This paper examined the current status of mass support for American political parties with particular attention to the possibilities of partisan realignment. There is considerable agreement among analysts on the changing patterns of electoral behavior in the 1960s and 1970s. but much less agreement on the implications of these changes for the future of American` politics. There are three topics to be developed here:

1. a summary of the most relevant generalizations about electoral behavior,
2. a review of several expectations about realignment by other analysts, and
3. our own interpretation of these circumstances

Some themes appear in almost all recent work on electoral behavior and political parties and should appear familiar when restated here. In the last decade there has been an increase in political independence and a decline in partisan attachment, an increase in ideological thinking and issue orientation among Americans. a decline in the reliance placed on political parties, and a disintegration of the New Deal coalition. The most general and most important conclusion- drawn from these studies in the possibility that political parties have become useless and irrelevant as governing mechanisms. The evidence for the generalizations and this conclusion must be examined carefully because, by any standards, these are crucial statements about contemporary politics.

V. 0. Key first called attention to the phenomenon of sharp and durable changes in the underlying partisan division of the electorate, disruptions which he termed "critical elections.." [2] His notion was elaborated upon by Campbell. et al. and later by Poimper who developed classifications of election types with realigning elections defined as those characterized by a lasting change in the underlying partisan division of the electorate which produces a new majority party.[3]

Charles Sellers pointed out the cyclical nature of these realignments, a conceptualization that has been considerably elaborated in recent years by others, most notably W. Dean Burnham. [4] The following outline of the realignment process draws freely on the work of these scholars, but is an effort to integrate their ideas rather than to restate them precisely. Realignments are seen by Burnham and others as the product of a crisis, some societal upheaval that crystallizes political divisions in the electorate around some new issue dimension, a dimension relevant to the precipitating crisis. This galvanizing experience ushers in an era of partisan stability in which most elections are contested primarily on the basis of the realignment issue, with the electoral coalitions held intact by leadership appeals emphasizing the social cleavages and attendent party loyalties growing out of the realignment crisis.

More recent discussions have emphasized that as time passes, the coalitions born of the realignment begin to decay.[5] New issues arise that no longer follow so neatly across the divisions of the past. New generations of both voters and political leaders emerge for whom the slogans of past battles seem outmoded. Increasingly the old political alignments appear irrelevant to the concerns of the present. This detachment from old loyalties in evidenced in the occurrence of third party movements centered around new issues not adequately captured by the existing major parties, increasing volatility in the behavior of the electorate, both in splitting their tickets in a single election and in vote switching from one party to the other from election year to election year, and in an increase of voters unaffiliated with the major partisan competitors. This increasingly independent behavior on the. part of a substantial portion of the electorate sets the stage for the next realignment; indeed, without this disengagement from the alignment of the past, a crisis is unlikely to result in a full-scale rearrangement of the lines of cleavage in the electorate. The approximate length of time for the sequence of electoral phenomena to come full circle is thirty to forty years.

The initial view of realignments assumed, somewhat casually, that critical elections were the occasions for massive conversions of voters from one party to the other, with the abruptness of these partisan desertions mitigated, perhaps for some, by a stop at one of the third-party "half-way houses". More recent analysis, most notably by Anderson, Beck,. and Converse, [6] based on survey materials and focusing on the New Deal period lend credence to a theory of generational change as the basis for most of the shifts occurring during a realignment. While the number of voters who converted from one party to the other during the 1930s is greater than at other times, the bulk of the voters that permanently swelled the Democratic ranks were new voters, either young or previously non-participating, who entered the electorate for the first time in the early 1930s overwhelmingly on the Democratic side and remained there. Kristi Anderson makes a convincing case that the accumulation of young voters, women, immigrants. and the children of immigrants who had become eligible to vote during the 1920s but did not, provided the harvest of adherents that the Democratic Party reaped in the 1930s.

Both the cyclical nature of the realignment process and the generational foundations of partisan charge emphasize the importance of the existence of a pool of voters or potential voters, unattached to either party and thus available for capture by one or the other of them. Only when the decay of the old alignment has a substantial portion of the electorate free of party loyalties can a crisis act as catalyst to restructure the partisan divisions in the society Because partisanship is such a stable characteristic over the lifetime of the individual, the pool of voters available as new adherents to a party during realignment will come disproportionately from the young, from older independents, from the newly enfranchised and from the apolitical. More than anything else, perhaps, it is the evidence in recent years that this pool of available voters is growing, that increasing numbers of voters and potential voters are without a firm partisan mooring that has led to recurring reports of the decay of the New Deal alignment and the expectation that another realignment is imminent. The continued failure of the electorate to realign has led other analysts to surmize that the system has changed so drastically that another realignment will not occur.

What are, then, the particular changes in contemporary electoral behavior on which speculation is based about the prospects for realignment? Perhaps the most incontrovertible piece of evidence supporting the onset of a disintegration of the old partisan alignment is the impressive rise in the proportion of independents in the electorate. Between 1964 and 1974 the percentage of
independents increased from 24 percent to 38 percent, with the consequent decline in partisans coming more heavily from the Democrats (although Democrats and Republicans have lost similar proportions of their respective strengths over this period). Most of this increase is concentrated in the younger age cohorts; new voters are more likely to be independent and to stay independent than in the past.

Concomitant with this rise in thenumber of independents among young people is the apparent decrease in the "efficiency", with which parental party identification is transmitted to the younger generation. Paul Beck, in seeking to build a theoretical foundation for the empirical observation that realignments occur once every thirty to forty years, has argued that the parents of this newest cohort of young voters did not themselves experience the tribulations of the New Deal era as adults.[8] Thus the socialization of the youngest age group in the emotions and issues of the New Deal has been essentially indirect -- with consequent slippage in its effectiveness.

The increasing lack of holding power of the New Deal alignment is evidenced in other ways. Trilling has found a decline in the use of New Deal symbols and issues in the candidate and party evaluations of respondents in studies conducted by the Survey Reacarch Center and the Center for Political Studies. [9] This decrease is most evident among the younger age groups in the electorate. Other scholars, notably Nie, Verba and Petrocik and Ladd and Hadley, argue that the social group basis of the New Deal coalition in disintegrating. [10] While this argument has several facets --for example, the departure of the white south from the Democratic presidential party -- the most interesting aspect is the evidence suggesting a disengagement along generational lines of the class basis of partisanship in the north. To put it succinctly, the children of the middle class appear to be considerably more liberal and less Republican than their parents. Ladd and Hadley, particularly, develop the thesis that the intelligentsia, of a post-industrial society is a profoundly different political animal than the entrepreneurs of previous generations, despite our tendency to treat them both analytically as "middle class". [11] As a result, class becomes less of a determinant of political tendencies than it was at the height of the New Deal era.

Most of these changes have the greatest effect on new voters -- those just coming of age politically. As the trends continue, of course, these younger age groups will become an increasing proportion of the electorate. One set of changes. however. appears to affect all political generations of voters. Voters of all ages are more willing to respond to short-term forces in casting ballots for particular candidates. The great swings in electoral fortunes that produced landslides in opposite directions in presidential politics in the past two decades are one indication of this greater electoral volatility. Equally important is the reported tendency of voters to split their tickets in state and local contests. As the Survey Research Center and Center for Political Studies show, split-ticket voting in state and local contests has increased from a low of 32 percent of all voters in 1960 to 65 percent in 1972 (although it appears possible that 1956, 1960 was atypical in the small amount of ticket-splitting reported). In the 1960s and 1970s even self-declared partisans found themselves free to deviate temporarily from their loyalties.

Another. very visible set of trends is emphasized by some analysts as an integral part of the electoral changes taking place in the 1960s and 1970s. Since 1964, the level of trust and confidence in government and political leadership has steadily eroded, with an acceleration of this decline accompanying the revelations of Watergate. This disaffection, recorded by a variety of polling organizations using many similar but not identical indicators, appears to have reached all groups of citizens (though to varying extents) and covers a wide range of political, economic and social institutions, including the presidency, Congress, the courts, and the political parties.

Taken together, this evidence suggests a considerable growth in the size of the pool of voters without well-developed partisan loyalties. Most analysts, and we would include ourselves, see this as indicative of an increasing availabilty of potential new partisans in a realignment. Others have combined these findings with a set of somewhat more controversial generalizations to conclude that this detachment from party loyalties signals the end of parties as we know them in the American political system.

Several analysts have noted the increase in the prominence of issues in the consciousness of the American voter since the mid-sixties. This has taken a variety of forms. Nie and Anderson have shown, for example, that the level of issue "constraint" has risen rather remarkably from that observed by Converse in the 1956 and 1960 panel study. [13] Their overall index of attitude consistency, combining both domestic and foreign policy issues, rose from a gamma of 0.14 in 1956 to 0.38 for the 1972 data. Along the same lines. Nie, Verba and Petrocik have traced an increase in the use of ideological terminology in respondents' evaluations of the political parties and candidates, beginning in 1964. In general, these findings have been interpreted to mean that the parties and, even more, the candidates in the 1960s and 1970s have distinguished themselves ideologically and built connections between issues more overtly than in the "issueless" 50s. Once the political landscape was clearly marked out with issues, the public has been able to respond to them in kind.

Somewhat more difficult to untangle is the contention that the rise in the importance of issues signifies a decline in the relevance of partisanship for American voters. Several scholars have noted an increase in the relationship between issue positions and presidential vote choice since 1964, increasing in the proportions of individuals voting consistently with their issue positions, when issues and partisanship are incongruent, and an increase in the contribution of issue position to explaining vote choice in regression analysis. [16] As Converse has pointed out, much of the argument about the relative impact of party and issues on vote choice depends on the extent of congruence between partisanship and issue positions and the ability to uncover the temporal ordering of the two determinants. [17] The responses to open-ended questions show the decreased frequency of references to party ties in the evaluation of candidates, the declining relationship between individuals' evaluations of the parties and their vote choices, and the overall increase in the negative remarks among all voters. These patterns are even stronger among young voters. While not universally accepted these trends in the reactions of the electorate to issues and parties are taken by many as further evidence both of the continuing detachment of the voter from the partisan alignment of the New Deal and from political parties themselves.

These findings have served as the basis for a variety of assessments of the probable future of American electoral and party politics. We will consider four versions of these possible developments: [19]

• a continuation of the New Deal alignment;
• the realignment of party loyalties along a new issue dimension;
• the dealignment of partisan loyalties; and
• the disappearance of political party loyalties.

Continuation of the New Deal alignment. This perspective argues that there is more stability than change in present electoral behavior and/or that the volatility of voting behavior is a temporary departure. The admitted disruption of electoral politics by Vietnam, race and other non-economic issues, according to this view, had done little to affect voting except for President. Sundquist argues that the New Deal alignment is still substantially intact and Converse appears to agree with this interpretation. [20] Similarly, the increase in issue constraint noted since 1964 is interpreted by Converse as a reflection of the more explicit clustering of issues by the parties and other reference groups. [21] On the basis of similar data from other systems he suggests a return to lower and more "normal" levels of constraint in the future. In the absence of evidence of a new alignment and with only limited indications of deterioration of the old loyalties, the conclusion remains that the New Deal alignment is still intact although not an strong as it once was.

Realignment of Party Loyalties With New Issues. This perspective on electoral change is nearest to the familiar image of a realignment in which some new cross-cutting cleavage rearranges party loyalties and establishes a new, dominant coalition of electoral groupings. There are several versions of this perspective, usually focusing on the issue cluster of Vietnam, law and order, race and social permissiveness an the likely realignment issue. Phillips sees the emergence of a new Republican majority based on a coalition of voters similar to Nixon's support in 1972 or Nixon and Wallace's combined support in 1968. [22] The white south, the growing urban and suburban areas of the west and southwest, and elements of the northern working class disaffected from the Democrats on civil rights might join with the traditional segments of the Republican party to form a new and winning coalition. Scammon and Wattenherg, who coined the term "the Social Issue" for the collection of issues dominating politics in the late 1960s also see this outcome as possible, but only if the Democrats are so inept as to abandon the political center to the moderate wing of the Republican party. [23] On the other hand, should the Republicans move to the Right on this issue, the Democrats could reap the benefits of a realignment, by staking out a center position. Or, alternatively, the New Deal alignment will be prolonged if the parties neutralize the Social Issue by taking similar, center positions on it.

In contrast, Miller and Levitin suggest the emergence of a new Democratic majority based on a similar "New Politics" dimension which juxtaposes New Liberals to the Silent Minority. [24] Their view that the issues of the late sixties will ultimately favor the liberal side is based on 1970 and 1972 data that show the growth of the New Liberals and the contraction of the Silent Minority, even in the midst of the McGovern debacle.

In such an environment, the disaggregation of the party system will continue unabated. Nie, Verba and Petrocik conclude there is "little prospect for the emegence of a new party system from the disarray of the present system."[29] Burnham, in his most recent summary of these factors, argues:

"[The American electorate] is undergoing a critical realignment of radically different kind from any in American electoral history. This critical realignment, instead of being channeled through political parties as in the past, is cutting across older partisan linkages between rulers and ruled. The consequence in an astonishingly rapid dissolution of the political party as an effective intervenor between the voter and the objects of his vote at the polls." [30]

Among independents there is a substantial relationship between leaning toward a party and vote choice for President and Congressman from 1952 to 1972. As shown in Table 1 the strength of relationship between the three categories of independents and vote choice ranges around Somer's d s of 0.4 and that level is only somewhat below the relationship between the full range of partisan categories and vote choice. Thus it is not necessarily warranted to contend that all Independents totally lack guidance from partisan identification in forming opinions or casting votes. The majority of Independents may indeed be guided to some degree by party cues, and quite possibly in recent years Independents are as strongly influenced by party as weak partisans.

A similar point can be made about the evidence of an increase in split-ticket voting. Although the proportion of voters reporting that they split their tickets in state and local elections has increased substantially since 1956, two-thirds or more of the ticket splitters say they vote predominantly for one party. [33]

Neither of these points deny the contemporary increase in the size of the pool $ that has contributed to electoral volatility in recent years. The proportion of those who divide their tickets evenly has risen to almost one-fifth of the electorate. By the sa$ frequency of pure (non-leaning) Independents has increased from a little under $ to around fifteen percent by 1972. Even though 85 percent of the electorate ar$ degree, the availability of 15 percent of the electorate as relatively freely f$ parties provides more than enough voters for impressive electoral swings.

The more traditional views of realignment can tolerate and, indeed, depend on the increase in independent voting temporarily but those views also depend on the continued viability of party identification. There is no doubt that more voters are calling themselves "Independents" than in the past and this is a significant development in terms of the psychological detachment of individuals from political parties and their availability for realignment. But it is a more extrerne conclusion that the parties are becoming irrelevant to the choices of voters, that they no longer provide them with cues for organizing political information, and that they are on the verge of disappearance due to their failure to serve any useful purpose beyond the nomination of candidates. Such a conclusion seems unwarranted in light of evidence that most independents do perceive differences between the parties, state a preference between the parties when pressed, and behave in the voting booth in a manner consistent with that preference.

Party and Vote Choice. The other form of evidence that is commonly used to show the decline of party is the lessening relationship between partisanship and vote choice and the increased impact of issues on voting. However, the most extensive analysis of these trends, for example, Nie, Verba and Petrocik's examination of issue voting, is based almost exclusively on evidence from presidential voting. [34] If we compare the impact of party on vote choice in presidential voting with that in congressional voting, it can be shown that the decline of party as a guide to voting and the breakup of the New Deal coalition are not so obvious in the lower office.

Tables 2 and 3 present two uncomplicated ways of looking at the impact of party on voting behavior from 1952 to 1972. First, are two measures of association assessing the simple relationship between party and vote choice--gamma, which is most often use in analysis of this sort. and the more appropriate Somer'e d. [35] Secondly, we used the rate of defection from the party's candidate in elections for President and Congress. Both these displays show essentially the same thing: party loyal voting has decreased considerably less in congressional elections than in presidential voting. And one might also argue that the decline in party loyalty in presidential voting is not all that precipitous.

Even though the vote for President has become more free of both party influence and the social class influence associated with the New Deal alignment, it is still possible that the old coalition remains in voting for other offices. The "two-tier" party system refers to this independence of presidential voting from other patterns, but there is no evidence of the disintegration of the New Deal coalition or the disappearance of party voting below the presidential level. Actually the party identification of social groups belonging to the New Deal coalition as well an their voting for Congress has not changed much from 1952 to 1972, even as voting for President in these groups has changed significantly. [36]

If the New Deal coalition remains intact below the presidential level, there are several implications for the perspectives on realignments. The revitalization of Democratic dominance along the lines of the New Deal coalition appears a somewhat nore likely development. Also, for other forms of realignment to occur, the New Deal alignment must be disrupted and replaced. The many traumas of the 1960s and 1970s have not dealigned the American electorate as thoroughly as supposed. To a considerable degree a new, dominant coalition must still undermine, even today, the New Deal alignment.

The analysis of party voting has typically not made distinctions in voting patterns on the basis of geographic area. There are distinct differences between the north and the south, and Table 3 reveals, as would be expected, that defections in presidential voting are much more common in the south from 1952 through 1972. While defections in presidential voting have increased moderately in the north these years. the change is not great and over three fourths of the party identifiers support their party's candidate for President in every election.. Congressional voting is not distinctively different in the two regions. Though the increase in defection is not entirely a result of voting behavior in the south, the south contributes disproportionately to the increase in 1968 and 1972 as well as to a higher level of defection in all years except 1964. It is also the case that differences between the south and the rest of the country contribute to the increase in Independents in an unexpected way; in the 1960s the overall increase in self-identified Independents nationally is solely a result of the increase in the south. Later, to be sure. the increase appears in the north but the interpretation of the trend nationally does not usually emphasize the rejection of parties as exclusively southern through 1968. It is not easy to estimate how much of the political change in the nation is a peculiarly southern pattern but it is well documented that the southern electorate has been much more volatile in the past ten years. [37] Thus the most impressive evidence for the decline of party comes from the behavior of southern whites abandoning the Democratic presidential candidates to vote for Goldwater, Nixon and/or Wallace in conformity with their positions on, above all, civil rights.

There is no doubt, of course, that this is not party loyal voting. On the other hand, to argue that it is evidence of the growing irrelevance of party and their ultimate disappearance seems hardly appropriate either. Clearly it is a result of the disaffection of conservative southern Democrats from the liberal national party, but one can argue just as well that it is the growing coherence and strength of the northern wing of the party rather than its weakness that has led to this defection. Conditions may be ripe for a realignment based on regional and racial cleavages, but the disappearance of parties sui generis seems not to necessarily follow from the evidence.

A fundamental element in most discussions of the general decline of party loyalty has been the increasing impact of issues on vote choice. Discussions of issue voting are found throughout the recent literature on electoral behavior. Basically it is contended that beginning in 1964 the impact of issues has increased in presidential voting while the influence of party has declined. The evidence varies in complexity but usually shows that the simple or controlled relationships between issues, party and voting have changed over the last twenty years. Nie, Verba and Petrocik present the most sophisticated and cautious estimates of the unique and joint contributions of party and issues to vote choice from 1956 to 1972. [38] As Figure 1 which draws on their analysis shows, the unique contribution of party has declined greatly since 1960. The unique contribution of issues to the explanation of vote choice has increased from practically nothing to a noticeable share of the variance,11 percent. However. only in 1964 does the total impact, unique and shared, of issues rise as high as 25 percent and in all other years the impact of issues is distinctively less than party even if their shared influence is attributed entirely to issues.

What is more fascinating in Figure 1 is the decline during the period in the capacity of party and issues together to account for vote choice. About half the variance in vote choice is accounted for by these variables from 1956 through 1964 but by 1972 only a little over one third of variance is explained by these variables. Candidate orientation and all sorts of other short-term forces independent (statistically) of party and issues have assumed a larger role in determining vote choice. This is a strong argument for increasing electoral volatility in presidential voting and perhaps the disassociation of presidential voting from other behavior but not a convincing case for the dominance of issue voting.

Several analysts, but most notably Pomper, have pointed to the election of 1964 as the occasion for the rise of issues in political behavior. [39] The work on issue constraint by Nie and Anderson cited above also shows the dramatic change coming in 1964. [40] The common interpretation of this change emphasizes the role of political leadership in raising issue awareness and Pomper's data demonstrate the ability of the leaders to increase the issue alignment of parties. [41] The 1964 presidential campaign brings issue positions more strongly into line with party identifications and this issue-related role of party continues into the 1970s. These findings do not suggest the demise of parties but rather the increased distinctiveness of the parties on issues since 1964. These results from Pomper's study should be viewed as establishing the capacity of political party leaders to inspire issue awareness in the electorate and the central role of party in these attitudinal changes should not be overlooked. The increasing role of issues in vote choice is much more dependent on parties than it is in conflict with them.

In raising doubts about the empirical evidence on the disappearance of political parties several themes have been pursued. The weakening of party loyalties does not entail the abandonment of party as an object of identification or a cue-giving symbol. The increasing salience of issues in particular is a party related development. The dissolution of party ties associated with the New Deal has been noticeable principally in presidential voting and in the South.

Burnham, and Ladd and Hadley particularly have argued that the empirical trends are irreversible and spell the end of party systems because of several social changes that have occurred:

• the wide spread availability of mass media,

• the movement to a post-industrial society with its differing class structure; and

• the increasing distrust and cynicism toward social and political institutions. [42]

Just as the findings on issue voting are not necessarily destructive of the role of party in vote choice, the increasing attention to mass media is subject to various interpretations. That the media, television really, have come to be more important channels of communication appears undeniable, but it is less obvious that this development is at the expense of political parties. The political parties as organizations and the political party leaders use the media, again especially television, to communicate partisan messages both as news and as advertising. The effect is particularly strong during campaigns, perhaps, for conveying information. Patterson and McClure demonstrated the capacity of
campaigns to inform voters about issue positions without changing voter's attitudes.[43]

The availability of television to many candidates who would have lacked other equally effective means of reaching voters may partially account for ticket splitting among the most salient races. Factors like the impact of incumbancy are easily conveyed by the mass media and may diminish somewhat the impact of party and issues on vote choices. On the other hand, should support for party positions became a salient characteristic, as it would in a realignment of the traditional sort, the media would provide an opportunity for conveying candidates' support or non-support of party programs, endorsement or non-endorsement by party leaders to the voters. Thus in different circumstances, the availability of more information about candidates may actually encourage straight ticket voting rather than the reverse.

The media are channels to be used to enhance or undermine the impact of party depending mainly on the political leadership. The mass media have not caused the decline of parties although the media have carried information that has hurt the parties. The media and the media-based commentators are not independent political institutions; the media are not in a position to replace parties. The media represent an opportunity for political leaders, they provide a resource leaders can use if they are to realign the electorate now or in the future.

We are quite willing to accept the argument that the class structure in a post-industrial society will be different from that in a society in the midst of the industrial era. What we fail to find convincing is the notion that the cleavages in such a society would be so qualitatively different that they could not be reflected in a party system such as has existed in the past. The coalitions of support for the parties might well be different, the differences between (or among) the parties might reflect cleavages of a different type, but parties may remain the mechanism through which these cleavages are represented in politics. We did not see persuasive evidence that the parties have disappeared as objects of political significance to the voter at present and we see no commanding reason to expect societal changes that would necessarily render them irrelevant.

Another point ought to be made in support of the proposition that political parties will remain viable. First, the parties (to the distress of their critics) have survived an extraordinary amount of social and political change in the past. To the extent there is a theory of political parties, it suggests that both the structures and processes of American parties make them adaptable. If there is a change by some leaders in both parties that threatens this adaptability, it is the abandonment of integrative strategies and solutions in favor of more divisive, doctrinaire stands.

A third theme introduced as either a cause or effect of the decline of political parties is a loss of confidence and an increasing cynicism about them. It is certainly well established there is less trust in political parties and since this is presumably the public's reaction to the behavior of leaders, it is hardly surprising. By the same process. it can be imagined, skilled leaders, should they appear, could restore confidence and reduce cynicism, though it may he easier to destroy trust than create it. Furthermore this trend in attitudes is not limited to parties or political leaders, it extends to all of government and most social institutions. Cynicism may remain highbut it is surely not fatal to all institutions that it touches. Most informed observers are somewhatcynical about parties and politicians; probably the less active can handle cynicism too. The importance of the loss of confidence in political and social institutions in the last decade should not be forgotten, and in fact, can be viewed as a contemporary crisis of sorts. However, it appears unwarranted to conclude that existing levels and intensity of mistrust will necessarily prevent the parties from recovering. It is impossible to demonstrate with survey data, but the parties have survived several periods of disenchantment apparently as severe as the present.

There is little disagreement that the data of the 1960s and 1970s present evidence of the disaggregation of the old party alignment and of the increase in the pool of voters and potential voters detached from traditional party loyalties. We disagree. however, with those who see in the decline of the New Deal coalition the death of political parties. Their argument essentially is that the observed dealignment has progressed to a point from which the parties cannot recover, that many of their traditional functions will disappear and with them, the party system as we know it. In the paragraphs above we have attempted to show that the parties still remain referents for the voters, albeit they view them with less commitment than in the past. We see no evidence that some threshold has been reached from which the parties cannot recover as organizers of political information for the public or as aggregators of issues and interests linking the public to the governing process. Clearly. the parties are less effective in performing these functions than in the past. but we see no evidence that the parties are incapable of carrying out these functions if the political leadership and the electorate aligned themselves around some set of salient issues. In other words. we see no reason why the parties and the. electorate are not capable of a realignment in the traditional sense of the term.

Whether or not the party system has reached a point of no return is no more than speculation unless it has some grounding in theory. At the risk of over simplification, the reasoning of those who see the end of the party system in the current dealignment seems to be as follows:

• Conditions have been right for a traditional realignment since the early 1960s.

• None has occurred and there seems to be little evidence that one is about to occur in the near future.
• Therefore, the mechanism that by now should have produced a realignment must have gone awry.

There are several problems with this argument, but we are most concerned with bringing the conceptualization of the realignment process into sharper focus and paying greater attention to historical parallels. In the remainder of this paper we will offer a conceptualization of the realigning process, discuss successful and unsuccessful realignments from the past, and apply these views to the recent period.

The realignment process can best be thought of as encompassing three stages:

• a rejection of the party in power;

• policy action by the new party in power which is widely perceived as being successful; and

• an attitudinal adjustment on the part of enough voters to alter for a long time the balance of power of the parties in the electorate and perhaps the lines of cleavage as well.

While in no way contradictory to what others have said about the nature of realignments, it places emphasis upon the temporal ordering of these phases and draws attention to the possibility of an incomplete realignment.

This allows us to make some important points. First, there are many more. potential realignments than successful ones, more rejecting of the rascals in office than embracing of their replacements. Second, whether a rejection of one party becomes a successful realignment depends in large part, if not entirely, on what the party coming to power does with the opportunity. Only if they are perceived as successfully governing will the attitudinal adjustment of the third stage of realignment follow. The elections that we call deviating with the benefit of hindsight might also be seen as missed opportunities for realignment by the party of the leadershipin office.

This also requires that we emphasize another aspect of realignment--the need for the party coming to power to have unified control of government to accomplish a successful realignment. Given the separation of powers in the American system, a necessary (through clearly not a sufficient) condition for realignment is the attaining of control of both houses of Congress and the Presidency so that some image of successful governing can emerge.

The theoretical formulations concerning realignment sequences depend explicitly on the idea of the advantaged party capturing control of government. A realignment is precipitated by a crisis in which the current rascals are thrown out of office and a new group installed. A realignment depends upon a successful handling of the crisis, an outcome which solidifies the base of support of the new majority party and provides a set of symbols and slogans which they can use for years to come to rally the troops in election campaigns. If the party coming to power in a crisis does not have effective control of government--either because it lacks an ideological majority or for want of effective leadership--and if it does not manage to resolve the crisis, in a short time the party will lose control of one or more of the policy-making institutions to its opponents. No realignment would be said to have occurred.

Most of the empirical analysis of realignment has, however, concentrated exclusively on electoral changes and not considered whether these changes led to effective control of government by the advantaged party. [44] Whether today's changes in observed behavior are to become permanent depends in large part on what the party in government does or does not do. It is for this reason that we have argued elsewhere that unified control of the policy-making institutions for a sustained period of time should be made an integral part of the conceptual and
operational definition of realignment. [45]

The three widely acknowledged realignments: (1) the Civil War, (2) the realignment of 1896, and (3) the New Deal realignment do, of course, meet this test of involving a sustained period of unified control of government by the advantaged party. For present purposes, what is more important is the number of elections that included a rejection of one party and the giving of united control to the other and did not result in a realignment. Perhaps the clearest case is the election of 1892 in which the Democratic Party won the Presidency from the incumbent Republican and gained majorities in both houses of Congress. Historians have surmised that there was the beginning of a what might have been a long supremacy for the Democrats. [46] However, the recession of 1893 intervened to destroy the image of the Democrats as successful managers of government and put an end to the Democratic realignment-in-the-making. Similarly, we can view the 1850s as a missed opportunity for the Democrats to launch a realignment by solving or successfully delaying the slavery question with united control of government they held in 1852. Thus the incapacities of political leadership, and the course of events over which the leadership has no control, can turn an incipient realignment into just another deviating election.

The notion that the electorate responds to the government's handling of the crisis sometime after the initial rejection of the other party is more difficult to show historically because of the absence of survey data. We are hypothesizing a change in attitude--specifically an adoption of partisanship--after anddistinct from a change in voting behavior. With aggregate election data we are necessarily restricted to evidence of the change in voting behavior. The tendency has been, quite understandably, to date the realignment at the time of the initial change in behavior. Indeed, if the process worked exactly as we have outlined it, we would not necessarily expect any further change in behavior after that point. Only survey data would reveal the reinforcing of the behavior with an attitudinal change sometime afterward.

Evidence for this pattern is available for the New Deal realignment. While Roosevelt won a convincing victory in 1932, his support came from all quarters, not simply from the New Deal coalition that would emerge later. Shively uses Literary Digest poll data with its famous bias toward the middle class to show that support for Roosevelt was quite uniform across the electorate in 1932. [47] In 1932, as in the1920s, the Literary Digeat poll predicted the results of the presidential election quite accurately. Since the biased sample did not distort the predicted outcome, one can infer that the middle class behaved in these years much like the electorate as a whole. Not until 1936, Shively argues, did the class basis of the New Deal coalition become apparent in response to the program that the Roosevelt administration was able to enact with its unified control of the political branches of the government.

The campaign rhetoric of 1932 also did not suggest that a new set of symbols and appeals was in the making that would last for another thirty-odd years. Had, the Roosevelt programs not been perceived as successfully meeting the crisis of the Depression, there is no reason to suppose that the young and new entrants into the electorate in the 1930s would have developed lasting Democratic identification based on the symbols and issues of the era.

The emphasis on attitudinal change following and being a response to the performance of the party in power means that it is unlikely that public opinion will anticipate a realignment. The electorate will respond, or not respond to political leadership. [48 ] Important individual characteristics like party identification will be slow to change, lagging behind voting or attitudes toward issues. The absence of signals in the present electorate that foretell a realignment should not be treated as a significant indication about. the future. A thoroughly informed electoral analyst in 1932 would not have detected what in retrospect was the beginning of the New Deal realignment.

By 1968 it was evident that a Democratic realignment had not emerged and other analysts were discussing the pending Republican realignment. Nevertheless. from our perspective, the discussion of a potential Democratic realignment in the early 1960s was thoroughly appropriate. The Democrats had firm control of both Houses of Congress and the Presidency was occupied by a Democrat of heralded ability to deal with the legislative branch. President Johnson offered a policy- making agenda designed to revitalize the New Deal coalition and win new adherents in the age of affluence. Had the administration not become mired in Vietnam, had the ghettos not begun to burn, we can image a new generation of voters as attached to the Great Society as a previous generation had been to the New Deal. The failure of the incipient realignment of 1964 was a failure of leadership and the intrusion of unforeseen and perhaps uncontrollable events, not the failure of the willingness or capacity of the electorate to accept and hold loyalties to political parties.

Following the 1964 election both Burnham and Pomper saw the possibility of a Democratic realignment in the traditional style. [49] Voting patterns (if not many party identifiers) had shifted dramatically and on balance In favor of the Democrats. In aggregate terms the south shifted toward the Republican Party and much of the north moved toward the Democrats. In fact, the disruption of aggregate electoral patterns in presidential voting is much more dramatic than the electoral shifts in any of the realienments of the past. By 1964. 1968, and 1972 the Democratic presidential vote was negatively correlated with the previous century and a quarter of voting for President at the state level. Certainly the early 1960s looked like the beginnings of a realignment. In 1960 the "rascals were thrown out" of office and unified control of government was established by the Democrats. By 1964 an incumbent President enjoyed a landslide victory and was perceived as dealing effectively with the nation's problems. For the only time in decades the Democratic Party was viewed briefly as better able to handle foreign affairs although, of course, Goldwater was responsible for much of this favorable image.

The potential Republican realignment of 1968-1972, announced by Phillips and others, encountered different problems. The Nixon administration, as well as Republican hopes for retaining the Presidency, may have floundered on Watergate. But it was the unwillingness of a Republican President to transfer his electoral success to party members running for Congress that failed to bring a permanent shift in the partisan division of the electorate. Without the unified control of government and with the unwillingness to share credit for policy- making with the Republican Party, the attitudinal changes in partisanship were unlikely to follow the very real changes in voting behavior that occurred in1968 and 1972.

With this perspective in mind, we can look again at the prospects for a traditional realignment in the latter half of the 1970s. The conditions for realignment certainly appear to be present. The widespread disaffection with Washington, the decline in trust in government, and the increase in cynicism about incumbent officials may not qualify as a crisis of major proportions, but the level of dissatisfaction seems strong enough to lead to a rejection of the occupant of the White House. The decay of the New Deal alignment has progressed far enough to assure a rather plentiful supply of non- aligned voters and potential voters to basically change the balance of power between the parties should one or the other capture their imagination.

The third element necessary for the successful completion of a realignment requires more detailed consideration. We have argued that the attitudinal change that is associated with realignment is a response to actions by government actions that are seen by the electorate as successfully dealing with societal problems. Given the structure of American government, the achievement of successful policy-making in all, probability requires that one party have unified
control of the three policy-making institutions. If this is the case, the possibility for realignment in the late 1970s are definitely asymmetrical. Given the continued viability of the New Deal coalition at levels below the presidential and considering the numerical impossibility of the Republicans gaining a majority in the U. S. Senate in 1976 only the Democratic Party would appear to have a chance of gaining unified control of government at the present time.

Whether the Democrats in power could manage to govern successfully is, of course another question. Burnham, in particular has implied that the real crisis of the American political system is the insolubility of current problems that, in effect, no government can successfully govern America. Obviously if this turns out to be true, we ought not to expect a traditional realignment to occur, but that might be the least of our worries.

In our view the real uncertainty about the possibility of realignment lies in the current disassociation of the presidency and Congress. If this disjuncture continues in policy-making it would make it unlikely that successful governing by a majority party would occur and, as importantly, unlikely that the credit for successful governing would be given to the party in power. Unless the electorate makes the connection between policy-making that they perceive to be successful and candidates running for office under a party label, the adoption of partisan loyalty necessary for a realignment will not occur. As we have argued above, the widespread availability of mass media can be an asset, not a hindrance, in facilitating the transmission of information on these points. Whether policy-making is successful and whether successes are advertised as party efforts rather than personal ones depend upon the behavior of party and governmental leadership. As Gerald Pomper argues, the American voter is a responsive voter; [51] the response will be determined in large party by the stimulus.

What are the prospects, then for 1976? We see two possibilities:

• Realignment favoring the Democrats, following a Democratic victory in the presidential election, continued control of the House and Senate and policy-making that is perceived as successful by the electorate; and
• Continued dealignment following either a Republican presidential victory (but without unified control of government) or a Democratic electoral victory but failure to appear to be solving national problems. [52]

Whatever happens will depend partially on what the electorate does in November, but even more on what the leadership does after that.