8

Designing an Amendment Process

Sanford Levinson

Imagine two written constitutions.1 One sets out a standard series of polit-
ical structures and governmental empowerments and limitations; it con-
cludes with a clause saying: “Anything in this constitution can be changed
by the passage of ordinary legislation as spelled out in this constitution.”
Were this “parliamentary sovereignty” model – found, for example, in the
Austrian Constitution2 – present in the United States Constitution, then
constitutional amendments could come about by agreement of majorities
in both houses of Congress and assent by the president or by two-thirds
vote in each house overriding a presidential veto. Our second constitution
comes to a radically different conclusion: “And the Articles of this con-
federation shall be inviolably observed by every state . . . ; nor shall any
alteration at any time hereafter be made in any of them; unless such alter-
ation be agreed to in a congress of the united states, and be afterwards
confirmed by the legislatures of every state.”3 Not only does the second
constitution require assent by a different institutional layer from that of
the national political assembly, in this case, the constituent states of the
union; it also requires that this latter act of assent be unanimous. By def-
inition, as with Poland’s (in)famous liberum veto, this allows one holdout
state to countermand the desire of every other state (and, presumably, the
I am grateful to Jim Fleming, Doug Laycock, Hans Linde, Scot Powe, and Eugene Volokh
for comments on earlier drafts of this essay. A version of this essay was published as
“The Political Implications of Amending Clauses,” Constitutional Commentary 13 (1996):
107–23.
1
The adjective is important, for it is obvious that all political systems can be said to have
“constitutions” in the sense of constitutive conventions of practice and tradition. Yet most
– all but seven current states, in fact – have chosen to have written constitutions, and this
chapter concerns only such systems and some of the problems attached to “putting it in
writing.”
2
This appears to be the basic rule adopted in the Austrian Constitution, the easiest of all
national constitutions to amend. See Lutz 1995, 260.
3
Articles of Confederation, Article XIII, in Kurland and Lerner 1987, 26.

271
272 Sanford Levinson

national legislature) for constitutional change. Both of these examples are

taken from real political life, even though the latter constitution, the U.S.
Articles of Confederation, lasted only six years. Theoretically, though, one
can imagine an even stricter constitution, such as John Locke’s draft con-
stitution for the Carolinas in which he wrote that “[T]hese fundamental
constitutions shall be and remain the sacred and unalterable form and rule
of government . . . forever.”4 Such delusions of unamendable grandeur
are, however, found in no present constitution. Still, the scheme established
by the Articles, or even the substitute structure found in Article V of the
Constitution that replaced the Articles, presents a dramatic enough con-
trast to such relaxed structures as that found in Austria to generate more
than enough food for thought.
As to Austria, one might be tempted to say that the polity described
really does not have a “constitution” at all, at least if a “constitution” is
in some ways supposed to stand “above” and in some sense even “outside”
the everyday system of ordinary political decision making. Thus Mark
Tushnet has written that “[p]erhaps some degree of institutional stability
is required for a system to warrant the name constitutional, which suggests
that it should not be too easy to amend all of a constitution’s provisions,
or perhaps any of its basic institutional prescriptions” (Tushnet 1995, 223,
225). Lynn Baker has similarly written that a “constitution is, after all, a
fundamental law, which we therefore expect – and want – to be more per-
manent than the ever-changing, ordinary, statutory law” (Baker 1995, 143,
146). From this perspective, then, Austria may have less of a “constitu-
tion,” whatever its title, and more of a functional initiating statute that
gets the political game underway but is otherwise thoroughly “inside” the
ordinary political order that it establishes.
It is “inside” in a double sense: first, its mechanism for change differs
not at all from the standard-form politics of legislation; and, second, only
those already inside the political system – that is, elected officials – par-
ticipate in the decision-making process. To be sure, even this constitution
might in fact be difficult to change insofar as ordinary legislation is itself
difficult to pass, as is the case in the notoriously complex system estab-
lished by the United States Constitution, with its bicameral legislature and
independent role for the president (not to mention the political implica-
tions of federalism and consequent hindrances to the establishment of a
truly united national party system). But one can easily imagine alterna-
tives to our given political structure, which has, indeed, been adopted by
no other country in the world. One would, for example, predict both more
legislation and more amendment in unicameral than in bicameral systems;
similarly, one assumes that passage would be easier if the president (or

4
From The Fundamental Constitutions of Carolinas, sec. 120 (drafted by John Locke).
 Designing an Amendment Process 273

monarch) played no role, especially if the formal system, as in the United

States or France, for example, tolerates the possibility of an executive and
legislature controlled by different political parties.
Perhaps even this “minimalist” constitution would have some special
prestige because of the identity of its authors that would lead to a certain
cultural hesitation to amend it. There is no logical reason why it could not
receive the “veneration” thought by James Madison to be so important to
the constitutional enterprise (Levinson 1990, 2443), though, as an empir-
ical matter, it may be that such “veneration” is a function of the difficulty
of amendment. Cognitive dissonance theory might predict, for example,
that one will tend to adjust and even find merit in structures that are in
fact difficult to change, and the absence of difficulty might lead to a
reduced level of effective commitment.
In any event, this constitution presents no special obstacles to its own
change. A bicameral system could, for example, become unicameral if
both houses agreed to the change; the presidency could simply be abol-
ished and replaced with a prime minister drawn from the legislature; or
life tenure for judges could be eliminated. Should we discover, after a suit-
able passage of time, that this constitution had remained unchanged we
would, I think, be entitled to offer the lack of change as evidence of very
high satisfaction, at least on the part of ruling elites, in regard to the
original scheme.
The second constitution is, of course, at the opposite extreme. Its
unanimity condition makes significant change highly unlikely, at least
empirically, outflanked only by the Lockean constitution, which, by
making amendment theoretically impossible, means that most modest
change – at least where “change” is defined as formal amendment – would
presumably require its “overthrow.” Evidence of lack of formal change
could be plausibly submitted only for the proposition that dissatisfaction
had not risen to such a fever pitch that regime overthrow was found to be
preferable to continuation of the system established by the constitution.
It would, however, be foolhardy to view nonamendment as any more pos-
itive an endorsement or reaffirmation than that. (Any such claims, indeed,
should be subject to the same discount as given statements by leaders in
dictatorial one-party states that the absence of a competing party denotes
unproblematic acceptance by the population at large.)
But a system that makes formal change exceedingly difficult (even if not
theoretically impossible) forces us to confront a subject of great theoreti-
cal importance: how precisely do we identify constitutional amendments?
As Stephen Griffin, among others, has well pointed out, an ever present
alternative to formal amendment is informal amendment.5 Least likely in

5
See Griffin 1995b, 37–62. See also Lutz’s (1995) superb essay.
274 Sanford Levinson

a dynamic political system is no amendment, whether or not these changes

take canonical textual form. Noam Zohar, an Israeli philosopher, has ana-
lyzed the theoretical problem of amendment within Jewish halacha, which
is absent any formal process of amendment. After all, halacha is based on
divine revelation, and it is untenable, both practically and perhaps even
theologically, to suppose that God would have committed errors subject
to correction by fallible humans. Yet Zohar points out that there is cer-
tainly a great deal of significant change in Jewish law, though it is almost
never described as “amendment” (Zohar 1995).
It is thus problematic to identify “amendment” only as formal textual
additions (or subtractions). Only the most atheoretical person, I believe,
can confidently assert that the U.S. Constitution has exactly twenty-seven
amendments.6 Indeed, one function of such almost literally thoughtless
confidence is to blind us to the reality of non–Article V amendments
within our own constitutional system.7 In that sense, emphasis on Article
V as the source of all amendments is truly ideological, reinforcing a certain
kind of political understanding and promoting a false consciousness about
our political reality.
Bruce Ackerman has been the most notable proponent of the presence
of non–Article V (and nontextual) amendments within what any well-
trained lawyer would today identify as “the United States Constitution”
(see, e.g., Ackerman 1991; 1998). Although he earlier focused on changes
in the “domestic” power of the national government surrounding the
post–Civil War period and the New Deal, most recently he has turned his
attention to foreign affairs. He thus argues that the authorization of the
North American Free Trade Agreement and the General Agreement on
Trade and Tariffs by majorities of both houses of Congress, rather than
by two-thirds of the Senate, where the 1787 Constitution reposed the
power to ratify treaties, is evidence of a profound “structural amendment”
provoked by World War II and its aftermath.8
Still, even if one believes that it is foolish to assert that only formal
changes count as amendments, it would be perverse to reject the impor-
tance of formal amending structures or of the formal additions to, and
6
Professor James Fleming has suggested to me that a theoretical person certainly can suggest
this, on the ground that a “bad theory” is, “nonetheless, a theory.” There may be some-
thing to this. It is the case, though, that most people who offer the “confident” assertion
suggested in the text are no theorists at all: they are simply repeating a conventional wisdom
that has been insufficiently theorized. At best, they are mindlessly repeating someone else’s
bad theory.
7
See Levinson 1995. For our purposes, I put to one side the difficulties in counting the
Eighteenth and Twenty-first Amendments and the question of whether the Twenty-seventh
Amendment was properly ratified. On this latter issue, see Levinson 1994, 101.
8
See Ackerman and Golove 1995, 799. Harvard’s Laurence Tribe (1995, 1221) scathingly
attacks the Ackerman-Golove analysis, and Golove 1998, 1791 replies.
 Designing an Amendment Process 275

subtractions from, constitutional text that such structures uniquely allow.

At the very least, any legal culture, like that of the United States, that
includes textual argument among the array of lawyerly rhetorics9 must
appreciate the advantage of being able to refer to specific text rather than
have to make what some analysts would dismiss as appeals to unwritten
general traditions or conventions. Opponents of President George Bush’s
policy in the Persian Gulf were certainly helped by the presence of the
declaration-of-war clause in Article I of the Constitution, and they would,
concomitantly, have been significantly hindered had the “power to declare
war” been placed in Article II.
Needless to say, to note the importance of text does not require that
one believe either that texts are self-interpreting or that textual argument
will necessarily prevail over, say, doctrinal or prudential argument. Still, it
is impossible to believe that anyone in our legal culture (or others with
written constitutions) believes that text is truly irrelevant. Otherwise why
would one care whether, for example, a balanced budget amendment or
any other proposal was in fact adopted? Similarly, the importance of text
in our constitutional tradition presumably explains why supporters of
women’s rights committed themselves to the equal rights amendment
(ERA), even though few were willing to explain precisely what rights it
would add beyond those already protected by (their version of ) the
Fourteenth Amendment. Sophisticated poststructuralist critiques of the
sufficiency of text do not in the least serve to negate the practical impor-
tance of texts within everyday political life.
Anyone thinking about constitutional design – consider, for example,
someone flying to Eastern Europe to offer advice about constitutional
design in Eastern Europe and elsewhere – must therefore address proce-
dures for amendment every bit as much as the standard topics of institu-
tional design.10 Indeed, few topics are more important, whether as a
theoretical or practical matter, than amendment clauses.
I return to our two modal constitutions that establish the two ends of
a spectrum. What might lead to adoption of one or the other of these two
admitted extremes? To adopt the title of a book that I have edited,
Responding to Imperfection, let me suggest that the authors of the first
constitution would be maximally modest as to their own capabilities and
maximally aware of their capacity for imperfect judgments. That is, they
would acknowledge the relatively high probability that their notions of
proper government, whether one is referring to institutional design or the

9
See Bobbitt 1981; 1991, which develop the notion of American constitutional law being
constituted by specific rhetorical modalities, one of which is reference to the text of the
Constitution.
10
For an excellent discussion, see Holmes and Sunstein 1995.
276 Sanford Levinson

authorization or prohibition of specific powers, are in fact subject to error,

given the complexities of political life.
Consider the statement of Virginia’s George Mason on June 11, 1787,
as he opened the debate in Philadelphia on what amending procedure the
delegates should adopt for the Constitution taking form that summer.
“[T]he plan now to be formed will certainly be defective,” he told his fellow
delegates. “Amendments therefore will be necessary, and it will be better
to provide for them, in an easy, regular and Constitutional way than to
trust to chance and violence.”11 Indeed, I drew the title Responding to
Imperfection from a comment by Mason’s fellow Virginian George
Washington to his nephew, Bushrod Washington: “The warmest friends
and the best supporters the Constitution has do not contend that it is free
from imperfections, but they found them unavoidable and are sensible
if evil is likely to arise there from, the remedy must come hereafter”
(Kammen 1986, 83).
Those who accentuate the possibility of imperfection and adopt what
might be termed a “statutory” mode of amendment accept the twin like-
lihoods as well, first, that future generations are likely to recognize the
existence of these imperfections and, second, that these generations will be
sagacious enough to correct them. In turn these successor generations will
presumably also be wise enough to realize that they, too, will be imperfect
in their political judgments and thus leave it open to their successors to
engage in the same presumptively progressive response to imperfections,
and so on ad infinitum. Thus all successor generations would presumably
feel empowered to change the constitutional rules whenever that seemed
to be a good idea. Needless to say, such confidence was not expressed by
Washington or his colleagues in Philadelphia.
The second constitution would presumably be authored by persons who
had both an inordinate confidence in their own political wisdom coupled
with a perhaps equally inordinate lack of confidence in successor genera-
tions. Locke’s use of the word “sacred” may be telling, moreover, insofar
as it suggests a self-perception by the framers of themselves as (at least)
demigods, whose work is entitled to the same awesome respect as that
given “real” gods.
Although one might think that the first constitution is maximally open
to change, that is not the case. One can imagine a third constitution that
concludes not only with the postulated sentence but, in addition, states
that “this constitution can also be amended by majoritarian popular
referendum on initiatives propounded by 5 percent of the population.”
What political presuppositions might account for this addition?

11
Farrand 1937, 1:202–3 (emphases added).
 Designing an Amendment Process 277

One obvious answer is that the framers are familiar with contemporary
public choice theory and its concern about agency costs. That is, to limit
constitutional amendment only to what gains the assent of those already
ensconced within governmental institutions is, almost by definition, to
lessen the possibility that the occupants of political office will be amenable
to proposals that would significantly affect their interests or, in the lan-
guage of public choice, diminish the possibility of engaging in successful
rent seeking for themselves and their supporters (Boudreaux and Pritchard
1993, 111).
One does not have to be a modern to have this insight. Again one can
turn to George Mason, who vigorously opposed initial drafts of Article V
that placed exclusive power to initiate amendments in the hands of
Congress. “As the proposing of amendments is . . . to depend . . . ulti-
mately, on Congress, no amendments of the proper kind would ever be
obtained by the people, if the Government should become oppressive, as
he verily believed would be the case” (Kurland and Lerner 1987, 577). It
was therefore vital to create alternatives to Congress as the possible pro-
genitors of constitutional amendment. Mason’s mirror image was New
York’s Alexander Hamilton, who trusted the states no more than Mason
trusted Congress. After all, said this highest of Federalists, “The State
Legislatures will not apply for alterations but with a view to increase their
own powers” (ibid.) and, presumably, weaken those of the national gov-
ernment. The solution found in Article V seems to address both of their
concerns: Hamilton won the right of Congress to propose amendments,
but Mason won the right of states to initiate a new constitutional con-
vention upon petition of two-thirds of the states.
Needless to say, this latter mode of amendment has never in fact been
attempted. Most mainstream analysts seem frightened to death by the very
possibility, though I confess I do not share this view. In this regard, it would
be especially helpful to recognize that the United States in fact includes
fifty-one constitutions within its territory and to study the propensity of
states to subject their constitutions to the more-or-less frequent scrutiny of
constitutional conventions. Some analysts have recently argued that state
constitutions are not “real” constitutions precisely because they are so little
blessed by the “veneration” visited upon the national constitution, but this
obviously begs the question as to how precisely we identify something as a
constitution (see, e.g., Gardner 1992, 761; 1993, 927).
It is also worth mentioning in this context the powerful argument of
Akhil Reed Amar (1995) that the American idea of popular sovereignty
requires the possibility that the U.S. Constitution be amendable as well by
a majority of voters in a popular referendum, in addition to the super-
majoritarian procedures set out in Article V. Even though, as a practical
278 Sanford Levinson

matter, Amar’s method is not only untried but also, for most Americans,
I suspect unthinkable, his argument is noteworthy insofar as it is built on
seeing the role of Article V as protecting the people en masse against the
corruption of their political agents rather than necessarily endorsing the
political status quo itself. Thus for Amar the Article V requirement of
extraordinary majorities in both houses of Congress and state legislatures
is far less a commitment to the perfection of the existing constitutional
scheme than an expression of the deep mistrust of political actors and fear
that too-easy methods of change would provide simply a royal road to rent
seeking.
Even the Article V possibility of convention-by-call-of-the-states does
not entirely overcome the agency problem, for one might imagine circum-
stances when all political officials, regardless of the level of office held,
would have interests in common against the civilian populace. Thus
perhaps the western states – most (in)famously California – teach an essen-
tial lesson by offering the possibility of amendment by direct initiative and
referendum of the sovereign people themselves, freed of any requirement
to beseech political intermediaries for their substantive approval (though
these intermediaries must still presumably agree to place the measures on
some official ballot and then to count the votes).12 For what it is worth, it
is not only the fevered states of the American West who have rejected the
monopolization of the amendment process by state officials. Switzerland,
for many a symbol of boring stability, has been described as “the only
nation in the world where political life truly revolves around the
referendum. . . . the great political moments of modern Switzerland have
occurred not in the following of bold statesmen but in the national debates
that have drawn the masses to the polls to decide their country’s future”
(Kobach 1994, 98).
What must one believe to endorse direct initiative and referendum as a
mechanism of constitutional amendment? One possibility is that one
possess an unusually high, some might say paranoid, mistrust even of
popularly elected agents, who will presumably be corrupted once they take
their seats in Washington, Albany, or Budapest, coupled with an equally
remarkable, neo-Rousseauian faith in an uncorrupted people. To adopt an
initiative-and-referendum system like California’s is, in addition, to reject
the importance placed by Madison on representative government and on
multiple filters between the mass of the electorate and ultimate political
outcomes. These filters, of course, range from the ostensibly virtuous char-
acters of those likely to be elected to office to the encouragement of certain
kinds of deliberation by the rules or practices of our political institutions.

12
See, e.g., California Constitution, Art. II, sec. 8; Arizona Constitution, Art. IV, pt. 1,
sec. 1; Montana Constitution, Art. XIV, sec. 9.
 Designing an Amendment Process 279

Indeed, it is just this escape from the filtration of republican deliberation

that has led Hans Linde to suggest that the practice of initiative and
referendum in the western states at least on occasion violates the “repub-
lican form of government” clause in Article IV of the Constitution (Linde
1993, 19).
One response to such arguments is that presented by UCLA professor
Eugene Volokh, himself a participant in the political wars over California
Proposition 209, by which the electorate prohibited affirmative action in
that state. “The legislative process,” he argues, will often allow

minorities who care deeply about a subject to overcome the will of majorities who
care less deeply. This might, for instance, be why the anti–affirmative action cam-
paigns are being waged more on the initiative front than the legislative front: If
the minority that is pro–affirmative action is very passionate about its support, and
the majority that is anti is more lukewarm in its opposition, then a legislator – a
repeat player – may prefer to vote for affirmative action and mildly alienate the
majority than vote against and strongly alienate the minority. And, of course, the
minority will be able to cut legislative deals that it can’t with the initiative.
This feature of the legislative process may often be a good thing, but I can see
someone arguing that it isn’t always – that there ought to be a mechanism for
majority sentiments, even those most weakly held, to prevail over strongly held
minority sentiments.13

There is certainly something to Volokh’s argument, which, of course, raises

the general issue of the relevance of preference intensity within a demo-
cratic theory committed to the formal equality of every voter or, more to
the point in this instance, of every representative, even if they owe their
actual elections to a particular single-issue constituency that is, in fact, not
at all typical of the actual distribution of public opinion. Of course, one
criticism of referenda is that they may give great (and presumptively un-
deserved) advantage to highly intense, well-organized, and well-financed
groups as against the ordinary mass of the polity. If one views referenda
more as Madisonian nightmares than Jeffersonian dreams, the question is
whether the nightmare would be caused by the theory of initiative and
referendum or, rather, by the empirical circumstances of its practice within
the United States (as opposed, say, to Switzerland).
My impression – for I confess that I have not sufficiently studied
the matter14 – is that popular referenda have supplanted legislative
decision making as the preferred method of constitutional change
in California and, perhaps, in others of the western states. Imagine if the
U.S. Constitution were like the California Constitution and allowed

13
E-mail from Eugene Volokh to Sanford Levinson, December 6, 1995.
14
See, though, Cronin 1989.
280 Sanford Levinson

amendment by popular initiative and referendum in addition to the pro-

cedures set out by Article V. What would the likelihood be that proponents
of balanced budget or term limits amendments – or, indeed, of any other
amendment that profoundly changed the political status quo – would
invest in electoral politics of the more-or-less Madisonian variety (i.e.,
focusing on representative government and gaining the required superma-
jorities at both national and state levels of politics) rather than in what
might be called plebiscitarian politics (i.e., focusing on direct, unmediated
democracy free of the filters provided by representation or of the partic-
ular kinds of deliberation fostered by institutional political processes)?
Given the immense difficulty of amendment through Article V procedures
– to win the “amendment game,” for example, one must win the approval
of two national legislative bodies plus no fewer than seventy-five state
bodies (assuming that one of them is Nebraska’s unicameral legislature) –
is it not readily predictable that rational agents would focus their resources
on initiatives and referenda? For better or worse, recent displays of con-
gressional consideration of balanced budget amendments might become
a thing of the past, as “We the People” instead were summoned to decide
about the wisdom of constitutionalizing a particular theory of political
economy.
That may as yet be only a dim specter on the horizon, but let me suggest
that one highly thinkable outcome of the frustration over failure to gain
the final vote needed to propose the balanced budget amendment – or
of the earlier failure, in the 1970s, to gain ratification of the ERA in spite
of the fact that a majority of the states representing a majority of the
population had given it their assent – will be the amendment of Article V
itself. The juxtaposition of the ERA with the balanced budget amendment
should illustrate, incidentally, that dissatisfaction with the requirements of
Article V does not necessarily assume a particular political coloration.
What is protected by Article V is the status quo, whether liberal or
conservative.
Stephen Griffin has recently suggested that the worst feature of the
current U.S. Constitution is indeed Article V, precisely because it makes
formal change so inordinately difficult. Amendatory change is often
masked as “constitutional interpretation,” at immense costs in intellectual
cogency or candor. This also gives to judges both responsibility and power
that one might well think they are unsuited for, yet another political impli-
cation of such a rigorous amending clause. Even worse, perhaps, is that
highly desirable change is stifled because one cannot in fact figure out an
alternative to use in place of the formal procedures.
Griffin made his comment in a symposium asking only for identifica-
tion of the worst (or “stupidest”) aspect of the Constitution (Griffin
1995a); it did not ask for positive recommendations as to how to cure the
 Designing an Amendment Process 281

suggested defects. But that is clearly the next step for anyone who does
accept the view that the United States is not well served by its amending
procedure. One might even make the radical suggestion that one might
find desirable alternatives through study of the constitutions of the
American states, not to mention foreign constitutions, none of which have
such difficult schemes of amendment. Actually, the preceding sentence is
incorrect in one small respect. One country did have a more complicated
scheme of amendment than that bequeathed us by the Philadelphia
Convention of 1787. That country was Yugoslavia.15
In any event, I think it is worthwhile to take the next step and to imagine
the changes one might advocate for Article V, as well as their political con-
sequences. One set of changes might simply involve greater specification
of the answers to a number of important conundrums suggested by our
actual political history. The easiest example concerns the right of states
to rescind their ratification of a proposed amendment, at least prior to a
declaration by the National Archivist that sufficient numbers of ratifica-
tions have been received to make the proposed amendment “part of this
Constitution.”16 Because the ERA never gained the assent of sufficient
numbers of states under any theory, we never had to face the question of
the constitutionality of Idaho’s attempted rescission.
My own view is that a state does indeed have a right to change its mind.
Imagine that the balanced budget amendment had been successfully pro-
posed by the 1995 Congress and that several states had rushed to ratify it.
(Indeed, New Jersey attempted to ratify such an amendment even before
proposal, so eager were the legislators to have New Jersey become the first
state to endorse it.) Imagine also that, like the ERA, it ran into some
trouble, and that the electorate began to realize that a constitutional
amendment was unnecessary in order to achieve the goal (and might have
significant side costs as well). Does one really want to argue that a state

15
See Lutz 1995, 261. Lutz developed an “Index of Difficulty” based on the complexities of
amending procedures. Yugoslavia was highest, at 5.60. The United States follows with 5.10.
Next come Switzerland and Venezuela with 4.5. Austria and Sweden have the “easiest”
constitutions to amend, with rates, respectively, of 0.80 and 1.00. This does not, obviously,
include those few countries that do not have written constitutions, like Great Britain, New
Zealand, or Israel.
It would be absurd to argue that the current difficulties in the former Yugoslavia are
due to the formal difficulty of constitutional amendment. Can one, however, be entirely
confident, in the absence of detailed study, that it played no role in making necessary polit-
ical changes, following Tito’s death, simply too difficult to realize through ordinary polit-
ical processes? Even if formal constitutional process explains no more than 2 percent of
the variance in accounting for contemporary South Balkan politics, is that not still a
damning indictment, unless the rigidity in fact helped to purchase forty years of relative
ethnic peace during the Tito years?
16
See Article V. On rescission, see Rees 1977, 896.
282 Sanford Levinson

should not be entitled, as a constitutional matter, to change its collective

mind on a matter of such profound import? After all, a vote not to ratify
does not prevent a future legislature from deciding to endorse a proposed
amendment. Why does the option to switch work in only one direction?
Whatever one’s views on the merits, does one really want to leave this
hanging as an open question, to be decided either by Congress or the
Supreme Court as one’s jurisprudence dictates?17 The better course, it
seems to me, is to come to some decision, while there is no amendment
pending, and to codify it in the Constitution itself.
The most spectacular failures of Article V to provide any genuine guid-
ance come in regard to the convention that could be called on petition of
thirty-four states. As noted, this is as yet only a theoretical problem, but
so long as we wish to leave open this possibility of a convention, it cannot
be wise to leave open as well such fundamental questions as whether states
would vote by unit, as in Philadelphia, or by individual delegates or
whether the agenda of the convention could be limited by the states or
Congress, on the one (actually two) hand, or only by choice of the “sov-
ereign” convention itself, on the other. As should be obvious, once one gets
started, a “corrective” Article V devoted only to filling in some of the
blanks – and not even touching, for example, the basic structure of requir-
ing supermajority votes in Congress and ratification by three-quarters of
the states – could well turn out to be almost as long as the existing
Constitution itself.
But surely the most basic disputes would concern the basic structure
of Article V. Griffin’s complaint is not merely that Article V has some
stunning lacunae in it, but rather that it clearly disserves the polity. So
what direction might “reform” of Article V take?
Would anyone, for example, suggest a process by which national-level
officials alone could amend the Constitution? Consider, for example, a
proposal to allow amendment by vote of two-thirds of each house of
Congress and presidential approval, or by vote of three-quarters of each
house (in order to prevent, for example, an absolute presidential veto of
modification of presidential power itself). I assume that in fact few of us
would be tempted by such a proposal, and I assume that what would
animate most of us in our opposition would indeed be the lack of any
formal state role and/or the lack of popular participation as through a
referendum. But what underlies that mistrust of a national power would,
I think, be some version of the agency argument. Otherwise, if one’s objec-
tion to the present Article V is its role in preventing vitally needed changes,

17
See, e.g., Dellinger 1983, 386, arguing for judicial review of such questions; Tribe
1986, 433, arguing that Congress has “plenary power” to decide on the legitimacy of
rescissions.
 Designing an Amendment Process 283

it is hard to see why one would not endorse simply eliminating the par-
ticipaton of the states.
If one does endorse continued participation by states, that simply forces
one to confront the question of how many states should have to endorse
an amendment before it is accepted as part of the Constitution? Would
anyone seriously defend, as we enter a new millennium, the present system
that in essence allows one house of thirteen states to block the desires of
the remaining public? That is, the “amendment game” gives victory to
those who can win thirteen such houses against the side that prevails in
seventy-five houses in thirty-eight states other than Nebraska. Quite
frankly, I can think of no defense for the present rules of this particular
game unless one is committed simply to making it as difficult as practi-
cally possible to engage in formal amendment.
Does an alternative number suggest itself ? One possibility, obviously,
is a simple majority of states. The major problem with that is the theo-
retical possibility that such a majority could be gathered by aggregating
states that themselves contain substantially less than a majority of the
American public. Is it adequate to overcome this fear to point out that any
amendment must first gain the support of two-thirds of the House of
Representatives, which is, of course, apportioned on the basis of popula-
tion? One might respond, of course, that the apportionment is scarcely
independent of political factors, ranging from incumbency to race, and
one might not really believe that the support of two-thirds of the repre-
sentatives necessarily translates even into majority support by the public
in general.
Unless one is a “high federalist” in a distinctly modern sense – that is,
someone who really does accept the metaphysical integrity of Idaho qua
Idaho, and so on – it seems hard to argue that actual population ought
not play some role in the ratification process. One might well, therefore,
adopt a version of the Australian rule, which is to require a majority of
the states and that this majority include a majority of the national popu-
lation. Note well, though, that this allows for the possibility that even if a
minority of states containing a majority of the national population sup-
ported an amendment that, by stipulation, has gained the approval of
Congress or a national convention, the amendment would nevertheless
remain unratified.
At this point, then, we have to ask ourselves why we would care that
even a majority of states ratify an amendment. The answer, presumably,
would be to return to some of the original debates of 1787–8, where one
finds rampant mistrust on the part of small states in regard to the
potential conduct of large states. Still, we might ask ourselves why the
organization of the Senate, based as it is on formal state equality, does
not offer (more than) enough protection to “states qua states.” How
284 Sanford Levinson

much protection are Wyoming, North Dakota, Alaska, and Rhode Island
entitled to against the wishes of, say, California, Texas, Florida, and
Michigan? Is it only the fact that I am a Texan (of sorts) that makes me
unsympathetic to continuing the remarkable power given small states
within our political system?
If one remains justifiably suspicious of exclusive national amendment,
but is equally suspicious of maintaining the role of states in ratification,
is an acceptable alternative the national referendum on amendments first
proposed by Congress or by a national convention? Resistance to this
notion could be based on fear by, say, people living in Mountain and
Upper Midwestern states that they would simply be swamped by their
fellow citizens who have chosen life in the mega-city. But, of course, more
fundamental objections would be based on some of the earlier-expressed
fears either about the corrupting role of money in politics, including
national referendum campaigns, or about the inability of ordinary citizens
to think reflectively about the kinds of issues appropriate for constitutional
placement.
I have, up to now, been assuming the necessity of two-thirds vote in
each house of Congress. But why maintain the supermajority requirement
at all, especially if one maintains a sufficiently strong role for states in the
ratification process to guarantee some kind of barrier against a “rush to
judgment”? Or, why not require a congressional supermajority only if the
president formally opposes the proposed amendment? Otherwise, I’d be
inclined to take my chances with congressional majorities plus presiden-
tial approval plus ratification by sufficient states to comprise a majority of
the population or popular ratification.
I have also been assuming that a constitution establishes a single rule
for constitutional amendment. There is clearly no necessity that this be the
case, as demonstrated by Article V itself, which varies the difficulty of the
amendment process with the importance of given issues. Thus the drafters
of Article V explicitly exempted two issues from the general rules regard-
ing amendment set out at the beginning of the article. First, unanimity is
required to change the rule of equal representation of states in the Senate.18
Second, any amendment concerning congressional abolition of the slave
trade prior to 1808 appears to be precluded. Whatever one might think of
these specific precedents, the latter one of which certainly points to one of
the most horrific aspects of the Constitution, they nonetheless point to the
possibility of requiring a more difficult process of amendment for things
18
Contrary to what is sometimes asserted, the Senate clause is not “unamendable” as a
matter of theory, though, as a practical matter, that is almost certainly the case, given the
extreme unlikelihood of, say, Wyoming agreeing to give up its excess of power in the
Senate. What is less clear is whether Article V could be amended to change the unanim-
ity requirement by less than a unanimous vote. See, e.g., Linder 1981, 717.
 Designing an Amendment Process 285

we define as “basic rights of the people” than for all other constitutional
provisions.
There may be good reason to require very high supermajorities before
limiting rights of freedom of speech and the like. Does any such reason
suggest itself in regard to term limits, whether of legislators or the presi-
dent, or, for that matter, to any other structural feature of the Con-
stitution? For me the question is rhetorical, for I can think of no good
reasons to support the formal stasis engendered by Article V. No doubt
the adoption of such a two-tier system would lead to significant wrangling
about what might count as a “basic right,” but such wrangling seems a
small price to pay for what would be a distinct improvement in overall con-
stitutional design and the increased possibility of cogently responding to
significant structural imperfections.
Ultimately, though, all such discussions take us back to our simplistic,
but not, I hope, simpleminded, models outlined at the very beginning of
this chapter. That is, to what extent do we first acknowledge the possibil-
ity of imperfection and then have faith in our fellow citizens to respond
adequately to such imperfections? Our answers to these questions, whether
we are conscious of them or not, ultimately dictate where along the spec-
trum of possibilities we choose to place (and defend) our own procedures
for constitutional amendment. I use the words “we” and “ours” on the
assumption that most readers are Americans concerned to assess the impli-
cations of our own amendment process for the quality of our polity. But
it should be obvious that the issues raised go beyond such parochial
concerns, whether one is wrestling with the writing of constitutions
elsewhere in the world or simply engaging in detached reflection about
the consequences for political practice of various institutional designs.


Ackerman, B. 1991. We the People: Foundations. Cambridge, Mass.: Harvard
University Press.
1998. We the People: Transformations. Cambridge, Mass.: Harvard University
Press.
Ackerman, B., and D. Golove. 1995. “Is NAFTA Constitutional?” Harvard Law
Review 108: 799–929.
Amar, A. R. 1995. “Popular Sovereignty and Constitutional Amendment.” In
Sanford Levinson, ed., Responding to Imperfection: The Theory and Practice
of Constitutional Amendment, 89–115. Princeton: Princeton University Press.
Articles of Confederation. 1987. Reprinted in P. B. Kurland and R. Lerner, eds.,
The Founders’ Constitution: Major Themes, 23–26. Chicago: University of
Chicago Press.
Baker, L. A. 1995. “Constitutional Change and Direct Democracy.” University of
Colorado Law Review 66: 143–58.
286 Sanford Levinson

Bobbitt, P. 1981. Constitutional Fate. New York: Oxford University Press.

1991. Constitutional Interpretation. Oxford: B. Blackwell.
Boudreaux, D. J., and A. C. Pritchard. 1993. “Rewriting the Constitution: An
Economic Analysis of the Constitutional Amendment Process.” Fordham Law
Review 62: 111–62.
Cronin, T. E. 1989. Direct Democracy: The Politics of Initiative, Referendum, and
Recall. Cambridge, Mass.: Harvard University Press.
Dellinger, W. 1983. “The Legitimacy of Constitutional Change: Rethinking
the Constitutional Amendment Process.” Harvard Law Review 97: 386–
432.
Farrand, M. 1937. The Record of the Federal Convention of 1787. 4 vols. New
Haven: Yale University Press.
Gardner, J. A. 1992. “The Failed Discourse of State Constitutionalism.” Michigan
Law Review 90: 761–837.
1993. “Roundtable: Responses to James A. Gardner.” Rutgers Law Journal 24:
927–56.
Golove, D. 1998. “Against Free-Form Formalism.” New York University Law
Review 73: 1791–942.
Griffin, S. 1995a. “The Nominee Is . . . Article V.” Constitutional Commentary 12:
171–73.
1995b. “Constitutionalism in the United States: From Theory to Politics.”
In S. Levinson, ed., Responding to Imperfection: The Theory and Practice of
Constitutional Amendment, 37–62. Princeton: Princeton University Press.
Holmes, S., and C. Sunstein. 1995. “The Politics of Constitutional Revision in
Eastern Europe.” In Sanford Levinson, ed., Responding to Imperfection: The
Theory and Practice of Constitutional Amendment, 275–306. Princeton:
Princeton University Press.
Kammen, M., ed. 1986. The Origins of the American Constitution: A Documentary
History. New York: Penguin Books.
Kobach, K. W. 1994. “Switzerland.” In D. Butler and A. Ranney, eds., Referendums
around the World: The Growing Use of Direct Democracy, 98–153.
Washington, D.C.: AEI Press.
Kurland, P. B., and R. Lerner. 1987. The Founders’ Constitution. Chicago:
University of Chicago Press.
Levinson, S. 1990. “ ‘Veneration’ and Constitutional Change: James Madison
Confronts the Possibility of Constitutional Amendment.” Texas Tech Law
Review 21: 2443–60.
1994. “Authorizing Constitutional Text: On the Purported Twenty-seventh
Amendment.” Constitutional Commentary 11: 101.
1995. “How Many Times Has the United States Constitution Been Amended?
(A)<26; (B)26; (C)27; (D)>27.” In Sanford Levinson, ed., Responding to
Imperfection: The Theory and Practice of Constitutional Amendment, 13–36.
Princeton: Princeton University Press.
1996. “The Political Implications of Amending Clauses.” Constitutional
Commentary 13: 107–23.
Linde, H. A. 1993. “When Initiative Lawmaking Is Not ‘Republican Government’:
The Campaign against Homosexuality.” Oregon Law Review 72: 19–40.
 Designing an Amendment Process 287

Linder, D. 1981. “What in the Constitution Cannot be Amended.” Arizona Law

Review 23: 717–33.
Locke, J. 1909. The Fundamental Constitutions of Carolinas. Reprinted in F.
Thorpe, ed., American Charters, Constitutions, and Organic Laws, 1492–1908,
5: 2772–86. Washington, D.C.: Government Printing Office.
Lutz, D. S. 1995. “Toward a Theory of Constitutional Amendment.” In Sanford
Levinson, ed., Responding to Imperfection: The Theory and Practice of
Constitutional Amendment, 237–74. Princeton: Princeton University Press.
Rees, G. 1977. “Comment, Rescinding Ratification of Proposed Constitutional
Amendments – A Question for the Court.” Louisiana Law Review 37:
896–925.
Tribe, L. 1986. “A Constitution We Are Amending: In Defense of a Restrained
Judicial Role.” Harvard Law Review 97: 433–45.
1995. “Taking Text and Structure Seriously: Reflections on Free-Form Method
in Constitutional Interpretation.” Harvard Law Review 108: 1221–303.
Tushnet, M. 1995. “The Whole Thing.” Constitutional Commentary 12: 223–5.
Zohar, N. 1995. “Midrash: Amendment through the Moulding of Meaning.” In
Sanford Levinson, ed., Responding to Imperfection: The Theory and Practice
of Constitutional Amendment, 307–18. Princeton: Princeton University Press.