. But an initial justification for deadly force does not create an unlimited license to continue using it after the threat has ended.

Action–Reaction Is Not a Separate Legal Doctrine

“Action versus reaction” is a useful human-performance concept, not an independent Fourth Amendment test. Courts still apply Tennessee v. Garner, Graham v. Connor, Plumhoff v. Rickard, Barnes v. Felix, and applicable circuit precedent. Human-performance evidence can help explain why officers may need to act before an attack is physically completed, why a person initiating movement may act before an observer can perceive and respond, why officers may fire more than one shot before perceiving that the threat has changed, and why stopping an already-initiated motor response is not instantaneous – but it does not replace the constitutional requirement of objective reasonableness.

Police Need Not Wait for Injury to Occur

The Supreme Court’s deadly-force framework is preventative. Under (Tennessee v. Garner, 1985), deadly force may be reasonable where the officer has probable cause to believe the person poses a threat of serious physical harm to the officer or others, and the threatened harm need not already have occurred.

In (Elliott v. Leavitt, 1996), a handcuffed arrestee produced a concealed handgun and pointed it toward officers. The Fourth Circuit rejected the argument that police had to wait to see whether he would actually fire, explaining that the Fourth Amendment does not require police officers to wait until a suspect shoots to confirm that a serious threat of harm exists. The Eleventh Circuit reached a comparable formulation in (Long v. Slaton, 2007), stating that the law does not require officers in a tense and dangerous situation to wait until the moment a suspect uses a deadly weapon to act to stop the suspect. In (Estate Of Larsen ex rel. Sturdivan v. Murr, 2008), addressing an edged-weapon encounter, the Tenth Circuit considered commands to drop the weapon, hostile motions with the weapon, distance, and manifest intentions before rejecting the idea that an officer must necessarily wait for the final visible movement of the blade before acting defensively.

Threat Imminence Is Not the Same as Completed Attack

A threat can be sufficiently immediate before physical contact occurs. Relevant facts in assessing imminence include a weapon drawn or readily accessible, threatening movement, rapid closing of distance, prior attack, express threat, ignored commands, movement toward a third party, physical positioning, and limited time or space for response. The operative question is whether a reasonable officer would perceive a sufficiently immediate threat of death or serious bodily harm – not whether the suspect had already completed the attack.

Edged Weapons and Anticipatory Defensive Force

This distinction is especially important with edged weapons. A knife attack requires physical closing of distance, but an officer’s response also requires time, and the fact that a person is not yet within actual blade-contact distance does not automatically mean the threat is non-immediate. Relevant facts include distance, speed, direction of travel, weapon position, barriers, cover, terrain, footing, third-party proximity, whether the officer’s weapon is already deployed, whether the officer is moving, and what the subject has already done.

There is no federal constitutional rule under which presence inside twenty-one feet automatically justifies deadly force or presence outside that distance automatically precludes it. Distance is one variable in a dynamic encounter, and the better analytical question is how much usable time and tactical margin existed before the threatened person could be reached and what response options were realistically available within that interval.

Third-Party Threats

Action–reaction issues are not limited to officer self-defense. Where a knife-armed person is moving toward a civilian, the relevant distance for timing analysis is subject-to-civilian rather than subject-to-officer. (Ventura v. Rutledge, 2020), illustrates this: Ventura had reportedly assaulted the woman he was approaching, held a knife, ignored repeated commands to stop, ignored an express warning that the officer would shoot, and came within approximately ten to fifteen feet of the third party. The Ninth Circuit decided the case on qualified immunity rather than constitutional merits.

Human Performance: Why Reaction Is Not Instantaneous

Before a person can respond to a changing stimulus, the nervous system must detect the stimulus, direct attention, recognize its meaning, select a response, initiate motor activity, and execute movement – a sequence researchers describe through terms including perception time, decision time, reaction time, response initiation, movement time, and response inhibition. These processes take measurable time and vary across individuals, tasks, uncertainty, stress, visual complexity, movement complexity, training, fatigue, divided attention, and whether a response has already begun. No single reaction-time value should be treated as a universal police-use-of-force constant.

Blair et al.: Armed Suspect Action vs. Officer Reaction

Blair and colleagues conducted a police reaction-time experiment in which officers faced a simulated armed person whose gun was initially pointed downward. The officer had a firearm aimed at the subject and ordered the subject to drop the gun; the simulated subject then either surrendered or attempted to shoot. The study found that officers generally were not able to fire before the armed subject who initiated the shooting movement, demonstrating the structural timing advantage possessed by the person who initiates an unexpected action. The study does not establish a universal minimum safe distance, does not establish that every weapon movement constitutes an imminent deadly threat, and does not establish that anticipatory shooting is always justified.

DOI: 10.1177/1098611111423737

Decision and Motor Preparation Matter

Police shooting studies also show that response time can be affected by visual attention, threat recognition, firearm position, experience, cognitive load, and task uncertainty. Taylor’s experimental study with active law-enforcement officers found that muzzle position affected aspects of shoot/no-shoot performance, supporting the broader conclusion that “reaction time” is not one fixed biological number.

DOI: 10.1177/1098611120960688

Starting Force Requires More Than Awareness

A useful analytical distinction separates recognizing “I should act” from the physical act occurring. An officer may cognitively recognize a threat before drawing, raising the firearm, aligning it sufficiently, pressing the trigger, moving to cover, or executing another force response. The sequence – threat recognition → decision → motor initiation → force – involves distinct intervals even where each is very short.

Stopping Force Also Requires a Sequence

The reverse principle is equally important. An officer does not stop physical action at the exact instant the external world changes; stopping requires that the threat change be perceived, its significance recognized, a decision made to inhibit the response, and the physical response actually ceased. A video may establish the frame at which a knife begins falling, a gun leaves a hand, a subject starts collapsing, or a vehicle clears an officer, but that frame does not necessarily establish the exact moment when a reasonable officer could perceive the change, understand its significance, decide that force should stop, and physically terminate the response.

Response Inhibition

Human-performance research distinguishes response initiation from response inhibition. Once a motor response is underway, stopping it requires active inhibitory control, and shoot/don’t-shoot research supports the proposition that inhibitory control contributes independently to whether participants prevent inappropriate or continuing shooting responses. This does not establish a legally permissible number of additional rounds; it supports the narrower proposition that stopping an initiated physical response is itself a neurocognitive and motor task rather than an instantaneous event.

No “Reaction-Time Safe Harbor”

Human-performance science should not become a liability safe harbor. It does not establish that officers are automatically entitled to one extra shot, two extra shots, a fixed number of milliseconds, or an additional volley after the threat changes. The legal question remains what a reasonable officer in that position would have perceived when the justification for force had ended, and whether continued force was reasonable under the rapidly evolving circumstances.

Plumhoff v. Rickard: Force May Continue While the Threat Continues

In (Plumhoff v. Rickard, 2014), the Supreme Court rejected the argument that the number of shots alone made force excessive, holding that if officers are justified in firing to end a severe threat to public safety, they need not stop shooting until the threat has ended. That limiting phrase is critical: Plumhoff does not authorize continued deadly force after the justification disappears.

Waterman v. Batton: The Threat Can Change Within Seconds

(Waterman v. Batton, 2005), provides one of the clearest analyses of the start/stop problem. Officers fired at an approaching vehicle they reasonably believed presented an immediate threat, and the Fourth Circuit held the initial shots justified. The court separately considered shots fired after the vehicle had passed the officers, holding that force justified at the beginning of an encounter is not justified even seconds later if the justification for the initial force has been eliminated, and a jury could find the later shots unconstitutional because the vehicle had passed and the immediate threat had ended. Waterman therefore presents three distinct timing questions: whether force was justified when it started, when the objective threat ended, and when a reasonable officer should have perceived that change.

Zion v. County of Orange: Threat Termination Is Not Suspect Termination

(Zion v. County Of Orange, 2017), is especially important for edged-weapon cases. Zion had attacked and stabbed an officer with a knife, run from that officer, and been shot by another deputy; after Zion fell, the deputy approached and fired another series of shots. The Ninth Circuit emphasized that terminating a threat does not necessarily mean terminating the suspect – a person can remain alive, moving, conscious, or capable of some movement without continuing to present an immediate deadly threat. The court reasoned that if a wounded knife-armed person is on the ground and no longer presents an immediate threat, a reasonable officer may need to reassess rather than continue shooting. A knife differs from a firearm in this context: a wounded person on the ground holding a knife may have a much more geographically limited ability to harm others than a person retaining a firearm, which does not eliminate danger but does change the threat geometry.

Spencer v. Glaser: Reassessment Does Not Mean Amnesia

(Spencer v. Glaser, 2026), provides an important counterweight. Selby initially advanced while holding a knife and was shot; he fell and dropped the knife; approximately twelve seconds later he attempted to rise and moved toward the officer. The Fourth Circuit held that the earlier knife-wielding advance remained relevant to how a reasonable officer could interpret the later movement. A changed circumstance matters, but immediately preceding threat conduct can remain part of the totality.

Start and Stop Are Asymmetric Problems

The decision to start force and the decision to stop force involve related but different information. When starting force, the officer may be predicting what will occur next – specifically, whether there remains enough time to prevent serious harm if the next observable movement is waited out. When stopping force, the officer is processing evidence that the danger has changed and whether the threat has ended sufficiently that continued force is no longer justified. Both require perception and judgment, and neither should be assessed as though the officer has instantaneous knowledge of future or changing events.

Multiple Shots Are Not Simply Independent Reset Decisions

Each application of force must remain justified, but this does not necessarily mean every trigger press represents a fully independent conscious decision from a cognitively reset baseline. In rapid shooting, visual information continues changing, recoil occurs, attention remains focused, motor activity may already be underway, trigger presses occur in sequence, and the officer must perceive and inhibit the response when the threat changes. Later shots can become unreasonable once the threat ends, but the transition should not be assumed to occur at the exact same millisecond as an externally visible change.

Volleys Can Be Analytically Useful

Courts sometimes examine the temporal structure of force – initial volley, pause, second volley, and subsequent force – which can be more informative than merely counting rounds. Relevant factors include the duration of pauses, subject movement during pauses, weapon status, officer movement, commands between volleys, whether the person falls, whether the person attempts to rise, whether the person remains capable of attacking, and whether the officer changes position. A pause may create an identifiable opportunity to reassess, but no universal pause duration automatically establishes constitutional liability.

The Number of Shots Is Not Dispositive

A high round count can be important evidence without being independently determinative. The more useful question is what threat existed when each materially distinct application or sequence of force occurred. An officer can fire multiple rounds at a continuing severe threat without the number alone establishing excessive force, while a single later shot can be unconstitutional if fired after the justification for deadly force has clearly ended.

Edged-Weapon Threat Geometry

Action–reaction analysis is especially dependent on geometry in edged-weapon cases. Relevant distance variables include the subject’s proximity to the officer, a civilian, a victim, a doorway, a dropped knife, or another weapon. Movement variables include whether the person is stationary, walking, running, lunging, falling, rising, or changing direction. Barrier variables include whether a locked door, vehicle, wall, fence, counter, shield, or other obstacle is present. Weapon-position variables include whether the knife is raised, lowered, at the side, directed at self, directed at another, dropped, or within reach. Officer-posture variables include whether the firearm is holstered or drawn, whether the officer is moving backward, behind cover, trapped, attending another person, or attempting less-lethal force. These variables collectively affect usable response time.

Movement After Being Shot Is Not Automatically a Continuing Attack

A person who has been shot may fall, stumble, rotate, recoil, collapse, extend an arm, attempt to catch themselves, exhibit involuntary movement, or continue pre-existing momentum. Movement alone should not automatically be coded as continued attack or active resistance; the correct question is what the movement objectively communicated in context.

Falling Does Not Automatically End the Threat

The opposite categorical assumption is equally unsafe. A person may fall and still retain a firearm or knife, remain within reach of a victim, attempt to rise, regain the weapon, or continue attacking. That a subject fell does not automatically mean the threat ended.

Action–Reaction and Officer-Created Jeopardy

The fact that officers may need to act before an attack is completed does not eliminate analysis of earlier police conduct. An officer may lawfully defend against an immediate knife attack while a separate question remains whether earlier tactics unnecessarily compressed time and distance. After (Barnes v. Felix, 2025), courts may consider relevant preceding circumstances rather than applying an artificial moment-of-threat cutoff. Whether the threat was sufficiently immediate when force was used and whether earlier police conduct bore on the totality of reasonableness remain analytically distinct questions.

Action–Reaction and Commands

Commands can affect timing. A command to drop the knife can communicate a lawful expectation, test responsiveness, create an opportunity for peaceful resolution, provide warning, and clarify intent. Officers are not always constitutionally required to wait for a complete response, however, where the person is simultaneously advancing, attacking, or creating an imminent threat. Conversely, where a person is stationary and contained, the opportunity to allow compliance can be considerably greater.

Action–Reaction and CEW Use

CEW encounters present a special timing problem. An officer may deploy a CEW, obtain neuromuscular incapacitation, command the subject to drop a knife, observe the knife remain in hand, interpret retention as resistance, and escalate force – but successful neuromuscular incapacitation can itself interfere with the subject’s ability to voluntarily release the weapon. Timing analysis must therefore account for the active electrical cycle, recovery period, voluntary movement capacity, weapon accessibility, and whether the subject is actually reinitiating attack behavior rather than merely failing to release under incapacitation.

Human-Performance Evidence Should Not Decide the Legal Question

Reaction-time studies are informative but not dispositive. Laboratory or simulator studies differ from real police encounters in stress, unpredictability, visibility, background complexity, injury, movement, auditory conditions, stakes, participant expectations, and available responses. Their strongest contribution is mechanistic – demonstrating that perception, decision, initiation, and inhibition consume time – while their weakest use is attempting to declare that a specific shooting was reasonable because an average laboratory reaction time was a particular number of milliseconds.

Evidence Limitations

The action–reaction literature requires caution. Potential limitations include simulator rather than field conditions, artificial or anticipated stimuli, small or selected samples, military rather than police samples in some studies, variable definitions of reaction time, equipment differences, firearm-position differences, limited representation of edged-weapon movement, possible researcher or institutional conflicts, and difficulty generalizing group averages to one officer in one encounter. The Lab should avoid presenting a single reaction-time number as settled fact.

Research-Library Coding Implications

The Lab should consider coding the temporal structure of firearm force rather than only total shot count.

1. Threat onset

• exact/estimated timestamp
• indeterminate

2. Threat type

• firearm
• edged weapon
• vehicle
• physical attack
• other

3. Threat target

• officer
• civilian
• mixed
• self only
• indeterminate

4. Officer force onset

• timestamp
• latency from observable threat where measurable

5. Weapon status at force onset

• held
• raised
• advancing
• dropped
• unclear

6. Number of rounds

• total
• by officer

7. Volleys

• one continuous sequence
• two
• three+
• indeterminate

8. Pause between volleys

• exact duration if measurable
• approximate
• none discernible

9. Threat-changing event

• subject falls
• weapon drops
• movement stops
• person turns away
• barrier intervenes
• third party moves clear
• subject attempts to rise
• subject re-arms
• other

10. Time from threat-changing event to cessation of force

• exact where measurable
• approximate
• indeterminate

11. Continued voluntary threatening movement

• yes
• no
• disputed
• indeterminate

12. Possible involuntary / force-induced movement

• yes
• no
• indeterminate

13. Reassessment opportunity

• clear pause
• possible
• none apparent
• indeterminate

14. Command during or between force

• yes
• no
• indeterminate

15. Weapon accessibility after fall

• retained
• within reach
• separated
• unknown

This allows a sequence of threat begins → officer perceives/responds → force starts → threat changes → officer perceives/reassesses → force stops, rather than assuming the entire event is one indivisible moment.

Leading Decisions – Chronological

Tennessee v. Garner

The Court held that deadly force may be reasonable where an officer has probable cause to believe the person poses a threat of serious physical harm to the officer or others. The decision establishes that the deadly-force doctrine is preventative and that injury need not already have occurred.

---

Graham v. Connor

The Court held that force is judged from the perspective of a reasonable officer on the scene, with allowance for split-second judgments in tense, uncertain, rapidly evolving circumstances. The decision is the core doctrinal basis for considering the temporal limits of human decision-making.

---

Elliott v. Leavitt

The Fourth Circuit held that officers confronting a serious firearm threat do not have to wait until the suspect actually shoots. The decision provides a clear articulation of anticipatory self-defense.

---

Waterman v. Batton

The court held that force justified initially may become unjustified seconds later when the threat has been eliminated. The decision is one of the clearest start/stop force analyses in the circuits.

---

Blanford v. Sacramento County

The Ninth Circuit held that deadly force was reasonable where a man carrying a large sword ignored repeated commands and warnings and continued toward a residence whose occupants were unknown. The case demonstrates anticipatory protection of possible third parties in an edged-weapon context.

---

Long v. Slaton

The Eleventh Circuit held that officers need not wait until the precise moment a deadly weapon is used before acting to prevent threatened harm. The decision is a frequently cited formulation of anticipatory defensive force.

---

Estate of Larsen ex rel. Sturdivan v. Murr

The court held that immediate-threat analysis in an edged-weapon encounter can include commands, hostile weapon movement, distance, and manifest intent, rejecting a simplistic requirement that police wait until actual weapon contact is imminent in a literal last-instant sense.

---

Garczynski v. Bradshaw

The Eleventh Circuit held that under the circumstances presented, police did not have to wait until the armed person completed the final aiming movement before defending themselves.

---

Plumhoff v. Rickard

The Court held that officers justified in using deadly force to end a severe threat need not stop firing until that threat has ended. The authorization lasts only while the severe threat continues.

---

Zion v. County of Orange

The Ninth Circuit held that once a threat has ended, continued deadly force can become excessive, and that terminating the threat is not synonymous with terminating the person. The decision is particularly relevant to edged-weapon post-incapacitation reassessment.

---

Kisela v. Hughes

The Court applied qualified immunity where precedent did not clearly establish the unlawfulness of shooting a knife-holding woman near another person, demonstrating the importance of weapon, distance, third-party proximity, commands, and factual specificity to the clearly-established analysis.

---

Ventura v. Rutledge

The Ninth Circuit held the asserted unlawfulness was not clearly established where a knife-armed person continued toward a recently assaulted third party despite commands and an explicit shooting warning. The court did not decide the constitutional merits.

---

Barnes v. Felix

The Court held that courts may not artificially restrict the Fourth Amendment analysis to the final instant immediately preceding force, requiring that threat timing be considered in the broader relevant chronology.

---

Spencer v. Glaser

The Fourth Circuit held that earlier weapon and threat conduct can remain relevant to how a reasonable officer interprets renewed movement after circumstances change. Reassessment must account for relevant recent history rather than treating later movement as occurring in a vacuum.

What the Rule Does Not Establish

This body of law and science does not establish that:

• action always defeats reaction;
• officers may shoot merely because reaction takes time;
• a person with a knife is automatically an immediate deadly threat;
• twenty-one feet is a constitutional boundary;
• officers must wait until a knife is within touching distance;
• officers may continue firing after the threat ends;
• every post-threat shot is automatically excessive;
• every visible movement after a shooting is voluntary;
• a person must become completely motionless before force stops;
• a fixed number of milliseconds is constitutionally protected;
• a fixed number of additional rounds is constitutionally permissible;
• reaction-time averages determine what a specific officer perceived;
• or human-performance science overrides the Fourth Amendment totality analysis.

Policy Implications

Evidence-informed policy can recognize both sides of the timing problem. Useful principles include not requiring officers to wait for actual injury before responding to an imminent deadly threat, avoiding rigid distance rules, training officers to understand movement and reaction timing, emphasizing continuous threat reassessment, training officers to distinguish falling, involuntary movement, and renewed attack, recognizing the need to stop force once the threat is reasonably perceived to have ended, emphasizing reassessment after pauses or volleys where circumstances permit, using video review to examine threat transition rather than merely round count, and avoiding treatment of laboratory reaction-time averages as case-specific legal presumptions. These are policy considerations, not categorical constitutional mandates.

Relationship to Other Backgrounders

This Backgrounder should be read with the Federal Constitutional Excessive-Force Standard, Commands, Opportunity to Comply, and Timing of Force, Officer-Created Jeopardy and Pre-Force Conduct, What Is Clearly Established Law?, Qualified Immunity, Emergency Aid, Welfare Checks, and Warrantless Entry, and Failure to Provide or Render Medical Aid Following Use of Force. The combined temporal framework is: pre-force context → threat develops → perception → force begins → threat changes → perception/reassessment → force ends → medical response.

Research-Library Conclusion

Fourth Amendment doctrine does not require police to wait until the final instant of a deadly assault – or until someone is actually struck, stabbed, shot, or run over – before taking defensive action. The relevant question is whether a reasonable officer could perceive a sufficiently immediate threat of serious physical harm. Human-performance research provides a plausible mechanism for that rule: perception, threat recognition, decision-making, and motor execution consume time, which gives the person initiating an unexpected action a temporal advantage. The same principle applies in reverse. When the threat changes, an officer must perceive the change, recognize its significance, inhibit an ongoing response, and physically stop – a process that also takes finite time. Neither law nor science creates a reaction-time safe harbor, however. Plumhoff permits continued firing only while the threat continues, while Waterman and Zion demonstrate that force can become excessive when the threat has ended and a reasonable officer should have recognized that change. For edged-weapon research, the most defensible approach is dynamic and chronological: measure distance, movement, weapon status, third-party exposure, force onset, threat-changing events, pauses, reassessment opportunities, and force cessation rather than applying a fixed distance or reaction-time rule.

Not Legal Advice

This Backgrounder is for research and educational purposes. It summarizes federal constitutional doctrine and selected human-performance research, and should not substitute for jurisdiction-specific legal research by qualified counsel.

Key Human-Performance Research

Blair, J. P., Pollock, J., Montague, D., Nichols, T., Curnutt, J., & Burns, D. (2011). Reasonableness and reaction time. Police Quarterly, 14(4).
DOI: 10.1177/1098611111423737

Taylor, P. L. (2021). “Engineering Resilience” Into Split-Second Shoot/No Shoot Decisions: The Effect of Muzzle-Position. Police Quarterly, 24(2).
DOI: 10.1177/1098611120960688

Vickers, J. N., & Lewinski, W. (2012). Performing under pressure: Gaze control, decision making and shooting performance of elite and rookie police officers. Human Movement Science, 31(1), 101–117.
DOI: 10.1016/j.humov.2011.04.004

Horn, R. R., Lewinski, W. J., Heidner, G. S., Lawton, J., Allen, C., Albin, M. W., & Murray, N. P. (2024). Assessing between-officer variability in responses to a live-acted deadly force encounter as a window to the effectiveness of training and experience. Ergonomics, 67(8), 1035–1050.
DOI: 10.1080/00140139.2023.2278416

References

Barnes v. Felix (2025). 605 U.S. 73.

Elliott v. Leavitt (1996). 99 F.3d 640.

Estate Of Larsen ex rel. Sturdivan v. Murr, 511 F.3d 1255 (10th Cir. 2008) (2008). 511 F.3d 1255, United States Court of Appeals, Tenth Circuit. https://casetext.com/case/estate-of-larsen-v-murr

Long v. Slaton (2007). 508 F.3d 576.

Plumhoff v. Rickard (2014). 572 U.S. 765.

Spencer v. Glaser (2026).

Tennessee v. Garner (1985). 471 U.S. 1, U.S.

Ventura v. Rutledge (2020). 978 F.3d 1088.

Waterman v. Batton (2005). 393 F.3d 471.

Zion v. County Of Orange (2017). 874 F.3d 1072.

How to cite

Edged-Weapon Response Lab. (2026). Threat Timing, Action–Reaction, and the Start and Stop of Police Force. In EWRL Law & Policy Analysis (Version 1.0). https://droptheknife.org/law-and-policy/action_reaction