¹ Pitkin, Hanna, The Concept of Representation (Berkeley: University of California Press, 1967)Google Scholar.

² The field research for this paper was conducted between June, 1965 and August, 1966. AH statements refer only to community action program developments prior to August, 1966. Semi-structured interveiws with over 175 political actors in the three cities and in Washington, D.C were held during the course of the research. Respondents included representatives of the poor, public officials, heads of public and private welfare agencies, civil rights leaders and newspaper reporters. In addition, the research involved detailed inspection of newspaper coverage, review of minutes of community action agency committee meetings, attendance at neighborhood and citywide poverty meetings, and examination of public and private documents. Statements in the text are thus based on the close familiarity with the program in the three communities that is achieved through discussion with a wide range of political actors. In all cases, reports of the behavior of actors is based either on written evidence or on oral evidence from several actors speaking from various perspectives. A more detailed description and analysis of the community action program in these three cities can be found in Peterson, Paul E., City Politics and Community Action: The Implementation of the Community Action Program in Three American Cities” (Unpublished Ph.D. Dissertation, University of Chicago, 1967)Google Scholar.

I wish to express my appreciation to the Russell Sage and Woodrow Wilson Foundations for assistance which made this research possible. I am indebted to J. David Greenstone, Theodore Lowi, Duncan MacRae, and Michael Lipsky for their helpful comments on earlier drafts of this paper.

³ Pitkin, op. cit., p. 11.

⁴ Ibid., p. 43.

⁵ Ibid., Ch. IV. Pitkin includes accurate reflection of political opinions as well as social characteristics within her definition of descriptive representaton. Since we will confine our use of descriptive representation to considerations of social characteristics, we will call this socially descriptive representation.

⁶ Ibid., p. 111.

⁷ Actual representation may be distinguished from symbolic representation. As Pitkin argues, “We distinguish practical activity rationally directed toward bringing about ‘real’ goals … from expressive, symbolic actions.” Ibid., p. 102. This latter form of representation, she says, “need have little or nothing to do with … enacting laws desired by the people.” p. 106. Insofar as all formal representatives have the capacity to stimulate some favorable response among their constituents, symbolic representation is, to some extent, always present. Thus, where actual representation is almost nil, as it was in Chicago's community action program, it could be characterized by “only symbolic representation.” In order to avoid introducing still another form of representation into the analysis, we have instead classified Chicago simply as “very low” on actual representation.

⁸ Ibid., pp. 209–210.

⁹ Ibid., Ch. IX–X. Also, see Flatham, Richard, The Public Interest (New York: John Wiley & Sons, 1966)Google Scholar, passim.

¹⁰ In attempting to identify the public interest with the interests of institutions, Samuel Huntington presupposes an argument along the lines I have set forth. Huntington argues that “Institutional interests differ from the interests of individuals who are in the institutions …. Individual interests are necessarily short-run interests. Institutional interests, however, exist through time; the proponenent of the institution has to look to its welfare through an indefinite future.” ”Political Development and Decay,” World Politics, XVII (04, 1965), 412–413Google Scholar. But institutions are only complexes of role relations which individuals wish to sustain. In order to attribute interests to institutions, one must initially distinguish between personal and role interests. Speaking of the interests of institutions independent of the preferences of those with the authority to speak for the institution is thus perfectly parallel to speaking of role interests independent of the opinions of those occupying the social role.

Another solution to the problem of defining interests is to use the concept for heuristic purposes only. It can be argued, as does Ralf Dahrendorf, that it is analytically useful to postulate objective role interests in order to explain social conflict and social change. Dahrendorf, Ralf, Class and Class Conflict in Industrial Society (Stanford, California: Stanford University Press, 1959), pp. 173–179Google Scholar. This solution rests on the explanatory power of making such an assumption about role interests. Although J. David Greenstone and I have adapted Dahrendorf's assumptions to the analysis of race conflict in our study of Politics and Participation, to be published by the Russell Sage Foundation, in this paper I am arguing that one can identify objective role interests on quite separate grounds. As Pitkin has pointed out, ordinary language usage justifies a definition of interest that is not simply dependent on the subjective feelings and desires of the individual.

¹¹ See, for example, descriptions of the machine in Hofstadter, Richard, The Age of Reform (New York: Random House, 1955)Google Scholar and Meyerson, Martin and Banfield, Edward C., Politics, Planning and The Public Interest (New York: Free Press of Glencoe, 1955)Google Scholar.

¹² The analysis of the political determinants of the arrangements for formal representation, which we are only able to sketch here, can be found in Greenstone, J. David and Peterson, Paul E., “Reformers, Machines and the War on Poverty,” in Wilson, James (ed.), City Politics and Public Policy (New York: John Wiley & Sons, 1968), pp. 267–292Google Scholar.

¹³ Donovan, John C., The Politics of Poverty (New York: Pegasus, 1967), pp. 41–43Google Scholar.

¹⁴ U.S. Congress, An Act to Mobilize the Human and Financial Resources of the Nation to Combat Poverty in the United States, Public Law 88-452, 88th Cong., 2nd Sess., 1964, p. 9Google Scholar.

¹⁵ Office of Economic Opportunity, Community Action Program Guide (Washington, D.C., 1965), I, p. 18Google Scholar.

¹⁶ U.S. Congress, House Committee on Education and Labor, Hearings, Examination of the War on Poverty Program, 89th Cong., 1st Sess., 1965, p. 483Google Scholar.

¹⁷ Emphasis added. Philadelphia Evening Bulletin, Februry 8, 1965.

¹⁸ The proper name for these councils varied from city to city. For purposes of simplicity and in order to emphasize inter-city comparisons, we shall refer to all of them as “neighborhood councils.”

¹⁹ Again, I am avoiding the use of proper titles.

²⁰ Procedures in the other four areas were even less formally representative; community leaders came together, formed an organization, and were recognized by the City as the policy-makers for the community action program in the neighborhood.

²¹ Toward the end of my research, in the spring of 1966, it seemed that at least one representative from each neighborhood to the city poverty council would soon be chosen by the council instead of the director.

We are substituting a more general descriptive name for the proper title of Chicago's service centers in order to emphasize inter-city comparability.

²² The declining political strength of the New York City mayor, who must contend with a virile reform movement, is described in a number of penetrating analyses of New York politics. See Sayre, Wallace and Kaufman, Herbert, Governing New York City (New York: W. W. Norton and Co., 1965)Google Scholar; Levi, Theodore J., At the Pleasure of the Mayor (New York: Free Press of Glencoe, 1964)Google Scholar; Wilson, James Q., The Amateur Democrat (Chicago: University of Chicago Press, 1966)Google Scholar; and Costikyan, Edward N., Behind Closed Doors (New York: Harcourt, Brace and World, Inc., 1966)Google Scholar. The relationship between the political strength of the mayor and the development of the community action program in all three cities is analyzed in Greenstone and Peterson, “Reformers, Machines …,” op. cit.

²³ The power of the mayor in Chicago politics is well described in Banfield, Edward, Political Influence (New York: Free Press of Glencoe, 1961)Google Scholar.

²⁴ Although the style of politics has changed somewhat since this book was written, the best published introduction to Philadelphia's politics is Reichley, James, The Art of Government (New York: Fund for the Republic, 1959)Google Scholar.

²⁵ This analysis is based on research in three of Chicago's seven neighborhood service centers, including a white and Puerto Rican community on the North Side, a better organized, more stable Negro community on the South Side, and a disorganized slum providing the first home for southern Negro migrants on the West Side. I wish to thank Isaac Balbus, Marguerite Barnett and Rennie Davis for granting me their permission to use material they have gathered on Chicago's community action program. More supporting evidence for the argument presented here can be found in Peterson, op. cit., Ch. III.

²⁶ In addition to observation of citywide developments in Philadelphia, research was conducted in three randomly-selected neighborhood council areas. This research is reported in further detail in ibid., Ch. IV.

²⁷ Shostak, Arthur B., “Containment, Co-option, or Co-determination?” The American Child, XLVII (11, 1965), 17Google Scholar.

²⁸ Since the representatives of the poor continued to have some influence over employment of “nonprofessionals,” it is an exaggeration to say, as did one close observer of the program, that the “poor were rendered as powerless as ever.” Shostak, Arthur B., “Urban Politics and Poverty,” Paper presented at the annual meeting of the American Sociological Association, 1966, p. 2Google Scholar (Mimeographed.) Yet the fact is that this same observer had earlier reported such a glowing picture of poor people participating in policy-making that the statement clearly discloses the declining representation of the poor in Philadephia. See the article in footnote 27 for the earlier, optimistic analysis.

²⁹ Because of the need to examine operating programs in a city where the community action program was slow in being implemented, we selected for investigation the first three neighborhoods in which a community action program was operating. Thus, although data on citywide developments were collected, this report draws largely upon developments on the Lower West Side of Manhattan, in the Bedford-Stuyvesant area of Brooklyn, and in East Harlem. New York's community action program is discussed in greater detail in Peterson, op. cit., Ch. V.

³⁰ A detailed analysis of the attempts by representatives of the poor to change the structure and policies of East Harlem's educational system can be found in Peterson, Paul E., “The Politics of Educational Reform,” in Lutz, Frank (ed.), Dynamics in Urban Education (Columbus, Ohio: Charles Jones, 1970)Google Scholar.

³¹ Young Men's Christian Associations.

³² Supra, p. 121.

³³ Key, V. O., Southern Politics (New York: Alfred Knopf, 1950), p. 37Google Scholar.

³⁴ While records were kept on the number of voters for each CAC area as a whole, the number of voters for each voting section within the areas could not be obtained directly. Although the number of votes cast for each candidate within each voting section was available, the sum of these votes could not be accepted as a direct measure of the number of voters because each person was permitted to vote for as many as twelve candidates. Neither was the solution to divide the total number of votes cast in an election by twelve, since many people did not fully utilize all twelve votes. The best, though far from satisfactory solution, we felt, was to assume that at each of the four polling places within a CAC area the voters used equal amounts of their total voting power. Such an assumption enabled us to formulate the following proportion:

where S_v equals the total number of voters in the voting section, Sb equals the total number of ballots cast for all candidates in the voting section, A_v equals the number of voters in the CAC area, and A_b equals the total number of ballots cast for all candidates in the CAC area. Since three of the four quantities were known, we were able to solve for the fourth, giving us an estimate of the number of voters in each voting section.

Where more candidates ran for office in a section, it is possible that voters used more amounts of their voting power in order to vote for several of their friends and neighbors on the ballot. If so, this would mean that our assumption that voters used equal amounts of their voting power in estimating the number of voters biases the findings in the direction of estimating more voters in sections that had more candidates. Some of the high correlation between number of candidates and voter turnout, therefore, may be attributable to the exercise of greater voting power by voters in those sections that had more candidates. Even if this were the case, however, it would in no way be contrary to our basic contention that voting behavior was affected by friends and neighbors politics.

³⁵ The size and characteristics of the total population in the sample areas were estimated from data made available by Richard H. Uhlig of the Philadelphia Health and Welfare Council. Since areas (and sections within areas) did not correspond with census tracts in a number of cases, all calculations based on these data are subject to some error. In addition, our estimate of the population was based on all people eighteen years and older reported in the 1960 census data. Since voting was limited to those twenty-one years and older, our figures underestimate the percentage of those eligible that voted. However, unless the age distribution for the three years eighteen to twenty-one varies substantially from one neighborhood to another, it ia unlikely that this affects inter-area comparisons.

³⁶ Two candidates were excluded from the analysis, because it was impossible to ascertain in which voting section they lived.

³⁷ See, for example, McConnell, Grant, Private Power and American Democracy (New York: Alfred Knopf, 1966)Google Scholar, passim.

³⁸ Both the level of income which was the upper limit of the poverty zone and the percentage of poor on neighborhood and citywide councils fluctuated over time. In general Chicago's social representation increased somewhat from a very low level, whereas New York's decreased in 1966. More detailed information can be found in Peterson, , “City Politics and Community Action,” pp. 56–59, 133–153Google Scholar.

³⁹ There is some evidence that enforcement of this rule was not excessively rigid, but it was nevertheless apparent that the overwhelming majority of council members had incomes below the upper limit and the remainder were in only slightly better economic circumstances.

⁴⁰ In three randomly selected areas, the distribution was 72 percent Negro, 25 percent white and 3 percent Spanish-speaking. A close observer of the program estimated that in the city as a whole about 80 percent of elected representatives were nonwhite. Shostak, , “Urban Politics and Poverty,” p. 1Google Scholar.

⁴¹ Community action program officials in Oakland also eliminated middle class neighborhood leadership seeking to provide universalistic interest representation by requiring that representatives be socially descriptive of their constituency (i.e. having annual incomes of less than $3000). Masters, Nicholaset al., Politics, Poverty and Education: An Analysis of Decision-making Structures, Report submitted to the Office of Economic Opportunity, Washington, 02, 1968, p. 253Google Scholar.

^{page 507 note 1} Duncan MacRae provided considerable assistance in the development of the index of localism.

^{page 507 note 1} The index of over-representation has been used widely by political scientists and sociologists. See, for example, Matthews, Donald R., U.S. Senators and Their World (Chapel Hill, North Carolina: University of North Carolina Press, 1960), pp. 273–74.Google Scholar