First, let me clarify what I mean by "Liberal society." It can mean many things, but for the purposes of this blog post, it means a legal regime that recognizes a person's legal right to pursue any lawful calling and to quit his or her job whenever he or she wants. A corollary states that the liberal state, as part of its obligation to protect its citizens those others who reside in its jurisdiction, has the sole prerogative to use force and thereby denies to others the use of force to constrain others' choice of where and whether to work. No society, including the United States, is perfectly "liberal" in this sense, but I wager that the U.S. is at least liberal enough to fit the bill for the average case. I say "liberal society" and not "liberal legal regime" because "liberal society" sounds better and "liberal legal regime" is too toothy and cumbersome. But in this blog post, I mean "liberal legal regime."
Now, on to strikes. During a strike, the union's goal is usually to convince its members to abstain from work, to withhold their labor collectively, in order to convince the employer to negotiate with the union or if the employer agrees in principle to negotiate, to convince the employer to adopt certain contract terms the union finds acceptable. Another, closely related goal, is to advertise to others that there is a labor dispute and, especially if the employer is involved in a customer-oriented industry, to induce customers to boycott the employer's product.
This description is quite a wide brush stroke, because there are shades of other reasons people and unions strike: "hate" strikes (when, for example, white workers refuse to work alongside black workers); "syndicalist" strikes (workers strike over one or even several points of dispute, but don't necessarily seek a contract, preferring instead to use shop-floor revolts to assert their preferred labor conditions); "sympathy" or "secondary" strikes (striking to support workers in another dispute or refusing to cross picket lines simply because they are picket lines even if the workers who so refuse are not involved directly in the dispute). I think these are all related to the general phenomenon of strikes I am writing about, but I will focus on the simpler case of workers in a shop trying to withhold their labor to force a concession from their employer.
To address the strikers' problem, I am positing a situation that does not always occur, namely, a situation in which the employer tries to use replacement workers. Sometimes, of course, the employer prefers to wait out the work stoppage. (In some cases, the employer may welcome the work stoppage, such as when business is slow and keeping people on the payroll represents more an added expense, or when, especially during an industry-wide strike, the work stoppage creates a shortage in supply of whatever good is produced and creates a corresponding increase in demand.) In such a situation, where the employer introduces replacement workers, he or she is trying to undermine the union's attempt to withhold labor.
What's the union to do? In a liberal society, not much, at least not much that is legal. The union pickets can shame the replacement workers (e.g., call them "scabs"), can try to reason with the replacement workers (e.g., explain why they are striking in an effort to convince them not to cross the line), can try to offer inducements not to work (e.g., during a smelter strike in Denver in 1899,the union offered food vouchers, redeemable at sympathetic stores, to replacement workers who pled hunger as a reason for crossing the line. The goal was to encourage these workers to honor the strike).
What strikers cannot do in liberal society, inasmuch as the state's monopoly over "legitimate coercion" is concerned and inasmuch as the state is truly "liberal" in the sense I discussed above, is use "coercion" to prevent the replacement workers from crossing the picket lines. Strikers cannot legally, for example, threaten replacement workers who cross the line, and (I assume) they cannot legally form some sort of "human barricade" to prevent free passage of the replacement workers. The state in the liberal society, again inasmuch as that society is liberal and the state aligns itself with that liberalism, is charged with ensuring that people's right to pursue any lawful calling. It might use its own "legitimate" mechanism of force--the police or national guard--to ensure that right when it is allegedly endangered by strikers.
In case it isn't clear from what I wrote above, I am hinting toward the specter of "labor violence." In American labor history, and most other "[Name-the-nation] labor histories," there have been occasional and spectacular instances of violence arising in the midst of strikes. For America, think of the 1877 railroad strikes, the Pullman Strike in 1894, and the "Herrin Massacre" during the 1922 coal strike. There are also more punctuated and more targeted instances of violence that might be ascribed to labor radicals, such as the attempted assassination of GM CEO Alfred Sloan. But I'm focusing more on the instances where the goal of violence is to discourage strikebreaking.
These instances illustrate a potential problem for strikers. If the state does its job to protect people's right to work where and when they want, it must at some point draw a line against activities that can be described as coercive. We can probably mostly agree that certain activities fall among the proscribable activities within the competence of the state qua state. Taking a baseball bat and battering a replacement worker, for example, is "violence" by almost every recognized sense of the word "violence." I suspect that even the very union-friendly labor historians whom I know would acknowledge such battery to be "violence," even if they (or some of them) insist that the violence is somehow justified. It is also, I believe, at least understandable that the state sees its responsibility in such a situation as to prevent the violence. Even those same labor historians who might claim that in some cases striker-initiated violence be justified might recognize that the state has an interest (even if they say it's not a defensible interest) in asserting its prerogative to control or prevent violence.
There are some instances where the line is fuzzier. Some of the tactics I mentioned above that strikers might use might conceivably be considered coercion. Shouting "scab!," for instance, might imply a threat that something worse than being called a name will happen to the replacement worker. Setting up pickets to "inform" passersby that a labor dispute is going on might at least by implication have an effect that is more strident and more robust than simple education on current events at the factory. Sending a "conscience committee" to represent "the situation" to recalcitrant replacement workers is to my mind even iffier
And herein lies one problem. Where and how ought the state to draw the line? In the early 20th century, up to about 1930, the state sometimes drew the line at places we of today's sensibilities might bristle at. During a smelter strike in Denver in 1903, for example, a state judge issued an injunction that forbade strikers not only from threatening replacement workers, but from advertising the strike. The context for this injunction, it should be noted, was a very destructive move by the union in question when it commenced its strike (the workers left the smelters without drawing down the furnaces, an action that essentially ruined the part of the property in which the melted ore froze over; one might also mention incidents of outright fighting between union members and replacement workers.) Still, extending the prohibition the mere advertising of the strike likely seems too far to most of us in the U.S. in 2012. 
And another problem for strikers, it seems to me, is that they--most of them, with some exceptions as always and allowing for the inconsistency of thought-in-practice that afflicts almost all humans--sincerely buy into the notion that "informing" and "persuading" and "shaming" are as far as they ought to go. I'll even venture the hypothesis that whatever strikers' actual feelings, which might be a historically contingent fact, in practice they have usually acted this way. I suspect in other words that if one took a tally of every single event in history that could count as a strike, one would find the vast majority to include no instances of what most people would identify clearly and unambiguously as "violence."
By saying this is a "problem," I do not intend to interject a criticism. Some apologists for violence and some champions of the working class and some advocates for "labor militancy" see such a disinclination to violence as a sort of "false consciousness."  But I'm inclined to see it as a sincere respect for some standard of right and wrong. And I happen to agree with that standard. I'm no pacifist, but I do think that violence, domestic or foreign, ought to be either a last resort, or an only resort, to obtain "a vitally important end." As desperate and "vitally important" as a dispute over wages and working conditions on jobs can seem--and I admit, I have never myself been in the severe and degrading circumstances that have in the past been the context for much of labor violence--I don't think these issues necessarily become therefore "a vitally important end" in the sense that I intend it.
I'm saying, rather, that (most?) strikers' notion that violence is unacceptable is a "problem" in the sense that it's a conceptual difficulty strikers deal with. They might speak as if the jobs they are striking for are "their" jobs, with perhaps a whiff of a belief in a proprietary interest in these jobs. They might even speak as if the jobs are extensions of themselves in a way that, say, a mere piece of non-real property is not an extension of oneself. They may believe, in short, that something approaching 'self-defense" is on the line, and yet I posit that most of them, at least today, would not endorse actual, according-to-Hoyle, physical violence.
What I'm talking about is more than the difficulties of simple line-drawing, although the difficulties of line-drawing are an issue. I'm suggesting also that there is a conflict in values, and a conflict over the legitimacy of those values most of us in the U.S. in 2012 are inclined to accept. We place a value on life and safety from violence. We place a value on peaceable assembly. We place a value on freedom of association and the attendant freedom of speech to promote one's views. We place a value on personal autonomy. All of these values enter the picture when it comes to strikes.
I ask my pro-union friends and my fellow labor historians to keep in mind some of the costs involved when they promote "labor solidarity" or even more modest, institutionalized tools to help unions, such as union shop laws (which, for lack of a better term, is my name for what are called "right to work" laws).
 Perhaps this is more of a feature of "states" in general, whether liberal or not. Max Weber supposedly elucidated this concept--a monopoly on legitimate coercion or a monopoly on legitimate violence--as the defining feature of the state. The one source that people often link to in order to demonstrate his definition is an essay he wrote on "Politics as a Vocation." [click here] I must not have read it closely enough or understood it as well as I ought to have, but although Weber in that essay does mention the state as that which observes a monopoly on violence, he does not seem to theorize the concept all that much in the essay, or at least not as much as I would expect from the instances in which his name has so often been invoked in support of the concept. I have some reservations about the concept, especially the "legitimate" part of "legitimate violence." And I also wonder about the many instances of what can be called "violence" that are in some way tolerated by (or more accurately ignored by) the state, as well as other practices that might be deemed "coercive" but nevertheless accepted (or ignored) by the state. There seems to be a question-begging aspect to this definition that bothers me. Again, as I'm not a theorist or a political scientist, and I don't mean to question that definition overmuch. And I recognize that I may simply be misunderstanding Weber's essay. I do think, however, that any "liberal state" worthy of the attribute "liberal" would protect people's basic prerogative to pursue any lawful calling and, especially, to quit their jobs. My own definition of "liberal society" is question-begging in this way (how is a calling "lawful" in a way that doesn't invoke the definition of "lawful" in the first place).
 I'm not a fan of the word "scab." People who serve as replacement workers are often in very marginal positions and I wager that as often as not are in worse situations than the strikers. To smear them as "scabs" represents, in my view, a striking lack of compassion or empathy. Of course, maybe I'd feel differently if I were the striker and someone was crossing picket lines to take "my" job.
 I'm no economist, and perhaps I am using "shortage of supply" and "increase in demand" incorrectly, as far as the discipline of economics is concerned. However, I wager that in some cases, employers and sometimes even unions think along the lines of limiting supply and increasing demand. This way of thinking appears to have been true of the administration of the United Mine Workers and the coal operators' associations of the "central competitive field" of bituminous mines (mines located primarily in Illinois, Indiana, Ohio, and western Pennsylvania) from c. 1898 through c. 1910.
 I should note I am purposely bracketing, for the sake of this blog post, state and quasi-state violence that goes over the line. One might cite President Cleveland's zealous use of soldiers--over then Governor John Altgeld of Illinois--to break the Pullman Strike on the disingenuous pretext that the federal government was merely protecting the free-flow of the mails (I say "disingenuous" because the union offered to facilitate the free-movement of rail cars carrying U.S. mails.) One might also cite the violent liberties various "citizens alliances" took during the Colorado "labor wars" of 1903-1905 (and in other states in other years), under the tacit--and sometimes explicit, with the aid of national guardsmen--approval of the governor, George Peabody. Any discussion of "the state and the unions" ought at least to acknowledge that the state goes (sometimes? often?) overboard in its exercise of powers against workers.
 I name all three separately--"apologists for violence" and "champions of the working class" and "advocates 'labor militancy'"--because I see them as discrete positions. I submit that even if those who hold these positions sometimes overlap, they need not necessarily overlap.
 Maybe sometimes it can be. I can imagine scenarios--company town, company store, little possibility of exit or "voting with one's feet," robust and arbitrary enforcement of "vagrancy" and "criminal breach of contract" laws--where the resort to violence approaches something that can be justified. Even in those instances, I believe that violence is to be considered an altogether regrettable, even if necessary, action. At any rate, it behooves me, as jester at scars who never felt the exact same wound, to keep in reserve at least some of the moral judgment I might mete out in such situations.
 Per my discussion in footnote  above, I have concerns about what is and is not "coercion" in the sense entailed by "monopoly on legitimate coercion." It seems to me if one takes the definition attributed to Weber very far, one must claim that the state has at its core pretensions that are essentially totalitarian. Maybe that's the point and maybe it's supposed to be a thought-provoking parable on the dangers of (too much) statism. I confess that my thoughts have been so far provoked, but I'm not sure what else I can do with the concept other than reassert the definition.
 I suggest that even at the points where such values cause such disagreement--abortion; the death penalty; time, place, and manner restrictions on some forms of speech; campaign finance reform; the PPACA mandate--people on most sides of these issues, at least when pushed gently, acknowledge the existence of these values as legitimate and state either that the conflict is not rally a conflict of values, or that exceptional circumstances require the circumscription of these values.