The author is grateful for the help of Colette Chambellard, Dan Clawson, Sam Cohn, Stanley Engerman, Ken Fones-Wolf, Herb Gintis, Claudia Goldin, Michael Hanagan, Carol Heim, Debra Jacobson, Bruce Laurie, Mike Podgursky, John Stifler, Judith Stone, Paul Swaim, Charles Tilly, Jeffrey Williamson, two anonymous referees, and participants at the Harvard University Economic History Workshop and the 1986 Cliometrics Conference.

¹ The socialists include over 40 union leaders who served in the Chamber of Deputies. Many union leaders earned a place on the infamous carnet B list of revolutionaries to be arrested in case of war. See Maitron, Jean et al. , Dictionnaire biographique du mouvement ouvrier francais, 6 vols. (Paris, 1973–1977);Google ScholarBecker, Jean-Jacques, Le Carnet B (Paris, 1973);Google ScholarMaitron, Jean, Le Mouvement anarchiste en France (Paris, 1975).Google Scholar

² The Dutch delegate to the 1910 International Trade Union Congress quoted in Gompers, Samuel, Labor in Europe and America (New York, 1910), p. 133.Google Scholar

³ Ridley, F. F., Revolutionary Syndicalism in France (Cambridge, 1970), pp. 11–15.Google Scholar

⁴ The long debate on American socialism is summarized in M. Laslett, John and Martin Lipset, Seymour, eds., Failure of a Dream? (Garden City, 1974).Google Scholar Also see, Wilentz, Sean, “Against Exceptionalism: Class Consciousness and the American Labor Movement,” International Labor and Working Class History, 26 (Fall 1984), pp. 1–24;CrossRefGoogle ScholarSombart, Werner, Why Is There No Socialism in America (1st edn., 1906; 1st English edn., White Plains, 1976)Google Scholar; Perlman, Selig, A Theory of the Labor Movement (New York, 1928);Google Scholar and Martin Lipset, Seymour, “Radicalism or Reformism: The Sources of Working Class Protest,” American Political Science Review, 77 (03 1983), pp. 1–18.CrossRefGoogle Scholar

⁵ See Stearns, Peter, “National Character and European Labor History,” Journal of Social History, 4 (Winter 1970), pp. 95–124;CrossRefGoogle Scholar and Stearns, Peter, Revolutionary Syndicalism and French Labor (New Brunswick, 1971)Google Scholar; Shorter, Edward and Tilly, Charles, Strikes in France (Cambridge, 1974).Google Scholar

⁶ Wilentz, , “Against Exceptionalism,” p. 18.Google Scholar

⁷ Aveling, Edward and Marx-Aveling, Eleanor, The Working-Class Movement in America (London, 1891);Google Scholar the preface to the 1887 English edition of Friedrich Engels, The Condition of the Working Class in Manchester in Marx, Karl and Engles, Friedrich, Letters to Americans, 1848–1895: A Selection (New York, 1953).Google Scholar

⁸ See Gallie, Duncan, Social Inequality and Class Radicalism in France and Britain (Cambridge, 1983).Google Scholar

⁹ Engels, to Sorge, , 12 2, 1893,Google Scholar in Marx, and Engels, , Letters, p. 258.Google Scholar

¹⁰ Established unions can conduct collective bargaining without strikes, since both employers and unions know from experience of the other's ability to inflict costs in a labor dispute. This was rarely the case in pre-Worid War I labor relations in the United States or France, however. Collective bargaining was unusual in both countries, and employers and workers had reasonable doubts about the ability of new and young unions to influence strike outcomes. See Bureau, Paul, Le contrat de travail (Paris, 1902);Google ScholarLaroque, Pierre, Les Rapports entre patrons et ouvriers (Paris, 1938);Google Scholar and Conell, Carol, “The value of Union Sponsorship to Strikers” (Ph.D. dissertation, University of Michigan, Ann Arbor, 1980).Google Scholar

¹¹ See Shefter, Martin, “Trade Unions and Political Machines,” in Katznelson, Ira and Zolberg, Aristide, eds., Working-Class Formation (Princeton, 1986), pp. 197–276;Google ScholarMarks, Gary, “Unions and the Organizational Revolution: The Experience of Britain, Germany and the United States” (unpublished manuscript, University of North Carolina, 03 1987).Google Scholar

¹² This work is based on data for more than 7,000 French strikes 1895–1899 and 1910–1914 collected by Shorter, Edward and Tilly, Charles from the annual publication, France, Direction du travail, Statistique des grèves (Paris, 1895–1914)Google Scholar. I have collected similar data for over 2,000 individual American strikes 1881–1894 from the United States, Commissioner of Labor, Third Annual Report (Washington, D.C., 1888)Google Scholar and Tenth Annual Report (Washington, D.C., 1895)Google Scholar. I have also collected summary data on strikes by state for 82 industries in 1903 from ibid., Report (Washington, D.C., 1906)Google Scholar. These data have been linked with other data on regional and industrial conditions described in Friedman, Gerald, “Politics and Unions” (Ph.D. dissertation, Harvard University, 1986). Data on French strikes in 1895 and 1898 have been matched with archival records for the number of arrests and financial support given to strikers in Archives Nationales de France, series F12 boxes 4676−78 and 4685−86.Google Scholar

¹³ The dependent variable in the strike success regressions for individual strikes (France 1895–1899 and 1910–1914 and the United States 1881–1894) is dichotomous, equaling 0 in strikes where the strikers do not gain any concessions and I where they gain some or all of their goals. Logit regressions were estimated using the LIMDEP program. Elasticities have been calculated at the mean values of the independent variables. Only grouped data are available for American strikes in 1903. The dependent variable, the proportion of successful or compromised strikes in a state and industry, is a continuous variable bounded by 0 and 1. Weighted OLS regressions have been estimated for the log odds ratio following the procedure used in Berkson, J., “Application of the Logistic Function to Bio-Assay,” Journal of the American Statistical Association, 39 (09 1944), pp. 357–65;Google Scholar and Berkson, J., “A Statistically Precise and Relatively Simple Method of Estimated Bio-Assay with Quantal Response Based on the Logistic Function,” Journal of the American Statistical Association, 48 (09 1953), pp. 565–78.Google Scholar Strikes are counted as a success in the government reports if the workers gained all of their demands, a failure if they gained none, and a compromise if they gained some but not all of their demands. Although the process of reaching a compromise settlement—found in over 30 percent of French strikes but only 15 percent of American—is different from that followed in a pure success, I count them together, since workers in both cases gained concessions from the employer through collective action. There are also practical reasons for counting successes and compromises together. There is considerable random noise in the data for compromises reflecting the varying extent of demand inflation prior to the settlement. Counting compromises and successes together avoids a bias in favor of “moderate” unions who “win” strikes by limiting their demands as against “aggressive” unions who make more demands but then accept a compromise which may include greater gains than the “moderate” union's success. See the discussion in Cohn, Sam, “Moderation, Social Structure and Strike Success” (unpublished manuscript, University of Wisconsin, Madison, 02 1986), pp. 74–80;Google ScholarShorter, and Tilly, , Strikes, pp. 66–70; and Conell, “The value of Union Sponsorship”.Google Scholar

¹⁴ The dummy variable for union involvement can measure the effect of union involvement on strike success because as late as 1914 many strikes in both countries were conducted without the involvement of a formal labor union. In addition to the union variable, these regressions also include the industry in which the strike occurred, the strike's location, the workers' demand, and the size of the establishment being struck.

¹⁵ Regressions have been also estimated for the union effect on strike size, strike participation (the share of workers joining strikes), strike duration, strike rates, and the probability of an arrest or of outside mediation in French strikes. The results of these regressions are available upon request; the union impact on each of these variables is presented in Tables 2 and 3.

¹⁶ It is important to separate the effects of organization from the effect of other characteristics of the individuals in these organizations. Even without a union, workers who belong to unions may have been among those workers best able to gain concessions from employers, as is shown by their ability to maintain an organization. This may not mean that they were in a good position to make further gains, however. They may have already received whatever concessions employers were ready to concede easily.

¹⁷ Gompers, Samuel, Labor (New York, 1910), pp. 249, 252.Google ScholarDuncan, James in AFL,Proceedings, 1911 (Washington, D.C., 1911), pp. 149–50.Google Scholar

¹⁸ Griffuelhes, Victor, “Les grèves et le syndicalisme francaise,” Mouvement socialiste, 178 (03 15, 1906), p. 254;Google ScholarJouhaux, in Fédération des ouvriers des métaux, Congrès unilaire … 1909 (Paris, 1909), p. 9;Google ScholarYvetot, Georges, ABC syndicaliste (Paris, 1908), p. 33.Google Scholar

¹⁹ Note that this result is after controlling for strike location, industry, establishment size, and strike issue. French unions did not win strikes by restricting their demands to issues employers were ready to concede easily. Over 28 percent of French union strikes 1895–1899 and 1910–1914 were over work rule and personnel issues compared with only 19 percent of nonunion strikes. Strikes over these issues were particularly threatening to employers used to being “master in my own house”. Reflecting this greater employer resistence, strikes over these issues were significantly less successful than were those over wage and hour demands. By raising these issues, unions lowered the positive effect of union involvement on strike success. See, Friedman, , “Politics,” pp. 223, 228, 302.Google Scholar

²⁰ Mediation was usually conducted under an 1892 law entitled “Loi sur la conciliation et l'arbitrage en matiere de differends collectifs entre patrons et ouvriers ou employes.” This generalization does not, however, include all cases where police or troops were used to protect property or the right to work. An arrest occurred in nearly 5 percent of strikes in 1895 and 1898, reducing the probability of success by nearly 24 percent. Also note that union involvement is associated with a reduction of over 20 percent in the probability of arrests.

²¹ Letter from prefect of Saone-et-Loire to Prime Minister Waldeck-Rousseau, , 06 7, 1901, in Archives Nationales de France, series F7 box 12782.Google Scholar

²² Quoted in Shorter, and Tilly, , Strikes, p. 43.Google Scholar See also Néré, Jean, “Aspects du déroulement des grèves en France,” Revue d'histoire économique et sociale, 34 (09 1956), pp. 286–302;Google Scholar and Laroque, , Rapports, pp 146–47.Google Scholar

²³ The remaining cases were due to the state official's own initiative; see Shorter, and Tilly, , Strikes, p. 30.Google Scholar Contemporaries noted that workers were more interested in state mediation than were employers; see Halévy, Daniel, Essais sur le mouvement ouvrier en France (Paris, 1901), p. 88;Google ScholarPerrot, Michel, Les ouvriers en grève (Paris, 1974), p. 701–3;CrossRefGoogle Scholar Jean Jaurès in the Chamber of Deputies on 21 11 1895 in France, Chamber of Deputies, Journal Official (Paris, 1895), p. 2437;Google Scholar and Fagnot, F., “La Loi sur la conciliation et l'arbitrage,” in Gide, Charles, ed., La Droit de grève (Paris, 1909), pp. 199–232.Google Scholar On employer attitudes see Laroque, , Rapports, p. 135;Google Scholar and Raynaud, Barthélemy, Le Contrat collectif de travail (Paris, 1901), p. 74;Google ScholarVogel, David, “Why Businessmen Distrust their State,” British Journal of Political Science, 8 (01 1977), pp. 45–78.CrossRefGoogle Scholar

²⁴ This result is from logit regressions including the strike's issue, industry, location, socialist electoral strength in the department, as well as the strike's size, duration and union involvement.

²⁵ Dues levels in most French unions were under 50 centimes a month, well under 0.5 percent of an average worker's earnings. Even a relatively large union, the Building Trades Confederation, had only 10 centimes (2 cents) per worker in its reserve fund in 1907; see du Batiment, Fédération, Compte rendu du congrès unitaire … 1907 (Paris, 1907), p. 35.Google Scholar

²⁶ Data on the amount of aid received by strikers have been collected for strikes in 1895 and 1898 from the government strike reports in the Archives Nationale de France, file F12 boxes 4676–78 and 4685–86. The amount provided is not reported for half of the cases where aid is mentioned, probably because it was minimal. Estimates of the amount of aid are based on cases where quantities are given. The Metalworkers Union reports that between 05 1, 1907 and 04 15, 1909, 29 strikes were aided with 57,207 francs provided for a total of 121,470 days unemployment, or 47 centimes a day; see du Métallurgie, Fédération, Compte rendu du congrès … 1909 (Paris, 1909), pp. 32, 43.Google Scholar Also see P. Hanagan, MichaelThe Logic of Solidarity (Urbana, 1980), p. 91.Google Scholar For aid given in earlier strikes see Perrot, , Les ouvriers, p. 524; and Néré, “Aspects,” pp. 288–89.Google Scholar

²⁷ Griffuelhes, , “Les Grèves,” p. 255.Google ScholarPelloutier, Fernand, “Grèves,” L'Ouvrier des deux mondes, 1 (02 1897), p. 131.Google Scholar

²⁸ The share of workers in struck establishments joining strikes is significantly lower in strikes conducted by unions with benefit funds; Friedman, , “Politics,” pp. 243–44.Google Scholar

²⁹ Pannetier, , du Batiment, Fédération, Congrès (Paris, 1907), p. 56.Google Scholar

³⁰ The contrast with American unions around 1900 is striking. See Janes, George. The Control of Strikes in American Trade Unions (Baltimore, 1916);Google ScholarFriedman, Gerald, “Union Strikes and Striking Unions” (unpublished manuscript, University of Massachusetts, Amherst, 09 1987).Google Scholar

³¹ See Griffuelhes, Victor, Le Syndicalisme revolutionnaire (Paris, 1909), pp. 14, 16.Google Scholar There is also some evidence to the contrary that French unions may have improved strike timing. Over 26 percent of union strikes began in the high-employment August-October period, compared with only 19 percent of nonunion strikes. See France, , Direction du travail, Salaires et durée (Paris, 1897), pp. 706ff;Google ScholarPerrot, , Les ouvriers, pp. 103–14. High success rates in French union strikes are not due to the involvement of workers in a relatively strong bargaining position. Data on the average wage paid strikers are available only for strikers in 1898 and 1913. In those years union strikers were drawn from the better paid part of the labor force with average earnings 44 percent above those of nonunion strikers in 1898 and 15 percent higher in 1913. These high earnings may have reflected the involvement in unions of skilled workers and of previously successful strikers. High wages are associated with lower rates of strike success, however, perhaps because high wage strikers had previously exhausted their employers' readiness to make concessions. The average daily wage of successful strikers in 1913 and 1898 was 12 percent below that of unsuccessful strikers.Google Scholar

³² Chagrin, , “Le mouvement syndical en Amérique,” Revue syndicaliste, 65–66 (10–11 1909), pp. 194–95.Google Scholar

³³ Trautman, William, Why Strikes are Lost (Chicago, 1911);Google ScholarC. Ford, Earl and Z. Foster, William, Syndicalism (Chicago, 1911);Google ScholarSaint-John, Vincent and Trautman, William, “Les ‘I. W. W.’,” La Vie ouvrière, 43 (07 5, 1911), pp. 28–29;Google Scholar and V. Debs, Eugene, “Working Class Politics,” International Socialist Review, 11 (11 1910).Google Scholar

³⁴ See Riker, William, Theory of Political Coalitions (New Haven, 1962).Google Scholar David Montgomery made this point in noting that skilled workers sometimes opposed the organization of common labores, since it “made it diffcult for employers to grant concessions to their craftsmen at the expense of helpers and laborers.” See Montgomery, David, “Workers' Control of Machine Production in the Nineteenth Century,” Labor History, 17 (Fall 1976), p. 499.CrossRefGoogle Scholar

³⁵ At the KOL's 1886 peak, union strikes were 5 percent less successful on average than were nonunion strikes.

³⁶ Journal of United Labor (03 25, 1885), p. 944.Google Scholar

³⁷ Montagu, F. J. in a letter in the Journal of United Labor (09 15, 1883), p. 562. Unions in 1881 to 1886 have only half as much effect on the probability of strikes receiving financial aid as they do in later years, and such aid replaces only 35 percent of lost wages 1881–1886 compared with 80 percent replaced 1887–1894.Google Scholar

³⁸ See for example Jules Oestreicher, Richard, Solidarity and Fragmentation (Urbana, 1986).Google Scholar

³⁹ Quoted in Ware, Norman, The Labor Movement in the United States, 1860–1895 (New York, 1929), p. 146.Google Scholar

⁴⁰ Legislation authorizing state mediation of labor disputes was enacted in 25 states and on the federal level by 1905. In most cases this legislation was a “dead letter.” Arbitration was earned out in only 8 states with a significant amount of state intervention in only 2. As in France, mediation was associated with larger srikes and with higher levels of strike success. Only 1.61 percent of strikes and lockouts were settled through mediation in 1901–1905, however—too few to significantly affect patterns of American strike success. See Hatch, George, Government Arbitration of Labor Disputes (Washington, D.C., 1905), pp. 570–79, 620–28.Google Scholar

⁴¹ These are not all strikes but also include other cases of civil disorder. On the national guard, see Skowronek, Stephen, Building a New American State (Cambridge, 1982);CrossRefGoogle ScholarSmith, Goldwin, “The Labour War in the United States,” Contemporary Review, 30 (09 1877), pp. 529–41;Google Scholar and H. Riker, William, Soldiers of the State (Washington, D.C., 1957).Google Scholar On the army see Wilson, Frederick, Federal Aid in Domestic Disturbances, 1787–1903 (New York, 1969);Google ScholarCopper, Jerry, “The Army as Strike Breaker,” Labor History, 18 (Spring 1977), pp. 179–98;Google ScholarHacker, Barton, “The U.S. Army as a National Police Force,” Military Affairs, 33 (04 1969), pp. 255–64;CrossRefGoogle ScholarFusfeld, Daniel, The Rise and Repression of Radical Labor, U.S.A. 1877–1918 (Chicago, 1980);Google Scholar and Goldstein, Robert, Political Repression in Modern America (Cambridge, Mass., 1978).Google Scholar

⁴² Le Révolté (26 02 1887),Google Scholar quoted in Murphy, Marjorie, “And They Sang the Marseillaise,” International Labor and Working Class History, 29 (Spring 1986), p. 35.Google Scholar

⁴³ See Goldstein, , Political Repression, pp. 11–15;Google ScholarCouvares, Frank, The Remaking of Pittsburgh (Albany, 1984), pp. 62–63;Google ScholarFrankfurter, Felix and Green, Nathan, The Labor Injunction (Gloucester, 1963), pp. 71–72;Google Scholar and Shefter, , “Trade Unions,” p. 245.Google Scholar French observers were amazed at the readiness of American employers to use violence against strikers. They regarded the use of professional strikebreakers as an American innovation; see Sayous, A.-E., “Le Patronat et la grève,” in Gide, , ed., Le Droit, pp. 138–39.Google Scholar

⁴⁴ Quoted in D. Burnham, Walter, “Periodization Schemes and Party Systems,” Social Science History, 10 (Fall 1986), p. 308.CrossRefGoogle Scholar

⁴⁵ Friedman, Lawrence, A History of American Law (New York, 1973), pp. 86–87.Google ScholarFrankfurter, and Green, , The Labor Injunction, pp. 21, 80, 96;Google ScholarHiller, E. T., The Strike (Chicago, 1928);Google ScholarPaul, Arnold, Conservative Crisis and the Rule of Law (Ithaca, 1960);Google ScholarTomlins, Christopher, The Slate and the Unions (Cambridge, Mass., 1985);Google Scholar and U.S. Commissioner of Labor, Special Report: Labor Laws of the United States (Washington, D.C., 1986).Google Scholar Even more threatening to unions than the injunction was the revival (first in an 1867 New Jersey case State vs. Donaldson) of conspiracy indictments against labor unions. See the discussion in Montgomery, David, Beyond Equality (New York, 1967), pp. 146–47.Google Scholar

⁴⁶ Quoted in Kaufman, Stuart, Samuel Gompers and the Origins of the American Federation of Labor (Westport, 1973), p. 201.Google Scholar See also Mink, Gwendolyn, Old Labor and New Immigrants in American Political Development (Ithaca, 1986), p. 186.Google Scholar

⁴⁷ Leaders of the I.W.W., for example, traced their organization's lineage back to the KOL; see Haywood, William and Bohn, Frank, Industrial Socialism (Chicago, 1911), p. 41.Google Scholar Some CGT leaders were involved in the KOL's French branch; see Dommanget, Maurice, Les chevaliers du travail en France (Lausanne, 1967);Google ScholarJulliard, Jacques, Fernand Pelloutier (Paris, 1971).Google Scholar

⁴⁸ American craft unions promoted strike success despite their involvement in growing numbers of strikes over work rule and job control issues. Because changes in technology and in the organization of work in the late nineteenth century challenged the position of workers in many skilled crafts, a growing share of American strikes were conducted over work rule and personnel issues, especially union strikes. While fewer than 25 percent of union strikes between 1881 and 1886 were over work rule issues, such strikes account for nearly half of all union strikes after 1886. As in France, American employers resisted demands for changes in work rules especially fiercely, and such strikes were significantly less likely to succeed than were those over wage and hour issues. See Friedman, , “Politics,” pp. 225, 229, 302;Google ScholarMontgomery, , “Workers' Control,” Chandler, Alfred, The Visible Hand (Cambridge, Mass., 1977).Google Scholar

⁴⁹ On strike rates and American unions, see Friedman, “Union Strikes and On Striking Unions,” The ability of conservative craft unions ability to restrain rank-and-file militancy sometimes made them useful allies of businessmen seeking to restrain labor militancy; this point is emphasized in Shafter, , “Trade Unions,” pp. 261–62.Google Scholar See also Weinstein, James, The Corporate Ideal in the Liberal State (Boston, 1968);Google Scholar Vincent Saint-John and W. Trautman, “Les ‘I.W.W.’”; Foster, William, “Le Congrès de ‘l'American Federation of Labor’,” Vie Ouvrier, 54 (12 20, 1911), pp. 779–82;Google Scholar and Mink, , Old Labor, p. 166. The smaller size of craft unions may have made it easier to settle disputes without strikes by lowering the cost to employers of making concessions; see fn 34.Google Scholar

⁵⁰ This is because they are most valuable as allies where they are hard to replace during a strike. The value of unskilled allies also depends on the elasticity of substitution of unskilled for skilled labor. I have made no attempt to measure this for either country, but some labor historians have argued that changes in management techniques raised the elasticity of substitution of unskilled for skilled labor in the United States. These new techniques may have raised the value of the unskilled as allies to American skilled workers. See, for example, Braverman, Harry, Labor and Monopoly Capital (New York, 1974);Google ScholarEdwards, RichardContested Terrain (New York, 1979).Google Scholar

⁵¹ Wages in four skilled occupations (printers, carpenters and cabinetmakers, metal smiths, and stone cutters) were compared with the wage of day laborers for the principal town or city in each French department in 1893 and 1906 using data in France, Salaires et durée and France, Direction du travail, Salaires et durée … en 1906 (Paris, 1907). Skilled French workers earned 48 percent more than did the unskilled in 1893 and 41 percent more in 1906.Google Scholar Wages of American compositors, cabinetmakers, carpenters, machinists, molders, and bricklayers in 1890 and 1903 were compared with those of building laborers, street laborers, foundry laborers, and hod earners using data for 97 cities in the United States, Commissioner of Labor, Annual Report, 1904 (Washington, D.C., 1905). Skilled Americans earned 80 percent more than did the unskilled in 1890 and 85 percent more in 1903.Google Scholar

⁵² Among more recent works, see Mink, , Old Labor, p. 50;Google ScholarCottereau, Alain, “The Distinctiveness of Working-Class Cultures in France,” in Katznelson, Ira and R. Zolberg, Aristide, eds., Working Class Formation (Princeton, 1986), pp. 140–42.Google ScholarL. Engerman, Stanley makes a related argument regarding the role of abundant land in social development in the United States and elsewhere in “Slavery and Emancipation in Comparative Perspective: A Look at Some Recent Debates,” this JOURNAL, 46 (06 1986), pp. 317–39.Google Scholar

⁵³ American wage differentials may have widened in the late nineteenth century. Although wage differentials were 80 percent in 1890 and 85 percent in 1903, the average daily wage of skilled workers in 1880 is listed as only 61 percent above that of unskilled workers in a sample of 882 cases from the United States, U.S. Census Office, 10th Census. vol. 20: Report on the Statistics of Wages (Washington, D.C., 1886).Google Scholar American differentials were also widest in the craft unions' urban strongholds in 1890 and 1903, but this is not the case prior to the rise of craft unionism in 1880. On trends in American wage differentials, see Shergold, Peter, Working Class Life: The American Standard in Comparative Perspective, 1899–1913 (Pittsburgh, 1982);Google Scholar and Lindert, Peter and Williamson, Jeffrey, American Inequality: A Macroeconomic History (New York, 1980).Google Scholar

⁵⁴ The average French differential is 1.40 or less than 80 percent of the average American differential of 1.80. Increasing the French differential by 0.40 lowers the effect of size on success by 0.0500, or around 11 percent of the strike size coefficient of 0.4682. Lowering American differentials to the French level raises the effect of size on American success only to 0.0553 in 1881–1894 and 0.1444 in 1903.

⁵⁵ Other changes in American management in the late nineteenth century, such as the growth of the multi-plant firm and the development of new management techniques, may have also weakened unions and make strikes more difficult to win in the United States than in France. These changes cannot, however, explain the relative success of craft unionism in the United States. By lowering the value of specialized skills and control over local production, these developments probably affected skilled workers more than the unskilled, lowering the relative effectiveness of restrictive craft unions in the United States—a point argued by critics of craft unionism in both the United States and France. The collapse of unions in the American steel, textile, paper, and chemicals, and much of the metal-fabrication industries reflects the growing power of the emerging giant firms in these sectors, and the inability of American craft unionism to respond effectively. Note that French unions flourished in some of these same industries after 1900, often with government assistance. See Friedman, , “Politics,” pp. 378–86;Google ScholarBrody, David, Steelworkers in America: The Nonunion Era (Cambridge, Mass., 1960);Google ScholarMontgomery, David, The Fall of the House of Labor (Cambridge, 1987).CrossRefGoogle Scholar

⁵⁶ See the discussion of the sources of Socialist electoral support in Friedman, Gerald, “French Republicanism in Theory and Decline” (unpublished manuscript, University of Massachusetts, 03 1987), pp. 29–31.Google Scholar

⁵⁷ The preference of many nonsocialist republicans for an alliance with the left over one with the right reflected the sharp division of the French electorate dating back to the French Revolution. See, Auspitz, Katherine, The Radical Bourgeoisie (Cambridge, 1982);CrossRefGoogle ScholarGoguel-Nyegaard, Francois. La Politique des partis sous la IIIe République (Paris, 1946);Google Scholar and Siegfried, André, Tableau des parties en France (Paris, 1913).Google Scholar

⁵⁸ Of 182 second-round contests with a Socialist candidate (SFIO) in the first-round, the Socialist withdrew in 106 cases, leading to the election of 72 centrists and Redicals. Centrists withdrew in 55 cases and Radicals in 46 cases, leading to the election of 17 and 19 Socialists. On this election, and on voting in the Chamber elected in it, see Friedman, , “French Republicanism,” pp.26–29.Google Scholar Also see Loubère, Leo, “The French Left-Wing Radicals and Their Views on Trade Unionism,” in International Review of Social History, 7 (07 1962), pp. 202–30.CrossRefGoogle Scholar

⁵⁹ Shorter, and Tilly, , Strikes, p. 39.Google Scholar

⁶⁰ Ibid., p. 43.

⁶¹ The Paris Commune of 1871 ended with the death of 25,000 communards and the arrest of 40,000 more; Adolphe Thiers cabled his prefects, “The ground is paved with their corpses; this terrible spectacle will be a lesson to them”; see Goldstein, Robert, Political Repression in Nineteenth-Century Europe (London, 1983), p. 125.Google Scholar

⁶² Auspitz, Radical Bourgeoisie, provides a superb analysis of these republicans. See F. Stone, Judith, “The Radicals' Ambivalence toward the State, 1871–1914” (unpublished manuscript, Western Michigan University, 04 1987);Google ScholarSewell, William H. Jr, Work and Revolution in France (Cambridge, 1980).CrossRefGoogle Scholar Among the more avilable contemporary sources are Durkheim, Emile, De la division du travail social (Paris, 1893);Google Scholar and Bourgeois, Léon, Solidarité (Paris, 1902).Google Scholar

⁶³ Waldeck-Rousseau's memo is reprinted in Pelloutier, Fernand, Histoire des bourses du travail (Paris, 1971), pp. 270–71. Also note his memos to prefects concerning strikes on 27 Feb. 1884 in Archives Nationales de France, file F7 box 12773. Waldeck-Rousseau emphasized that “strikes are the exercise of a right and [officials] misunderstand their obligations if they act to restrict them,” He urged officials to note situations likely to develop into strikes and to discuss with employers measures necessary to avoid them.Google Scholar

⁶⁴ See Julliard, Jacques, Clemenceau: briseur des grèves (Paris, 1965).Google Scholar

⁶⁵ Lorwin, Val, The French Labor Movement (Cambridge, 1966), p. 36.Google Scholar The mildness of French legal repression was also recognized by some French union leaders; see Keufer, A., “Ce qui est licite et illicite en cas de grève,” in Gide, , ed., Le Droit, pp. 86–87;Google ScholarFinance, Isidore, Les Syndicats ouvriers aux Etats-Unis (Paris, 1894).Google Scholar

⁶⁶ Clemenceau joined others in the center and right in attacking the CGT in parliamentary debates, but he opposed banning it. He hoped that moderates within the CGT—whose strength he exaggerated—would be able to take control of it. See the Chamber of Deputies debate May 7 to May 14 1907, France, Chamber of Deputies,Journal Official (1907), especially Clemenceau's speech on 14 05 1907 in ibid., pp. 1002–7. For a different view of the relationship between the state and French labor, see Schottler, PeterNaissance des bourse du travail (Paris, 1982).Google Scholar

⁶⁷ Around 1900, nonfarm employees made up 57 percent of the American labor force but only 36 percent of French labor force; see Friedman, , “Politics,” p. 101. Despite the size of the American proletariat, American Socialists before World War I were never able to elect more than one representative to Congress. The SFIO elected over a hundred deputies to the Chamber of Deputies elected in 1914.Google Scholar

⁶⁸ See Kleppner, Paul, The Third Electoral System, 1853–1892 (Chapel Hill, 1979);Google ScholarLichtman, Allan, “Political Realignment and ‘Ethnocultural’ Voting in Late Nineteenth Century America,” Journal of Social History, 16 (Spring 1983), pp.55–83.CrossRefGoogle Scholar

⁶⁹ Both of America's pre-World War I Socialist congressmen, for example, were elected in close three-way races but then defeated for reelection by fusion candidates. Rather than an alliance with the Socialists, Republicans and Democrats in New York and Milwaukee preferred to support their traditional rival's candidate. See Fink, Leon, Workingmen's Democracy: The Knights of Labor and American Politics (Urbana, 1983);Google ScholarStave, Bruce, ed., Socialism and the Cities (Port Washington,1975);Google ScholarHunter, Robert, Labor in Politics (Chicago, 1915).Google Scholar

⁷⁰ Norman Thomas, who had considerable direct experience in socialist politics, makes a similar point in “Pluralism and Political Parties,” in Laslett, and Lipset, , Failure of a Dream, pp. 654–61.Google Scholar

⁷¹ Decision in Farmer's Loan and Trust Co. v. Northern Pacific R.R. Co., 60 Fed. 803 (E.D. Wisc.), quoted in Paul, Arnold, Conservative Crisis, p. 122.Google Scholar

⁷² Over 47 percent of union members in manufacturing industries in 1899 were foreign born, as were over 54 persent of strikers in 1903; see Friedman, , “Politics,”, pp. 385, 305.Google Scholar

⁷³ D. Burnham, Walter, “The United States: The Politics of Heterogeneity,” in Rose, R., ed., Electoral Handbook (New York, 1974), p. 718.Google Scholar Also see Lipset, “Radicalism”; Hartz, Louis, The Liberal Tradition in America (New York, 1955).Google Scholar

⁷⁴ After the Civil War, American capitalists also did not need a strong state to break down residual barriers to capitalist development; Vogel, , “Why Businessmen,” p.57;Google ScholarTilly, Charles, The Contentious French (Cambridge, 1986).CrossRefGoogle Scholar

⁷⁵ An unnamed businessman quoted in Vogel, , “Why Businessmen,” p. 62.Google Scholar

⁷⁶ These include stricter voter registration laws and the extension of judicial review over social welfare legislation. See the discussion in Burnham, Walter D., “The Changing Shape of the American Political Universe,” American Political Science Review, 59 (03 1965), pp. 7–28;CrossRefGoogle ScholarKleppner, Paul, Who Voted? (New York, 1982).Google Scholar

⁷⁷ For an analysis of the conflict between these two, see Bowles, Samuel and Gintis, Herbert, Democracy and Capitalism (New York, 1986).Google Scholar