The Third Electoral System, 1853-1892: Parties, Voters, and Political Cultures

By Paul Kleppner

This analysis of the contours and social bases of mass voting behavior in the United States over the course of the third electoral era, from 1853 to 1892, provides a deep and rich understanding of the ways in which ethnoreligious values shaped party combat in the late nineteenth century. It was this uniquely American mode of "political confessionals" that underlay the distinctive characteristics of the era's electoral universe.

In its exploration of the the political roles of native and immigrant ethnic and religious groups, this study bridges the gap between political and social history. The detailed analysis of ethnoreligious experiences, values, and beliefs is integrated into an explanation of the relationship between group political subcultures and partisan preferences which wil be of interest to political sociologists, political scientists, and also political and social historians.

Unlike other works of this genre, this book is not confined to a single description of the voting patterns of a single state, or of a series of states in one geographic region, but cuts across states and regions, while remaining sensitive to the enormously significant ways in which political and historical context conditioned mass political behavior. The author accomplishes this remarkable fusion by weaving the small patterns evident in detailed case studies into a larger overview of the electoral system. The result is a unified conceptual framework that can be used to understand both American political behavior duing an important era and the general preconditions of social-group political consciousness. Challenging in major ways the liberal-rational assumptions that have dominated political history, the book provides the foundation for a synthesis of party tactics, organizational practices, public rhetoric, and elite and mass behaviors.

• Language: English
• Publisher: The University of North Carolina Press
• Release date: Oct 10, 2017
• ISBN: 9781469639536

Paul Kleppner

Paul Kleppner is the author of The Cross of Culture: A Social Analysis of Midwestern Politics, 1850-1900.

Book preview

1 A Prologue to the Analysis of Past Politics

Political history is often sterile stuff, a collection of names, titles, and events beaded on a chronological string. Yet a time perspective can yield an awareness of dimensions of the political system that otherwise escape detection. The wars of domestic politics, like those between nations, are not events of a moment but extend through the years.

V. O. Key, Jr.

The analysis of past politics is replete with potential pitfalls and unique analytical opportunities. Conceived as a series of recurring, but individually unique, developments beaded on a chronological string, political history produces analytically sterile stuff. More broadly conceived, political analysis enables us to move beyond idiosyncratic happenings and to penetrate the relation between society and political system.

Traditional research strategies have encumbered that sort of analytical penetration. They have focused instead on exciting events, colorful personalities, and the dramaturgy of political notables. Pertinent evidence has often been skillfully blended to produce graceful, and even elegant, narratives. Yet for all of their informational value and analytical insights, these subordinate conceptual unity to chronological sequencing. Their central preoccupation has been telling the story.

The ways that events unfolded and the roles that particular elites played are neither unimportant nor irrelevant to an understanding of past politics, but they do not constitute the beginning, the middle, and the end of political history. Yet traditional research strategies typically have focused on these aspects of past politics to the virtual exclusion of mass electoral behavior. Most political narratives, of course, refer to elections and to their results, but those references occupy a distinctly secondary role. Far less attention and fewer pages of monographs are devoted to them than to elite machinations.¹ And the interpretive frameworks use the categories and terms of understanding peculiar to the elites. The result is a political history only tangentially concerned with mass behavior, and then only as refracted through the mental eyes of political notables.²

Approaches of this genre constrain the analytical potentialities of political history and at best provide a socially jaundiced view of the past. To obviate these limitations, one must broaden the conceptual framework and apply it within the context of a different type of research strategy. We need to treat mass electoral behavior, not as a coincidental afterthought, but as the primordial building block of the electoral system. And that treatment must begin, not with elite descriptions of how the masses reacted, but with an examination of observable voting behavior. This study was designed to apply such an approach to an analysis of late nineteenth-century electoral behavior.

The essential nature of political action during what Mark Twain dubbed The Gilded Age has not been an object of heated dispute among most historical analysts. Despite some shifts in emphasis and ostensibly fresh interpretations, accounts of late nineteenth-century party battles generally continue to be conceived within older frames of reference and analytical categories. The persuasive rhetoric of Lord Bryce and James Ford Rhodes, on the one hand, and that of Charles Beard and Matthew Josephson, on the other, continues to exercise a tyranny over current conceptions.³ These analyses differ in their particulars and underlying values, but their legacy is a shared and still dominating motif: the politics of the Gilded Age were dull and sterile; they were evasive of the real issues of the era, the questions arising from the massive economic and social transformations that were the hallmarks of nineteenth-century America.

There are distinguishable elements of credibility in that view. American electoral politics have never been organized exclusively around social-class polarities. Questions of political economy have not always and everywhere in the electoral past mobilized mass constituencies and determined election outcomes. Nor, for that matter, has most electoral behavior been issue oriented. Yet even granting these elements of credibility does not enhance the explanative power of the larger view, for that view ultimately relies on the a priori assumption that political action lacking such characteristics must necessarily and inherently be evasive and irrelevant. That assumption, in turn, makes the analyst's norms the criteria against which the rationality of past behavior is to be assessed. Analyses conceived within this framework reveal little of the attitudes and values of those uncounted millions who were the involved participants. Yet they do provide considerable insight into the shared values and a priori assumptions that have shaped most historical descriptions of past politics. Although it is not appropriate here to digress into an extended conceptual critique of those formulations, it is useful to draw attention to some of their essential and common elements.

First, they reveal a shared but implicit commitment to the desirability and to the operation of a liberal-rational model of public-opinion formation. That view posits a political system composed of an atomized collection of individuals, each more or less informed about public issues and possessing views about them. This preconception produces a picture of social conflict organized along issue alignments as individuals independently form their opinions. Applied to mass electoral behavior, the model explains voting decisions as the result of a process in which men ponder the logical relations between their own political wants or goals and the alternative policy stances of the contending parties. The ideal political universe is thus presumed to be one disproportionately populated by active, informed, and issue-oriented citizens.⁴

Second, and because it is both a logical corollary and requisite condition of the liberal-rational model, most historical descriptions reflect consensual agreement on the proper role of political parties. Political parties should articulate unequivocal stances on issues and shape these into distinctive and logically constrained programmatic packages; they should function as instruments of policy articulation and integration. Only by performing in this way would political parties provide issue-conscious citizens with partisan alternatives that were consonant with unambiguous policy distinctions. And only then would elections constitute the outcomes in particular, the series of referenda or policy that the liberal-rational model assumes they should be.⁵

Whenever concrete party behaviors fail to conform to these desiderata, the analytical judgments are harsh. On the one hand, when historians have perceived that parties failed to present distinctive policy alternatives, they have often internalized and reiterated the complaints of the nineteenth-century political notables that party politics were devoid of live issues.⁶ That assessment, however, relies directly on the normative assumption that the exclusive nature of party is programmatic; and that assumption, in turn, is a deduction from the liberal-rational model of public-opinion formation and not one derived from observation and analysis of past behavior. On the other hand, that same model produces an even more astringent assault on appeals to party loyalty. Nothing more directly undermines the model, after all, than a standing decision. Frequently echoing and invariably citing the political notables, historians regularly denigrate party loyalty and wistfully allude to the voter who wished to follow his own conscience.⁷ Reflecting and perhaps sharing in the antiparty animus of the political elites, analysts regularly devote more energy to a normative critique of party loyalty and the mechanisms sustaining it than to an understanding of either its social origins or of the role of party as a reference symbol.

Finally, traditional descriptions of past politics offer evidence of a broad diffusion among historical analysts of the assumption that social-class distinctions are, or ought to be, the invariable substructural linchpins of political behavior.⁸ Other types of social-group differences are portrayed either as temporary distractions or as superstructural images of these underlying realities. Thus, if mass politics are seen as failing to pivot on what the analyst regards as the real (i.e., substructural) distinctions, it follows axiomatically that the prevailing bases of cleavage must be evasive or derivative. That conclusion, of course, is nothing more than a deduction from the starting premise. It is, moreover, a deduction with starkly elitist implications, for ultimately it hinges on the analyst's projecting his own values on past men and arguing implicitly that they should have perceived the public concerns to which those values are addressed as the most important ones.

I do not mean to suggest that most historians consciously intend to interpret the past in terms of their own values but only that their norms and implicit assumptions shape their research strategies and often, perhaps unconsciously, influence their receptivity to evidence that reinforces their preconceptions. Those research strategies aim at an objective reconstruction of specified aspects of past experience. The tactical means to that end entail careful and painstaking examinations of traditional sources, such as correspondence, diaries, newspapers, and public utterances. These types of sources, if construed literally, provide researchers ample grist for their issue-oriented mills.

Political notables such as George William Curtiss, E. L. Godkin, and Carl Schurz—as well as radical elites such as Eugene Debs, Ignatius Donnelly, and Robert Schilling—were active, well informed, and articulate. Theirs were political worlds of principles and policy, information and issues. They frequently and cogently articulated their concern with specific matters such as tariffs, civil service, and currency expansion or monopolies, wages, and labor organization. Moreover, they linked these separate ideas into a larger frame of reference. They had knowledge, not only of specific, but of interstitial information that enabled them to discern what went together and why it did, for in their political belief systems discrete ideas were glued together, or were psychologically constrained, by a few crowning postures. Those crowning postures provided judgmental yardsticks with which to measure and make sense of a wide range of information. To describe and analyze their political values is to glimpse the type of belief system characteristic of those at one level of the political structure.⁹

However, we cannot assume that the linkages that characterized belief systems at that level extended to all other levels of the political structure. To interpret mass behavior through a reconstruction of elite belief systems is to present a typology of past political concerns that mirrors only the thinking and issue-oriented focus of relatively small, self-conscious, and self-enclosed groups of abnormal citizens.¹⁰ To project elite cognitions to the mass public requires assuming that most nineteenth-century contemporaries possessed the same levels of awareness and involvement, concrete and interstitial information, as the political notables. It implicitly assumes that most voters perceived their political worlds in the same ways, and in the same terms, as the political elites. That assumption ignores the enormous differences in kind between mass and elite cognitions. Most people generally lack the levels of contextual information that are common among the elite. As a consequence, mass belief systems are more loosely constrained, they encompass a narrower range of relevant ideas, and the character of those elements central to the mass systems is typically less generic or abstract than in elite belief systems.

To point out these distinctions is not to claim that only elites have opinions or beliefs, or even that elite belief systems are qualitatively superior to those of nonelites. It is rather to assert that analysts must aim at reconstructing the belief systems of a wide variety of groupings within the social and political structures and that in so doing we should not automatically assume that whatever beliefs ‘go together’ in the visible political world (as judged from the attitudes of elites and the most articulate spectators) must naturally go together in the same way among [the] mass public.¹¹

If elite cognitions cannot be assumed to be wholly reflective of mass opinions, how then do we reach beyond politics at the top?¹² How do we determine what public questions mattered to most citizens, what touched their daily lives and energized their commitments? It is useful to tackle that conundrum by presenting an outline of the conceptions and research strategy that underlie this book, as well as an explicit statement of its assumptions.

Studies of mass politics must begin with observations of mass behavior. How a group voted—whether Democrat or Republican, how strongly, and under what conditions—can be resolved, not by deduction or inference, but only by thorough, systematic analysis of the relevant empirical data. For purposes of this study, that has required the collection, processing, and statistical analysis of large quantities of county and minor civil-division voting and demographic data over an extended temporal sequence. Through such an approach it is possible to describe the regular patterns of a social group's partisanship and to notice breaks in those patterns.

Of course, it is considerably easier to assert that studies of mass voting behavior should begin with descriptive statements of the central tendencies of group partisanship than it is to arrive at those statements. To provide credible descriptive statements requires avoiding deterministic models in which both the research design and the conclusions are simply deductions from a priori assumptions. It is not the substantive nature of the conclusions but the logic of the research design that indicates employing a deterministic approach. To avoid that pitfall, research designs must test the effects of a variety of potential determinants of voting behavior. This study has made use of a wide variety of economic, ethnic, and religious indicators in order to describe group voting behavior. These have been employed in a number of statistical ways, ranging from simple cross-tabulations to select units or groups of units for illustrative purposes, to multiple-regression models to assess the relative simultaneous contributions of distinctive types of variables. I have also tested for the impact of contextual effects, demographic as well as political, and for observable differences dependent on levels of aggregation. Although most of the negative findings are not presented in statistical detail, the descriptions provided derived from a multivariate data analysis.

Systematic observation and description of social-group behavior is the requisite first step. That constitutes the only credible basis for insight into the beliefs, attitudes, and values of a group. Only from behavior can these be inferred.¹³ And it is equally important to emphasize that they can only be inferred. Arrays of statistical data do not demonstrate, nor even address, the problems of human motivation. Why people responded politically as they did cannot be demonstrated statistically. Analytical and factual questions are of different logical orders and accordingly require different designs of proof. How a group voted is a factual question requiring an empirical response. Why it voted that way involves an analytical question that can be posed meaningfully only after the factual one has been answered empirically.

To move beyond statistical aggregates and penetrate human motivation, the researcher must search extensively through the society to identify group antagonisms and conflicts and determine which of these became sources of party combat. The past political world must be reconstructed in terms of its political subcultures. Culture, or any specified aspect thereof, is one of those vague conditioning concepts, or catchwords, whose rhetorical value becomes perishable in practical applications.¹⁴ To be of analytical use, the involved sequence must be specified in more detail and must be linked with an explicit conception of political party.

Involvement with a group shapes members’ beliefs, attitudes, and values; it imparts a perspective to assist them in organizing their perceptual fields.¹⁵ This group-anchored perspective is something of a screen, or filter, through which the members process and compare external stimuli. Technically, that process is called cognitive interaction, and through it a response, or behavior, is shaped. The collective behavioral manifestation of that shared, internalized, group-imparted perspective is designated by the term group subculture. In turn, political culture is the political aspect of that group subculture: a historical system of widespread, fundamental, behavioral, political values . . . classified into subsystems of identity, symbol, rule, and belief. A group's political culture consists of its politically relevant purposive desires, cognitions, and expressive symbols. It refers to the internalized expectations in terms of which the political roles of the group's members are defined and through which their regularized patterns of political behavior come into being.¹⁶

Political subcultures are not created by political parties. The causal flow is in the opposite direction. Party oppositions and alliances are formed from antagonistic relations between (or among) group political cultures. Yet not every social-group conflict translates into party distinctions. It is not enough to point to the existence of conflict and then assert its relevance to party struggle. Historical explanations cannot rely on reference to the wondrous workings of some unspecified process of political immanence. Some conflicts may have no relevance to the political system; others may generate demands on the system through channels other than party, e.g., through pressure groups and bureaucracies. And even among those conflicts that are expressed through party, some will be more intense than others because they involve ideas more central to the belief systems of the combatants.

When antagonistic political cultures do generate party oppositions, attachment to the group comes to involve a set of political orientations that includes a partisan proclivity. Initially, of course, the reasons for the party attachment are alive in the members’ minds. As memories fade, as membership turns over, the situation- or object-specific reasons become, in the neutral sense of the word, stereotypes, the subtlest and most pervasive of all influences, the pictures in the mind through which the members of the group impart order to their political worlds. In time, through habituated and group-anchored partisan response, the party as institution itself becomes a reference group.¹⁷ Thus, political party, as a social grouping, can be thought of as a very special case of the much more general group-influence phenomenon. It is a special case of the phenomenon, not simply because its political salience in any electoral situation is exceedingly high and unidirectional, but because individuals come to psychological membership in it through the intermediary of involvements with other social groups. The influence of a group leads members to identification with a party, and the latter serves as a kind of self-steering mechanism that guides and shapes responses to new political stimuli.¹⁸

Political parties, of course, are not only reference groups or symbols; they are also concrete and complex entities. That quality of dynamic complexity has eluded traditional historical analysis of party struggles. Those accounts generally avoid an explicit conception of party and rely instead on working definitions deduced from implicit theories of political system. Given some conception of political system, analysts can readily infer that certain functions ought to be performed, and they search for the concrete institutions that carry out those requisite functions. Thus political parties emerge largely by indirection and deduction. This approach perpetuates misconceptions that becloud the analysis of past politics. It encourages analysts to view parties as somehow detached from their sociopolitical contexts, as reforming missionaries from some far-off places set loose in the political system to impart order and coherence to it.¹⁹ Second, it underlies the view of parties as the exclusive structures through which contending and competing social demands are articulated and aggregated. Conceiving this intermediary role of party as an invariant given deflects analytical attention both from analogous roles played by other durable structures within the political system and from the ways in which party's role may have changed across time. Third, by not explicitly designating the meaning of the term party, analysts readily blur the distinctions among different types of party activity and thus fail to recognize that each involves particular cognitions and behaviors. Fourth, concentration on the functions of party leads researchers to a common descriptive metaphor: the party as electoral machine.²⁰ Concealed beneath that literary device, combat among parties in traditional accounts has been emptied a priori of its social, cultural, and ideological content. It is one thing to think of parties as, inter alia, mobilizing agencies but quite another to present them as sufficient causes of that mobilization. Exaggeration of the electoral-machine metaphor does the latter and thus precludes analysis of the extrapartisan sources of party struggle.

If parties and their activities are proper objects of analysis, then we must provide explicit conceptions of what those terms designate. A political party may be defined as a specialized institution within the boundaries of the political system that, through its dual capacity both to articulate and to aggregate interests, provides direct links between the decision makers and large numbers of interest groups outside the boundaries of the political system.²¹ The definition remains incomplete, however, until we indicate the specific ways in which political party may be distinguished from other institutions assigned to the same genus.

Party differs from other specialized political institutions that both articulate and aggregate interests (e.g., the bureaucracy) in that it alone selects candidates and contests elections, attempts to organize elected governmental decision makers, and attempts to win converts or recruits. Analogous to these three clusters of distinguishing activities, party alone among the specialized political institutions is characterized by a distinctive tripartite pattern of organization. As Frank Sorauf has suggested, we can think of party as being composed of three elements or sectors. First the organization proper, which has an internal life of its own—the party officials, the hyperpolitical activists, the purposeful, organized, initiating vanguard of the party. Second the party in office, whether executive or legislative, which includes the caucuses, floor leaders, whips, and patronage networks—those who have captured the symbols of the party and speak for it in public authority. Third the party in the electorate—the least stable, least active, least involved, and least well organized sector; those who attach themselves to the party either through habituated support at the polls or through self-identification; those who are ‘of’ the party because it is for them a symbol that provides cues and order to the political cosmos.²² Party, of course, is no one of these elements but all in interaction: Party exists when activists, officials, and voters interact and when they consciously identify with a common name and symbols.²³

This tripartite structure is party's telling characteristic, for unlike other specialized political institutions the object of party's mobilizing efforts—its clientele—is not external to it. Rather, party includes its own clientele, as well as a tangible political organization and sets of officeholders. Thus, party's membership encompasses the widest possible range of involvement and commitment. It is at once a well-defined, voluntary political organization and an open, public rally of loyalists.²⁴ By its activists and officeholders party may well be seen as a programmatic instrument, but by its mass supporters it is only necessary that it be perceived as an object of habitual loyalty, a reference symbol. However, since party's activity within all three categories of membership is goal oriented and since the goals depend ultimately on electoral victories, the party must act to attain the goals of its activists, its officeholders, and its clientele.²⁵ Indeed, in arguing from cleavages over policy among political elites to partisan distinctions among social groups within the mass electorate, historians have implicitly assumed an absolute identity of goals among all three sectors of party. There is neither reason nor need to make as sweeping an assumption.

Nor need we assume that the relations of dominance and subordinate among the sectors are invariant across time or space. There is no standard mix, no predetermined proportions and intensities among the elements, that we need accept either as a sine qua non or as a criterion. We should expect different mixes, different patterns of relations across both time and space. Description and analysis of these patterned relational distinctions, at all levels of governmental action and within varied sociopolitical contexts, and of their permutations across time are prerequisite for understanding the development of the political system.

Yet just as the presence of party requires interaction among its three sectors, so that implies the existence of some relation among them. If that relation need not be one of programmatic congruence, it must at least be one of psychological rapport. That is, I assume, as a minimal condition, that the three elements are glued together through a shared common latent value continuum.²⁶ Analysis that tries to reconstruct the essence of those shared values is crucial to probing beyond politics at the top. Understanding the nature of the psychological bridges that linked a party's officeholders and activists with its loyal followers enables us to penetrate the political worlds of those who composed the mass electorate, for in the nature of that rapport between a party's character and the dispositions of its constituent groups are clues to the concerns and emotions that permeated the latter, energized their commitments, and transformed social-group antagonisms into party oppositions.²⁷

Analysts of past politics cannot recapture that shared bond of psychological rapport in the same ways as political scientists and sociologists do with modern parties and voters. We do not enjoy the luxury of being able to administer a structured questionnaire to a past electorate to inquire into its perceptions and motivations. Instead, we must use the materials that are available, knowing and acknowledging that they are at best crude and imprecise surrogates for modern instruments that measure public opinion. Historical sources—elite sources, such as newspapers, speeches, and party platforms—allow us to probe mass opinion only in-ferentially, and then only if we move beyond their manifest content and search for the evocative code words and symbols, the emotions and values with which rhetoricians knew their listeners or readers to be already imbued.²⁸ When congruent with systematic description of observable behaviors, that analysis enables historians to posit sets of structured inferences concerning which public matters resonated emotionally with group political characters and, thus, about why groups voted as they did.

This is the conceptual framework that underlies the research strategy of this book. That strategy involves two assumptions that have only been implied thus far. First, I assume that a number of social variables as experienced and internalized by individuals are equally capable of shaping their voting decisions and partisan commitments. Second, I assume that for most citizens the experiencing and internalizing are group processes, that the cleavages that involve mass publics are [shaped] in terms ... of loyalties to competing groups.²⁹ The concrete result of applying such a strategy to an analysis of past politics differs from the products of traditional political history. It depicts party combat in terms of antagonistic political cultures; it views elections as integral parts of a larger and ongoing social process.

1. H. Wayne Morgan, From Hayes to McKinley, devotes pp. 57–117 to discussing the preelection-day events in 1880; he confines his analysis of the outcome and references to the group bases of politics to pp. 118–21. That relative allocation of space and analytical concern is a common feature of historical accounts of past politics.

2. The term political notables is a designedly ambiguous one; as a matter of stylistic convention, it will be employed here synonymously with the term political elites. Neither term, however, is meant here to imply that those so designated exercised decision-making power, that they influenced those who did, or that they necessarily were at the same time economic and/or social elites. In other words, I am here simply using these terms to refer to that group of prominent contemporary figures who have been regularly quoted and cited by historians as sources of information concerning late nineteenth-century politics.

3. Thomas C. Cochran, The ‘Presidential Synthesis’ in American History, p. 374, observes that history probably suffers more than any other discipline from the tyranny of persuasive rhetoric. The rhetorically persuasive works that have shaped interpretations of late nineteenth-century politics are James Bryce, The American Commonwealth, vol. 2; James Ford Rhodes, History of the United States from Hayes to McKinley, 1876–1896; Charles A. Beard and Mary R. Beard, The Rise of American Civilization, especially 2:285–343; Charles A. Beard, The American Party Battle, pp. 76–93; Matthew Josephson, The Robber Barons; idem, The Politicos, 1856–1896.

4. The quotation is from V. O. Key, Jr., Public Opinion and American Democracy, p. 60. Key here characterizes what he calls classical treatments of public opinion; it is my judgment that this description can be used to categorize the public-opinion model that historians have employed. For specification and discussion of the conditions for issue-oriented voting, see Angus Campbell et al., The American Voter, pp. 169–87.

5. Theodore J. Lowi, Party, Policy, and Constitution in America, pp. 238–76, uses the phrase (p. 263) and perceptively analyzes American electoral politics as representing outcomes in general that have imparted shape to large clusters of public policy.

6. See, for example, John A. Garraty, The New Commonwealth, 1877–1890, p. 240; and the claim by John M. Dobson, Politics in the Gilded Age, p.43, that the lack of issues produced a "political scene that closely resembled that seldom-achieved hypothetical situation in which all other things were equal." Emphasis in the original.

7. The quotation is from Dobson, Politics in the Gilded Age, p. 43. For the attitudes of the political notables toward party spirit, see John G. Sproat, The Best Men, pp. 60–66. And for an example of a historian's internalizing the notables’ view of party, see Geoffrey Blodgett, The Gentle Reformers, p. 40.

8. Sometimes this results from confusing the concept of economic interest group with social class, as in Morgan, Hayes to McKinley, pp. 120, 318. At other times it reflects the simple, though deceptively phrased, economic determinism called for by James R. Green, Behavioralism and Class Analysis, p. 98, and given more detailed expression by David Montgomery, Beyond Equality. For a critique of the logic underlying these approaches, see Lee Benson, Turner and Beard, and especially pp. 151–60.

9. For the concept of belief system, see Philip E. Converse, The Nature of Belief Systems in Mass Publics, pp. 206–61. And for an understanding of the properties and organization of belief-disbelief systems, see Milton Rokeach, The Open and Closed Mind, pp. 31–70.

10. The characterization is from Cochran, ‘Presidential Synthesis,’ p. 376.

11. Converse, Nature of Belief Systems, p. 230. This should not be construed to imply that mass belief systems are illogical and, therefore, inferior to elite belief systems. Logical systems are only a subclass, a special type, of psychological system; see Rokeach, Open and Closed Mind, pp. 33–35. Converse's essay has sparked considerable research on the separate but related questions of issue constraint in mass belief systems and issue-oriented voting. For analyses that point to increased levels of both after the 1950s, see Gerald M. Pomper, From Confusion to Clarity, pp. 415–28; Norman Nie and Kristi Andersen, Mass Belief Systems Revisited, pp. 540–91. And for a superbly conceived analysis of the problem of issue voting, see Michael Margolis, From Confusion to Confusion, pp. 31–43. It is worth noticing that even at its increased level mass attitudinal consistency across issue domains remains quite weak in an absolute sense, and as Margolis demonstrates, it falls significantly short of the requisite criteria for issue-oriented voting.

12. The expression is from Morgan, Hayes to McKinley, p. vi.

13. Rokeach, Open and Closed Mind, p. 32; idem, Beliefs, Attitudes, and Values, p. 2.

14. On the role of catchwords in the social sciences, see J. P. Nettl, Political Mobilization, pp. 42–53.

15. My understanding of the psychological processes involved owes a great deal to three works by Milton Rokeach. In addition to the volumes cited in notes 9 and 13, above, see The Nature of Human Values.

16. The quotation is from Donald J. Devine, The Political Culture of the United States, pp. 14–18; the following two sentences are adaptations of the definition of political culture given by Harry Eckstein, A Perspective on Comparative Politics, Past and Present, p. 26. For the formulation that has directly shaped my thinking about the concept of political culture, see Ronald P. Formisano, Deferential-Participant Politics, pp. 473–87.

17. The quotation is from Walter Lippman, Public Opinion, pp. 59–60. On the concept of party as a reference group, see Robert E. Lane, Political Life, pp. 299–300. And see the perceptive comments on the party name as an image’ that shades imperceptibly into the voluntary realization of its meaning in Graham Wallas, Human Nature in Politics, p. 104.

18. See the discussion of Secondary Groups, the Political Party, and the Influence Process, in Campbell et al., The American Voter, pp. 327–31; also see Converse, Nature of Belief Systems, pp. 229–30, 234–35, 239–41.

19. Frank J. Sorauf, Political Parties and Political Analysis, pp. 48–55; the quotation is on p. 49. For perceptive discussion of the analytical need to view parties as intervening, dependent, and independent variables, see David E. Apter, The Politics of Modernization, pp. 181–82.

20. Bryce, American Commonwealth, 2:76–153; Henry Jones Ford, The Rise and Growth of American Politics, pp. 141–49, 294–310; Dobson, Politics in the Gilded Age, pp. 51–52. The view of party as electoral machine typically assumes hierarchical organization with the flow of authority running from the top downward. For an extensive refutation of the notion that late nineteenth-century parties were organized and controlled from the top down, see Robert D. Marcus, Grand Old Party.

21. For this dual capacity of party, see Gabriel A. Almond and G. Bingham Powell, Jr., Comparative Politics, pp. 100–104. For an explicit conception of political system and the necessity of delineating the boundaries of that system, see Lee Benson, Political Powerand Political Elites, pp. 282–87.

22. For the activities of party and the discussion of its tripartite structure from which the quotations have been taken, see Sorauf, Political Parties and Political Analysis, pp. 46, 36–38, respectively; for the description of party officeholders, see Frank J. Sorauf, Party Politics in America, p. 11.

23. Formisano, Deferential-Participant Politics, p. 475.

24. The discussion is based on Sorauf, Party Politics, pp. 10–12; the quotation is on p. 12.

25. Goal orientation, of course, does not automatically presuppose an exclusive focus on the allocation of tangible resources. Furthermore, it is important to emphasize that party loyalists may, under specified conditions, see party as more than a reference symbol; they may be as programmatically oriented as activists and officeholders. In saying that they must at least see party as a reference symbol, I mean only to draw attention to the minimal necessary condition.

26. The quoted phrase is from Murray Edelman, The Symbolic Uses of Politics, p. 155. Also see Paul F. Lazarsfeld and Neil W. Henry, Latent Structure Analysis, pp. 15–44; Robert E. Lane, Political Ideology, pp. 10–11; Robert E. Lane, Political Thinking and Consciousness.

27. For works that develop the concept of poltiical character, see Robert E. Lane, Political Character and Political Analysis, pp. 115–25; Ronald P. Formisano, Political Character, Antipartyism, and the Second Party System, pp. 685–88. Also see David Riesman, Nathan Glazer, and Reuel Denny, The Lonely Crowd.

28. Ronald R Formisano, The Birth of Mass Political Parties, p. 13.

29. Key, Public Opinion, p. 60.

2 The Third Electoral System: An Empirical Perspective

To this day many of our comrades still do not understand that they must attend to the quantitative aspect of things—the basic statistics, the main percentages and the quantitative limits that determine the qualities of things. They have no figures in their heads and as a result cannot help making mistakes. . . . In all mass movements we must make a basic investigation and analysis of the number of active supporters, opponents and neutrals and must not decide problems subjectively and without basis.

Mao Tse-tung (1949)

Periodization is both a tool and an objective of historical analysis. By demarcating a specified period of time and assigning a particular label to it, historians mean to denote that it has certain qualities or characteristics setting it apart from other periods of time. And the assigned label suggests the nature of those qualities—for example, the Era of Good Feelings, The Gilded Age, The Progressive Era. These semantic tags are employed as shorthand surrogates for concepts; each concept, in turn, refers either to a class of phenomena or to certain aspects or characteristics that a range of phenomena have in common.¹ That is, the concepts integrate and thus impart analytical unity to a variety of discrete occurrences whose aggregate quality is posited to differ from that of the occurrences in other specified periods of time. Often, however, historical periodization proceeds subjectively and without basis. Researchers tend to ignore the fact that concepts are logical abstractions that refer to underlying empirical phenomena. They regularly write of ages, epochs, and eras without first attending to the quantitative aspect of things.

To avoid that mistake, one must begin with an overview of the quantitative limits. Within a nation of continental proportions, even such a summary review of the basic statistics of a time segment of the voting universe requires more than a national perspective. To implant the figures in our heads, we must examine longitudinally and cross-sectionally the relevant data at the national, regional, and state levels.² And tactically it is useful first to clarify the integrating framework and then to structure the empirical inquiry to bring into focus several important qualities of the late nineteenth-century party system.

Electoral Systems and Realigning Sequences

Elections are recurring events reflecting underlying and ongoing sociopolitical processes. To abstract that dimension of similarity, however, does not imply that all elections are precisely similar. Some differ from others in behaviorally and systemically significant ways.

Empirical studies of the American electoral universe point to two broad categories of elections that differ from each other in kind and not merely in degree. Some elections primarily maintain the electoral system. The outcomes of these elections reflect some mixture of traditional party identifications and short-term, or surge, factors associated with current events or particular candidates. The relative importance of these long- and short-term components may vary, and on the basis of their valences in any particular mixture, we can distinguish three subtypes of elections. In maintaining elections the prevailing coalitional structure continues to dominate with only marginal changes from the earlier pattern. In such elections existing partisan loyalties are the primary attitude influences that shape voting decisions. In deviating elections, short-term attitude influences produce an outcome that differs significantly from the normal distribution of party loyalties. In these elections immediate events or personalities deflect party loyalties and produce an outcome that is inconsistent with the underlying balance of partisan attachments. In counterdeviating, or reinstating, elections the normal patterns reassert themselves following the expiration of the deviation-producing surge factors. However, none of these subtypes of elections drastically or permanently alters the coalitional bases of mass politics. The standing decision that structures the basic division of party loyalties among the mass electorate remains essentially intact.³

However, in some elections that standing decision is overturned. These are realigning, or critical, elections. Occurring relatively infrequently, such elections have distinct characteristics that set them apart from electoral system-maintaining elections. Realigning elections involve a large-scale and durable reshuffling of the coalitional bases of party voting support that extends to all levels of partisan elections. They evidence abnormally high levels of intensity—which is reflected in the process of selecting a candidate, unusually sharp issue polarization, and higher-than-normal levels of voter participation. Because they constitute a broadly diffused negation of the earlier standing decision, such elections change the shape of the voting universe.

Sequences of realigning elections are infrequent occurrences, but they are not random happenings. Integrally related to the essentially constituent nature of American parties, they emerge directly from the dynamics of that function. That is, realignments arise from emergent tensions in society which, not adequately controlled by the organization or outputs of party politics as usual, escalate to a flash point. When such tensions are felt intensely and diffused broadly among the mass public, they shatter the existing coalitional structure of mass politics by providing new bases for electoral mobilization. In turn, that produces significant transformations in large clusters of policy as well as alterations in the roles played by institutional elites. Realigning sequences, then, are involved with redefinitions of the universe of voters, political parties, and the broad boundaries of the politically possible.⁴ Their cyclical recurrence allows us to describe the predominant rhythm of mass partisan politics over the past two centuries as one of long periods of electoral stability punctuated by infrequent, short, and intense bursts of electoral reorganization. Each burst of electoral reorganization has produced a discrete party system. Each party system has shared common elements with the others, but each has also displayed its own characteristic patterns of voting behavior, of elite and institutional relationships, and of broad system-dominant decisions.⁵ Viewed across time, American electoral politics have witnessed no fewer than five national party systems.

The first party system, from 1789 to about 1820, was marked by an experimental, hesitant, sometimes nonpurposive, and never fully successful groping toward the creation of durable partisan structures. It might more properly be termed a pre party party system. The second party system, from 1828 to the realignment of the 1850s, grew from a host of local political alignments; it was characterized, inter alia, by a fully nationwide two-party competition. The third party system, or the Civil War party system, from 1853 to 1892, emerged from the collapse of one of the major parties, the breakdown of nationwide partisan competition, and the reorientation of both the party system and its policies along explicitly sectional lines. The fourth party system, from 1893 to about 1932, emerged from the incapacity of the third system to accommodate the emergent demands of the cash-crop agrarians of the South and the West coupled with the onset of an urban-industrial depression in early 1893. Unlike the other party systems, this one was characterized by the absence of viable partisan competition both at the national level and in a majority of the states. It was marked, too, by a steep decline in the high rates of voter mobilization that had been a hallmark of the third electoral era. Finally, the fifth party system, or the New Deal party system, resulted from the depression and realignment of the 1930s. It was marked by the gradual emergence of two-party competition in areas where that had not existed for decades and by the addition of class cleavage to the traditional mixture of voting alignments.

This terse overview of the nation's party systems draws attention to two features of American electoral politics. First, we can usefully and meaningfully periodize American electoral history in terms of a succession of discrete party systems. In doing that, however, it is essential to bear in mind that the term party system refers to voting systems or electoral-politics systems and does not relate to organizational structures. More specifically, the term party system, or electoral era, designates a set of maintaining, deviating, and reinstating elections bounded by realigning sequences.⁶ Expressed in the terms appropriate to the analogous conception of a voting cycle, we can describe each electoral era as involving a stable (or equilibrium) phase bounded by fluctuation phases.⁷ Second, since party systems can be demarcated by realigning sequences and since realigning elections, by transforming the basic shape of the electoral universe, differ in kind from other types of elections, we can both time the emergence of each party system and describe its structure empirically.

However, in employing the party-system conception as a periodiz-ing and integrating tool, we should not expect each discrete party system to be a pure type. Transformations in the contours of empirical data seem invariably more abrupt and sharper in their outlines than underlying transformations in political culture. Thus, we should expect each electoral era to exhibit mixtures of the old and the new, to be marked by syntheses of mixed forms and practices, cognitions and behaviors. Only by being attuned to the possibility of such mixtures, and the need to analyze the valences of their components, can we take into account the temporally and spatially uneven development of the electoral system.

Contours of the Third Electoral System

The transition from the second to the third electoral era entailed more than a simple substitution of the Republicans for the Whigs as the major opposition to the Democrats. It involved as well a change in the balance between the contending major parties and a shift in their regional sources of support. Longitudinal measures of partisan strength and balance provide convenient ways of summarizing these dimensions of the transition. Table 2.1 presents the mean strength of each of the major parties over the sequences of presidential elections from 1836 to 1852 and from 1856 to 1892 and the arithmetic difference between those means within each time segment.⁸

At the national level the swing appears to have been a mild one. The minuscule Whig lead that prevailed over the last five presidential elections of the second electoral era was transformed into a larger, but still mild, Republican lead. However, underlying that marginal national transformation were enormous sectional swings. Those shifts in partisan balance within regions, when aggregated to the national level, tended to counterbalance each other. The Midatlantic and East North Central areas swung sharply to the Republicans and provided that party with partisan leads that pale only in comparison with the level of Republican dominance in New England. In these areas the relatively tight partisan balance of the second party system gave way to much wider Republican leads. Countermovement in two other regions somewhat balanced these anti-Democratic swings. The Democrats opened a wide lead in the eleven-state Confederacy and converted the border region from its earlier Whig attachment.¹⁰

TABLE 2.1 Longitudinal Measures of Partisan Strength, 1836–1852 and 1856–1892⁹

This two-way movement across the regions of the country gave the third party system a shape reflective of sectional polarities that were unknown during the second electoral era. We can explore that suggestion and develop additional insights into the structure of the third electoral system by examining the Democratic percentage of the total presidential vote and its associated measures of central tendency and dispersion for separate elections over a span of time encompassing all of the third and parts of the second and fourth party systems (see Table 2.2).

The data afford two views of the third electoral system. First, an election-by-election inspection of the level of Democratic percentage strength reveals the changes in that party's political fortunes across time. The Democracy reached its nineteenth-century low point with its sectional split in 1860. Despite its apparently rapid resurgence in 1864 and 1868, Democratic strength remained somewhat artificially depressed through its postwar nadir in 1872. Over the twenty years after 1872, however, Democratic strength was obviously and rather consistently higher than it had been in the elections from 1856 to 1872. By implication, the data simultaneously point to the rather precarious national position occupied by the major anti-Democratic party. The Republicans won the presidency in 1860, but they polled only 39.7 percent of the total vote cast. The subsequent secession of eleven southern states reduced the size of the electoral universe, postwar Reconstruction measures changed the character of the southern electorate, and the Republicans polled majorities of the total vote in the 1864, 1868, and 1872 presidential elections. However, the party was unable to replicate that feat in any subsequent election, presidential or congressional, prior to 1896. In fact, in the sequence of presidential elections from 1876 to 1892, the Republicans were weaker than the Whigs had been over the sequence from 1836 to 1852.¹¹ Yet declining Republican strength did not translate automatically into surging Democratic majorities. In the presidential contests between 1880 and 1892, neither party was able to command the voting allegiance of as much as 50.0 percent of the total participating electorate.

TABLE 2.2 Distribution of the Democratic Presidential Vote

The measures of central tendency and dispersion provide another dimension of information and a better view of the shape of the party system. The variance data (σ²) require some preliminary clarification. Variance is a measure of dispersion about the mean of a distribution. Each item in a series is subtracted from the mean of the series and each remainder is squared; those products are summed and then divided by the number of items in the series. That quotient, the variance of the distribution, indicates the average squared deviation of the individual items from their own mean. The higher the variance, the greater the average (squared) distance of the dispersion of the items around the mean of the series. Applied to cross-sectional election data, low variances indicate relatively even distributions of support for a party across the reporting units. In comparison, increases in variance indicate some mixture of very high and very low levels of support across the units in the series.¹²

An examination over time of the means and variances presented in Table 2.2 points to a sectional rupture of the partisan distributions as a prime factor underlying the collapse of the second party system. That sectional disruption occurred in two distinct stages. In the first stage the sectionally based Republican party was substituted for the nationally oriented Whig party as the major anti-Democratic opposition. The systemic impact of that substitution is dramatically evident in the sharp increase in variance (and decrease in mean percentage) for the major anti-Democratic party between 1852 and 1856. The Whig vote in 1852, whether calculated by states or by sections, was one of the most uniformly distributed in American electoral history. In contrast, the Republican votes in 1856 and 1860 were among the most unevenly distributed.¹³ This upsurge of variance mirrored the Republican party's nonexistence below the Mason-Dixon line coupled with its preponderant strength in New England. Although the variance of the Democratic distribution by states more than doubled over the interval from 1852 to 1856, the even sharper increase in its variance by sections provides evidence of the underlying strains that shortly thereafter tore the party (and the nation) asunder. The nearly fivefold increase in the variance of the Democratic distribution by sections reflected the party's collapse in New England combined with marginal increments in its level of voting support in the states that later formed the Confederacy. The second stage of this process of sectional polarization came between 1856 and 1860. It is apparent in the sharp reductions in the Democratic mean figures and in yet another more than doubling of the variance both by states and by sections. The collapse of the Douglas Democrats in the Confederate and border regions underlay these changes.

With the violent disruption of the Union and postwar legal changes in the character of the voting universe, the variances of each of the partisan distributions were reduced, only to begin slow but uneven increases from their shared 1872 lows. From that point forward (with the exception of 1884) the variances increased from one election to the next through 1892. The sharp upturn in the measures of Democratic variance in 1892 resulted from the party's virtual collapse in the Mountain states and its sharp drop in the West North Central area. The Republican distribution became less homogeneous across time, as that party's efforts to develop a southern base collapsed and as its strength remained relatively intact (at least through 1888) in the Mountain states. As with the Democratic distribution Republican variance surged upward in 1892, marking the beginnings of the sharp North-and-East vs. South-and-West political polarization that reached its dramatic climax (though by no means its termination) in 1896.

Each of these angles of view provides insight into the structure of the third party system. Viewed across time, the Democratic percentages suggest something of a transition (if not transformation) in electoral state at some point between 1872 and 1876. It is at least clear that from 1876 through 1892 the Democrats drew consistently higher levels of voting support and that there was less election-to-election fluctuation in the amount of that support than during the period from 1856 through 1872.¹⁴ Viewed cross-sectionally (by states and by geographic regions), the data point to immense imbalances in the geographic distributions of party strength: sectional polarities of a nature and magnitude that distinguish the third from the second party system. That succession of party systems involved the post-1852 collapse of national parties and their displacement by parties that drew disproportionate support from one or more sections of the country.

Although the variance data bring into focus the sectional polarities underlying late nineteenth-century party oppositions they also draw attention to another significant feature of the third party system. The Democrats tended to gain strength over time in those parts of the country in which they had initially been weak. Although the Democratic regional mean in 1868 was nearly identical with that of 1888, the variance of the regional distribution had been reduced by more than half over the intervening twenty years.¹⁵ And the pattern is a relatively consistent one. Excluding the 1872 and 1892 deviating elections, higher means and lower variances characterize the post-1876 Democratic distributions when compared with those of the sequence from 1856 to 1868. And again, as with the percentage strength data, 1876 seems to stand out as something of a turning point.

Phases of the Third Electoral System

Any electoral era consists of a set of electoral system-maintaining elections bounded by realigning sequences. Yet it is necessary to isolate the breakpoint between the realigning and maintaining sequences. Then we can divide the system into its behaviorally distinct parts and describe the empirical qualities of each.

To address this problem, one must shift the angle of view from cross-sectional to longitudinal comparisons. Although a number of distinct ways exist to do that, each involves calculating a statistical measure for one sequence of adjacent elections and comparing that with a similar measure for the next sequence. For example, Walter Dean Burnham has statistically compared the first five of a sequence of ten successive presidential elections with the second five to identify ‘cutting points’ of transition between one system of electoral politics and another.¹⁶ Using presidential data at the national level, Burnham found an unusually sharp cutting point associated with the midpoint 1874, separating the 1856— 72 presidential sequence from the 1876–92 sequence. We can follow Burnham's lead, but with two modifications. First, we will evaluate longitudinal means based on sequences of five biennial elections. Second, the evaluation will be executed at the national and regional levels. Specifically, that procedure involved calculating a longitudinal mean for every possible five-election sequence of biennial elections for the nation and for three regions from 1848 to 1904.¹⁷ Second, I used a t test for differences between means to compare each mean with that for the next sequence to determine whether the observed difference was greater than that in their standard deviations would lead us to expect. Leaving aside the breaks in the series associated with the realignments of the 1850s and 1890s, the array in Table 2.3 indicates the peak values of t and the midpoints separating the two sets