Liberalism, Primary Goods, and National Educational Standards
Charles HowellSyracuse University
The movement for national standards and testing has elicited both support and opposition
from across the political spectrum in the United States. The Reagan administration got
the ball rolling in 1983 with publication of A Nation at Risk; a decade later,
Clinton guided enabling legislation for Goals 2000 through a Democratic Congress.
Teachers' groups, civil rights groups, Republican and Democratic leaders, and Evangelical
Christian activists have all expressed reservations about the movement.1
A subset of the arguments for and against Goals 2000 focuses on its national scope. National as opposed to state-level standards are praised as progressive or denounced as an infringement of states' rights. The virtue of national standards is that they would raise the educational level of some states and thus reduce inequality among all states. Their vice is that they would impose the preferences of a national electorate on states that do not share them. As one advocate of national standards notes, citizens are worried that federal intrusion will corrupt their values.2 In due course I shall examine these arguments more closely. Before doing so, however, I will consider the problem of state versus national authority from the standpoint of political theory. Because of the parties' conflicting concerns about freedom and equality, I shall focus on liberalism, and in particular on John Rawls as its leading contemporary exponent.
At first glance it appears that Rawls would have little to say about the distribution of authority within a federal political system. He does not address the subject directly in Political Liberalism. The major structural elements of his theory - justice as fairness, the original position, the overlapping consensus - could apply equally to smaller political jurisdictions or to larger ones. The theory makes no provision for choosing between them.
Several writers who have commented on Rawls, however, do deal with regional policy differences that are in some respects similar to those displayed in the standards debate. Will Kymlicka discusses the attempts of aboriginal peoples in Canada to protect their way of life by restricting property rights and thus limiting commercial development.3 Charles Taylor deals with the attempts of the Quebecois to maintain French culture by limiting English-language signs and restricting access to English-language education.4 Both groups have used their political power to resist the homogenizing pressures of modern commercial society and to maintain a distinctive culture, which amounts to what Rawls calls a comprehensive moral doctrine - an organic conception of the good that orders priorities for those who adhere to it.
Granted, the parallel is not exact. Minority-dominated political jurisdictions in the United States are not now trying to opt out of constitutional provisions. But in principle any large group or coalition sharing basic values could influence state educational policy. That is the scenario implied by states' rights objections to national educational standards.
In Political Liberalism, Rawls does not discuss conflicts between different levels of democratic government, and it is not clear whether his model can accommodate them. In a unified state, groups that embrace different comprehensive doctrines create an "overlapping consensus," whereby the majority of citizens agree on political values which dictate fair terms of cooperation and guarantee the right of citizens to pursue different visions of the good life. The consensus is supposed to represent a set of shared values, not a compromise between competing interest groups. Thus, a group that joins the national consensus would share a set of values endorsed by the national electorate and would have no reason to resist them through state-level politics. Moreover, the consensus on basic institutions and values is supposed to be constructed without regard to citizens' other commitments. In the so-called "original position," the hypothetical meeting place in which political principles are formulated, citizens deliberate behind a "veil of ignorance" which enables them to disregard their personal interests and circumstances, including religious and cultural loyalties that might play a role in shaping regional policies.5
The basic Rawlsian picture, then, is supposed to eliminate self-dealing in the development of basic institutions and standards of justice. This might or might not rule out conflict between national and state-level educational standards based on group representation, depending on whether the dissent is extended to basic political values. What it does appear unequivocally to rule out is the dissent of the Quebecois and the aboriginal peoples, precisely because these involve basic constitutional questions - rights of property and free speech. This is Taylor's and Kymlicka's point: that the theory of Political Liberalism is incomplete because it cannot accommodate the legitimate aspirations of disadvantaged minorities.
Before accepting this indictment, however, I want to probe further. There are some points of flexibility in Rawls's theory that deserve exploration. One promising feature is that the veil of ignorance is not opaque. The "original position" is not literally a meeting place, but rather a "device of representation" - that is, a constraint on political rhetoric. The objective is not to forget one's own circumstances and interests, but rather to formulate political arguments without reference to them, since other citizens will not find self-pleading persuasive (PL, I:4, 24-25). With these qualifications in mind, the Quebecois and aboriginal peoples could satisfy Rawls by reformulating their dissent in neutral terms, arguing that justice as fairness entails preservation of cultures, quite independent of what culture one happens to belong to.
This approach, however, has two disadvantages. First, Rawls rules out preservation of culture as a goal of the liberal state. On his view, not all comprehensive doctrines are equally worthy of citizens' support. Those that do not get it are not entitled to preferential treatment by the state, for that would amount to sectarianism (PL, V:6, 197). Second, in both the Canadian situation and the states' rights scenario, we would need to ask why an ostensibly non-sectarian argument fails nationally but succeeds locally. The only relevant difference appears to be the proportion of voters belonging to the group making the argument. The disproportionate result suggests that the neutrality of the argument is only rhetorical.
There is, however, another feature of Rawls's account that seems to allow, at least to a limited extent, the pursuit of group values through political means. Citizens in a pluralist political order, according to Rawls, often suffer "strains of commitment," which he defines as "strains that arise in [a liberal] society between its requirements of justice and citizens' legitimate interests its just institutions allow." There is tension between "the political conception of justice and permissible comprehensive doctrines" (PL, I:3, 17-18). An interest in cultural continuity does seem to be one of these "legitimate interests" of citizens; the cultures qualify as comprehensive doctrines, and barring some radically illiberal turn, they do seem to be permissible ones. Citizens are not required to set aside these comprehensive doctrines in their conduct of political life; on the contrary, it is precisely the goal of liberal theory to afford citizens the means to pursue them.
To some unspecified extent, then, dissent of the kind exemplified in states'-rights arguments and Canadian constitutional debates may be permissible. If so, Rawls's theory may allow for regional differences in public policy, including educational policy. Further, since strains arise because of the comprehensive doctrines that citizens hold, policy differences are more likely to emerge where comprehensive doctrines are disproportionately represented in subnational political jurisdictions that enjoy some degree of autonomy.
Regional diversity, in short, is possible, and it may be permissible. But is it likely? The Canadian examples suggest that disproportionate regional representation is not just a matter of chance. A large group of like-minded citizens can enact public policies that change demographic patterns in ways that enhance their influence. Quebec's restrictions on education and commercial speech are reported already to have stimulated out-migration of Anglophones. In the aboriginal territories, restrictions on land holding by outsiders will tend to have the same effect - at any rate, that is one of the acknowledged purposes of their policy on land ownership and development. In both cases, the policies in question may also tend to discourage out-migration of members of the locally dominant group and in-migration of those representing non-favored cultures.
Given that regional political diversity is possible, likely, and to some limited extent permissible, and given its prominence in recent political developments, in North America and elsewhere, it might seem surprising that Rawls does not discuss it. Why, more generally, has modern American political philosophy shown so little interest in federal arrangements?
Perhaps a tentative answer can be found in the unappealing historical resonance of regionalism in the United States. Champions of states' rights supported slavery and Jim Crow laws. It is a mistake, however, to suppose on this account that oppressive policies are an essential feature of decentralized political systems. In the United States, that association depends on an historical accident: the dominance of Northern over Southern states from the mid-nineteenth century onward. If it had been the Northern states that were in the minority and were forced to secede to protect their political values, then our modern associations of regionalism and states' rights would undoubtedly be different.
Oppression, then, is not an essential feature of decentralized politics. But are there any positive reasons for liberal theorists to embrace federalism? Two reasons can be suggested. First, ruling out regional diversity leads to unacceptable consequences for the structure of the liberal state. For some public purposes, centralized control is unworkable. If authority in these areas is granted to local democratically-elected governmental bodies, then it would be unreasonable to expect citizens in a region not to act on their own views of justice, even where these diverge from the views of the national electorate. Freedom of movement makes it at least marginally more likely that differences will persist and intensify through self-selection. Equity and efficiency, presumably valued in liberal states for reasons having nothing to do with federalism, thus entail the possibility and perhaps the likelihood of regional diversity in public policy.
Second, the federal arrangement deserves the approval of liberal theorists because it promotes one of the fundamental goods they claim the liberal state offers. Rawls maintains that the liberal system, in permitting diversity of comprehensive doctrines, offers citizens a wider choice of ways to pursue a good life than would be available under a more restrictive regime. Critics of liberalism have charged to the contrary that commerce homogenizes culture, atomizing social groups and eroding community-based comprehensive doctrines. The examples of the aboriginal peoples and the Quebecois make clear that a federal political system offers tools with which to resist that erosion. Why should liberal theory refuse to make use of such a resource?
I have offered a sketch of an argument for including federalism as a component of liberal theory. I shall now develop the proposal further through an analysis of the standards debate. Suppose Rawls agrees to write federalism into his theory. He will need some sort of criterion for determining when states can make their own policy choices, and when the nation as a whole must decide. Let us examine the standards debate to see if any resources of the theory can assist in this enterprise.
Proponents of national standards generally argue that the choice should not be left to states for two kinds of reasons. First, there are a range of practical considerations involving efficiency, competitiveness, and economic development. I shall not discuss these points, since they can be settled without recourse to political theory.
A second set of arguments is relevant to political theory because it focuses on equity. The idea here is that without national standards, unacceptable differences would arise in the quality of public education. These differences are not the "savage inequalities" between urban and suburban schools within each state, but rather the somewhat subtler differences in average educational level from one state to another. One might argue that savage inequalities ought to take precedence in any reform effort, but this position is neutral regarding national or state-level standards, and thus does not bear on the issue at hand.
Whether or not liberalism requires that different states offer education of the same level and quality is not at all clear. Rawls has relatively little to say about education: His brief discussion focuses narrowly on the knowledge and skills citizens need to be self-supporting and to exercise political rights (PL, V:6, 199f). This somewhat sketchy account, however, can be filled out by consulting Rawls's analysis of primary goods.
The concept of primary goods, as Rawls presents it, is linked to "persons' needs as citizens," which in turn are the basis for the "kinds of claims it is appropriate to make when questions of political justice arise." It is designed to avoid two kinds of difficulties that arise in applying distributive principles: first, problems of interpersonal comparison associated with principles based on utility and welfare, which commonly involve a subjective element; and second, the problem of sectarianism when distributive policies favor one or more comprehensive doctrines at the expense of others. To this end, primary goods are defined as whatever distributable goods it is rational to want, whatever else one wants. Among the primary goods, Rawls lists "basic rights, liberties, and opportunities," along with "all-purpose means such as income and wealth" that can be used to further citizens' various interests and projects (PL, V:3, 178-90).
Based on this account, it seems reasonable to add education to Rawls's list of primary goods. It is linked in the requisite way to citizens' needs, and thus gives rise to equitable claims on social institutions. Conservatives and liberals agree that this is the sort of good that should be provided at public expense; they differ mainly as to the kinds of institutions in which it should be provided. There is broad agreement that education ought to be available to everyone, though of course what "available" means is contested. Our reasons for treating education in this way generally line up with the reasons that underwrite Rawls's primary goods: Education enables us to exercise our political rights and take advantage of economic opportunities, and education is an all-purpose means that can advance a wide variety of comprehensive doctrines.
To the extent that the standards-and-testing movement relies on arguments based on equity, it generally follows Rawls's line about how primary goods are to be distributed. The movement promises criteria that will allow citizens to make claims on social institutions based on their publicly recognized needs. The claims will be compelling, first because the thing claimed has been recognized as a primary good that should be available to every citizen, and second because significantly unequal test results would be prima facie evidence of unfair distribution.
Up to this point, Rawls's account of primary goods has been informative. We can see that the standards argument generally reproduces the line of reasoning he lays out. But to assert that education is a primary good does not settle the question of national versus state-level standards. We need to examine Rawls's account of primary goods more closely to see if it includes any relevant distributive principle.
There are two categories of primary goods - call them Type A and Type B. The first enable citizens to exercise their political rights and fulfill their political responsibilities - to vote, to participate in public life, to be self-supporting. A-type goods are introduced at the most abstract level of the Rawlsian model. They do not depend on any contingent facts about individual citizens or groups, but are rather conceptual necessities without which any liberal society would be unworkable. It is therefore inconceivable that they should vary from one political jurisdiction to another. Regionally based differences in their distribution would represent a fundamental defect in the political order. If education is an A-type good, equity arguments in favor of national standards are valid.
Type B, by contrast, covers goods that can be considered "all-purpose means" to advance the goals of various permissible comprehensive doctrines. Just how many comprehensive doctrines must be accommodated in this way, Rawls does not specify; but the idea is that a B-type good appeals widely enough to win public support and avoid sectarianism. Unlike the Type A list, the Type B list reflects contingent facts about the political jurisdiction that provides the goods. What items qualify will likely depend to some extent on which comprehensive doctrines are represented in the political body that is making the choice. Voters in Massachusetts and voters in Oregon may reach different conclusions about what should be included in the list. Further, justice does not dictate equal distribution, as it does for Type A goods. Rawls's favored distributive principle for Type B goods allows inequalities, provided they benefit the least advantaged representative citizen. Who this citizen is and how inequalities are to be arranged in order to benefit him or her will obviously vary with social and economic circumstances. Thus, in a heterogeneous federal union, justice does not require uniformity among geographic regions. If education is a B-type good, then regional variation in educational standards is not inequitable and does not justify imposition of national standards.
What sort of good is public education? We are committed to viewing it as a primary good, distributable equally to all citizens, but is it an A-type or a B-type good?
We can answer this question quite simply by looking at the kinds of programs schools offer. Some programs, like arithmetic, reading, and driver education, are provided to all students regardless of ability or interest. Others, like calculus, auto mechanics, or varsity basketball, are distributed on the basis of interest, ability, or occupational plans.
Clearly, these two sets of programs are treated differently. Programs of the first kind address basic needs common to every citizen: Everyone needs to be able to read, perform arithmetic operations, and follow the rules of the road. By contrast, not everyone needs to be able to replace the brakes on a car, sink a lay-up, or solve a partial differential equation. Programs of the second kind are "all-purpose means" - they advance specific projects that some citizens choose, not because they are essential to citizenship itself, but because they advance one or more determinate conceptions of the good. Programs of this sort, reflecting variable individual conceptions of the good life, thus mirror the array of comprehensive doctrines represented within the school district and/or state that provides them.
This analysis suggests the outline of a Rawlsian answer to our question about the appropriate geographic scope of educational standards, at least for programs that fall unambiguously into one of the two categories. Programs that address citizens' basic needs and are provided for everyone are A-type goods. Without them, citizens would be unable to participate in the life of the community. Regional variation in the quality of such programs is unacceptable. In this case, arguments for national standards based on considerations of equity are sound.
In the case of elective programs, however, this specific line of argument is untenable. Elective programs are clearly B-type goods. They reflect a local mix of views about the kinds of projects that are likely to advance the good life. Consensus on what programs to offer and at what level depends on contingent facts about who lives within the jurisdiction that is to provide the program. Regional variation is therefore permissible, perhaps even inevitable, given state-to-state differences in the representation of comprehensive doctrines. The equity-based argument for national standards is therefore invalid.
If this analysis is accurate, we should expect to see its themes reflected in the actual debate about standards. The mixed character of educational programs suggests that the case for national standards is not as straightforwardly compelling as proponents might wish. We should expect caution in advocacy, and caution is in fact what we find. In a 1994 conference on standards sponsored by the Brookings Institution, a number of participants discuss fears of federal control.6 Undersecretary of Education Marshall Smith, for example, cites "traditions of local control" and a belief that federal involvement would bring "the teaching of 'progressive values'" as two main sources of opposition to Goals 2000.7 Colorado Governor Roy Romer reports having predicted at the very outset "that the fight over standards was going to be one of federalism and, at state level, what was state and what was local."8 Theodore Sizer asks, "Why has a national or federal role been so quickly accepted...? Why is this centralization being accepted without searching discussion? ...Who are we to take charge? Who is the 'we?'"9
As might be expected, the proponents of national standards have various strategies for trying to assuage these fears of federal control. One approach is to emphasize the voluntary character of national standards. Marshall Smith, for example, points out that states are free to adopt them or not. The standards movement, he claims, should not be construed as a bid for federal "control" over an area traditionally within the purview of state authority. He emphasizes that the Clinton administration sees the proper "locus of reform" as the state rather than the national government, and consequently views standards as "useful exemplars" rather than directives. National leaders have "explicitly rejected the idea of common national standards that should be required of all the states."10 The position is politically astute: A Republican Congress had recently barred any link between federal education funding and state compliance with the standards.
A somewhat different approach is to defend national standards but to restrict their scope. Sizer takes this approach, urging that only three standards be adopted on a "mass basis": "resourceful reading, clear writing and speaking, and computational mathematics." This may look like a purely prudential argument, but note how it is aligned with Rawls's analysis. Instruction in reading, writing, and computation are A-type goods - goods that are required for citizenship and economic participation and must be distributed equally, regardless of local values. "Mass" standards here would not pre-empt legitimate local authority.
Roy Romer illustrates the point by describing a tour of a cable television facility. Seeing technicians working at consoles, he asks, "What does a fourth grader need to know and be able to do in math to get that job?" For Romer, this is the kind of question that can help to overcome resistance to standards, and it turns out to be precisely the approach one would recommend on the basis of Rawls's analysis. Romer clearly senses the political force of an argument based on primary goods of the first kind. "Can't we find a way," he asks, "that we can define what a youngster needs to know and be able to do in math in the fourth grade without getting caught up in a debate about outcome-based education and whether it will corrupt our values?" Romer maintains that "math and science need to be moved on first" because they are less likely to provoke this debate.11
Yet another approach is represented in the opportunity-to-learn (OTL) standards, the purpose of which is to assess "the resources, practices, and conditions" needed to "provide all students with an opportunity to learn the material in voluntary national content standards or state content standards."12 OTL standards are intended, in the words of Andrew Porter, to provide an "antidote to the potentially negative effects of high-stakes testing" on disadvantaged students.13 In effect, they expand the argument for national standards to cover "savage inequalities." The provision has gained Democratic support for standards and testing, but at the same time it has intensified Republican resistance. Using Rawls's analysis of primary goods, it is easy to see the logic of both positions. On the Democratic side, OTL standards focus attention on equity within each state, and thus gain support from the principle that all primary goods (not just A-type goods) ought to be distributed fairly within the jurisdiction that provides them. On the Republican side, OTL standards illegitimately shift the focus of the debate. Inequities within states, they would argue, are irrelevant to the debate about whether standards should be implemented at the state or national level.
As we have seen, the debate about national education standards rather accurately follows Rawls's analysis of the two kinds of primary goods. Education includes some A-type goods and some B-type goods, as well as others that do not fit neatly into either category. Its mixed character partly explains the divergence of views about national standards. It would be simplistic to reduce the whole debate to a difference in focus, with one side concentrating on mandatory universal programs and the other on voluntary programs suited to specialized interests and tastes; the comments of Romer and Sizer, however, clearly demonstrate that this is one fault line in the controversy. As we might expect, worries about equity loom large throughout the discussion: The OTL standards demonstrate a concern that in our zeal for equity between states, we not lose sight of the glaring disparities of educational opportunity within states. It is not clear that national standards would address these inequities any more effectively than state-level standards; it is understandable, however, that proponents would try to find grounds to claim that they would do so.
Analysis of the standards debate also yields some tentative conclusions about the role of regional diversity in liberal theory. A liberal order ought, in principle, to tolerate variations in policy among political subdivisions. Yet there are limits to the level and kind of regional differences that can be permitted. Liberals cannot allow differences in the levels of A-type primary goods either between jurisdictions or within jurisdictions. By contrast, B-type goods must be distributed fairly within the jurisdiction that provides them but need not be equalized between jurisdictions. Liberal theory provides resources to distinguish between the two types of goods, which is one reason it can usefully be applied to the debate over national educational standards and to other disputed issues of regional autonomy.
For a response to this
essay, see Sison.
MY THANKS to Nisha Gupta and Emily Robertson, both of Syracuse University, for their helpful comments on an earlier version of this essay.
1. For recent arguments on both sides, see Diane Ravitch, ed., Debating the Future of American Education: Do We Need National Standards and Assessments? (Washington D.C.: The Brookings Institution, 1995); John O'Neil et al., "The Challenge of High Standards," Educational Leadership 50, no. 5 (1993), 5-89; and Larry Cuban, "Four Stories about National Goals for American Education," Phi Delta Kappan 71 (1990), 265-71.
2. Roy Romer, "Explaining Standards to the Public," in Ravitch, Debating the Future, 66.
3. Will Kymlicka, Liberalism, Community and Culture (Oxford: Clarendon, 1989), chap. 7.
4. Charles Taylor, "The Politics of Recognition," in Multiculturalism and "The Politics of Recognition," ed. Amy Gutmann (Princeton, NJ: Princeton University Press, 1992), 52-60.
5. John Rawls, Political Liberalism (New York: Columbia University Press, 1993), I:4, 22-23. This book will be cited as PL in the text with page numbers for all subsequent references.
6. Published as Diane Ravitch, ed., Debating the Future of American Education: Do We Need National Standards and Assessments (Washington, DC: The Brookings Institution, 1995).
7. Marshall Smith, "Education Reform in America's Public Schools: The Clinton Agenda," in Ravitch, Debating the Future, 22.
8. Romer, "Explaining Standards," 66.
9. Theodore Sizer, "Will National Standards and Assessments Make a Difference?" in Ravitch, Debating the Future, 35.
10. Smith, "Educational Reform," 13.
11. Romer, "Explaining Standards," 67-69.
12. H. Report. 103-446, Congressional Record, 3/21/94, p. H1626.
13. Andrew Porter, "The Uses and Misuses of Opportunity-to-Learn Standards," in Ravitch, Debating the Future, 40.