Harknett and Yalcin (H&Y) replace the struggle for power (or the struggle for power and peace, as Morgenthau's famous subtitle has it) with a 'struggle for autonomy'. Rather than assuming that states always want power or security, H&Y say that the only motive that can be derived from structure -- i.e. from 'anarchy' (lack of a central authority) and the distribution of capabilities -- is a desire "to possess the capacity to act in a sustained manner that preserves and enhances [units'] capacity to act into the future -- they merely want to remain autonomous" (p.506; italics omitted). However, if units (states or non-state actors) don't survive they can't remain autonomous, a point to which I'll return.
Before getting into the weeds, I'll try to give a brief overview of H&Y's approach. Some of their points, when put simply, seem cogent clear enough, if debatable. A problem with the article, though, is that the points are usually not put simply but wrapped up in verbiage and often inelegant sentences. (In some -- not all -- scholarly IR journals, there seems to be no copyediting to speak of, which means that stylistically everything is left to the authors.)
H&Y argue that states respond to shifts in the distribution of power by adjusting their behavior, based on a calculation of what they want and what they think it's possible to get. When one state has a much larger share of power than all the rest, the other states will not embark on a likely fruitless revisionist quest to change the system. Rather, they will maneuver within it, seeking to enhance their freedom of action without directly challenging the leading state. (This, incidentally, is close to the situation Stephen Walt describes in Taming American Power, a book H&Y don't cite.) Conversely, when the distribution of capabilities is more even, states, H&Y say, will seek to advance their positions more directly and less subtly, leading to the likelihood of intensified security dilemmas and increasing the chances of major war. Put in this way, the argument is plausible enough to be tested, at least in a loose sense, through historical inquiry. The authors don't undertake such a test, though they do have a brief section at the end applying their theory to the Cold War, which they split into three periods characterized, they say, by different distributions of power between the U.S. and USSR.
The authors distinguish their approach from the extant varieties of structural realism, i.e. defensive and offensive realism, on the grounds that those approaches supposedly assume invariant state motivations (i.e., states always seek to maximize either security or power). According to H&Y, both defensive and offensive realism have erred by taking motives as given, rather than seeing motives as shaped by structure. They write:
[S]tructure shapes not only behaviors but also identities and orientations of agents. In offensive and defensive realist theories, state identities and motivations are defined and assumed independently from the shaping power of structural factors. States are taken to have a specific motivation whatever the constraints and opportunities of structural conditions. Structure does not affect the survival motivation in neorealist theories of IR, which assume that even if there is no direct threat to state survival, it is a survival instinct that is driving action. In contrast, in the structural autonomy theory developed here, units rearrange not only their behaviors, but also their identities and motivations in response to the distribution of power (p.502; italics in original).The description of structural realist theories in this passage is not, I think, entirely accurate. There is more going on in offensive realism than "a survival instinct." True, Mearsheimer does write that "the only assumption dealing with a specific motive that is common to all states says that their principal objective is to survive" (The Tragedy of Great Power Politics, p.32), but in an endnote (p.414, n.8) he writes: "Security concerns alone cannot cause great powers to act aggressively. The possibility that at least one state might be motivated by non-security calculations is a necessary condition for offensive realism, as well as for any other structural theory of international politics that predicts security competition."
H&Y insist that "one could assume a survival motivation" only in "an environment with no opportunity but full of threats" (p.509). The problem here, however, is that the word "motivation" has more than one meaning. "Motive," according to the dictionary I have at hand, can mean "some inner drive, impulse, intention, etc. that causes a person to do something or act in a certain way," but "motive" can also mean simply a "goal." And units can have a goal of surviving regardless of whether the environment is threatening or non-threatening. Indeed, as I mentioned before, a unit that doesn't survive obviously can't remain autonomous. In this sense autonomy presupposes survival. Now it is true that the rate of state death in the present international system is very low and thus survival is not often a motive in the causative meaning of "motive"; but survival can remain a goal even when the desire to survive is not an immediate driver of behavior. H&Y seem wrong to assert that the "traditional realist-ascribed motive of survival implies a logic in which helpless states would eventually require the delegation of state autonomy to a higher authority in a fearful environment populated by units wishing for survival" (p.509). Where do they get the notion that states in a "self-help" system are or may become "helpless"? It doesn't follow.
H&Y proceed to distinguish between "diffuse" and "concentrated" power structures. In diffuse power structures, a category that includes both bipolarity and multipolarity, gaps in power between units are relatively small; in concentrated power structures, as one might guess, power gaps are large and one unit is clearly more powerful than the rest. Diffuse structures are more war-prone, they argue, and against Waltz they maintain that bipolar systems (and balanced multipolar systems) are not more stable than the alternatives. They write:
In what in Waltzian terms would be considered a balanced system, states will feel less threatened (or more secure), but this does not translate into behavior promoting stability. Since the primary motivation is not security (or survival) but autonomy, "balanced" systems will not necessarily be stable, but rather have strong structurally induced incentives to change the power structure (and the relative distributions within it) to gain autonomy. The Cold War behavior of the two superpowers became more change-oriented during periods in which their power was more "balanced" with each seeking a breakout capacity via military technology, additional allies, exploitation of minor states (the competition over the Third World), or expanded realms of competition (the Space Race). The structure, itself, induced intense change-oriented policy, not stability-seeking on the part of the superpowers. (p.514)This is interesting, but there is at least one problem with the argument in this section: H&Y maintain that a diffuse power structure decreases autonomy ("the very existence of other actors with equal capabilities decreases the level of autonomy for all" -- p.514), but it's not clear why this should be the case given their definition of autonomy, quoted above, as "the capacity to act in a sustained manner that preserves and enhances...capacity to act into the future" (p.506). What seems to happen in this part of the argument is that H&Y take autonomy to mean the degree of a unit's freedom of action, but that's not how they define autonomy at the beginning of the piece.
Reading "The Struggle for Autonomy" raises, among other questions, the issue of what is the most promising direction for realist (or realist-inflected) theorizing about international politics. In their first footnote the authors say that "neo-classical realists with their multi-causal and multi-level frameworks are increasing the number of factors across the realist paradigm. This is a degenerative process from a structuralist perspective." Yet neoclassical realism emerged precisely because structural realism is limited both in what it can explain and how well it can explain it. Waltz's main substantive generalization -- that given states-under-anarchy there is a strong tendency toward balancing -- has received sustained and quite persuasive criticism, and Mearsheimer's view that supposed uncertainty about intentions pushes great powers to act aggressively (albeit calculatedly) to each other is, if anything, even more questionable. Harknett and Yalcin think the problem, in effect, is that Waltz and Mearsheimer are not structural enough. But is this indeed the problem, or does it rather lie in structural realism's inability to take into account factors that matter -- domestic politics, regime type, ideology, bargaining, to name several? Parsimony is not desirable if parsimonious theories can't explain important outcomes.
Neoclassical realists bring in perceptions and domestic politics not because they want a "hybrid" theory for its own sake but in order to explain outcomes more satisfactorily. Even sticking with H&Y's example of the Cold War, as Lobell et al. observe (see the introduction to this edited volume) the system's structure by itself can't explain why the U.S. after 1945 opted for containment rather than "competitive cooperation" with the USSR. (In the interest of keeping this post shorter than a mini-treatise, I will not go into H&Y's discussion of the Cold War in detail. Suffice to say it is open to criticism.)
In short, it's far from clear, at least to me, that a "refinement of structural realism" (H&Y, p.499) in an even more structural direction is the way to go. To be more direct, I think it's not. Those interested in these matters can of course read the article and reach their own conclusions.
Added later: There is another issue (well, lots of them but one I should have mentioned): The decreased likelihood of major war in the present period may have very much less to do with structural factors (i.e. with the "concentrated" power structure) and much more to do with long-term trends and factors that are not structural. Since I've discussed this fairly extensively elsewhere on the blog (see the "decline of war" label under topics), I won't harp on it further here.
Added still later: For Waltz on the survival motive, see Theory of International Politics, pp.91-2.