¹ Let me summarize the usage I have tried to follow as I have devised a terminology for talking about these things. The basic notion is a set of cases that have shared features and that exert influence upon one another. Obviously these are relative judgments; we are in the realm of more and less. I use the term populations to refer to sets that are very large and long-lasting, such as automobile injury cases, divorce cases, and collection cases. It may be that the members of such populations have less influence on one another than do the members of the more specific groupings that I call congregations (cf. Rheingold, 1982:3). I can also imagine a set of very closely related cases (for example, involving a single party or incident) that might be called a family of cases.

² Product liability cases have been counted separately in the Annual Reports since 1974; asbestos cases have been counted separately since 1984. See United States Administrative Office of the United States Courts, Annual Reports of the Director for various years, Table C-2. Conversely, waning congregations may cease being of interest and cease to be counted separately—e.g., the Administrative Office's Annual Report stopped counting selective service cases after 1978, as displayed in Figure 1 below.

³ This and some other categories of federal litigation have the special virtue of providing a comprehensive count of all cases of a particular type. The example suggests that the explanatory power of the underlying activity hypothesis may be most clearly displayed when there is a sudden and dramatic decline in a type of activity.

⁴ Munger (1988) and McIntosh (1983) elaborate and test this curvilinear model, which is found in the work of Friedman (1976) and others.

⁵ Some major contributions to this perspective, which is summarized in Galanter, 1986b, are Felstiner, Abel and Sarat, 1980–81; Miller and Sarat, 1980–81; Mather and Yngvesson, 1980–81.

⁶ This example points back to the ambiguities of the underlying activity hypothesis. I think it would be generally agreed that this was a period of decreasing discrimination in some general sense. As segregation declined there may have been more occasions for individuals to experience discrimination practiced against them personally by identifiable individuals; certainly there were more occasions to experience discrimination that was perceived as legally remediable.

⁷ A statute (34 U.S.C. sec. 3404) passed just after the Civil War provides that fees for attorneys' services in claims under laws administered by the Veterans Administration “shall not exceed $10 with respect to any one claim.” This has conferred on the VA a unique immunity from having its benefit determinations challenged in the courts. This limit, the Supreme Court has recently held, does not violate due process of First Amendment Rights (Walters v. National Assn. of Radiation Survivors, 473 U.S. 305, 87 L. Ed. 2d 220, 150 S. Ct. 3180 (1985)).

⁸ The presence of a similar pattern in Ontario suggests to Flanagan (1987: 41) that the solution to the puzzle “is not related to the idiosyncratic aspects of the administration of the NLRA [National Labor Relations Act] but may instead be rooted in behavioral response to the particular approach to regulation of labor relations adopted in both countries.”

⁹ Rules regulating fee shifting and triple damages which clearly affect stakes are usually enacted by legislatures but are interpreted and applied by courts. Other features of litigation also affect stakes. For example, in Flanagan's unfair labor practice complaints, the low level of formal sanctions imposed on an employer for a violation never puts that employer in a worse position than if he had complied initially (Flanagan 1987: 82). Flanagan wonders why there is any compliance at all. Part of the answer must be in the transaction costs of the proceedings and in the sanctions that workers and unions can apply (slowdown, low morale, turnover, sabotage, strike). Litigation frequency can be so responsive to the stakes in the underlying transaction precisely because the litigation itself does not attach new stakes to offset the perverse incentives that he describes.

¹⁰ Of course, these messages do not neatly segregate information on substantive rights and obligations from information about diffuse costs and risks. Thus businesses may engage in various defensive maneuvers designed to reduce exposure to suit. See, e.g., Lindsey (1985) (“defensive architecture,” restaurant identification of intoxicated customers, elimination of home counselling visits by ministers, etc.); Johnson (1985) (reluctance to provide references for former employees); Purnick (1985) (delay of construction to permit archaeological dig: “What we're doing is to prevent a lawsuit...”).

¹¹ These preventive effects are often lumped under the heading of deterrence. I have followed Gibbs (1975) in trying to separate out a number of distinct effects. See Galanter (1983b).

¹² An earlier Conference Board survey of the risk managers of major United States corporations reported that product liability and the related crisis of insurance availability “have left a relatively minor dent on the economics and organization of individual large firms” (Weber, 1987: 2).

¹³ Cf. a 1987 survey of psychologists, in which only 1 percent responded that it was not ethical to break confidentiality if a client was homicidal (Goleman, 1987).

¹⁴ Friedman and Ladinsky (1967). Cf. Stookey's (1986: 299) depiction of the introduction of workers' compensation as the decline to zero of the ratio of tort cases to mining accidents.

¹⁵ Virginia Law Review (1947); Illinois Law Review (1936). The extensive literature generated by this legislation does not include any estimate of the amount of such litigation. But there was a general concurrence with the view of a pioneer proponent of such legislation that “90 percent of the heart balm suits ... are blackmail suits ... which attempt to capitalize on some one's indiscretions” (Indianapolis News, 1935).

¹⁶ Fedders and Guttenplan (1980); Allen (1987: 1) observes that “[m]any of the shredding programs hinge on the premise that in the event of a lawsuit nearly any corporate document—including executives' diaries and seemingly innocent memorandums—can become dynamite in the hands of an opposing counsel.”

¹⁷ A nice example of educative effect of earlier cases is Brickey and Miller's (1975) description of the way that traffic court defendants whose cases are heard later in the sequence learn by observing earlier cases.

¹⁸ I have come across several striking instances of such “buyouts” of the legal resources of a class of claimants, foreclosing or making difficult recovery by the bulk of the class. One was the agreement by plaintiffs' lawyers in the litigation arising from the Hawk's Nest Tunnel disaster to surrender all case records to the defense (Cherniack, 1986: 56, 65, 72). Similarly the earliest asbestos litigation against the Johns Manville Corporation ended in a 1933 agreement that the attorney would not participate in bringing any new suits against the company (Brodeur, 1985: 113–14, 164). According to Brodeur there was no further asbestos disease litigation against Johns Manville for more than ten years—that is, until after the immense wartime exposures that eventually brought the company to bankruptcy.

¹⁹ Gans and Rheingold (1984: 16A-97). Interestingly, the authors justify the plaintiffs lawyer “cashing out” on the ground that his “primary obligation ... is to do the best job for his present clients, and if he can command a premium for their cases or settle their cases earlier than otherwise on the basis of this agreement it would seem his duty to do so.” Brodeur (1985: 92) reports an agreement extracted from successful asbestos plaintiffs' lawyers in 1977 “not to take on any new clients from the Tyler factor work force ... [and] that whatever evidence we had obtained in the course of our discovery ... would not be handed over to subsequent plaintiffs.”

²⁰ This distinction between effects on the underlying activity and on disputing behavior is not absolute, for many activities (for example, loans, franchises, and leases) routinely incorporate—and may even be said to be constituted largely of—arrangements for potential breakdown and dispute.

²¹ I was reminded here of the situation of assigning a grade to the first of a pile of examination papers, when the grader is committed to apply a uniform standard but can only surmise the contours of the whole congregation.

²² Chin and Peterson (1985: 48) report that Cook County (Ill.) juries award multiple plaintiffs 27 percent less than comparable single plaintiffs. Kakalik et al. (1984: 44) report that recovery by asbestos plaintiffs declined from an average of $88,000 in single-plaintiff lawsuits to $63,000 with 2 to 10 plaintiffs to $52,000 with 11 to 25 plaintiffs to $21,000 with 26 or more plaintiffs. See also Coffee (1987: 915–17) on this discount in class actions.

²³ As this example reminds us the incentives for such anticipatory investment may be distributed unequally (see Galanter, 1974a).

²⁴ For example, Rheingold (1982: 2) suggests an interesting distinction between what we might call epidemics (like asbestos) and disasters—congregations that come into being with a single event. If a set of cases starts gradually, he observes, “a few lawyers have already done considerable work and are thus less disposed to share the data they have created.”

²⁵ After the first award of punitive damages against the manufacturer of the morning-sickness drug Bendectin, the plaintiff's lawyer, who had spent six years litigating Bendectin cases, observed:

“we learn more in every one of these cases,” said Mr. [Barry J.] Nace, who has yet to earn any contingent fees pending the resolution of appeals. “After working on it for six years, we are still finding new documents. This subject is filled with difficult areas to learn, and we are finally able to explain it.” (Strasser, 1987)

²⁶ MER/29 was a drug marketed in the early 1960s as an agent for lowering cholesterol (see Rheingold, 1968).

²⁷ Rheingold (1968: 137–38) points out that this was in part a shift in decisionmakers. Cases outside metropolitan areas tended to come up for trial (and thus for settlement) earlier than cases in crowded metropolitan courts.

²⁸ On the course of the litigation, see Brodeur (1985); Hensler et al. (1985); Dungworth (1988: 35–38).

²⁹ But note that even as it was taking its fatal plunge into bankruptcy, Manville had won nineteen of twenty-eight tried cases (Brodeur, 1985: 255).

³⁰ Of asbestos cases from 1977 to 1986, 3 percent went to trial; of personal injury product liability cases in ten courts from 1980 to 1986, 9 percent went to trial (Willging, 1987: 25, 27).

³¹ Alternatives to the High Cost of Litigation (1988). In 1988 the Trust was joined by another trust with assets of $2.4 billion to administer the claims of Dalkon Shield victims (Cooper, 1988; Freudenheim, 1988).

³² The figures given here are only for federal cases. On the asbestos cases in the state courts, see Hensler et al. (1985). On product liability cases generally in the state courts, see U.S. General Accounting Office (1988, 1989). The Administrative Office reports did not count asbestos separately until partway through the 1984 statistical year. Estimates of asbestos filings in earlier years are presented in Dungworth (1988: 36).

³³ The drop in nonasbestos product liability filings presumably reflects changing trends in the outcome of such cases. In a pioneering analysis of outcomes of product cases, Henderson and Eisenberg (1990) found that after the early 1980s plaintiffs were less successful at trial and defendants won an increasing portion of doctrinal victories.

³⁴ This seems to fit neatly with Abel's (1973) notion of the mix of dependence and autonomy that characterizes highly differentiated dispute institutions.