Global Nuclear Politics
Department of Politics, University of York, Autumn term 2020-21
Background briefing for lecture 3: limits of nuclear deterrence
This note outlines provides further detail on two of the lecture’s themes: the
theory of ‘normal accidents’ and the political psychology of leadership decision-
making in relation to nuclear threats and nuclear deterrence. The lecture will
focus on the limits of nuclear deterrence and will touch on these themes but, as
with previous preparatory notes, the overview below provides important
arguments and context.
Normal accidents theory
The development of organisational theory and how very serious accidents can
happen was first applied to nuclear weapon systems and organisations in the
1990s. This raised specific concerns about nuclear weapons accidents in general
(of which there have been many) and the effects of accidents and other
organisational failures, such as false alarms, during nuclear crises or in a war
between nuclear-armed states and the effect on the practice of nuclear
deterrence.
‘Normal’ or ‘system’ accidents are terms coined by the American sociologist
Charles Perrow. Perrow studied the control of hazardous technologies such as
nuclear power plants in his 1984 book Normal Accidents: Living with High- risk
Technologies. Perrow provides ample evidence to show that organisations can
struggle to deliver persistently safe outcomes with what are called complex,
tightly coupled systems. This can result in disasters like the Challenger space
shuttle explosion in 1986, Deep Water horizon oil rig disaster in 2010 and
Fukushima in 2011.
Perrow argued that that the failure of individual components or items is a
ubiquitous feature of almost all systems. However, in complex systems
component failures can create complex interactions, which are unfamiliar, or
unplanned and they can create unexpected sequences in the system that are
either not visible or not immediately comprehensible to operators. This is
exacerbated by complex systems are ‘tightly coupled’: when ‘there is no slack or
buffer or give between two items. What happens in one directly affects what
happens in the other’ and this can precipitate a rapid can cascade of failures, as
Perrow puts it (p. 15).
In 1993 US political scientist Scott Sagan published his research on ‘normal
accident’ theory and nuclear weapons operations in a book called The Limits of
Safety: Organizations, Accidents, and Nuclear Weapons. Drawing on Perrow, he
describes the two schools of thought in the scholarly literature on complex
organisations as follow: “The first is the optimistic view [...] ‘high reliability
theory’, whose proponents argue that extremely safe operations are possible,
even with extremely hazardous technologies, if appropriate organizational
design and management techniques are followed. The second school, [...]
‘normal accidents theory’, presents a much more pessimistic prediction: serious
accidents with complex high technology systems are inevitable” (p. 46). Sagan
found that the United States’ nuclear weapons control complex had all of the
characteristics of the type of ‘normal accident’ organisation. His research “raises
1
, serious doubts about the central assumption that a nuclear war could not occur
unless political leaders decided it was in their states’ interests [...] the belief that
nuclear deterrence can prevent nuclear war under all circumstances should be
seen as exactly that: a belief, not a fact” (p. 262).
Moreover, his research detailed many incidents of ill-discipline, alcoholism, drug
abuse and mental illness among personnel working with nuclear weapons as
some of the sources of risk even elite organizations find difficult to screen out,
and which could contribute to system accidents. A string of recent incidents and
scandals involving US military personnel with nuclear weapons responsibilities
serves to highlight the absence of robust internal learning mechanisms in place
within the US nuclear weapons complex.
At the heart of the argument is the fact that nuclear weapons organisational
systems are complex, tightly coupled systems. The development, production,
storage, maintenance, deployment, safety and security of nuclear weapons
involves a system of organisations, technologies and procedures that are very
complex. This is in addition to the complexity of systems involved in the
detection of and response to an incoming nuclear attack. For nuclear deterrence
to function as envisaged, these systems must interact in order for nuclear
weapons to be available to use at the desired time. This is actually a daunting
challenge, as former US ICBM officer Bruce Blair details in his 1993 book The
Logic of Accidental Nuclear War. Because of the way in which nuclear command,
control, communications, and intelligence systems are tied together, unexpected
technical and human failures can quickly multiply and interact in ways that no
one can predict, or respond to quickly or effectively enough to avert disaster
because of their incomprehensibility for a time to the operators, especially in
high-alert situations. Blair and other experts and governments have called for
nuclear de-alerting (see here, here, and here) to make the alert and launch
control systems for nuclear weapons less responsive, although, for some, this
undermines the practice of nuclear weapons.
Eric Schlosser, an American investigative journalist, documented many accidents
involving nuclear weapons in a recent book called Command and Control that
focussed on United States command and control of its arsenal during the Cold
War A recent study by Chatham House catalogued many historical instances in
which organizational or technical systems failed or offered ambiguous signals, so
that individual judgment at odds with those systems was all that stood in the
way of nuclear weapons use. Russian academic Pavel Podvig has also discussed
nuclear close calls and nerve-wracking experiences in the USSR and now Russia.
This situation still pertains. The end of the cold war and nuclear arms reductions
do not fundamentally alter the risk of accidental or inadvertent launch of nuclear
weapons in peacetime or in a crisis beyond the intent and control of political
leaderships. The United States and Russia each still have approximately 800
strategic nuclear warheads deployed on high alert. The number of accidents,
false alarms, unauthorised military actions or unknown to the political leadership
during the Cuban Missile Crisis is indicative of the challenge of ‘normal accidents’
(see here for a good summary).
Leadership psychology and nuclear weapons decision-making
In the 1970s US political scientist Robert Jervis began to explore threat
“perception” and “misperception” and the variance between what leaders
2
Department of Politics, University of York, Autumn term 2020-21
Background briefing for lecture 3: limits of nuclear deterrence
This note outlines provides further detail on two of the lecture’s themes: the
theory of ‘normal accidents’ and the political psychology of leadership decision-
making in relation to nuclear threats and nuclear deterrence. The lecture will
focus on the limits of nuclear deterrence and will touch on these themes but, as
with previous preparatory notes, the overview below provides important
arguments and context.
Normal accidents theory
The development of organisational theory and how very serious accidents can
happen was first applied to nuclear weapon systems and organisations in the
1990s. This raised specific concerns about nuclear weapons accidents in general
(of which there have been many) and the effects of accidents and other
organisational failures, such as false alarms, during nuclear crises or in a war
between nuclear-armed states and the effect on the practice of nuclear
deterrence.
‘Normal’ or ‘system’ accidents are terms coined by the American sociologist
Charles Perrow. Perrow studied the control of hazardous technologies such as
nuclear power plants in his 1984 book Normal Accidents: Living with High- risk
Technologies. Perrow provides ample evidence to show that organisations can
struggle to deliver persistently safe outcomes with what are called complex,
tightly coupled systems. This can result in disasters like the Challenger space
shuttle explosion in 1986, Deep Water horizon oil rig disaster in 2010 and
Fukushima in 2011.
Perrow argued that that the failure of individual components or items is a
ubiquitous feature of almost all systems. However, in complex systems
component failures can create complex interactions, which are unfamiliar, or
unplanned and they can create unexpected sequences in the system that are
either not visible or not immediately comprehensible to operators. This is
exacerbated by complex systems are ‘tightly coupled’: when ‘there is no slack or
buffer or give between two items. What happens in one directly affects what
happens in the other’ and this can precipitate a rapid can cascade of failures, as
Perrow puts it (p. 15).
In 1993 US political scientist Scott Sagan published his research on ‘normal
accident’ theory and nuclear weapons operations in a book called The Limits of
Safety: Organizations, Accidents, and Nuclear Weapons. Drawing on Perrow, he
describes the two schools of thought in the scholarly literature on complex
organisations as follow: “The first is the optimistic view [...] ‘high reliability
theory’, whose proponents argue that extremely safe operations are possible,
even with extremely hazardous technologies, if appropriate organizational
design and management techniques are followed. The second school, [...]
‘normal accidents theory’, presents a much more pessimistic prediction: serious
accidents with complex high technology systems are inevitable” (p. 46). Sagan
found that the United States’ nuclear weapons control complex had all of the
characteristics of the type of ‘normal accident’ organisation. His research “raises
1
, serious doubts about the central assumption that a nuclear war could not occur
unless political leaders decided it was in their states’ interests [...] the belief that
nuclear deterrence can prevent nuclear war under all circumstances should be
seen as exactly that: a belief, not a fact” (p. 262).
Moreover, his research detailed many incidents of ill-discipline, alcoholism, drug
abuse and mental illness among personnel working with nuclear weapons as
some of the sources of risk even elite organizations find difficult to screen out,
and which could contribute to system accidents. A string of recent incidents and
scandals involving US military personnel with nuclear weapons responsibilities
serves to highlight the absence of robust internal learning mechanisms in place
within the US nuclear weapons complex.
At the heart of the argument is the fact that nuclear weapons organisational
systems are complex, tightly coupled systems. The development, production,
storage, maintenance, deployment, safety and security of nuclear weapons
involves a system of organisations, technologies and procedures that are very
complex. This is in addition to the complexity of systems involved in the
detection of and response to an incoming nuclear attack. For nuclear deterrence
to function as envisaged, these systems must interact in order for nuclear
weapons to be available to use at the desired time. This is actually a daunting
challenge, as former US ICBM officer Bruce Blair details in his 1993 book The
Logic of Accidental Nuclear War. Because of the way in which nuclear command,
control, communications, and intelligence systems are tied together, unexpected
technical and human failures can quickly multiply and interact in ways that no
one can predict, or respond to quickly or effectively enough to avert disaster
because of their incomprehensibility for a time to the operators, especially in
high-alert situations. Blair and other experts and governments have called for
nuclear de-alerting (see here, here, and here) to make the alert and launch
control systems for nuclear weapons less responsive, although, for some, this
undermines the practice of nuclear weapons.
Eric Schlosser, an American investigative journalist, documented many accidents
involving nuclear weapons in a recent book called Command and Control that
focussed on United States command and control of its arsenal during the Cold
War A recent study by Chatham House catalogued many historical instances in
which organizational or technical systems failed or offered ambiguous signals, so
that individual judgment at odds with those systems was all that stood in the
way of nuclear weapons use. Russian academic Pavel Podvig has also discussed
nuclear close calls and nerve-wracking experiences in the USSR and now Russia.
This situation still pertains. The end of the cold war and nuclear arms reductions
do not fundamentally alter the risk of accidental or inadvertent launch of nuclear
weapons in peacetime or in a crisis beyond the intent and control of political
leaderships. The United States and Russia each still have approximately 800
strategic nuclear warheads deployed on high alert. The number of accidents,
false alarms, unauthorised military actions or unknown to the political leadership
during the Cuban Missile Crisis is indicative of the challenge of ‘normal accidents’
(see here for a good summary).
Leadership psychology and nuclear weapons decision-making
In the 1970s US political scientist Robert Jervis began to explore threat
“perception” and “misperception” and the variance between what leaders
2
