Introduction: origin stories—the field’s and mine

Many scholars have told origin stories of the socio-legal studies field, situating its beginnings in American legal realism, British law-in-context studies (Twining, 1978, 2012; Nelken, 2009), the American Sociological Association’s sociology of law subdivision (Levine, 1990; Garth and Sterling, 1998), American desires to add social science rigor to legal and political movements for social change in the 1960s (see earliest issues of the Law & Society Review, 1966–1974) and the observation by both legal and social science scholars that law’s impacts and institutions needed to be studied empirically to assess, among other things, the “gap” between the “law on the books” and the “law in action” (Pound, 1910; Friedman and Macaulay, 1969; Abel, 1973a, 2010; Macaulay, 1984; Friedman, 1986; Menkel-Meadow, 1990; Trubek, 1990; Calavita, 2016). Others have attempted to document a canon of classic works which define our field (Seron, Coutin, and Meeusen, 2013; Morrill and Mayo, 2015) or key concepts that define what we are studying: for example, Weberian “rationality” of the law (in regulatory and bureaucratic institutions; Hawkins, 1984; Ayres and Braithwaite, 1995); Durkheimian social statistics, patterning and other positivist empirically based claims (crime and divorce rates, court statistical studies and political attitude studies; e.g., Gibson, 2009); Marx-informed studies of class and power in law (e.g., Galanter, 1974); and more subtle influences of culture and everyday practices and understandings of law (e.g., Ewick and Silbey, 1998) as constituting a multiplicity of simultaneously existing systems of rules and norms, known as legal pluralism (Merry, 1988). Though economics was somewhat present at the multidisciplinary “birth” (mostly in the form of economic history; Hurst, 1964; Horwitz, 1977, 1992), more recent co-optation of socio-legal concepts and research by behavioral economics and psychologists has added some forms of empiricism to the study of the “efficiency” of laws and legal institutions (Posner, 1973, 1995; Langevoort, 1998, 2001).

Each of us who choose to work within these fields, concepts and ideas can probably remember what first drew us to the multidisciplinary study of law and legal institutions (see Garth and Sterling, 1998), which is how I define socio-legal studies (Sterett, 2015), or as one commentator famously put it “real law” (Calavita, 2016). Below, as illustrations, I recount some of the ways I conducted my own socio-legal research (primary research, secondary research, meta-analyses, synthetic analyses and some “mid-level” theory development in dispute resolution and negotiation). In the sections that follow I discuss a few notable examples of “good” uses of socio-legal research, in empirics, theory, policy and doctrinal treatments, and a few “abuses” or failures to properly use empirical data or other forms of socio-legal analysis.

For me, that first moment of “Durkheimian epiphany” (Menkel-Meadow, 1990) came when, as a young poverty lawyer, trained as an undergraduate in sociology by some of the leading scholars in that field at Columbia University, I was preparing for my first trial. My social scientist husband was studying regression analysis, time series and other empirical methods in the University of Michigan-Interuniversity Consortium for Political and Social Research, and I attended some of his classes (circa 1972). While preparing for trial and considering how a judge or a jury might draw conclusions from the evidence I was presenting, I read Kalven and Zeisel’s now classic empirical study of The American Jury (1966). I fell in intellectual love. How could lawyers prepare a case, any case, without understanding the patterns and dynamics of the different kinds of decision makers they might appear before. Yes, every case has its own evidence, but judges and juries bring with them the sum of their human experiences. They are located in communities, in professions, in societies and in relationships (familial, religious, workplace, friendships) that are likely to affect how they process the evidence they hear and the decisions they make in any case. ¹ How could anyone only read statutes, cases and documents, or even only interview witnesses as their legal preparation for a case? All cases are situated in ongoing systems, institutions, communities, states, localities, societies and particular legal systems. My study of sociology told me these things would influence decision making in cases and would vary in different matters, cases and locales. I was hooked. From that moment on, I worked to add social science to the rest of my own legal education (continuing in graduate sociology courses), later my practice as a poverty and civil rights lawyer, and then to my own teaching of trial advocacy, negotiation, discrimination law, civil procedure, legal ethics and the legal profession, and now international law.

From the moment I began to teach and research (and still practice) law, I always asked socio-legal (multidisciplinary, informed by different disciplines) questions:

1. With what assumptions are we working? How do we “know” what we think we know? What theories, ideas or frameworks are guiding our thinking (Mannheim’s sociology of knowledge, 1929)? Where do they come from?
2. What are the statistical/quantitative patterns in this legal issue, case, matter, process (Durkheim, 1893, 1897)?
3. Who stands to gain/lose by particular processes, substantive laws, institutions (Galanter, 1974)?
4. What human processes, besides the desires for principle (jurisprudence; Hart, 1997) and rationality (Weber, 1954), inform and affect what people actually do inside legal institutions (psychology, emotion, culture, ethics, religion) as they act individually or in social groups (structural-functionalism; Parsons, 1951; neo-institutionalism)?
5. What other fields of multidisciplinary knowledge help us understand the role of law and legal institutions embedded in social interactions? Beyond the classic constituent fields of socio-legal studies (law, sociology, political science, psychology, history, anthropology, economics), what do we now learn from game theory, decision theory, organizational development and newer methods of quantitative (e.g., artificial intelligence and mass data mining) and cultural-qualitative analysis?

So, when, as a young legal scholar, I planned my first major research project (with the same aforementioned social scientist husband) to study how overworked poverty/legal aid lawyers allocated their work priorities, without fees as a motivating factor (US legal aid lawyers are salaried employees), we designed a triangulated study of several methods (primary data collection) to understand the possibility of a variety of different factors in work allocation decisions (see Meadow and Menkel-Meadow, 1983, 1985; Menkel-Meadow, 1998). With a research grant awarded by the United States National Science Foundation, we studied two sets of legal aid lawyers (one urban, the other rural), by conducting unstructured interviews, doing observations (we and trained research assistants “trailed” each lawyer for three random days of work and recorded, not privileged content, but particular tasks undertaken) and analyzing data from time sheets filled out by the studied attorneys for one random complete week of work. Multiple data sources provided several windows on work allocation decisions, with both aggregate and statistical analysis of most common work tasks undertaken, narrative explanations (both by the subjects of the study and by “external” observers) and analysis of formal work rules, supervisory and other standards of work. Though I was advised not to conduct this kind of lengthy, labor-intensive, data-driven, co-authored and non-doctrinal research before obtaining tenure in an American law school, I did it anyway.

I then encountered a first “abuse” of socio-legal studies. The Reagan administration in the 1980s sought vigorously to defund American legal aid, and a former law school classmate of mine in the federal government sought to have me testify about my data to demonstrate that legal aid lawyers were deciding “on their own” to sue the very government that was funding them and were choosing “leftist” political issues (e.g., school desegregation, immigration, abortion, consumer law reform and other kinds of class actions) to engage in social and legal redistribution projects rather than sticking to individual case processing (evictions, divorces, hearings for government benefits, etc.). Legal aid was not completely defunded, but at the time government funding was seriously reduced and has not yet, as of this date, recovered federal funding at equivalent levels, despite several Democratic administrations since. I was able to deflect “invitations” to testify as I knew my data would be “re-interpreted” by hostile political forces, but I voluntarily suppressed publication of some of my work due to fear of its misuse.

As a young scholar, bridging law and several other disciplines, my major early work was a normative theory of legal negotiation, drawing from work in political science, diplomacy, philosophy, game theory, sociology, social and cognitive psychology, anthropology, economics, decision theory and planning theory, in fact derived from observations I had made in my early poverty law practice, later expanded to all lawyering (Menkel-Meadow, 1984). I argued from these many different disciplines that the lawyer’s default conception of scarce resources and distributive bargaining produced adversarial behaviors that in turn produced brittle, and often binary, outcomes which were often not efficient, social welfare-enhancing or appropriate for client needs. Instead, I proposed and outlined another “framework” (middle level theory) of conceptualization of integrative bargaining and behavior (itself derived from the early interdisciplinary work of Mary Parker Follett, 1995). This work, which coincided with larger cultural developments—the publication of Getting to Yes (Fisher and Ury, 1982) and later the fall of the Berlin Wall (see: Menkel-Meadow, 2013b; Menkel-Meadow, Love, Schneider, and Moffitt, 2018)—set the ground for my life’s work as a conflict resolution scholar and practitioner.

My next Durkheimian epiphany came from secondary analysis. In what was, and continues to be, one of the most successful multinational socio-legal projects, Rick Abel (of UCLA) and Philip Lewis (Oxford Centre of Socio-Legal Studies) convened, over many years, a group of scholars of the legal profession to undertake a study of the legal professions in common, civil and “mixed” legal systems (18 different countries on five continents). It was finally organized into a three-volume study of Lawyers in Society (Abel and Lewis, 1988, 1989, 1995), and my task was to collate and analyze the data of the work of women lawyers in all countries participating in the study. I remember the “Eureka!” moment when, in completing my analysis, I noticed that women were present in some form as lawyers, judges or other legal professionals in all of the represented countries, but, when one analyzed the rank and status of particular occupations, women were always found in the least-valued or least-prestigious segment of the profession, evidencing the kind of occupational segregation being observed in most occupational sociology studies of the time (Menkel-Meadow, 1989). Today, these Durkheimian patterns of gender differences in the profession are ardently debated as scholars of the legal profession continue to report at national and gender levels of data collection (see Schultz and Shaw, 2003; Abel, Sommerlad, Hammerslev, and Schultz, 2019–2020) and debate the social meanings of gender (and now other) diversity in the legal profession. As a scholar of negotiation behavior, I continue to be involved in this project, arguing that women legal actors still work from gendered assumptions as perceived by others and enacted by themselves (Menkel-Meadow, 2012).

In a later project of a meta-analysis of empirical work on courts and civil procedure reform (which I co-authored with Bryant Garth, 2010), we explored the “uses” and “abuses” of the “commissioning” of empirical studies by courts and other interested legal bodies (both plaintiff and defense attorneys), seeking to justify or support particular policy outcomes and civil procedure rule changes. I was a “victim” of such demands when I conducted a study of the effectiveness of a mandatory arbitration program for a federal court, which was then used to lobby in Congress for the continuation of the program by interested officials seeking to reduce dockets, even though I felt the results were not sufficiently rigorous to publish them in a peer-reviewed social science journal. As another example, when I conducted a rigorous study of multiple forms of dispute resolution in a local city, suggesting some issues with the use of mandatory mediation in landlord–tenant cases, my report was not publicly revealed by the commissioning organization which wanted to promote the continuing use of mediation (without, in my view, adequate legal protections for some litigants appearing without counsel).

I relate these stories of my own social science and law experiences, as both a scholar and a commissioned consultant, to demonstrate what the rest of this essay explores more fully—socio-legal studies are essential for true understanding of how law and legal institutions operate, as embedded in other social processes and institutions. But, even with rigorous methods and testable theories, courts, political stakeholders and advocates are able to “use” socio-legal studies for theory development, data analysis and policy initiatives, but also socio-legal research may be ill used for advocacy or, more disturbingly, not used at all, or ignored when empirical claims are at issue. In the sections which follow I provide some examples of both.

“Good” uses of socio-legal studies: verifiable patterns of socio-legal behavior—ideas, empirics and policy

The best of socio-legal interdisciplinary work provides theories, concepts, testable hypotheses and robust empirical findings to understand the interaction of laws, legal actors (judges, lawyers, policy, juries, litigants and lay people) and legal institutions with the people and other institutions that are affected by law. Socio-legal scholars can study law as the dependent variable (law as affected by other social forces, processes and institutions) or as the independent variable (how does law affect other social institutions and the “acted upon” by law; Epstein and King, 2002; Epstein and Martin, 2014). Data collected to study law and legal processes can be quantitative—large aggregate data (as in court statistics, survey data or newer forms of large scale text analysis, as in the Comparative Constitutional project and “word cloud” analysis; see e.g., Law, 2017) or smaller sets of properly (randomly) selected samples of data. Data collection can be qualitative too, as in narratives, interviews, analysis of texts, use of focus groups and other methods. Those who study law and culture are often concerned with the mutually constitutive and interactive relations of law and the social (Mezey, 2015). Those who engage in socio-legal studies may be law trained or trained in one of the relevant social sciences (political science, sociology, psychology, anthropology, economics, history) or these days in both (joint degrees are increasingly common; George, 2006). Many scholars now identify as “socio-legal scholars” as an independent, if not methodologically uniform, discipline. ² “Good” socio-legal studies have contributed new ideas or tropes about law and its enactment in social life, reported on relatively robust empirical patterns of behavior and been useful in some policy initiatives. I review a (non-systematic) sample of those here.

“Misuses” (or lack of use) of socio-legal studies

From the time of my first Durkheimian epiphany as a law student, I began to be hypersensitive to how often judges made empirical claims and statements in their opinions without any reference to empirical data. Others took note of this in a variety of textbooks and articles (see, e.g., Loh, 1984; Heise, 2002; Epstein and Martin, 2014; Monahan and Walker, 2014) designed to teach law students, lawyers and judges how to do, and understand the significance of, empirical research. Others have criticized whatever use the Supreme Court has made of empirical data in constitutional decision making or challenged the very notion that empirical data are important for judicial decision making (see, e.g., Zick, 2003; Edwards and Livermore, 2009).

In what has been the biggest rejection of rigorous social science in the courts thus far, the United States Supreme Court in 1987 rejected data from years of detailed study of the disparities in the use of the death penalty. In McCleskey v. Kemp, the Supreme Court refused to credit the decades of work done by David Baldus and his colleagues to document racial differences in death penalty punishment rates. In the case, a Georgia African-American man challenging his death penalty conviction asserted it was a “cruel and unusual punishment” (under the 8th Amendment) and a denial of “equal protection” (under the 14th Amendment of the US Constitution), where data robustly demonstrated that black men convicted of killing white victims were more than four times more likely to receive the death penalty than those convicted of killing black victims. Other racial disparities in death penalty sentencing had been demonstrated with high degrees of significance, but the court concluded that none of this met the constitutional standards of requiring “proof” of actual intent (rather than disparate treatment inferred from data) of discrimination (Ellsworth, 2012; Barnes and Chemerinsky, 2018). For a different approach, in Canada, see Perryman (2018). And, more recently, much of the litigation in constitutional same-sex marriage cases in the United States presented social science evidence about family relationships and childrearing (see, e.g., Hull, 2017).

For several years I taught a law and society seminar in a law school and, after introducing some of the classic works of the field and having papers presented by some of the leading American socio-legal scholars, asked students to read critically some major American decisions which relied on empirical statements about the world, but which did not examine any empirical data at all. Judges instead relied on their own “experiential” data and unsystematic personal observations rather than on rigorously and scientifically collected empirical data. Students were then asked to design a research plan for how to actually study the phenomenon at issue, to conduct a literature review of what empirical work had already been done on that question, and then to either (1) attempt a small-scale study of the phenomenon under review or (2) rewrite the opinion with proper references to existing empirical studies. ¹⁰ This seminar’s purpose was for students to understand the role of empiricism in law and critically examine legal judgments, opinions and decisions for the rest of their lives and to notice when data are absent, ignored or misapplied.

Some have criticized the methodological weaknesses of much of what passes for empirical legal research (Epstein and King, 2002) and have urged that all law students and judges need to be better versed and educated in rigorous assessments of data analysis. This is extremely important for many reasons, but particularly because, as I have alluded to above, many of us have been “pulled by the policy audience’s” request to “create” studies to “prove” predetermined desired policy goals, as in requests to justify various forms of legal process (e.g., mediation or litigation) or procedural rule reform (both criminal and civil), and especially with regard to desires to increase efficiency and reduce costs of some legal process without sufficient attention to other values (see Menkel-Meadow, Love, Schneider, and Moffitt, 2018, Ch. 14). Others accuse (rightly in some instances) some social scientist legal (forensic) experts of being “bought” (or, worse, “whores”) for particular “sides” of litigation and policy debates (see, e.g., Kousser, 1984).

More significant are the difficult epistemological issues of different standards of “proof” in the physical and social sciences and legal standards. Many expert witnesses and scholars have commented on the mismatch of the “beyond a reasonable doubt” standard in criminal law (in prediction testimony for example), and even the “preponderance of the evidence” standard in many civil matters does not square with the probabilistic standards of claims (depending on relative confidence levels) that can be made from empirical studies (see, e.g., Robbennolt, 2002; Greiner, 2008; Haack, 2014). Causal analysis in the social sciences does not match up with the legal standards for causation in any of its domains (torts, discrimination, environmental and epidemiological harm), and so empiricists and experts are disappointed they are not fully credited in legal proceedings, and lawyers and judges are often frustrated at their inability to use empirical research in a dispositive way. ¹¹

At the same time, there is the concern that socio-legal (and all social science data studies, as well as the newer forms of “algorithmic” logic with deep data mining), coding and quantification may tend to assimilate in coding boxes and classifications that which is not really uniform. I recall my own experience in listening excitedly to an early classic study of socio-legal studies at an LSA meeting (Susan Silbey and Sally Merry’s classificatory study of mediators, 1986) and being influenced by their conceptualization of “settlers” (task-oriented agreement-seeking mediators) versus those who were more “therapeutic” and interpersonal in their orientations. Many decades later, as an experienced mediator of hundreds of matters, I find this classification system an artifact of simplistic conceptualization and the need to “codify” and reduce a great variety of mediator orientations, moves, techniques, goals and behaviors into simple, binary categories. As a mediator, I use many tools, techniques, goals and behaviors, and these depend greatly on context, the kind of matter, what is at stake, the parties, whether they have lawyers, and their many kinds of needs—I, and most of my mediator colleagues, do not fit neatly into these “either/or” boxes. I had similar reactions (in print) to the coding of so-called “similar” case types and courts in the RAND studies (Kakalik, Dunworth, Hill, McCaffrey, Oshiro, Pace, and Vaiana, 1996) of the effectiveness of ADR in the federal courts (Menkel-Meadow, 1997). As a practicing lawyer, I had litigated in both the Eastern District of Pennsylvania and the Central District of California, and these courts had very dissimilar caseloads and local legal cultures. From my having been inside both of these court systems, the comparisons in the large aggregated study of case treatments seemed inapposite to me.

Too much reliance on aggregate data for “predictive” models is also a potential misuse of empirical data in the legal system. Virginia Eubanks studied the use of algorithms and mass data sets to diagnose and predict who is likely to be neglecting and abusing children, failing to cooperate with electronic welfare and health benefit information requests, and eligible for housing benefits and assistance for homeless populations (Eubanks, 2018). This important work flags the epistemological, as well as methodological, differences in the algorithmic “predictive” thinking of our new age of artificial intelligence decision making, as contrasted to the legal standards of due process and individual adjudication “after,” not “before,” the facts have been “proven.” Recently, the United States Supreme Court refused to grant review (certiorari) of a claim that it was a denial of due process to use algorithmic predictive models for risk assessment in criminal law sentencing (see Loomis v. Wisconsin, cert denied June 26, 2017). Thus, judicial decision making can be problematic from both “too little” and “too much” data—depending how data are used and what legal standards are applied.

As I have come to be skeptical of judicial decisions that make empirical claims without empirical data and support, I now find myself sometimes equally skeptical of the empirical claims made in socio-legal studies that codify and classify phenomena, behaviors and people which I know are far more complex and context variable than empirical “measurements” permit in many studies. ¹² Thus, after many decades in the field, I often find that what is being described with data sets, both large and small (e.g., many claims about ADR or litigation generically), doesn’t ring true to me in situations where I have been a participant or witness to the phenomena described. ¹³ This has heightened my own respect for more nuanced ethnomethodological and participant observer studies, with long narratives and descriptions and fewer grander, more general claims. In other areas in which I work, such as gender difference in negotiation behavior, both aggregate and smaller case studies remain inconclusive, largely because different research designs in so many different contexts have produced such variable findings (see Menkel-Meadow, 2012, and Hollander-Blumoff, Chapter 12, this volume)—context really matters, and some things we study are highly interactive. Yes, ultimately, I still believe that all methods and different studies tell us something, but our field has not engaged in enough rigorous replication studies and iterated testing of findings. ¹⁴

One notable “misuse” of socio-legal studies has, for me, been the “co-optation” of early socio-legal research (e.g., group decision making as in jury studies, “non-rational” factors in decision making, empirical studies of legal behavior) by law and economics scholars who now claim a new field of “behavioral economics” (Jolls, Sunstein, and Thaler, 1998; Ulen, 2013). After decades of classical economics “assumptions” about human behavior (and of “maximizing utilities”), in the late 1990s, law and economics scholars and social and cognitive psychologists began to realize that not all “men” (I use that word deliberately) reasoned exactly the same way, under all conditions, and that—aha!—decision making about legal and other matters might vary with context (see also Klass and Zeiler, 2013). Much of the work of behavioral economics has been extremely useful for the study of law, legal actors and legal institutions in many settings (e.g., plea bargaining; negotiating; legal decision making; resource allocation; incentive structures for legal compliance, including its use in many policy settings; risk analysis; regulatory policy [Sunstein, 2011]; environmental law; labor relations; and contracting generally), but the failure to read, cite and properly give attribution to those socio-legal scholars (particularly sociologists, psychologists and anthropologists) who have harvested these concepts—often with extensive data!—is academically shameful. This appropriation of concepts and decades of empirical work sadly demonstrates the failure to integrate a truly multidisciplinary field—we all still seem to work within silos as we study the law from outside of doctrine. Legal economists seem to think that they have invented empiricism in law! ¹⁵

Implications: the limits of disciplinary thinking and a call for rigorous multidisciplinariety

From where will the Durkheimian epiphanies of the future come? This volume contains a set of many examples of law and … disciplines. Examples are drawn from research being conducted all over the world now, not just in Anglo-American spheres of influence (or intellectual colonialisms?). Theory, methods, data and sources of interpretation are diversifying as our socio-legal field moves from studies of legal pluralism to actual migration of thought and globalization. As this volume attests, there is so much interesting work being done, it is hard to keep up with it all.

The expansion of studies of law and … to so many new sites, with so many new kinds of data (major data sources from modern technology capture; Eubanks, 2018) and new lenses of interpretation, presents some challenges (e.g., Barnes, 2016)—some old, some new:

1. Does our field have (need) a canon? What are the key propositions of knowledge that characterize what we know about how law functions in society and how social forces structure law and legal institutions? What new concepts or tropes are we studying/should we be studying—for example, mixed “identity” and classifications (Mezey, 2003), “adjectival” constitutionalism, subnational and supranational legal systems (e.g., Shari’a law)?
2. Do we have clearly accepted methods (social science reliability, validity and causality standards, other measures of reliability)?
3. How much replication/cumulation of knowledge do we have/desire? (Should we be emulating the physical sciences, or do we have our own ways of evaluating knowledge production?)
4. How much should we commit ourselves to policy research? (Are we an applied science? How does this vary from legal culture to legal culture?)
5. Is the field committed to particular social/political/legal values (social justice, equality, democracy), or can it be a “neutral” or objective field of study? ¹⁶
6. With what sense of ethics and responsibilities should we approach those we study? Beyond the now required institutional review boards for study of human subjects in most major universities (Schrag, 2010), what else do we owe those we study—beyond consent, explanations of our work and assistance in solving the problems we see (lack of housing, racialized policing, “crimmigration” or injustice; Menkel-Meadow, 2009)? ¹⁷
7. With so many new comparative, global and diverse studies of socio-legal phenomena, what are our comparators? How do studies in legal regimes (constitutional, authoritarian), processes (litigation, alternatives to litigation, criminal, civil, administrative) and actors (lawyers, judges, police, tribunal officials) explain when so much of this is now hybridized, and nothing is a pure “ideal type”?

I will not belabor here the arguments I have made in many places (see, e.g., Menkel-Meadow, 2007) that essential to any modern legal education should be rigorous training in the constituent social sciences that comprise, affect and influence law, and the evaluation of its impacts. As our knowledge about law and legal institutions and their efficacy increases by study of more sites of norms and rules (internally and from “below”), as well as from outside (top–down for policy initiatives) and through the lenses of many more multidisciplinary fields (including hybrid fields such as urban planning, decision theory, artificial intelligence, cultural studies), no modern lawyer, judge, government official or bureaucrat should only study or know the “doctrine” of law, without understanding something about how that doctrine will (or will not) be used. Nor should they assume their “experiential” knowledge is superior to independent empirical findings. For those “outside” of the law (or the acted upon), it is even more imperative to know how the law on the books is still so far from the law (or justice) in action. In so many areas of legal and social policy, we need all the data and help we can get to learn what is happening on the ground and what policy or advocacy interventions work or do not work—consider current issues of worldwide migration; labor displacement and loss; “crimmigration”; ongoing discrimination and inequality on grounds of race, ethnicity, gender and class; human rights violations and enforcement; whether “rights” and legal movements are effective in producing more social change; and how the use of technology will alter our legal and social relations, perhaps forever. I look forward to many more studies, using many different methods, to expose the newest forms of Durkheimian epiphanies that should illuminate what the law promises us and when it fails.