A society of voluntary cooperation means that there is no legal initiation of violence. There is however legal violence when a criminal initiates violence against someone's person or property, and by doing so forfeits his own right not to be violated.I would definitely agree with Rod in that there are uses of violence I consider in some sense "legitimate", and others I consider in some sense "illegitimate". The problem with Rod's definition is it that includes property, and the interesting controversy is not about the initiation of violence, but rather about precisely what constitutes property.
Voluntary cooperation does not mean that violations of person and property will be absent or should go unpunished by a legal system. When a criminal initiates violence, the victim has the right to use violence for restitution and punishment.
Furthermore, if we take the literal meaning of the "initiation of violence" as "the first use or threat of physical force against a person" (few would constitute the use of physical force against a nonsentient object as a violation of that objects own rights; the use of force against "property" is a violation of the sentient owner's rights) then the acquisition of property entails the initiation of violence: I must be the first to use physical force against a person who attempts to expropriate or infringe on the use of my nonsentient property. To escape the conundrum, the Libertarian must define the "initiation of violence" as something other than the plain meaning of the phrase: It must mean the use of force in ways the Libertarian considers illegitimate. But that's the fundamental definition of every form of government: a type of government is defined by its distinction between the uses of violence considered legitimate and illegitimate. We have to wonder why Rod is trying to bullshit us by using an English phrase with an unambiguous and uncontroversial literal meaning to represent the abstract, non-defining case.
("Radical" Libertarianism, which does not include property, is coherent and has, as Eric Frank Russell describes, considerable appeal. It might or might not be viable or evolutionarily stable.)
There are only four definitions of "property". The first is the recognition of ownership by mutual consent. If, for example, I find your wallet on the street, I'm going to give it back to you because I do in fact agree that the wallet is your property. But if property is established by mutual consent, theft is logically impossible. If I actually do agree that some object is yours, why would I take it? Contrawise, if I take some object, by definition I do not agree that it is in fact yours. So on this definition there can be no "violence" against property rights; where violence exists, there is by definition no mutual agreement to establish property.
The second definition is "the objects a person can physically protect," i.e. those objects a person can successfully initiate actual violence to maintain possession or ownership of. This "might makes right" definition seems so obviously contradictory to Rod's formulation that we cannot attribute it to Libertarians, at least not the garden-variety sort.
The third definition is "the objects over which a person has ownership by virtue of a social process." This definition overcomes the problems of property by mutual consent and avoids the "might makes right" implications of the second definition. But the problem is that almost all modern societies, including our own, implement this definition of "property". (All societies everywhere throughout history implement this definition of property, if you include passive compliance with a king or other authoritarian regime as a "social process".) And if ownership can be granted by virtue of a social process, it can equally be taken away by that process. When the government taxes you, it is taking away ownership by the same means it granted it to you in the first place; it is inconsistent on this definition to assent to one and object to the other.
The fourth definition is "the objects over which a person has ownership by virtue of some natural right." This seems like the definition that most Libertarians would prefer to use. The problem with this definition is that any conception of "natural" rights are at least as complicated as any body of legislation. More importantly, there is no objective means to settle disputes over natural rights in the same sense that there is an objective means to settle scientific disputes. Trying to establish some basis for natural rights just moves the problem around: Why choose a basis that establishes a "natural" right to some object for one person over another that establishes a "natural" right for another person?
I'm in favor of the third option: property is ownership established by a social process. The question then becomes not about property per se or the initiation of violence, but what sort of social process we want, and what we want it to do or not do. If any Libertarian agrees with the third definition and would like to describe a social process that grants a "Libertarian" definition of ownership, I'm all ears, and I'll evaluate that candidate social process on the basis of whether it is in my own interests. If any Libertarian agrees with the fourth definition, and can give a coherent account of "natural" property rights that is at least as rigorously provable as scientific claims about natural law, I'm equally receptive.