Is Accidental Nuclear War Impossible?

By the decree of Pakistan's Foreign Minister, accidental nuclear war
between Pakistan and India cannot occur. In a statement to the APP on
November 29, Mr. Sartaj Aziz said emphatically "I see no possibility
of an accidental nuclear war between Pakistan and India. Pakistan has
an effective command and control system".

This categorical statement is shockingly unscientific because it
presumes complete fore-knowledge of all future crises and exigencies,
a complete understanding of all the possible mechanisms that could
lead to a nuclear exchange, and complete confidence in India's command
and control system as well as that of Pakistan's. Further, it asserts
that human error, misjudgment, and miscalculation are impossible.
Unfortunately, not a single assertion or presumption is logically or
scientifically sustainable. On the contrary, there have been numerous
tragic incidents in India and Pakistan that prove accidents and
miscalculations are far from rare. At best the Minister could have
argued that the probability of accidental nuclear war is small. Even
though this assertion too would have invoked many questionable
assumptions, nevertheless it could have been defended with some degree
of plausibility. As it stands, however, the statement is factually and
scientifically wrong.

It not my intent to split hairs on an abstract academic discussion of
the improbable versus impossible. The issue is far too serious for
that.

One need merely note that nuclear war by accident was never derided
and dismissed during the years of US-Soviet nuclear confrontation. On
the contrary, both sides took this possibility very very seriously. To
avert a false move during those five long decades, the two giants
spent trillions of dollars acquiring the most sophisticated forms of
intelligence gathering by satellites, aircraft, ships, and submarines.
The data from these were continuously analyzed using computers
equipped with artificial intelligence programs. This enabled both
sides to know each other's level of readiness for combat, and know in
advance preparations for a nuclear strike. Without such an elaborate
command and control system a doomsday nuclear confrontation may well
have occurred out of fear or suspicion.

Of course we know that a US-Soviet nuclear war did not occur, but the
danger had never been far away. In spite of every possible precaution
- and technology far more advanced than India or Pakistan can even
dream of - false information provided by radar and other detection
systems was a nightmare for the US and Russian militaries. There were
several serious false alarms causing much alarm, and this is true to
an extent even today. For example, it has recently become known that
on 25 Jan 1995 the Russians mistook a Norwegian scientific rocket for
Trident sea-launched warheads. This mistake lasted for a full eight
minutes - only two minutes away from the launch of Russian nuclear
missiles, which are 'launch-on-warning'. Today there exist fears that
although a nuclear launch is meant to be authorised by the Russian
President, the Defence Minister and the Chief of General Staff of the
Armed Forces, and subsequently by three officers at the missile sites,
nevertheless this chain of command can be bypassed. Russian officers
have been known to re-wire their systems to circumvent this and some
may have the ability to launch autonomously. It has also been reported
that sometimes only one officer remains on duty with the two keys and
the button at his disposal.

There are lessons here for all who care to learn from the experience
of others. First, even the best technology is not good enough when the
issue is whether or not to use nuclear weapons. Second, human
intervention - either through mal-intent, ideological fervour,
inexperience, or plain stupidity - can render the best plans and
technology impotent. The Pakistan-India nuclear confrontation brings a
special urgency to both sets of issues.

It is common to assert that since the US and Russia, each with tens
of thousands of weapons, were able to survive the Cold War therefore
there is no reason for Pakistan and India, which have far fewer
weapons, to feel alarmed. This is wrong reasoning. What may have been
considered good enough for preventing accidental US-Soviet war is
simply not good enough for us. Having a common border, and with
subcontinental missile trajectories of only 4-8 minutes, any type of
early warning system is useless. Even if the best satellites, cameras,
and computers in the world were miraculously made available to
Pakistan and India, this would achieve nothing. In this ridiculously
short time it is totally impossible to make a rational decision as to
whether the alarm is genuine, and whether the incoming missiles are to
be presumed as nuclear armed.

Because no early warning system against nuclear-armed aircraft or
missiles is possible, and because there is no way for Pakistan or
India to protect their respective command and control centres, there
is one and only one possible course of action. This is to disperse and
deploy nuclear-armed aircraft and (when available) missiles over as
wide a geographical area as possible under the command of separate
military units. Further each unit must necessarily be provided the
necessary authorization codes for arming and launching the nuclear
weapons in its possession.

Without providing autonomy to nuclear-armed military units, dispersal
makes no sense - a single bomb on the Rawalpindi GHQ would knock out
Pakistan's ability to mount a retaliatory strike. Even if the GHQ, or
some other command and control centre, were somehow fortified to
survive a nuclear blast in the vicinity, the electromagnetic pulse
which accompanies a nuclear blast would destroy all normal
telecommunications.

Hence autonomy of military units is an inescapable requirement for
maintaining a credible deterrent. But, at the same time, this has a
frightening cost because each unit, and not the PM and COAS, would
have the final say in launching a nuclear strike against India. Could
some ideologically charged Hindu-hating unit commander take destiny
into his own hands? Could deliberately falsified or "honestly wrong"
information reach a unit and result in its launching the weapons in
its possession? No one really knows, but the chances are certainly
not zero.

One could make a virtually identical argument about India. While it
is true that India is much larger, and Pakistan has fewer nuclear
weapons, the difficulty in setting up an Indian command and control
system that will not fail is almost equally severe. It would be stupid
to concentrate all nuclear decision-making in Delhi, and hence
dispersal of nuclear forces is equally important for India. But the
problems of dispersal are equally severe as well, and the possibility
of accidentally initiating nuclear war from that end exists to a
similar degree. We have no right to presume that the Indian command
and control system is any more reliable than ours is.

Are these fictional, exaggerated, fears? I wish it were so. But the
truth is that accidents, sabotage, and tragedy have frequently haunted
our two countries. India has seen the terrible Bhopal gas tragedy,
numerous nuclear reactor mishaps, dam collapses, and industrial
accidents. Pakistan has seen many tragedies too.

It was but ten years ago that an unending stream of shells and rockets
rained down from the skies of Rawalpindi and Islamabad, killing about
a thousand people and wounding and maiming many times that number. The
immediate reaction of most people around me, with whom I watched this
awesome display from my university 10 miles away, was that it was an
Indian attack. Others said that Kahuta had exploded. The government
behaved like a chicken with its head cut off and went around in
circles.

It was much later in the day that Radio Pakistan admitted that an
ammunition dump, located in the heart of the city, had blown up. To
this day, no official report of the Ojhri Camp disaster has been made
public and the cause remains secret.

The explosion of a nuclear device would be immeasurably more serious
than the blowing up of an ordinary ammunition dump. Indeed, thirty
years after Hiroshima US nuclear weapon designers became conscious of
the fact that in the event of fire or ordinary explosion, there is a
fair chance that a nuclear weapon could undergo nuclear detonation
even if it had not been readied for use. This could happen, for
example, if a bomber or missile were to crash upon one's own
territory. Subsequently there was a massive effort to make nuclear
weapons safer, as well construct the exceedingly elaborate electronic
and mechanical safety catches called Permissive Action Links (PALs).

One does not know whether Indian and Pakistani nuclear weapon makers
have put in the enormous effort need to "safe" their weapons and to
what extent they have succeeded or failed. But, generally speaking,
our national disposition is that of risk takers. The notion of fate
plays an important role in our poetry, language, and daily
behaviour. Therefore, for both India and Pakistan, safety has never
been an over-riding concern in driving cars and buses, disposition of
toxic wastes, construction of buildings, and so forth. Why should we
assume that it would be any different when it comes to building bombs?

A nuclear Ojhri is not, therefore, impossible. When Indian or
Pakistani nuclear weapon are assembled and deployed to operational
units, the danger will rise in direct proportion to their numbers. If
a nuclear explosion occurs for whatever reason, the natural assumption
would be that the device belonged to the other side. Even if the
device actually belonged to one's own side, a government, fearful of
public reaction, may commit the ultimate folly of attempting a
cover-up. The consequences of this could be various, including the
probable initiation of cross-border nuclear hostilities.

No one knows how probable accidental Pakistan-India nuclear war is,
no one knows what surprises Kashmir holds, and no mathematical
equation can provide the answer we want. But let us recall General
Zia-ul-Haq's famous remark, made soon after the crisis precipitated by
India's Brasstacks exercises along the Pakistani border in 1986,
"neither India nor Pakistan wanted to go to war but we could have
easily gone to war." Therefore, to create a false sense of security
in the post-nuclear age is an act of supreme folly.