For forty years I have studied the stunningly complex problem of enemy rationality, especially in certain earlier published writings concerning the particular nuclear threat from Iran. Almost by definition, strategic assessments of nuclear deterrence always assume a rational state enemy; that is, an enemy that values its own continued survival more highly than any other preference, or combination of preferences. But this assumption may sometimes be problematic.
There is no reason to assume that all prospective attackers of the Jewish state will always rank physical survival above all other possible options or even that such attackers would hew perfectly to careful, systematic, and transitive comparisons of all expected costs and benefits. As long as such enemies are capable of missile attacks on Israel, and as long as Israel is unable to intercept these attacks with near-perfect, or possibly even perfect reliability (no system of ballistic missile defense, including Israel’s Arrow and Iron Dome, can ever be leak-proof), any Israeli dependence upon nuclear deterrence could have genuinely existential consequence.
Where should Israel go from here? Recognizing the substantial limitations of the so-called Middle East peace process, Israel must seek security beyond the protections offered by nuclear deterrence. It must, as recommended by Project Daniel (2003), also prepare for possible preemptions against pertinent military targets. Although many will find even such preparations to be “aggressive” or “uncivilized,” and although it may already be very late operationally in certain relevant attack scenarios, the alternatives may amount to national suicide. Significantly, the right of preemption is well established under customary international law as “anticipatory self-defense.”
Among other purposes, Israel needs nuclear weapons to undertake and/or support various forms of conventional preemption. In making its preemption decisions, Israel must determine whether such essential defensive strikes, known jurisprudentially as expressions of anticipatory self-defense, would be cost-effective. This would depend upon a number of critical variables, including: (a) expected probability of enemy first-strikes; (b) expected cost of enemy first-strikes; (c) expected schedule of enemy unconventional weapons deployment; (d) expected efficiency of enemy active defenses over time; (e) expected efficiency of Israeli active defenses over time; (f) expected efficiency of Israeli hard-target counterforce operations over time; (g) expected reactions of unaffected regional enemies; and (h) expected U.S. and world community reactions to Israeli preemptions.
Regarding preemption options, Israel’s overriding question is this: As Jerusalem must plan for such forms of anticipatory self-defense, against which particular configurations of hard targets should they be directed, and when should they be mounted? If it is assumed that enemy states will only add to their chemical/biological/nuclear arsenals, and that these additions will make any effective Israeli preemptions more and more difficult, if not altogether impossible, rational Israeli strategy would seem to compel Jerusalem to strike defensively as soon as possible. If, however, it is assumed there will be no significant enlargement/deployment of enemy unconventional weapons over time, this may suggest a diminished strategic rationale for Israel to strike first.
Israel’s inclinations to strike preemptively in certain circumstances could also be affected by the steps taken by prospective target states to guard against any Israeli preemption. Should Israel refrain too long from striking first, enemy states could implement protective measures that would pose additional hazards to Israel. These measures could include the attachment of certain launch mechanisms to nuclear weapon systems, and/or the adoption of “launch-on-warning” policies. Such policies would call for the retaliatory launch of bombers and/or missiles on mere receipt of warning that a missile attack is underway. Requiring launch before the attacking warheads actually reached their intended targets; launch-on-warning could clearly carry very grave risks of error.
Ideally, Israel would do everything possible to prevent such enemy measures from being installed in the first place, especially because of the expanded risks of accidental or unauthorized attacks against its armaments and population centers. Yet if such measures should become fact, Jerusalem might still reasonably calculate that a preemptive strike would be cost-effective. This is because an expected enemy retaliation, however damaging, might still appear less unacceptable than the expected consequences of enemy first strikes.
Perhaps the single most important factor in Israeli judgments on the preemption option will be the expected rationality of enemy decision-makers. If, after all, these leaders could be expected to strike at Israel with unconventional forces irrespective of anticipated Israeli counterstrikes, deterrence, as we have already seen, might not work. This means certain enemy strikes could be expected even if enemy leaders understood that Israel had successfully deployed its own nuclear weapons in survivable modes, that Israel’s weapons were entirely capable of penetrating enemy active defenses, and that Israel’s leaders were fully willing to retaliate.
Faced with an irrational enemy bent upon unconventional aggression, Israel could have no effective choice but to abandon reliance on traditional modes of nuclear deterrence. Even if it is not faced with an irrational enemy, however, Israel will have to plan carefully for certain preemption options, planning that must take into account Jerusalem’s own nuclear weapons. In the course of such planning, it will be important to recognize that enemy capabilities and intentions are not separate and discrete, but interpenetrating, interdependent, and interactive. This means: (1) capabilities affect intentions, and vice-versa; and (2) the combined effects of capabilities and intentions may produce certain policy outcomes that are greatly accelerated, and/or are more than the simple sum of these individual effects.