In his seminal book The State ( 1998) Anthony de Jasay draws attention to the anomaly that a minimal state would have to devote its coercive power to no other purpose than to keep itself minimal. Once a state exists, those in control of its fundamental coercive power will use it for their particular redistributive purposes and go beyond the minimum redistribution necessary to maintain the minimal state. If anarchy is under the threat of being destabilized by the invention of statelike organizations (external or internal to it), so is the minimal state by its proclivity to grow beyond any limits.
Within a politically realist perspective, neither anarchy nor the minimal state forms a stable equilibrium of social interaction. Nevertheless, Jasay prefers anarchy to a minimal state as an ideal. He believes that we should not even in theory accept as legitimate what by its nature implies the use of fundamental coercive power of collectivities--that is, a power to which its subjects did not agree in prior explicit or implicit acts of free assent. Any philosophical approach that legitimizes collective coercive power in some way or other will pull down our defenses.
For instance, for Jasay, Herbert Hart's famous rule of recognition (see Hart 1961), the social acceptance of which explains how a legal order can exist in the first place, is in truth a rule of submission. Any standard that transmits any legitimacy to state coercion beyond acknowledging that coercion as a brute fact will pave the way for expanding the state's role beyond any proper limits. To expand the state's reach, those in control of the state will use the willing obedience of those who apply the rule of recognition.
Of course, Jasay acknowledges the fact that states do exist, but he commends that in adapting to this fact, we always stick to the anarchical reservation that their existence is fundamentally illegitimate. However sophisticated theories may be, there simply is no fundamental justification for fundamental coercive power.
Jasay's theory--like, say, John Rawls's theory of justice or any other sophisticated political philosophy--is not directly about the state but rather about how we should think about the state. In particular, Jasay resents those theories, most notably contractarianism, that describe the existence of what is decidedly not the outcome of free consent as if it were the result of voluntary agreement. This description of it as if it were freely agreed upon merely camouflages the state's dangerous coercive nature. It lets the powerful have their way more easily. Combining contractarianism with concepts of democratic legitimacy and the "logical foundations of constitutional democracy" (the subtitle of James Buchanan and Gordon Tullock's book I he Calculus of Consent [(1962) 1999]) amounts, in Jasay's eyes, to deploying weapons of ideological mass destruction of liberty by means of political theory. For him, the only legitimate route to creating social institutions is assent-based unintended growth of conventions (for a more precise model of such a Humean-Hayekian process, see Sugden  2005).
To understand Jasay's views on this issue, it is most important to see that for him--as for Thomas Hobbes--no "natural law" or system of "natural rights" preceded the evolution of institutions and the conventions underlying those institutions. (1) There is no a priori standard of reason against which the conventions can be normatively assessed. Jasay subscribes to the a posteriori nature of all obligations and rights. They are brought into existence only when social conventions are created. In the original process of convention formation, individuals cannot be blamed--there is no standard for that at this point--for agreeing to whatever seems fit to them (i.e., they cannot be blamed for creating any externalities). Before conventions, there is no nonconventional and, of course, no "conventional" normative standard that can be used to criticize what individuals do unto others or unto themselves. All normativity is created by conventions.
Jasay, like the rest of us, needs to come to terms with the facts of life. Though "the world until yesterday" (Diamond 2012) was anarchic, we live in a world of states today. The economies of scale in warfare make truly anarchic existence outrageously unrealistic in the world today. (2) In my subsequent tribute to the depth of Jasay's reflections on the state and on the necessity to uphold anarchist ideals in keeping the state--unavoidable as it may be--under control, I explore whether the ideal that I introduce under the name "defensive state" may help. Can we acknowledge the unavoidability of the state conceptually and still keep intact the anarchist reservation against the state's legitimacy?
The defensive state that I envision is akin to Chandran Kukathas's (2003) "liberal archipelago." It is not protective of preexisting rights--preceding the existence of conventions--nor is it protective of some set of state-sponsored rights and obligations. Instead of requiring that individuals should be left to make their own decisions within a predefined system of rights, as the adherents of the notion of such a "protective state" typically do, the ideal of the defensive state requires that individuals be left alone in the first place to develop their conventions defining obligations and rights. The existence of the defensive state that strives to keep itself minimal is, of course, as much an anomaly as that of a minimal state. And it is as unstable as anarchy.
As far as realism is concerned, anarchy, the minimal state, and the defensive state are all in the same boat. Acknowledging this weakness of the very concept of a defensive state, I intend to explore whether an ideal theory of anarchy or an ideal theory of a defensive state may serve us better in defending our liberties against intrusions by the collective bodies under which we of necessity must live. (3)
From Anarchy to the Defensive State
If there were no state, individuals would automatically and trivially be left alone by the state. However, in anarchy there is no guarantee that no state will emerge. As history has proved, a state or statelike institution may in fact emerge. The economies of scale in aggression and warfare are such that it takes a state to defend us against the state once the state has been created (somewhere). Therefore, though in anarchy there may be no justification to found the state in the first place, (4) once that state exists (somewhere) we need a state to defend us against it and other states.
The "reluctant archist" accepts the fact of the state's existence while evaluating it according to how well it approximates ideals of anarchic liberty in a world of states...