Skip to main content
SearchLoginLogin or Signup

The Promises and Pitfalls of Path Dependence Frameworks for Analyzing Penal Change

Published onOct 01, 2021
The Promises and Pitfalls of Path Dependence Frameworks for Analyzing Penal Change


Although the study of penal changes throughout history is central to punishment studies, the field has taken little from historical institutionalists’ theories of institutional change. One of the most relevant such theories is path dependence. This article outlines path dependence frameworks’ most fruitful elements for studying penal change. Drawing on foundational political science and historical sociology texts, as well as several punishment scholars’ works, this article highlights the advantages of thinking through stasis and change, mechanisms of inertia such as feedback effects, and exogenous shocks. While path dependence offers a powerful framework, it can also be an unsatisfying explanation at times, particularly when path dependence is itself a seemingly uphill battle, when apparent stasis hides ongoing change, or when institutions survive hypothesized mechanisms of change. This paper closes by discussing some ways in which punishment scholars can strengthen the path dependence framework by blending it with recent theoretical developments in the punishment studies field.


Explaining penal change—such as the rise of the prison, the decline of capital punishment, the popularity of a new justification or theory of punishment, or the adoption of a new template for existing punishments—is one of the longest-standing core concerns for punishment scholars. Penal change has been central to classic studies by Émile Durkheim (1893), Marxist scholars (e.g., Thompson 1975), and Michel Foucault (1977), among others, as well as to more recent field-defining scholars like David Garland (1985). Punishment studies of the last several decades have been consumed by questions about one of the biggest changes to punishment in recent history—the various penal changes associated with late modernity around the Western world, including the rise of mass incarceration in the United States. As the field has expanded, so have new theories about penal change throughout American history: Campbell and Schoenfeld (2013) have offered a political sociology of punishment, Goodman et al. (2017) have offered an agonistic theory of penal change, and I have offered a neo-institutional account of penal change (Rubin 2019d).

Many popular theories explaining penal change reflect the lasting and foundational impact of classic social theory.1 Such accounts often focus on the macro- and especially national or international level. They often locate penal change’s origins in some larger element of society—culture, economics, politics, etc.—or in significant subsets of society that cut across these domains—institutional fields—which then create the conditions for, or facilitate, change in the penal domain. A major motif within these theories is that punishment is not dependent on changes to crime, but rather on changes in society. In contrast, few works within the punishment literature rely on established theories of institutional change from historical institutionalism.

Theories of institutional change operate at a meso-level, focusing on changes to specific institutions and organizations—banking systems, the U.S. Congress, worker unions, political parties, or education systems (see, e.g., Campbell 2004). In the punishment sphere, these include capital punishment, the prison, individual prisons, probation, and individual probation agencies, for example. Notably, some kinds of institutionalist accounts have been relatively popular. Multiple punishment studies fall under the banner of institutional analyses by examining “institutional arrangements” and how they shape penal ideas, policies, and trends (Savelsberg 2014). Political institutionalist theories, which focus on the relationships between various political institutions (and thus overlap somewhat with conventional social theories of change), have been slowly shaping the field for several decades now (e.g., Barker 2006; Savelsberg 1994; 2018). These theories have been more popular than historical institutionalist theories, which share a focus on many of the same mechanisms but remain under-utilized (but see Campbell and Schoenfeld 2013; Lacey 2013; Schoenfeld 2014). Indeed, given the field’s emphasis on penal change, it is rather surprising how little scholarship in the field draws from historical institutionalist theories of institutional change.2

Historical institutionalist theories of institutional change emphasize the timing or sequences of change and the timing-related mechanisms of change and stasis, focusing attention both on what enables and what stymies change. Rather than identifying generalizable social variables (like labor market shifts or political change), these theories identify specific patterns of change (evolution, conversion, layering, displacement, drift) or elements (like contingency) and mechanisms of change (critical junctures, feedback effects, sequences). Specific instantiations of these theories applied to empirical examples often identify important social contextual factors relevant to the change; but in contrast to more social theoretical explanations, the analytical focus is the way the change occurs, not the specific social contextual factors.

Perhaps the most recognizable and formalized of these historical institutionalist theories is path dependence, which, at its most basic level, identifies the ways in which early conditions can create lasting and important consequences for a given institution or organization’s course of change or stasis. As an analytical framework, path dependence is fairly rare within punishment studies (for important exceptions, see Dagan and Teles 2014; Rubin 2016; Schoenfeld 2010; 2018).3 This article seeks to change that by encouraging more scholars to utilize the tools of path dependence frameworks.

This article offers a brief introduction to the various types of path dependence and what might be considered their most useful tools for punishment scholars interested in understanding penal change. It also highlights some limitations of path dependence and some ways in which this approach can be merged with other theories of penal change in ways that will strengthen both approaches. Throughout this article, I use examples from early and more contemporary US prison history. My underlying goal is to present path dependence frameworks as a toolkit of theoretical concepts, mechanisms, and explanations for change to help scholars better understand and analyze penal change.

An Introduction to Path Dependence

Path dependence can be treated as a family of similar, but distinct, theories and frameworks. This family of theories and frameworks emerged from political science, historical sociology, and (historical) economics (Mahoney 2000). Because path dependence has multiple definitions and uses (e.g., Beyer 2010; Mahoney and Schensul 2006), which vary over time and across disciplines, it is difficult to provide a single definition. Indeed, the choice over which path dependence theorists to prioritize can lead to markedly different theories or approaches to understanding historical change. This section identifies some of that variation before drilling down into the specific toolkit punishment scholars can create from this family of theories.

Early Versions of Path Dependence

As an approach to understanding historical change, path dependence can be traced to multiple historical studies in the 1960s and the 1970s. These studies used a fairly weak version of path dependence—one that can be summarized as “history matters,” or (elaborated slightly) “prior history matters,” for understanding subsequent developments (Mahoney 2000; Sewell 1996). Indeed, some scholars point to general works of historical sociology as foundational examples of path dependence because they demonstrate the lasting significance of prior history. Particularly famous examples include a study of how political parties stabilized or “froze” across Europe following splits between the church, the state, and labor (Lipset and Rokkan 1967); a study of the revolutions in France, Russia and China, each of which had certain conditions in common before the revolutions broke out (Skocpol 1979); and a study explaining capitalism’s origin in Europe instead of China because of their different agricultural systems (Wallerstein 1974). While these theorists did not use the term path dependence, and some contemporary scholars disagree over the extent to which any of these examples illustrate path dependence, they are often invoked as early illustrations of specific concepts used by path dependence scholars (e.g., Beyer 2010; Mahoney 2000; Pierson 2000a; Thelen 1999).

By the 1980s and 1990s, new studies emerged that helped identify specific mechanisms by which prior history’s causal effects or impact on subsequent developments could be illuminated (Mahoney 2000, 507). Two mechanisms are worth highlighting. The first mechanism, introduced by economic historians, is the first-mover advantage: akin to the sociological concept of the accumulation of advantage, the first mover, or whoever is first to market, is more likely to succeed in the long term because of “increasing returns” (David 1985; North 1990; see also Arthur 1989). The second mechanism is the role of feedback effects, or clearly related consequences of early policies that directly cause, facilitate, or constrain downstream efforts to change or introduce new policy (Orloff 1993).

In this period, studies began using the term path dependence, although often without elaborating what exactly that meant (Mahoney 2000, 507; Pierson 2000a). It was really in the late 1990s and early 2000s that a number of scholars began summarizing extant studies, formalizing various theories or types of path dependence, and elaborating how these and other mechanisms allowed the past to influence the future (e.g., Mahoney 2000; Pierson 2000a;b; Thelen 1999; 2003).

Mature Versions of Path Dependence

Although path dependence became more heavily theorized around the turn of the century, it has not congealed, still offering a wide variety of definitions. Among the loosest and vaguest definitions comes from William Sewell (1996, 262-263): “what has happened at an earlier point in time will affect the possible outcomes of a sequence of events occurring at a later point in time.” Perhaps the most strict and narrow definition comes from Mahoney (2000, 7): “those historical sequences in which contingent events set into motion institutional patterns or event chains that have deterministic properties.” Mahoney explains that path dependence analysis “therefore involves both tracing a given outcome back to a particular set of historical events, and showing how these events are themselves contingent occurrences that cannot be explained on the basis of prior historical conditions,” all of which requires “a theory-laden process” (Mahoney 2000, 7-8). Mahoney’s emphasis on contingency is important to his account; indeed, he notes that historical arguments that omit the contingency of initial conditions are not path dependence analyses (Mahoney 2000, 512), again illustrating a comparatively narrow view in contrast to Sewell’s more encompassing definition.

Moreover, many path dependence analyses often include a discussion of “inertia” (e.g., Pierson 2000a). Such discussions arguably reflect a third and more colloquial definition of path dependence: once someone or something sets down a path, they stay on that path. This is essentially the social-behavioral version of Newton’s first law of motion: an object in motion stays in motion (given a frictionless surface), and an object at rest stays at rest (in the absence of some external force). This colloquial version of path dependence has been formalized as “punctuated equilibrium theory” or “lock-in”: once a policy or organization is formed, it will continue on its initial path until some exogenous shock causes a change (Krasner 1988). Some scholars explain this lock-in by drawing on earlier economic accounts of feedback effects (now applied to political settings) to explain the way in which the consequences of starting on a course increasingly forecloses alternatives to that course (Pierson 2000a). Another version of this type of path dependence has also been articulated within neo-institutional theory from organizational sociology. As several scholars have pointed out (e.g., Beyer 2010; Kluttz 2019; Mahoney 2000), neo-institutional theory is a specific type of path dependence that specifies legitimacy as one mechanism that encourages a particular policy, practice, or formal structure’s lock-in.

These examples of path dependence frameworks are not a comprehensive enumeration of path dependence theories; instead, they illustrate the variety of path dependence arguments. In what follows, I do not adopt a single definition of path dependence, or consistently distinguish between distinct types or theories, simply because of the variations and overlap across these different versions. Instead, I treat path dependence as a loose collection of analytical tools with varying degrees of utility for punishment studies. In this sense, path dependence research can be thought of as a theoretical toolkit from which scholars can draw—in much the same way that we have selectively drawn from the social theories of Durkheim, Marx, Weber, Du Bois, Ellias, Foucault, and Bourdieu.

The Promises of Path Dependence

In this section, I identify important contributions made in the path dependence literature outside of punishment studies (primarily in political science, political sociology, historical sociology, historical economics, and historical institutionalism). In each case, I offer examples of how these contributions can be, or have been, adapted as analytical tools to better analyze penal change. I also evaluate these tools and identify their limitations as well as their utilities. Importantly, I draw from several approaches, but I focus primarily on historical institutionalism and draw more heavily on the work of Kathleen Thelen, James Mahoney, and Paul Pierson than that of other theorists. In doing so, I am emphasizing what I see as the most fruitful elements of the many different types of path dependence, including versions that have already been imported into our field. Consequently, this section does not necessarily describe path dependence in the way other scholars would, particularly scholars outside of punishment and society. In sum, rather than offering a coherent theory of path dependence, I enumerate its specific analytic tools or contributions, particularly those tools that are best suited for understanding penal change.

Change and Stasis

One of the path dependence literature’s most striking potential contributions to the penal change literature is its emphasis on stasis—the absence of change (or perhaps the apparent absence of change). While penal change research often implicitly concerns stasis, presumably identified as the periods before and after a change occurs, most of the attention (and excitement) is about the change—where/when does it begin, where/when does it end? Popular representations of penal history are often periodized into distinct eras (e.g., Churchill 2019; Goodman et al. 2017; Irwin 2005; Rubin 2019d) and penal change is often defined as the shift between whatever characterizes these eras as distinct: e.g., the rise of the prison, the innovation or popularity of a new type of prison, or the explosion in the frequency with which the prison is used. This model of penal change has been heavily criticized by scholars challenging the uniformity of any given era (e.g., Goodman et al. 2017; Hutchinson 2006; Matthews 2005; Maurutto and Hannah-Moffat 2006; O’Malley 1999; Rubin 2019d; Seeds 2017)—particularly when disaggregated by jurisdiction and taking the local level into account (e.g., Goodman et al. 2017; Lynch 2010; Phelps 2017; Rubin 2019d)—and the distinctiveness of a seemingly new era from what came before (e.g., Brown 2002; Goodman et al. 2017; Hallsworth 2002; Phelps 2011). Together, these critiques effectively undermine the extent to which we can even think of distinct eras bordered by major moments of change (Churchill 2019; Goodman et al. 2017; Rubin 2019d). Despite these critiques, penal change remains these accounts’ major focus in questions like: “To what extent has there really been change?” “Is this the rupture moment everyone thought it was?” “What sorts of continuity are mixed in with the change?”

Said in different theoretical language, punishment studies suffer from a disproportionate focus on “institutional innovation,” and an insufficient attention to “institutional reproduction” (Thelen 2003). As Rubin (2015, 366) put it, “Studies of penal change have primarily analyzed innovations or their widespread consequences,” but the diffusion of those technologies has not enjoyed “similarly enthusiastic examinations.” To fully understand penal change, we must study not only the moments of change, but also the periods of stasis—or not only the periods of innovation, but also the diffusion that sometimes follows (Rubin 2015; 2019d; see also Churchill 2019).

Path dependence research recenters the analysis to the periods before and after change. For example, it would ask, “After the prison is created, how does it persist?” “How does it survive myriad critiques and challenges?” “How did capital punishment survive, at least in some form, the many challenges to it (in the US)?” It also asks questions like: “When do we not see change?” “How is a period of non-change (stability) sustained?” It shifts our attention away from the (admittedly) more exciting “moments” of significant change that so enlivened the early social theoretical accounts of punishment (Garland 2018) to what happens next. Alternatively, it shifts attention from the question, “Under what conditions is change possible?” to the question, “Under what conditions is stasis possible?” While these questions are really two sides of the same coin, they place the emphasis of our analyses in different locations. Ultimately, then, path dependence shifts our gaze from the beginning and end of the story to the underexplored middle. More than reminding us to study the stasis following the change, path dependence research also gives us a theoretical framework and set of tools to analyze it.

Theorizing Inertia

The basic argument underlying many path dependence analyses is the idea that an organization, a policy, or a practice, once formed, will continue, essentially unchanged, until something stops or changes it. How? What explains the continuation? While most path dependence analyses allow for some degree of inertia, the more useful path dependence frameworks theorize the mechanisms by which stasis is possible so we are not in the uncomfortable position of applying the laws of physics to human behavior (see also Beyer 2010).

Initial Conditions and Critical Junctures

Among the theoretically weaker explanations are those that empirically demonstrate the importance of the initial conditions by identifying their downstream consequences. Some scholars use the term “critical junctures” to refer specifically to those moments of change where individual actors’ or groups’ decisions become locked in, meaning difficult to undo or challenge (e.g., Capoccia 2015; Mahoney 2000; Pierson 2000a;b).

We see the importance of initial conditions or critical junctures in a number of empirical studies of penal change. For example, in their study of the passage of hate crime legislation across the country, Grattet et al. (1998) demonstrate that states’ statutes varied substantially over which groups targeted by hate criminals were protected by these statutes. The authors demonstrated, these statutes, once passed, were infrequently changed; consequently, whatever combination of groups these statutes protected would remain, even while other states, particularly those passing their statutes relatively late (a decade or two after the first states) were protecting more categories of potential hate crime victims.

My study of post–American Revolution penal code revisions reveals a similar finding (Rubin 2016). There were substantial variations in capital statutes in the new penal codes adopted in this period of anti–capital punishment reform—some states restricted capital punishment to treason and murder while others restricted it to three-to-five crimes and others to ten or more. I found “most states made the substitution in one fell swoop; if the most serious offenses were not included in this list, they would remain capital—at least until the next substantial push for reform” (433). However, because penal code revisions were often several decades (or more) apart, and penal reformers settled for partial victories rather than continuing to push for further restrictions, capital punishment was able to persist in all states even as incarceration increasingly became the modal punishment.

First-Mover Advantage

A specific type of initial condition is the first-mover advantage. Popular among economic historians, this theory holds that the first to market—or, to move away from market metaphors, the “first out of the gate”—will ultimately dominate the competition. This has been powerfully demonstrated by the success of the practically inferior QWERTY keyboard, which remains our dominant keyboard today, while the practically superior Dvorak keyboard (which was created several decades after the QWERTY keyboard) remains a footnote in history. Once the QWERTY keyboard had been adopted by many, it became financially and practically difficult to switch to a later but more ergonomic and efficient keyboard design (David 1985). Importantly, the significance of being first derives from the benefits accrued specifically therefrom (the “increasing returns”). Indeed, the QWERTY keyboard was not the first design in existence, but was the first to become popular. Once it became popular, it was difficult to dislocate. Being first, but still unknown, is not in itself useful. If being first carries no advantages, then there is no first-mover advantage from a theoretical perspective.

We see this distinction vividly with the history of the early prisons. Among second-generation prisons, one model of prison design—the Auburn System of solitary confinement at night and congregate factory-style work during the day—became the norm within American prisons from the 1820s to the 1860s (and, outside of the South, arguably longer). We can trace this dominance to, among other things, the fact that the Auburn System had a working model in upstate New York in the early 1820s. Soon, other states had adopted the model, allowing for multiple working models. Its competitor approach, the Pennsylvania System of long-term solitary confinement, only achieved a fully functional working model when Eastern State Penitentiary opened in Philadelphia in 1829. It was another decade before there were four such prisons (two in Pennsylvania and two in other states). Penal reformers, prison administrators, prison commissioners, governors, and foreign dignitaries could and did tour these prisons, but they had earlier and greater access to Auburn-style prisons. More than providing a model, these prisons also concretely conveyed the Auburn model’s relative popularity, despite a raging theoretical debate over the two approaches (Rubin 2015).

This series of events, however, contrasts with the history of the first-generation prisons. The dominant model of prison in the 1790s into the 1820s was the penitentiary house first created at Philadelphia’s Walnut Street Prison (Meranze 1996) and improved at New York City’s Newgate Prison (Lewis 1965; McLennan 2008). However, Walnut Street was not the first state prison. Before Pennsylvania’s Walnut Street officially became a state prison in 1794, Massachusetts (1785) and Connecticut (1790) each created state prisons (Hirsch 1992; Rubin 2016; 2018). These prisons were the first to the proverbial market; however, they remained fairly unknown. Philadelphia was a governmental and cultural center; foreign dignitaries and statesmen from across the country passed through. Additionally, Walnut Street was widely promoted by an active penal reform society—the first formed in the new United States—that publicized the reforms that they had fought for. Arguably, Walnut Street had other practical advantages over its predecessors (Rubin ND). For all these reasons, it was able to overcome any first-mover advantage of the two older prisons. Thus, for first-mover status to predict longevity, it must clearly convey an advantage; obscure, under-promoted innovations rarely diffuse (Rogers 2003). But once diffused, Walnut Street became the model for American proto-prisons; it was not dislodged until the second-generation prisons offered a new vision of prison that was bigger, stronger, and more tightly controlled.

Exogenous Shocks as Change Agents

Staying the course may be the default state according to path dependence frameworks, but path dependence scholars also argue that an exogenous shock is capable of interrupting this inertia. Essentially, an organization or society stays on a path, or a policy persists, until it experiences an exogenous shock. Notably, this is one of our more common explanations for penal change. Indeed, many of our social theoretical bases prepare us to expect a change in punishment following a significant change to the society—whether its culture, politics, or economy (see Garland 1990).

Empirically, however, we also see numerous examples of exogenous shocks bringing about significant changes to penal structures. For example, England’s use of convict transportation to North America halted because of the American Revolution, only to resume after the war, with ships traveling to Australia instead (e.g., Maxwell-Stewart 2010). Opposition to capital punishment, and support for incarceration, in the United States increased dramatically following Revolution; although this was part of a long-term trend, multiple scholars identify how the Revolution accelerated these reforms (Banner 2002; Hirsch 1992; Masur 1989; Rubin 2018; 2019a). The Southern states diverged from the rest of the country’s penal trends after the Civil War: Southern states’ Auburn-style prisons, destroyed in the war, were replaced by convict leasing (Oshinsky 1997). The period after the Great Depression and Second World War is also associated with myriad penal changes, including the rise of rehabilitation, visible at the national level in the United States (e.g., Irwin 2005; Simon 1993). The United States’ use of mass incarceration is also frequently linked to social and political changes in the 1960s (e.g., Beckett 1997), although some accounts place the origins even earlier (e.g., Gottschalk 2006; Murakawa 2014; Simon 2007). Most recently, scholars have identified the impact—particularly within the United States—of the Great Recession of 2008 on the rising popularity of “Right on Crime” as well as a wide variety of economic arguments against mass incarceration (Aviram 2015; Dagan and Teles 2014). More generally, any major shift that creates a need for technological innovation can be the cause of significant formal change (Rubin 2019d).

Taking a longer view, the common model of penal change throughout US history as a series of pendular shifts from liberal to conservative, or rehabilitative to punitive, extremes also fits this narrative (but see Goodman et al. 2017; Rubin 2019c). War in particular seems to be a recurring exogenous shock—the American Revolution, the Civil War, the Second World War, and the Vietnam War each preceded major penal changes.

Feedback Effects

While initial conditions, first-mover advantage, and exogenous shocks can be empirically demonstrated, they remain fairly unsatisfying theoretical premises because they do little to explain how a path is maintained. Thankfully, path dependence scholars have also theorized stronger mechanisms to explain inertia: feedback effects. Feedback effects are essentially downstream consequences that encourage a particular course of action (Orloff 1993). More specifically, tracing feedback effects entails understanding “how a policy creates conditions that shape its own future” (Dagan and Teles 2014, 267). While most analyses seem to use feedback effects as a means of “locking in” a particular policy or practice—that is, making it increasingly difficult to depart from (e.g., Arthur 1989; Pierson 2000b)—scholars sometimes distinguish between positive and negative feedback effects. Positive feedback effects, such as increasing returns, lock in a policy or keep a society or organization on a particular path, while negative feedback effects make that policy or path increasingly untenable (Weaver 2010).4 One particular kind of positive feedback mechanism, frequently discussed within neo-institutional theory, is institutionalization. Institutionalization encourages further momentum along a particular path by bestowing legitimacy upon organizations (including jurisdictions) that adopt the institutionalized policy or practice and deducting legitimacy from those organizations that fail to adopt it (e.g., Meyer and Rowan 1977).5

Unlike the previously discussed mechanisms, feedback effects are both far more commonly found in the punishment literature and increasingly described explicitly as such. As Dagan and Teles (2014, 268) note, “The literature on the carceral state is replete with examples of positive feedback.” One reason for this popularity may be the centrality of political explanations, particularly among scholars of mass incarceration, including growing attention from political scientists and political sociologists who are leading the way in identifying feedback effects (e.g., Campbell and Schoenfeld 2013; Dagan and Teles 2014; Murakawa 2014; Schoenfeld 2010; 2018).

Among the first such scholars to identify the application of feedback effects to studies of penal change was Heather Schoenfeld. Schoenfeld relies heavily on the role of feedback effects in her study of the expansion of Florida’s “carceral state”—the number of prisons and prisoners, as well as the general infrastructure that made such large numbers possible (Schoenfeld 2010; 2018). Using Florida as her case study, Schoenfeld offers three common ways in which feedback effects affect penal policy: 1) “policy creates or institutionalizes meaning,” 2) “policy creates politics,” and 3) “policy creates new state capacities” (Schoenfeld 2018, 8). She shows how key decision moments in Florida’s history have lasting and self-reinforcing consequences. For example, reviewing a prison conditions lawsuit, a Florida court decided that bad prison conditions were directly related to overcrowding and issued an injunction relating to “prison capacity.” As Schoenfeld explains, this injunction made resolving the lawsuit a problem for the Department of Corrections rather than for other groups that determine the flows into and out of prisons, like the state legislature or the police. The Department of Corrections, which has a limited range of options available to it, solved this problem by building more prisons, which increased the capacity, satisfying the injunction. However, based on how the courts defined prison capacity, building more prisons allowed for more incarceration, which ultimately contributed to Florida’s mass incarceration (Schoenfeld 2018, Chapter 3; see also Schoenfeld 2010).

Returning to the example of America’s early prisons, we can see feedback effects as well. I have argued that capital punishment persisted despite a new alternative punishment available (incarceration) in part because the first-generation of prisons helped solidify capital punishment (Rubin 2016). When these “proto-prisons” began experiencing fires, riots, and general chaos, state legislators and penal reformers advocated (sometimes successfully) for an expansion of capital punishment by making some crimes once again capital offenses, or by making other new crimes (like prison arson) also capital. The chaos in the prisons also contributed to a punitive backlash to progressive reforms making it difficult to advocate for reducing capital punishment in the 1810s and 1820s.

Among the early prisons, we can also see the role of legitimacy as a feedback mechanism. Legitimacy was a major reason for the diffusion of the second-generation prisons, the “modern prisons” (Rubin 2015). I have demonstrated that the earliest states to adopt modern prisons were driven by technological motivations (replacing their imploding and overcrowded proto-prisons). Once prison had become institutionalized among the early adopters, it became a marker of legitimacy and a symbol of modern statehood. Later adopters adopted the prison (in the late 1830s to 1860s) for the image of legitimacy it conveyed: Southern states were driven by their desire to overcome their reputations as barbaric backwaters that still relied on slavery, and frontier states were driven by their desire to be taken seriously as true states, equal to their Eastern counterparts. Thus, once the prison was institutionalized, it created a momentum that allowed the prison to spread even to states that had no technological need for it (see also Rubin 2019d).

Ultimately, feedback effects allow for more complicated theories of change by moving away from the initial conditions. In this version of path dependence, initial conditions are important because they set a change of events in motion (see Mahoney 2000), not because those initial conditions become frozen in time. This version of path dependence also moves away from simplistic inertia-based arguments more generally. Indeed, feedback effects allow punishment scholars to benefit from an oft-repeated insight (within path dependence scholarship) from Arthur Stinchcombe: whatever started the change is not necessarily what keeps it going (Mahoney 2000; Stinchcombe 1968; Thelen 2003). This is a powerful insight for punishment studies.

Particularly among macro-level theories of penal change, some initial macro-social change is expected to be influential or determinative until a new macro-social change occurs. We especially see this tendency in pendular theories of penal change, where some macro-social change causes the pendulum to swing in the other direction until some other change causes it to swing back after some period of time (e.g., Goodman et al. 2017; Stuntz 2011). I have criticized this tendency, arguing, “[penal] scholars treat diffusion as a natural consequence of innovation. They implicitly posit that whatever zeitgeist caused the innovation was present elsewhere, albeit less intensely,” instead of “interrogat[ing] the [diverse] factors motivating different jurisdictions’ adoption of new practices” Rubin (2015, 366). This recognition becomes increasingly important when we see diffusion as a stretched-out process that takes place over time, as different jurisdictions’ adoptions, at different points in the process, are driven by different factors (Rubin 2019d; see also Churchill 2019).

The Pitfalls of Path Dependence

Path dependence offers a variety of theories and mechanisms for understanding penal change and—especially—stasis, but it is important to recognize the framework’s limits as well. These theories do not always accurately predict the course of history. As such, path dependence is a helpful toolkit of possible explanations and mechanisms for making sense of the past, but it should not be viewed as a series of testable predictions. Indeed, some versions of path dependence that emphasize initial conditions’ contingent nature make clear the difficulty of predicting the outcome ex ante (e.g., Mahoney 2000). The historical particularities of any penal phenomenon must be taken seriously when using the path dependence toolkit. This section illustrates the limitations of leaning too heavily on some elements of path dependence frameworks without understanding the specific context. I use several examples from a larger research project on Eastern State Penitentiary.6

Eastern was one of the second-generation prisons that was authorized and opened in the 1820s. Prisoners there were kept in long-term solitary confinement (often called “separate confinement”), during which time prisoners worked within their cells, received visits from prison staff and local penal reformers, and received mentorship and education—an approach to incarceration called the Pennsylvania System. Eastern was the second of Pennsylvania’s two prisons authorized and built, but the first to become fully operational; consequently, it became the showcase prison for the Pennsylvania System.

The Pennsylvania System is often treated as a major competitor to the Auburn System, just as Eastern is often treated as a competitor to New York’s Auburn State Prison. But while contemporaries hotly debated the Pennsylvania and Auburn Systems, the Auburn System rapidly (for prison development) became the dominant approach to incarceration in the United States. The Pennsylvania System, by contrast, even at its height, was adopted at only four prisons; by the end of the Civil War, it remained an outlier, used only at Eastern, where it would continue to be used until the late 1870s (and the legally mandated approach until 1913).

Throughout this period, Eastern—or rather its administrators—received heavy criticism for using what amounted to a system of solitary confinement, particularly following several high-profile disasters at other prisons using a more extreme version of solitary confinement. Eastern’s administrators themselves struggled to implement the system and often strategically deviated from the very system they personally promoted as superior to the Auburn System. Why did Philadelphia’s Eastern State Penitentiary retain its highly criticized (and difficult to implement) system of long-term solitary confinement between 1829 and 1879? While I answer this question in my book (Rubin 2021), here I discuss alternative answers to illustrate the limits of path dependence, particularly its more facile versions.

The Failures of Inertia

A path dependence analysis seems ideal for answering this research question. In the course of my research, a number of readers suggested path dependence–like hypotheses for me to consider, several of which boil down to some kind of inertia. The key assumption among such hypotheses was that once the Pennsylvania System was authorized at Eastern, it would persist—at least until some exogenous shock allowed or forced authorities to change the policy. However, this is not a compelling answer based on the historical record. For one, the three other prisons that adopted the Pennsylvania System (Pennsylvania’s Western State Penitentiary and the New Jersey and Rhode Island state prisons) each abandoned the Pennsylvania System—and then authorized the Auburn System. And each did so earlier than Eastern: six, twenty-five, and thirty-plus years after the initial authorization versus fifty years.

Another inertia argument is a kind of architectural lock-in: once an expensive and lengthy construction project was in place, designed uniquely for this type of incarceration, the state would be stuck with this mode of incarceration. However, Western State Penitentiary, which opened several years before Eastern, was razed entirely when it became clear the architecture rendered implementation of the planned use of solitary confinement impossible. When the capacity expected from Eastern’s initial architecture was deemed too small, the initial architectural plan was changed midway through construction. In both cases, state officials were willing to weather the expense of changing the original architecture to ensure they could use the Pennsylvania System. Finally, when the New Jersey, Rhode Island, and Western prisons converted to the Auburn System, none of them required significant reconstruction to the main prison buildings (just additional buildings), suggesting architecture was no constraint on the choice of system.

A final inertia-related argument can be summarized as some form of state pride in the eponymous prison system. That is, state-level decision makers and those who influenced them would be unwilling to abandon the prison system named for their state. Given the historical record, however, this too seems unlikely. For one, when the Pennsylvania System was initially instituted, it was a controversial decision that followed years of debate among different groups of Pennsylvanians, some of whom (including a committee authorized by the legislature) endorsed the Auburn System. For another, the legislature had passed a number of laws throughout the nineteenth century—and refused to pass others—that made it difficult to operate the Pennsylvania System, according to its supporters; ultimately, the legislature’s inaction even helped to dismantle the Pennsylvania System at Eastern. Moreover, as already noted, Western abandoned the Pennsylvania System voluntarily: its administrators requested this change and the state legislature authorized it. Finally, local reformers—whose organization had brought the Pennsylvania System into being—began lobbying the legislature (somewhat successfully) for changes that, they argued, would mitigate some of its harsher features; such changes, supporters noted, would effectively end the Pennsylvania System as they knew it. State pride was not enough to keep these groups from opposing their namesake prison system.

Thus, we have multiple examples in which inertia fails as a useful explanation. One possible explanation for this failure is that—in physics at least—the kind of continuous motion expected from inertia requires frictionless conditions. But penal policy is far from frictionless (Goodman et al. 2017). As noted, Eastern’s retention of the Pennsylvania System was highly contentious not only on the national stage, but also on the state and local stages. Thus, inertia does not help us understand the retention of penal policies that are highly criticized and that take work to maintain—policies like the Pennsylvania System. Inertia-based path dependence analyses assume stability or some sort of automaticity following the initial conditions, in this case, that adoption seamlessly becomes retention. Instead, Eastern’s retention of the Pennsylvania System is the policy equivalent of swimming upstream.

The Failures of Exogenous Shocks

As noted previously, the other part of an inertia-based argument is the idea that an organization, policy, or practice persists until an exogenous shock disrupts it. Perhaps the above explanations for the Pennsylvania System’s persistence emphasize inertia too heavily, but not the exogenous shocks that end it. However, exogenous shocks—wars and recessions or depressions—were fairly common in this period and the Pennsylvania System at Eastern survived each of them.

The closest we come to an effective exogenous shock is the prison population’s increase and consequent overcrowding that began once the Civil War ended. Eastern’s administrators were forced to double cell their prisoners; administrators saw this as a grave violation of their system, as its central tenet was solitary confinement. While some scholars have mistakenly said that overcrowding ultimately ended the Pennsylvania System at Eastern, the Pennsylvania System persisted for at least a decade after overcrowding began in 1866. (Legally, the Pennsylvania System continued until 1913, but a clear change took place over the course of the 1870s when internal policies and practices, as well as administrators’ public discourse, subtly shifted away from their long-held Pennsylvania System.) Indeed, the administrators begged the legislature for relief, both for funds to build new cellblocks and changes in the prison’s jurisdiction to send some prisoners to the less crowded Western State Penitentiary. As they had frequently done before, the lackluster legislature dragged its feet in responding to the administrators’ pleas; by the time both changes eventually occurred, they were too little, too late. To the extent overcrowding prevented the Pennsylvania System’s daily implementation, the effect was not as sudden as the exogenous shock theory label implies; but more importantly, there was little direct impact of overcrowding on the prison administrators’ commitment to their system, which was crucial for the Pennsylvania System’s retention at Eastern (an argument I develop in my book).

An exogenous shock–based argument focused on overcrowding also obscures other, more significant causes of the Pennsylvania System’s decline, some of which began before the overcrowding. As I argue in the book, the real cause was the gradual decline of administrative support for the Pennsylvania System. Throughout this period, Eastern’s administrators also continuously defended the Pennsylvania System from criticism and argued for its superiority over any other approach. (At the same time, these very administrators often privately violated the Pennsylvania System’s most central tenets to improve the prison’s metrics and reputation while denying public accounts of this behavior.)7 The Pennsylvania System ultimately failed at Eastern, I argue, because administrators stopped deriving personal benefits from defending an increasingly obscure approach to incarceration. Over the course of the 1850s and 1860s, the Auburn–Pennsylvania System debate subsided, as did criticism of the Pennsylvania System. Responding to this criticism had been, I argue, one of the major drivers of Eastern’s administrators’ commitment to the Pennsylvania System. With this criticism declining—and with it, the attention and opportunity to defend themselves and their prison by describing it and themselves in glowing terms—the administrators shifted their efforts from defending the Pennsylvania System to establishing themselves as professional “penal scientists” and challenging what they perceived as upstart reformers attempting to claim that mantel for themselves. In the process, Eastern’s administrators found a new source of status and the Pennsylvania System became “expendable.”

Alternatives from Historical Institutionalism

These limitations of path dependence are not new discoveries (see Beyer 2010). Thelen (1999; 2003) has noted that significant political institutions, like courts and legislatures, are actually strongly resilient against exogenous shocks—indeed, this is one of the characteristics that, she says, makes them interesting to so many scholars (Thelen 2003, 209; Mahoney and Thelen 2009). However, Thelen also notes that one would be hard pressed to look at any major political institution and argue that there is no difference between what it looks like today and what it looked like when it was first created. As she explains, “From the perspective of a punctuated equilibrium model” (or really, I would add, many path dependence analyses), “there often seems to be too much continuity through putative breakpoints in history, but also often too much change beneath the surface of apparently stable formal institutional arrangements” (Thelen 2003, 211). The same points can be said of prison (e.g., McLennan 2008; Rubin 2019b), capital punishment (Banner 2002), and parole (Simon 1993), each of which, though still recognizable as the thing it started out as, is also incredibly dissimilar from its original.

To address this apparent irony and weakness within path dependence frameworks, Thelen (2003) has developed a model of “institutional layering” and “institutional conversion” (see also Mahoney and Thelen 2009). The underlying idea is to move away from a focus on major breakpoints (such as exogenous shocks) or an overemphasis on inertia, and instead to recognize how institutions almost constantly change—or rather, “evolve” and adapt to new conditions.

The focus here is on how “institutional arrangements are renegotiated periodically in ways that alter their forms and functions” (Thelen 2003, 213). One mechanism of this renegotiation is layering, which involves “partial renegotiation of some elements of a given set of institutions while leaving others in place,” such as the addition of new laws while others remain in place (Thelen 2003, 225; Mahoney and Thelen 2009, 15). Another mechanism is conversion, where “institutions designed with one set of goals in mind are redirected to other ends” Thelen (2003, 228), or even their “strategic redeployment” (Mahoney and Thelen 2009, 16). Through these mechanisms, we can explain how institutions evolve over time without identifying major moments of rupture—and without emphasizing too greatly the inertia we might expect from some path dependence analyses.

This approach has great appeal within the study of punishment. For example, while the prison itself has persisted, it has evolved through multiple models or templates (Rubin 2019b). The same is true of capital punishment (Banner 2002). Thelen’s approach is also consistent with Goodman et al. (2017)’s recent theory of penal change. Goodman et al. (2017) argue that the appearance of intermittent ruptures or significant changes—captured by the pendulum metaphor—hide perpetual conflict under the surface. Like a system of plate tectonics, in which the plates’s constant movement is only apparent on the surface when earthquakes hit or volcanos erupt, our punishment system is constantly contested. This contestation lays the groundwork for the apparent “moments” of significant change.

Thelen’s approach is also the basis for my theory of “penal layering” (Rubin 2016). In this version, reform movements that fail to abolish or replace a particular type of punishment (e.g., capital punishment) with another (e.g., incarceration) instead produce an expanded array of available punishments (e.g., capital punishment and incarceration). Metaphorically, we can see different depths and textures in these layers across different jurisdictions, but each punishment continues to coexist (layered atop one another) rather than one replacing the other.

Further, Thelen’s framework for understanding institutional evolution explains how punishment in any given era remains “braided” (Hutchinson 2006), “assembled” (Maurutto and Hannah-Moffat 2006), or “volatile and contradictory” (O’Malley 1999), but at the same time still give an overall impression that fits pendular descriptions when one fails to look closely (Rubin 2019c).


This article provides penal change scholars a brief introduction to path dependence theories and frameworks. I have presented path dependence as a helpful toolkit containing multiple analytical tools, including an emphasis on stasis (not just change), ways of theorizing apparent inertia (including initial conditions and critical junctures, first-mover advantage, and feedback effects), and exogenous shocks. I have also identified some limitations, including examples of when inertia is an insufficient explanation for stasis—particularly in cases that require penal actors to (essentially) swim up stream—and when exogenous shocks are an insufficient explanation for change—particularly when the exogenous shock coincides with and obscures other, longer-term trends. Finally, drawing on established critiques of path dependence research and some recent theories of penal change, I identify ways of moving beyond some of path dependence’s pitfalls, including its failure to recognize the way in which what seems like stasis may hide constant change.

Ultimately, penal scholars can benefit from path dependence theories and frameworks in multiple ways. First, injecting this new approach can refresh the field, which has long revolved around social theories. Second, and relatedly, theories explicitly designed to understand temporal processes (stasis and change) provide both a mindset and a toolkit for understanding penal change, especially in ways that the field has often failed to do (with few exceptions). Third, the toolkit path dependence provides allows for greater theoretical development in the middle range: perhaps as an artifact of the field’s reliance on social theory, so much of our field is primarily concerned with macro- or micro-level approaches, while meso-level approaches remain underdeveloped (but slowly growing over the last decade or so). Finally, using established theories and frameworks from the disciplines provides a bridge between interdisciplinary punishment studies and other fields, particularly those that tend to overlook this literature as they begin to analyze penal phenomena.


Arthur, W. B. (1989). Competing technologies, increasing returns, and lock-in by historical events. The Economic Journal, 99(394):116–131.

Aviram, H. (2015). Cheap on Crime: Recession-Era Politics and the Transformation of American Punishment. Oakland, California: UC Press.

Banner, S. (2002). The Death Penalty: An American History. Cambridge, MA: Harvard University Press.

Barker, V. (2006). The politics of punishing: Building a state governance theory of american imprisonment variation. Punishment & Society, 8(1):5–32.

Beckett, K. (1997). Making Crime Pay: Law and Order in Contemporary American Politics. New York: Oxford University Press.

Beyer, J. (2010). The Same or Not the Same: On the Variety of Mechanisms of Path Dependence. International Journal of Business, Human and Social Sciences, 3(3).

Brown, M. (2002). The politics of penal excess and the echo of colonial penality. Punishment & Society, 4(4):403–423.

Campbell, J. L. (2021 [2004]). Institutional change and globalization. Princeton University Press.

Campbell, M. C. and Schoenfeld, H. (2013). The transformation of America’s penal order: A historicized political sociology of punishment. American Journal of Sociology, 118(5):1375–1423.

Capoccia, G. (2015). Critical junctures and institutional change. In Mahoney, J. and Thelen, K., editors, Advances in Comparative-Historical Analysis, pages 147–179. Cambridge University Press.

Churchill, D. (2019). History, periodization and the character of contemporary crime control. Criminology & Criminal Justice, 19(4):475–492.

Dagan, D. and Teles, S. (2014). Locked in? Conservative reform and the future of mass incarceration. The Annals of the American Academy of Political and Social Science, 651(1):266–276.

David, P. A. (1985). Clio and the economics of QWERTY. The American Economic Review, 75(2):332–337.

Durkheim, Émile. (1984[1893]). The Division of Labor in Society. New York: The Free Press.

Foucault, M. (1977). Discipline and Punish: The Birth of the Prison. New York: Vintage Books.

Garland, D. (1985). Punishment and Welfare. Brookfield, VT: Gower.

Garland, D. (1990). Punishment and Modern Society: A Study in Social Theory. University Of Chicago Press.

Garland, D. (2018). Theoretical advances and problems in the sociology of punishment. Punishment & Society, 20(1):8–33.

Goodman, P., Page, J., and Phelps, M. (2017). Breaking the Pendulum: The Long Struggle Over Criminal Justice. New York: Oxford University Press.

Gottschalk, M. (2006). The Prison and the Gallows: The Politics of Mass Incarceration in America. New York: Cambridge University Press.

Grattet, R., Jenness, V., and Curry, T. R. (1998). The homogenization and differentiation of hate crime law in the United States, 1978 to 1995: Innovation and diffusion in the criminalization of bigotry. American Sociological Review, 63(2):286–307.

Hallsworth, S. (2002). The case for a postmodern penality. Theoretical Criminology, 6(2):145–163.

Hirsch, A. J. (1992). The Rise of the Penitentiary: Prisons and Punishment in Early America. New Haven: Yale University Press.

Hutchinson, S. (2006). Countering catastrophic criminology: Reform, punishment and the modern liberal compromise. Punishment & Society, 8(4):443–467.

Irwin, J. (2005). The Warehouse Prison: Disposal of the New Dangerous Class. Los Angeles: Roxbury.

Kluttz, D. N. (2019). The path of the law review: How interfield ties contribute to institutional emergence and buffer against change. Law & Society Review, 53(1):239–274.

Krasner, S. D. (1988). Sovereignty: An institutional perspective. Comparative Political Studies, 21(1):66–94.

Lacey, N. (2013). The rule of law and the political economy of criminalisation: An agenda for research. Punishment & Society, 15(4):349–366.

Lewis, W. D. (1965). From Newgate to Dannemora: The Rise of the Penitentiary in New York, 1796–1848. Ithaca: Cornell University Press.

Lipset, S. M. and Rokkan, S. (1967). Cleavage structures, party systems and voter alignments: An introduction. In Lipset, S. M. and Rokkan, S., editors, Party Systems and Voter Alignments, pages 1–64. New York: Free Press.

Lynch, M. (2010). Sunbelt Justice: Arizona and the Transformation of American Punishment. Stanford, CA: Stanford University Press.

Mahoney, J. (2000). Path dependence in historical sociology. Theory and Society, 29(4):507–548.

Mahoney, J. and Schensul, D. (2006). Historical context and path dependence. In The Oxford Handbook of Contextual Political Analysis. Oxford University Press.

Mahoney, J. and Thelen, K. (2009). A theory of gradual institutional change. In Mahoney, J. and Thelen, K., editors, Explaining Institutional Change : Ambiguity, Agency, and Power, pages 1–37. Cambridge University Press.

Masur, L. P. (1989). Rites of Execution: Capital Punishment and the Transformation of American Culture, 1776–1865. New York: Oxford University Press.

Matthews, R. (2005). The myth of punitiveness. Theoretical Criminology, 9(2):175–201.

Maurutto, P. and Hannah-Moffat, K. (2006). Assembling risk and the restructuring of penal control. British Journal of Criminology, 46(3):438–454.

Maxwell-Stewart, H. (2010). Convict transportation from Britain and Ireland 1615–1870. History Compass, 8(11):1221–42.

McLennan, R. M. (2008). The Crisis of Imprisonment: Protest, Politics, and the Making of the American Penal State, 1776–1941. New York: Cambridge University Press.

Meranze, M. (1996). Laboratories of Virtue: Punishment, Revolution, and Authority in Philadelphia, 1760–1835. Chapel Hill: University of North Carolina Press.

Meyer, J. W. and Rowan, B. (1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83(2):340–363.

Murakawa, N. (2014). The First Civil Right: How Liberals Built Prison America. Oxford University Press.

North, D. C. (1990). Institutions, Institutional Change, and Economic Performance. Cambridge: Cambridge University Press.

O’Malley, P. (1999). Volatile and contradictory punishment. Theoretical Criminology, 3(2):175–196.

Orloff, A. S. (1993). The Politics of Pensions: A Comparative Analysis of Britain, Canada, and the United States, 1880–1940. Madison: University of Wisconsin Press.

Oshinsky, D. (1997). Worse Than Slavery: Parchman Farm and the Ordeal of Jim Crow Justice. New York: Free Press.

Phelps, M. S. (2011). Rehabilitation in the punitive era: The gap between rhetoric and reality in U.S. prison programs. Law & Society Review, 45(1):33–68.

Phelps, M. S. (2017). Mass probation: Toward a more robust theory of state variation in punishment. Punishment & Society, 19(1):53–73.

Pierson, P. (2000a). Increasing returns, path dependence, and the study of politics. The American Political Science Review, 94(2):251–267.

Pierson, P. (2000b). Not just what, but when: Timing and sequence in political processes. Studies in American Political Development, 14(1):72–92.

Rogers, E. M. (2003). Diffusion of Innovations. Free Press.

Rubin, A. T. (2015). A neo-institutional account of prison diffusion. Law & Society Review, 49(2):365–399.

Rubin, A. T. (2016). Penal change as penal layering: A case study of proto-prison adoption and capital punishment reduction, 1785–1822. Punishment & Society, 18(4):420–441.

Rubin, A. T. (2018). The prehistory of innovation: A longer view of penal change. Punishment & Society, 20(2):192–216.

Rubin, A. T. (2019a). Early US prison history beyond Rothman: Revisiting The Discovery of the Asylum. Annual Review of Law and Social Science, 15(1):137–154.

Rubin, A. T. (2019b). History of the prison. The Handbook of Social Control, pages 279–292.

Rubin, A. T. (2019c). Interrogating the penal pendulum: An introduction to the review symposium on Breaking the Pendulum: The Long Struggle Over Criminal Justice. Law & Social Inquiry, 44(3):791–798.

Rubin, A. T. (2019d). Punishment’s legal templates: A theory of formal penal change. Law & Society Review, 53(2):518–553.

Rubin, A. T. (2021). The Deviant Prison: Philadelphia’s Eastern State Penitentiary and the Origins of America’s Modern Penal System, 1829–1913. Cambridge University Press.

Rubin, A. T. (ND). Innovation and diffusion: Theorizing penal change before and after the ideal type. Manuscript in Progress (Available by Request).

Savelsberg, J. J. (1994). Knowledge, domination, and criminal punishment. American Journal of Sociology, 99(4):911–943.

Savelsberg, J. J. (2014). Institutional theories of punishment. In Bruinsma, G. and Weisburd, D., editors, Encyclopedia of Criminology and Criminal Justice, pages 2542–2552. Springer New York, New York, NY.

Savelsberg, J. J. (2018). Punitive turn and justice cascade: Mutual inspiration from punishment and society and human rights literatures. Punishment & Society, 20(1):73–91.

Schoenfeld, H. (2010). Mass incarceration and the paradox of prison conditions litigation. Law & Society Review, 44(3-4):731–768.

Schoenfeld, H. (2014). The delayed emergence of penal modernism in Florida. Punishment & Society, 16(3):258–284.

Schoenfeld, H. A. (2018). Building the Prison State: Race and the Politics of Mass Incarceration. Chicago: University of Chicago Press.

Seeds, C. (2017). Bifurcation nation: American penal policy in late mass incarceration. Punishment & Society, 19(5):590–610.

Sewell, W. H. (1996). Three temporalities: Toward an eventful sociology. In McDonald, T. J., editor, The historic turn in the human sciences, pages 245–280. University of Michigan Press.

Simon, J. (1993). Poor Discipline: Parole and the Social Control of the Urban Underclass, 1890–1990. Chicago, IL: University of Chicago Press.

Simon, J. (2007). Governing Through Crime: How the War on Crime Transformed American Democracy and Created a Culture of Fear. New York: Oxford University Press.

Skocpol, T. (1979). States and Social Revolutions: A Comparative Analysis of France, Russia and China. Cambridge University Press.

Stinchcombe, A. L. (1987 [1968]). Constructing social theories. Chicago: University of Chicago Press.

Stuntz, W. J. (2011). The collapse of American criminal justice. Harvard University Press.

Thelen, K. (1999). Historical institutionalism in comparative politics. Annual Review of Political Science, 2(1):369–404.

Thelen, K. (2003). How institutions evolve: Insights from comparative historical analysis. In Mahoney, J. and Rueschemeyer, D., editors, Comparative Historical Analysis in the Social Sciences, pages 208–240. Cambridge: Cambridge University Press.

Thompson, E. (1975). Whigs and Hunters: The Origins of the Black Act. New York: Pantheon Books.

Wallerstein, I. (1974). The Modern World System I: Capitalist Agriculture and the Origins of European World? Economy in the Sixteenth Century. New York: Academic Press.

Weaver, R. K. (2010). Paths and forks or chutes and ladders?: Negative feedbacks and policy regime change. Journal of Public Policy, 30(2):137–162.


  1. The continued popularity of social theories of change may be traceable to the field’s origins in sociology and criminology, and not, say, in political science or history. Indeed, there has been a close relationship between punishment studies and social theory (Garland 2018). Garland’s (1990) work summarizing classic social theories’ application to punishment and penal change was a significant contribution that helped consolidate the field of Punishment & Society. While the field has long benefited from multiple theoretical perspectives, it has only recently experienced the sizable influence of theories and theorists not included in Garland (1990).

  1. This omission is part of a broader problem with historical criminology’s engagement with history and historicizing (Churchill 2019).

  1. A search of several leading outlets for punishment studies confirms this impression.

  1. Mahoney (2000, Mahoney and Schensul 2006) also distinguishes between “self-reinforcing sequences” and “reactive sequences,” but I have found this distinction less useful and even less common in the punishment literature.

  1. Institutionalization would also be an example of increasing returns (Pierson 2000a): the more organizations that adopt the practice, the more legitimacy adopting that practice bestows.

  1. For a detailed discussion of my methods, see Rubin (2021).

  1. Savvy readers will recognize that the first part of this argument illustrates a positive feedback effect: maintaining Eastern’s deviance created incentives to its administrators to stay their course. In this case, however, the feedback effect persisted for several decades, relying on pressures in the penal field, and then it began to fizzle over time as conditions changed, rather than a classic example of a feedback effect that gains strength over time, making a course change increasingly difficult.


A previous version of this paper was presented at the British Society of Criminology’s Historical Criminology Network (HCNet) workshop on path dependence. I thank Thomas Guiney and Henry Yeomans for the invitation to speak and gather my thoughts on this important topic, as well as all of the participants for their thoughts and questions that helped sharpen this contribution. I am also grateful to David Dagan, Thomas Guiney, and Henry Yeomans for reading and providing helpful comments on a complete draft of this paper. All errors remain my own.

No comments here
Why not start the discussion?