Monday, November 14, 2011

Fight or Flight - Did Walter Cannon Get It Wrong?

Fight-or-flight dominates threat-induced behaviour thinking. This domination has led to certain ingrained assumptions about what to expect of ourselves and others in response to a perceived threat. Fight-or-flight is a concept which is often referred to within activities associated with preparing a person to survive a violent encounter.

Walter Cannon proposed the fight-or-flight model in the early 1920s to explain the automatic physiological reaction (fight-or-flight response, aka stress response) that prepares the body for action: fight to defend ourselves or flee to escape the source of a perceived threat.

Behaviourally, the fight-or-flight model is obviously limited. I'll just consider the F models and those proposed by Bruce Siddle and Dave Grossman.

Siddle is described as an internationally recognised authority on use of force training and the effects of survival stress (aka fight-or-flight response) on combat performance. Those activities associated with preparing a person to survive a violent encounter who refer to fight-or-flight are referring to Siddle's theories. Whether they acknowledge the source of their wisdom, or even know the source, is another matter. Siddle proposed 3Fs: fight, flight, freeze in place (hypervigilance). With regards to Siddle's hypervigilance behaviour: 'Siddle uses a unique law enforcement interpretation of the term hypervigilance. ... the term, hypervigilance is used to denote what others might call panic or hyperarousal.'

Grossman is described as being one of the foremost experts in the field of human aggression and the psychology of combat. Grossman appreciates that the fight-or-flight model needs updating: 'One of the roots of our misunderstanding of the psychology of the battlefield lies in the misapplication of the fight-or-flight model to the stresses of combat.' He argues that adding posture and submission to the standard fight-or-flight model (fight, flight, posture, or submit) helps to explain many of the actions on the battlefield. Sadly he didn't use F words to describe these behavioural options. Grossman suggests the first behaviour is posture, and if that is unsuccessful then fight, flight, or submission behaviours are elicited.

Jeffrey Grey, a British psychologist, proposed a 3F model: 'freezing (keeping absolutely still and silent), flight, or fight.' This freeze response is different to Siddle's freeze response as will be explained below.

In 2004 Bracha and some colleagues asked: 'Does fight or flight need updating?' The answer is, absolutely! That is the purpose of my Beyond Fight or Flight.

Bracha et al propose 4Fs: freeze, flight, fight, or fright. When referring to Cannon's 2Fs, they suggest that 'both the order and the completeness of Cannon’s famous phrase are suspect.' The order should be flight or fight reflecting the initial tendency to flee when threatened with fight only being elicited if flight is not possible. Freeze is the initial response which they describe as the stop, look, and listen response. Fright is a term they use to refer to tonic immobility (TI)- a behavioural response all people involved in preparing a person to survive a violent encounter should be aware of, but of which sadly most are ignorant. TI refers to an evolved defensive response which is an involuntary catatonic state. It is argued that TI is elicited when fight is unsuccessful. Bracha et al provide an explanation for the use of the term fright, however, I'd suggest they chose the term because 4Fs are catchier than 3Fs and a TI.

Not long after the above proposal, Bracha, alone this time, proposed 5Fs: freeze, flight, fight, fright, and faint. The different between fright/TI and faint is that in the former the subject is still conscious and processing the experience whereas in the latter they are unconscious and not processing the experience.

Schauer and Elbert propose 6Fs: freeze, flight, fight, fright, flag, or faint. Flag appears to be a half-way house between fight and faint where the body goes from being rigid to flaccid - it starts to flag.

There are other defensive behaviour models proposed that go beyond the F models, however for our purposes the F models are sufficient.

All of the above F models refer to threat-induced behaviours. All of the above models refer to fear motivated behaviours. As fear increases, the behaviours travel along a defence cascade starting with freeze and ending with faint. Actually, there is one more evolved behavioural response - death. A previous post explained that the dominance of the parasympathetic nervous system in the latter half of the defence cascade can literally scare a person to death.

All of the above models refer to fear motivated behaviours - all except Cannon's, the originator of the fight-or-flight concept, 2F model. Cannon associated flight with fear and fight with anger. Did he get it wrong? Or did the others get it wrong? And does it matter?

Blanchard and Blanchard distinguish between defensive attack and offensive attack behaviours. They suggest that defensive attack is associated with physical threats and is motivated by fear. Defensive attack is engaged when an existential threat is proximal and imminent and escape is impossible. Offensive attack, on the other hand, is associated with threats to resources and is motivated by anger.

Defensive attack and offensive attack are also referred to as defensive aggression and offensive aggression. Aggression has been categorised as reactive aggression (also called hostile, affective, subjective, angry, impulsive, irritable, emotional, or retaliatory aggression) and proactive aggression (also called instrumental aggression). Reactive aggression is motivated by an emotion. It would include both fear-motivated defensive aggression and anger-motivated offensive aggression. In contrast, proactive aggression, or instrumental aggression, is not motivated by an emotion. It is cold blooded, premeditated, calculated behaviour that is motivated by some other nonaggressive goal (obtaining money, territory, social dominance, object acquisition). Sympathetic activation is usually absent and there is minimal negative emotion.

Did Cannon get it wrong? Based on the above, it would appear he did. It would appear that his fight response was actually motivated by fear and not anger.

This misattribution of the fight response to anger helps explain why Cannon thought the same physiological response is activated with both anger and fear. Plutchik explains that the notion of physiological differences between emotions seemed to be pretty much taken for granted until the researches of Cannon during the 1920s which apparently indicated the autonomic changes observed in cats and dogs during fear and anger were the same. This led many psychologists to assume that the autonomic changes for all emotions in humans were also the same, a conclusion which he suggests was by no means justified by the evidence. Since that time an increasing body of research has begun to show the existence of definite, physiological differences between emotions. Different emotions = different physiological responses.

Does it matter? Society seems to think so. If your aggressive actions result in the death of another person, to take an extreme example, and you are charged with that homicide, the difference between fear-motivated defensive aggression, anger-motivated offensive aggression, and nonemotional instrumental aggression will determine your fate.

Fear-motivated defensive aggression, within the limits of the law, will result in a self-defence defence. The difference between angry aggression and instrumental aggression could mean the difference between manslaughter and murder respectively:
In American common law, the heat-of-passion defence reduces a charge of murder to manslaughter if the defendant successfully demonstrates that he or she (a) was first adequately provoked by his or her victim, (b) became emotionally disturbed as a direct result of said adequate provocation, (c) did not have enough time to cool off before killing his or her perceived provocateur, and (d) did not, in fact, cool off before committing the homicide in question. (Fontaine 2008)
Emotion, as conceived by the emotion discipline, is more than subjective feeling. It is a construct that includes appraisal, feeling, physiological, action tendency, and behaviour. Feeling activates the other responses.

Different emotions = different behaviours. All fight behaviours are not the same. The Blanchards described different defensive and offensive attack patterns. It has long been accepted that anger-motivated violence is more vicious than instrumental violence. Stephen (1883) suggested, impulsive killings such as heat of passion homicide are crueler, more dangerous, and more ferocious than its premeditated counterpart. The reason for the difference is the goal of the aggression. Anger-motivated offensive aggression's goal is to cause harm and suffering. Nonemotional-motivated instrumental aggression's goal is a nonaggressive goal, eg. obtain money, territory, etc. Interestingly enough, society's law punishes the crueler form of aggression less than the less cruel form of aggression.

The different forms of aggression have tactical implications. Disengaging from an aggressive encounter when the other party is engaged in fear-motivated defensive aggression should terminate the encounter. It will not if the aggression is anger-motivated offensive aggression where harm is the goal, or nonemotional instrumental aggression where the obtaining of an extrinsic goal is the goal. A different strategy is needed to extricate oneself from those encounters when the other party is engaged in anger-motivated offensive aggression or nonemotional instrumental aggression. One has to know the resource which is being challenged or the extrinsic goal being sought in order to terminate the aggression in those cases.

Anger and fear are negative emotions. I'm working on aggression that is motivated by positive emotions. I've seen hints of these, but they are not as well studied as anger and fear motivated aggression.

Every activity that is involved in preparing a person to survive a violent encounter which refer to fight-or-flight, any of the F models, Siddle's theories, and any similar models and theories by whatever name, are only referring to threat-induced, fear-motivated behaviours and physiological responses. This applies to stress training, stress inoculation training, and stress exposure training which are training programs designed to better prepare military and law enforcement officers for operational duties. These models and theories are focusing only on one small part of human aggressive and violent behaviour. These fact are not explicitly acknowledged by these activities, if they are understood by them at all. Nor are the limitations of these models and theories understood when attempting to use them to understand aggressive or violent behaviour and to better prepare a person to survive a violent encounter.


  1. During an event at the University of Toronto in 1973, I had the chance to take part in a discussion with Dr. Hans Selye. He was speaking about the "fight or flight" model of Walter Cannon and I, as a recently discharged U.S. Marine had experiences that had some bearing on his talk. Speaking as someone with combat experience in Viet Nam and then experience of another sort of stressful activity, namely dealing with the daily security at the U.S. Embassy in Paris at a point where student protests and violent demonstrations were taking place. Also, another stressor at the time was taking part in the preparations for and ongoing security of the Viet Nam Peace Talks. These experiences along with a long-time serious practice of the Japanese Martial Arts gave me some insight into stress and it's effects on human action/reaction.

    I told Dr. Selye that I thought there were three stages or possibilities instead of two in the fight or flight model. Fight, flight, or fibrillation as I called it at the time. From a psychological viewpoint, it seemed to me that there were various motivations and values to each. With regard to the use, in this article, of the term hyper-vigilance to denote "panic or hyper-arousal" I must add that the outcome of these conditions may be varied in well trained individuals. These variations are spaced all along the scale between negative and positive. Dr. Selye was very interested in a first hand viewpoint and we had a very powerful, but relatively brief discussion after his presentation.

    It seems to me that John Coles has taken his research and conclusions along similar directions, and has defined what I had been thinking in better terms than I did. His work will make it much easier to get this across to serious students of martial arts.

    Thanks for your work.

    Chuck Clark
    Jiyushinkai Aikibudo

  2. Really very interesting and very valuable information about the Fight for Flight nice work.
    Fight for Flight


Your comments make my work all the more relevant as I use them to direct my research and theorising. Thank you.