After the Cold War ended, the existence of nuclear weapons on both sides wasn’t enough to stop the US from expanding NATO’s borders ever eastwards towards Russia’s borders, contrary to the terms on which Moscow thought Germany’s reunification and the admission of a united Germany into NATO had been agreed. Several Western leaders at the highest levels had assured Mikhail Gorbachev that NATO wouldn’t expand even ‘one inch eastward’. In 1999, Russia watched helplessly from the sidelines as its ally, Serbia, was dismembered by NATO warplanes that served as midwives to the birth of an independent Kosovo.

But Moscow didn’t forget the lesson. In 2014, the nuclear equation didn’t stop Russia from reacting militarily to the US-backed Maidan coup in Ukraine—which displaced the pro-Moscow elected president with a westward-looking regime—by invading eastern Ukraine and annexing Crimea. In other words, the more or less constant US–Russia nuclear equation is irrelevant to explaining the shifting geopolitical developments. We have to look elsewhere to understand the rebalancing of US–Russia relations over the past decade and a half.

Closer to home, nuclear weapons didn’t stop Pakistan from occupying the forbidding Kargil Heights on the Indian side of the Line of Control in 1999, or India from waging a limited war to retake it—an effort that cost over 1,000 lives. If Mumbai or Delhi were hit by another major terrorist attack that New Delhi concluded had Pakistan connections, the pressure for some form of retaliation across the border might well prove stronger than the caution about Pakistan having nuclear weapons.

Nor do nuclear weapons buy immunity for North Korea. The biggest elements of caution in attacking it are its formidable conventional capability to hit the heavily populated parts of South Korea, including Seoul, and anxiety about how China would respond. Pyongyang’s present and prospective arsenal of nuclear weapons and its capacity to deploy and use them credibly is a distant third factor in the deterrence calculus.

If we move from historical and contemporary cases to military logic, strategists face a fundamental and unresolvable paradox in ascribing a deterrent role to the bomb. To deter a conventional attack by a more powerful nuclear adversary, each nuclear-armed state must convince its stronger opponent of its ability and will to use nuclear weapons if attacked—for example, by developing tactical nuclear weapons and deploying them on the forward edge of the battlefield. But if the attack does occur, escalating to nuclear weapons will worsen the scale of military devastation even for the side initiating nuclear strikes. Because the stronger party understands that, the existence of nuclear weapons will induce extra caution. But it won’t guarantee immunity for the weaker party.

For example, ASPI’s Andrew Davies believes that, while there is little realistic likelihood of an outright invasion by China, sea-based or air-launched long-range strikes against Australian targets are imaginable. Suppose that were to occur, and suppose further that we had acquired the sort of high-yield nuclear bombs and long-range delivery systems that Hugh White mentions. Would we really threaten China with nuclear retaliation? What if it didn’t find our threat credible and persisted with its strikes. Would we carry out nuclear first strikes against Chinese targets? If we don’t, China will have called our bluff on a non-credible threat. If we do, perhaps to maintain ‘credibility’, we will have entrapped ourselves in a posture of mutual nuclear suicide in the name of national defence. These scenarios, too, need to be thought through to their logical conclusions instead of being left in the realm of elegant abstraction.

And at what financial cost in an ever more competitive fiscal environment? A common but mistaken belief is that nuclear weapons enable defence on the cheap. To the contrary, not only is there no diminution in the need for and costs of full conventional capabilities, but there are additional costs related to the safety and security requirements that cover the full spectrum of nuclear weapons, material, infrastructure, facilities and personnel. And, as Britain and France have discovered, investment in the essentially unusable nuclear deterrent can take funds away from conventional upgrades and expansion that are actually usable.

We’ve made something of a fetish of our belief in the benefits and virtue of a rules-based order. The 2017 foreign policy white paper notes that the leash function of strong rules is ‘becoming more important to Australia as the distribution of power changes in the international system’. After its breakout in 1998, we strongly condemned India for violating the NPT-centred global nuclear order. We have backed international action to contain Iran’s suspected nuclear ambitions in the past, and we continue to demand that North Korea comply with the non-proliferation obligations under the NPT.

Australia, too, is firmly bound by NPT obligations, and for us they’re reinforced by obligations under the South Pacific Nuclear-Free-Zone Treaty. It would take us a long time to recover from the stench of hypocrisy if we were to discard treaty obligations as a mere inconvenience when we’ve consistently rejected security arguments by others as justifications for getting the bomb.

In 2003, the Bush administration caused a furore by proposing a new type of nuclear weapon, formally called the ‘Robust Nuclear Earth Penetrator’ (RNEP), to defeat hard and deeply buried targets. The term ‘mini-nuke’ quickly became associated with the plan because of the ‘dial-a-yield’ feature of the design. Critics contended that the mini-nuke with lower yields blurred the distinction between conventional and nuclear war and made the use of nuclear weapons more likely. It was a return to Cold War–style tactical weapons, like Davy Crockett (a nuclear recoilless gun whose 20-ton TNT equivalent yield generated a lethal radiation zone that exceeded its launch range).

The trend towards lower-yield weapons has continued. In 2016, the US began development of the B61-12, which has a precision glide capability comparable to the conventional Joint Direct Attack Munition (JDAM), delivering accuracy of around 30 metres. Once again, it has dial-a-yield feature, with a lowest setting of 0.3 kilotons (300 tons), and it has an earth-penetrating capability. So, with the B61-12, RNEP lives, and once again this raises the issue of enhanced usability.

Since the end of the Cold War, Washington’s attention to nuclear strategy has waned. Nuclear weapons have tended to be seen in the West in the context of arms control and disarmament. That stance was bolstered by the increasingly strident efforts of proponents of abolition and nuclear ‘global zero’ to rid the world of nuclear weapons, off the back of President Obama’s 2009 Prague speech:

So today, I state clearly and with conviction America’s commitment to seek the peace and security of a world without nuclear weapons. This goal will not be reached quickly—perhaps not in my lifetime. It will take patience and persistence. But now we, too, must ignore the voices who tell us that the world cannot change.

Obama added a caveat to this soaring rhetoric, noting that ‘as long as [nuclear] weapons exist’, the US would ‘maintain a safe, secure and effective arsenal to deter any adversary, and guarantee that defense to our allies’.

Neither Russia nor China seems ready to jump on the abolitionist bandwagon, as they are too busy with nuclear force modernisation of their own. Nor do the Indians or the Pakistanis, who nervously point their nuclear weapons at each other. Clearly, denuclearisation is the last thing on Kim Jong-un’s mind, and the Iranians, having reaped the financial benefits of the Joint Comprehensive Plan of Action, still have their nuclear option in their back pocket when the deal sunsets in 2025. In the interim, they can continue to collaborate on nuclear weapons technology and missile systems with North Korea.

In contrast, US nuclear forces are ageing out. If we want a credible nuclear deterrent for an uncertain future, it’s clear that capability development must proceed. The modernisation of tactical nuclear weapons such as the B61-12 must go hand in hand with strategic force development, including replacing the Minuteman III ICBM, building sufficient numbers of B-21 Raider bombers to make it an effective capability, and proceeding with the Columbia-class submarines that will replace Ohio-class boomers.

We must also get serious about nuclear strategy again, by re-emphasising deterrence as the basis for nuclear force development rather than seeing nuclear weapons through the prism of arms control only. Deterrence works by preventing an opponent from acting in a manner inimical to our interests. If nuclear deterrence is effective, weapons don’t need to be used. That in turn demands a high degree of assurance that nuclear forces are in fact usable. Deterrence isn’t about bluffing. We shouldn’t recoil from discussing nuclear warfighting in the context of reinforcing deterrence, because to proceed under the illusion that deterrence doesn’t imply credible, usable forces misconstrues the concept. Mini-nukes and the B61-12 make sense if they reinforce deterrence credibility and thus make nuclear war less likely.

With that thought in mind, there has been growing interest in the Trump administration’s new Nuclear Posture Review, which is reportedly considering a greater focus on the use of mini-nukes, particularly as a counter to Russia’s pre-emptive de-escalation doctrine (i.e. ‘escalate to win’). It’s unclear whether that would involve new nuclear weapons, with the potential requirement for new nuclear testing, or an intent to work towards operational deployment of the B61-12. But much of the critical commentary has focused on such weapons being usable because they imply lower yields. Critics refer to a recent Defence Science Board report that advocates offering the president ‘rapid, tailored nuclear options for limited use should existing non-nuclear or nuclear options prove insufficient’ (PDF, p. 24). That seems to imply a return to limited nuclear war. Groeteschele notes in Fail-Safe that limited nuclear war is ‘not theoretical’. We’ve never fought one, but we do real planning for such an eventuality, and for good reason.

Having options beyond immediate surrender or suicidal all-out thermonuclear war is logical. The question of whether a limited nuclear war could be fought remains unanswered, but embracing an illusion epitomised by nuclear taboos leaves us unprepared for a more dangerous future. Discussion of mini-nukes and limited options represents a return to thinking about nuclear strategy and how it will drive future capability. That is a welcome step.

With nuclear modernisation programs under way across a range of countries, Russia asserting its right to deploy nuclear weapons in the Crimea, NATO reviewing the role of nuclear weapons in the alliance, and a recent report in the US arguing for a more versatile arsenal of tactical nuclear weapons, it’s clear the world’s revisiting an old problem: how to build effective nuclear deterrence arrangements.

Since the end of the Cold War, thinking about deterrence issues has been mainly confined to the academic and think-tank world. But policymakers are now having to re-engage with those issues. And the problem has a new twist: we no longer enjoy the luxury of a bipolar world. Indeed, as Therese Delpech observed in her RAND monograph Nuclear deterrence in the 21st century, nowadays ‘the actors are more diverse, more opaque, and sometimes more reckless’.

Done properly, deterrence is a contest in threats and nerve, or—to use Thomas Schelling’s phraseology—‘the manipulation of risk’. (The chapter so titled in Schelling’s Arms and influence is a great starting point for anyone wanting to think through the broader deterrence problem.) That helps explain why some thought the concept ‘ugly’. It’s hard to make a policy threatening massive damage to societies and civilians sound noble and aspirational. Still, the bad news is that the alternatives are worse. And if deterrence is going to remain the dominant approach in nuclear weapon strategy, we need to fit the strategy to the contemporary geopolitical environment.

Historical experience of the deterrence problem is greatest in relation to two competing superpowers, separated by intercontinental distances, endowed with the resources to manage challenges, and both knowing well the costs of major war. We’ve had relatively little experience of nuclear deterrence in contests between giants and midgets (US v North Korea), between established and fast-rising powers (US v China), and amongst players in a multipolar system. Even our understanding of the role nuclear deterrence plays in relations between regional rivals (think South Asia) remains under-developed. It’s entirely possible that the old superpower deterrence model might not fit those new challenges well. Indeed, maybe the old model doesn’t even fit the US–Russian strategic relationship well these days: Russia’s no longer governed by a sclerotic CPSU.

Some years back INSS’ Elaine Bunn (now a senior official in the Obama administration) wrote a paper unpacking the notion of ‘tailored’ deterrence introduced in the 2006 Quadrennial Defense Review. True, deterrence has always been characterised by particular strategic wrinkles, but Bunn’s paper was an attempt to bring those wrinkles to the fore in relation to the possibility of a nuclear-armed North Korea, Iran, or transnational terrorist group. Her exploration of three different forms of tailoring—tailoring to specific actors and specific situations; tailoring capabilities; and tailoring communications—helps to illustrate the growing complexity of the deterrence challenge.

It now seems likely that we’re headed back into a set of complicated deterrence debates. A strategy that might make sense in one strategic setting—for example, a degree of restraint by a giant engaged in a conflict with a midget—might well risk flagging unintended messages in another. In the giant–midget case, almost any crossing of the nuclear threshold by the giant risks imposing a set of desperate choices on the midget’s leadership, and desperate choices tend not to be good ones.

Deterrence in the context of an established power versus a fast-rising power has a different wrinkle. One effect of a deterrence-dominated world is to reward passivity over initiative. As Schelling notes, in the world of the arthritic, passivity tends to be the default choice. But fast-rising powers aren’t arthritic. Turning one aside from a revisionist agenda will probably be more challenging than deterring another established player.

Multipolarity brings its own wrinkles, including a more mixed set of adversarial relationships, asymmetrical contests, inadvertent signalling, and third-party exploitation of bilateral rivalries. Capability issues become more vexed: actors require the capabilities to deter and defend against another, but also the residual capabilities to remain a player in other contests. The pressure must surely be towards larger rather than smaller arsenals. And reputational issues become more dominant: just as Margaret Thatcher fought the Falklands War in part to show the Soviet Union that the West wouldn’t buckle in the face of force, so too players in a multipolar nuclear world will want to show resolve in one contest because of its implications for others.

Finally, and perhaps most controversially, deterrence turns upon a credible threat to cross the nuclear threshold if push comes to shove. During the 1960s the US advocated a doctrine of flexible response, arguing for a model of deterrence that would fail in small packets rather than in one catastrophic breakdown. Notwithstanding the giant–midget problem outlined above, there’s usually good sense behind such a doctrine: it makes deterrent threats more credible, avoids global annihilation in any initial crossing of the nuclear threshold, maintains a degree of ‘intra-war deterrence’ from the options still on the table, and optimises prospects for negotiated war termination. But historically the doctrine invited questions about the relative balance between usability and credibility in US nuclear policy—questions buried rather than resolved by the end of the Cold War.

Tailoring, messaging, usability, credibility, and thresholds: I suspect policymakers will soon be thinking about all those questions again, across a range of deterrence relationships.