390 BOOK REVIEWS engaging in a vigorous debate over whether the relationship between courts and legislatures under the Charter can best be characterized as a dialogue between partners of comparable status, or a monologue in which courts dictate public policy to elected legislatures.42 Judicial Power and the Charter provides a good basis for continuing that interdisciplinary conversation. 42 JANET L. HIEBERT, CHARTER CONFLICTS: WHAT IS PARLIAMENT’S ROLE? (McGill–Queen’s Univ. Press 2002); Peter W. Hogg & Allison A. Bushell, The Charter Dialogue Between Courts and Legislatures, 35 OSGOODE HALL L.J. 75 (1997); Christopher P. Manfredi & James B. Kelly, Six Degrees of Dialogue: A Reponse to Hogg and Bushell, 37 OSGOODE HALL L.J. 513 (1999); ROACH, supra note 24. See also the contributions of Peter Hogg and Allison Thornton (formerly Bushell), Rainer Knopff, Frederick L. Morton, Sábastien Lebel-Grenier, and Lorraine E. Weinrib to the April 1999 issue of Policy Options, a leading public policy journal in Canada. Constitutions in a nonconstitutional world: Arab basic laws and the prospects for accountable government, Nathan J. Brown, Albany: SUNY Press, 2002. xv, 244 pp. Reviewed by Noah Feldman* 1. There is a tendency, especially prevalent in the developed West, to think of the rule of law and constitutional governance in binary terms: either you have them or you don’t. If due process applies except when it does not, or if constitutional strictures bind government actors except when they do not feel like following them, then it often may be empty or even deceptive to speak of the presence of the rule of law or of a meaningful constitution. One can imagine good reasons to think of these important phenomena in such absolute terms. For one thing, an almost inescapably normative tone accompanies most of our discussions of the rule of law and of constitutionalism. So if we speak of a partial rule of law or of an incomplete constitutional form of government, we run the risk of appearing to legitimate what may be convenient façades—Potemkin constitutions erected to convey a false sense of regularity and fairness where neither exists. For another, it is philosophically defensible to say that the rule of law in its deepest sense cannot exist so long as the possibility of arbitrary or discriminatory deviation from the law exists. To the extent that a written constitution aspires to embody rule-of-law principles by reducing norms of governance to a form recognizable as positive law, enforceable by and against government actors, it, too, would seem to be fundamentally defective if the law can be ignored by those it is intended to constrain. But of course * Assistant Professor of Law, New York University School of Law. BOOK REVIEWS 391 there is a serious drawback associated with the insistence that the rule of law and true constitutionalism exist only where legal norms are applied consistently and across the board. It makes it far more difficult to develop a vocabulary for speaking of systems in which legal and constitutional rules affect government actors and their decisions without binding them absolutely. An executive who normally enforces judicial interpretations of law but who can, if willing and able to pay the political price, occasionally refuse to enforce them, is in practical terms not unaffected by the law; he is partially constrained by it, even if escape from those constraints remains possible. Similarly, from the perspective of the governed, it may be possible to demand enforcement of public or private rights successfully some of the time, even if under some circumstances the rights turn out not to be enforceable for extraconstitutional reasons; the people under those conditions are being governed with partial regard for their constitution, not with no regard to it. Partial conditions like these obtain in much of the world. Serious students of comparative constitutional law therefore have a meaningful incentive to promote conceptual clarity by finding a way to talk about partial rule of law and partial constitutionalism. It is the fate of Nathan Brown, a political scientist who studies the Arab world and particularly Egypt, the most populous Arab state, to grapple constantly in his career with the challenge of encountering the partial in the spheres of law and constitutional affairs. In an earlier comparative study, The Rule of Law in the Arab World: Courts in Egypt and the Gulf,1 Brown took on the challenge of describing legal practices and institutions in two different contexts in which partial legality was present in different ways. His significant accomplishment there was to show how judicial cultures played an important role in shaping legal outcomes under norms of autocracy. Now, in a readable and instructive new volume, Constitutions in a Nonconstitutional World: Arab Basic Laws and the Prospects for Accountable Government, Brown must confront the phenomenon of the partial in a far greater range of contexts and in a field much more widely strewn with theoretical landmines. Once again he manages the task impressively, giving us an overview of Arab constitutional history and contemporary constitutional conditions in the Arab world that will become standard. To this he adds a well-accomplished, largely freestanding final chapter devoted to Islamic constitutional theory; it is one of the best things in the book, and probably the best short treatment of the subject in English. Brown’s remarkable depth and breadth of knowledge of Arab constitutional structures is less in evidence, however, when he shifts from thick description to comparative and theoretical argument. He makes two central claims in the book. First, he wants to argue that constitutions must be taken seriously by comparativists even where they “organize power without limiting it” (p. 12). Second, he proposes that such “nonconstitutional” constitutions can serve as a groundwork for possible future limited constitutional government, at least in the Arab world. The first of these claims—essentially that constitutions that do not limit power are important, too—is incompletely realized, largely for a theoretical reason: it turns out that it is very difficult to distinguish written constitutions that do not limit power from written constitutions that do. It is therefore not at all clear how one could tell the 1 NATHAN J. BROWN, THE RULE OF LAW IN THE ARAB WORLD: COURTS IN EGYPT AND THE GULF (Cambridge Univ. Press 1997). 392 BOOK REVIEWS difference between a written constitution that “organize[s] power without limiting it” and a written constitution that sets out to limit power, but whose power-limiting provisions are simply ignored. Brown’s second claim, or rather prediction, is certainly true in the abstract: written constitutional texts that today have little effect in limiting the exercise of government power might someday play a part in the establishment of accountable, limited government. But Brown’s own account of the weaknesses of parliaments and constitutional courts in the Arab states suggests that it is not so much constitutional text as the underlying political reality of autocracy that presently stands in the way of limited government there. It follows that what Arab states need in order to establish limited government is more basic than anything constitutional texts on their own can accomplish: they need an infusion of actual democratic accountability in order to supplant and eventually replace autocracy. In that process today’s constitutional texts may turn out to be of use, but they cannot take the lead. Accountable democracy must be brought about in Arab states both by pressure from within, in the form of popular demand for self-government, and pressure from without, in the form of changed Western policy toward Arab autocracies. It can happen gradually, through the strengthening of Arab political institutions like legislatures and constitutional courts, as Brown intimates; and constitutional texts will doubtless play a part. But the importance of written constitutions themselves probably should not be overemphasized in this context. What is needed more urgently is reform of the unwritten constitutions of the Arab states, “that assembly of laws, institutions, and customs . . . that compose the general system according to which the people have agreed [or perhaps, been compelled] to be governed.”2 The customs of political practice that help make up the unwritten constitutions of the Arab states enshrine single-party dominance and autocracy far more effectively and pervasively than do written constitutions. The process of encouraging limited government in the Arab world would benefit from acknowledging the preeminent importance of unwritten political practice in constituting public governance there. Such an acknowledgment would contribute to the study of the partial in the constitutional context; it might also point to potential avenues for research in the sphere of Arab constitutionalism, in particular, and in the field of comparative constitutional law, more generally. 2. The story of Arab constitutionalism that Brown skillfully tells can be summarized relatively briefly. The nineteenth century saw several short-lived experiments in constitution writing, shaped by relations between the European and Muslim worlds. Tunisia, Egypt, and the Ottoman Empire—the last of these not Arab but encompassing Arab lands—each enacted constitutional documents in the wake of crushing debt and foreign pressure to grant concessions. Brown presents all three as driven by an impulse 2 Henry St. John Bolingbroke, A Dissertation on Parties, in BOLINGBROKE: POLITICAL WRITINGS (David Armitage ed., Cambridge Univ. Press 1997). Bolingbroke’s definition of a constitution included “derived from principles of fixed reason”; I have in mind a definition without that normative element. Similarly, I do not mean to claim that the consent of the governed exists under contemporary Arab constitutions. BOOK REVIEWS 393 toward internal rationalization and taxation rather than an intention to grant “any of the traditional liberal rights” (p. 32). “The goal,” says Brown, “was to strengthen the authority of the state in the face of internal rebellion, fiscal crisis, and external penetration” (p. 33). Each of the constitutions created a council, parliament, or assembly, but in no case was this body powerful enough to withstand the rapid reassertion of power by the central authority that had been forced to acquiesce in the constitution in the first place. It might be said that Brown’s presentation understates the externally oriented purposes of these constitutions in favor of their internal purposes, a move consistent, we shall see, with his tendency to emphasize the domestic relevance of constitutions over their reception abroad. After all, it was in this same part of the nineteenth century that the West was insinuating its influence into the Middle East precisely by using the leverage of debt. In the case of Tunisia, Brown notes that the bey “waited until he could present the document to Napoleon III before promulgating it” (p. 16), but still argues that the document was primarily aimed inward. Nonetheless, the nineteenth-century experiments in constitution writing set the stage nicely for Brown’s account of the emergence of constitutions in the monarchies and republics of the Arab world. The constitutions of the monarchies of Egypt and Iraq, both developed under British dominance, are of mostly historical interest, since both monarchies fell to military coups in the 1950s. The drafting of the Egyptian Constitution featured a complex interplay between the khedive-turned-King, who naturally resisted a move toward popular sovereignty; the British, who had themselves never claimed imperial sovereignty over Egypt but wished to remain its masters nonetheless; and the nationalist Wafd party, which initially opposed the Constitution as insufficiently democratic, then later came to rely on it to limit monarchical power once it was in government. The Iraqi monarchical Constitution was imposed by the British from above, as indeed was the Arabian Hashemite King Faisal himself after he was displaced by the French from his putative Arab monarchy based in Damascus. In Jordan, where the Hashemites still reign, Faisal’s brother Abdullah began his rule in 1921 under a Constitution designed by the British that provided for a weak unicameral legislature with ministers responsible to the amir (later the king). Later a second chamber was added, and the Constitution was amended significantly after Abdullah’s assassination in 1951, rendering the monarchy more formally constitutional. Constitutional procedures were followed in removing Abdullah’s invalid son Talal and replacing him with King Hussein, but from 1961 until 1989 King Hussein suspended parliamentary elections, and the constitution did not play a significant role in government. Since then there have been several elections—although as of this writing King Abdullah II has not permitted such elections during his relatively short reign—and the Constitution seems to be taken more seriously by the technocrats who surround the young king. Nonetheless, many laws are still passed by royal decrees promulgated while parliament is out of session, a practice sanctioned by the Constitution. The royalist Moroccan Constitution of 1962, since modified four times, but according to Brown still in essence preserved, gives the monarch great power over the elected parliament, in addition to the power to issue legislation on his own when the legislature is not sitting. The upper house, added in 1996, had the effect of protecting the monarchy during a liberalization process that began near the end of King Hasan’s reign. His son Mohammed VI has cautiously maintained a liberalization course. 394 BOOK REVIEWS The Kuwaiti Constitution of 1962, rather remarkably, confers the legislative power entirely on the elected parliament, even as it describes Islamic law (shari’a) as “a main source of legislation” (p. 56). But the amirs of the Sabah family have on several occasions suspended the Constitution in its entirety. The parliament does not have a strong record of standing up for itself. The one occasion on which the legislature has rejected decrees issued extraconstitutionally by the amir occurred in 1999, when the amir dissolved parliament and took the opportunity to grant women suffrage. Kuwait remains the one country in the Muslim world where women are excluded from voting. In Saudi Arabia, of course, no one votes, and the Basic Law, which is arguably not a constitution at all, gives the king essentially unlimited power, including the power to appoint members of the consultative council. If this story is not particularly encouraging, the situation in the Arab republics— which are mostly de facto presidential dictatorships—is perhaps even less so. Early Syrian attempts at constitutional parliamentary democracy ran into the objections of the French who brought Syria under their sphere of influence after World War I. Subsequent constitutions promulgated after independence lasted only briefly, and the Constitution of 1973, still in force, grants tremendous power to the president, as does the 1971 Constitution of Egypt, on which it is based. That Constitution, Brown argues, reflected Egypt’s post-Nasserite uncertain commitment to socialism and ambivalence about Nasser’s essentially unchecked power to redistribute property (pp. 81–82). The result was a constitution “that gave with one hand and took away with the other” (p. 81), listing socialist aspirations but specifying that basic rights should be enforced by subsequent legislation. Brown also makes the subtle observation, obliquely connected to his earlier work, that the most important aspect of the Egyptian Constitution was its placing of a range of administrative functions under the supervision of the judiciary and the creation of a Supreme Constitutional Court that would eventually become an institutional advocate for constitutional values. Subsequent amendments “downgraded” socialism; abolished Nasser’s Arab Socialist Union, previously the sole permitted political party; and made Islamic law a source of legislation. Brown is surprisingly optimistic about constitutional prospects in Egypt. In North Africa, Brown treats Algeria and Tunisia. The former had at first no constitution after independence, then a standard-issue socialist constitution until the reforms of 1989. Free elections under that Constitution were on the brink of bringing the Islamist Front Islamique du Salut to power when the military called off elections and plunged the country into the horrifying civil war from which it has only gradually begun to emerge. A new Constitution of 1996 provides for some pluralism but unsurprisingly prohibits religious parties. In Tunisia, President Habib Bourguiba not only dominated the process of constitution writing but subsequently amended the Constitution to become president for life. When he was replaced by his prime minister in 1987, after decades of autocratic rule, some superficial amendments were introduced and presidency for life abolished, but otherwise little changed. Although Brown does not say so, the 1999 election of Zayn al-‘Abdin Ben ‘Ali was a wonderful example of autocratic pseudodemocracy: the incumbent Ben ‘Ali took 99 and 44/100 percent of the vote,3 a number that may not have been inspired by the old Ivory Soap commercials, but certainly said something about the cleanliness of the political system. 3 See http://www.electionworld.org/election/tunisia.htm (last visited August 27, 2002). BOOK REVIEWS 395 The Libyan Constitution gets brief notice: actually it is a declaration calling the Qur’an the country’s constitution and establishing a series of popular committees through which the Libyan Jamahiriya is meant to be run. Rounding out his assessment of Arab dictatorships, Brown comes to the Iraqi Constitution, which he tells us “clearly presents itself as an authoritarian document” (p. 87); by now this comes as no surprise. 3. Brown’s account of the Arab constitutions is valuable and thoughtful. His assessment of their significance, however, is a bit confounding. It is probably worth quoting some of his language in order to convey the flavor of his analysis: Arab constitutions have never been routinely violated façades or mere pieces of paper unconnected with political reality. Neither have they been effective bases for constitutionalist practice. Instead they have served a variety of other purposes less familiar to constitutional scholars but quite important to Arab rulers. (p. 91) Those “other purposes” include, for Brown, “signaling basic ideological or policy tenets” and “the desire to organize or augment state authority” (p. 92). Only after these, and subordinate to them, comes the Arab nations’ desire “to underscore their sovereignty in the international arena” (p. 91). For Brown, these are “the purposes of nonconstitutional constitutions” (p. 10). Before assessing Brown’s account of the purposes of Arab constitutional documents, it is worth pausing to ask on what basis he concludes that Arab constitutions “have never been routinely violated.” After all, in most of the countries he discusses, arrests are made without justification, freedom of speech and association are regularly denied, trials can be trumped-up, and extra judicial torture is far from unknown. Sources ranging from the reports of international nongovernmental human rights groups to the annual human rights report prepared by the United States Department of State confirm that this is the case. Yet the constitutions of the Arab world do not openly sanction such violations of human rights, nor does Brown assert that they do. Rather, he says, “[m]ost do provide for substantial rights and freedoms” (p. 154). In what sense, then, can it be said that Arab constitutions are not routinely violated? Brown’s point seems to be that, as written, Arab constitutions provide for significant leeway in delineating basic rights, for example, by conferring upon the legislature the authority to determine what the contents of rights shall be and how they shall be enforced,4 or by providing only weak checks on presidential authority. To the extent that the constitutional text does relatively little to ensure rights, violation of those rights need not be seen as a violation of the constitution itself. If a constitution, for instance, guarantees free speech in accordance with national security, then sharp limitations on free speech are not really extraconstitutional. Absent some sort of argument to this effect, it is difficult to understand what Brown might mean by saying that the constitutions of, for example, de facto dictatorships are not routinely violated. 4 For example, article 22(b) of the interim Iraqi Constitution of 1990 reads: “It is inadmissible to arrest a person, to stop him, to imprison him or to search him, except in accordance with the rules of the law.” 396 BOOK REVIEWS But here Brown runs into difficulties created by the implicit assumption that constitutional guarantees of basic rights in the West are somehow more firmly specified in constitutional texts than are the guarantees of rights in so-called nonconstitutional constitutions. It is, in fact, standard in many Western constitutions for the text to specify that individual rights are subject to a balancing with the interests of the state. The Canadian Constitution Act of 1981 is probably the most famous example of this formulation,5 but there are many others. And where, as in the United States, the text of the constitution does not specify balancing, the binding judicial interpretation of the constitution typically does. It is a commonplace of American constitutional law, for example, that the textual proposition that “Congress shall make no law abridging the freedom of speech” actually means that Congress shall make no such law unless there is a compelling state interest that the law is narrowly tailored to serve. If the state’s interest is great enough, freedom of speech may be abridged after all. Other balancing tests are common in the context of other individual rights. The same is true, incidentally, of the executive power under the U.S. Constitution— the text is extraordinarily vague, speaking only of the “executive power” vested in the president. The practice of American presidents, in complex interplay with Congress and the courts, has produced a set of constitutional practices that begin to delineate the scope of presidential power. But whatever creative interpretations of the words “executive power” have been proffered by lawyers and scholars over the last 200 years, it cannot plausibly be said that the text of the Constitution provides any limitation on presidential power in and of itself. That requires the active intervention of the legislature and the judiciary, as well as a public that confers legitimacy upon them. The alchemy of constitutional governance includes text as one of its elements, but the text itself, since it is not self-interpreting, cannot do all or even most of the work. So Brown’s unstated argument that violations of basic rights or overweening presidential power in Arab states are not evidence of constitutional violation would seem to rest upon a legal fallacy. If rights were systematically violated in Canada, as they are in Egypt, or if the power of the American president were exercised as autocratically as that of the president of Syria, an observer following Brown’s logic could easily conclude that the constitutions of Canada and the United States had not been violated either. Loose constitutional text is a pervasive feature of written constitutions, and the question whether a constitution has been “violated” can never be answered absent a welldeveloped theory of constitutional interpretation that is particular to the constitutional structure in question. In short, saying that Arab constitutions are not violated when rights are violated or when power is consolidated tells us little about Arab constitutional practice. 4. What, then, of Brown’s two stated purposes for nonconstitutional constitutions, namely ideological signaling and the organization of state authority? The former may 5 See Canada Constitution Act of 1981, part 1, title 1, section 1 [Limitation of Rights] (“The Canadian Charter of Rights and Freedoms guarantees the rights and freedoms set out in it subject only to such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society”). BOOK REVIEWS 397 have held some importance when Arab nationalism or socialism were vibrant forces in the Arab world, but in the current disillusioned climate, the self-conscious ideological purposes of older constitutional texts are likely to look thoroughly anachronistic. Even when considered in historical context, it is doubtful that the constitutional declarations of socialist and other ideologies, so common in socialist constitutions the world over, played a major role in diffusing ideology in the Arab world. After all, to whom, domestically, could they have been directed? The average citizen would have little or no knowledge of what the constitutions said, and in any case would have learned of the state’s official ideology through far more public channels of communication and propaganda. As for lawyers and other elites, these would already have a strong sense of the state’s ideological position prior to its embodiment in constitutional text. The one very important exception to this general suggestion about constitutional ideology is the amendment of several Arab constitutions to include language making shari‘a into a source of legislation or sometimes the source of legislation. Eagerly sought by Islamists and resisted by secularists, such language unquestionably has broader ideological significance accessible to the public via the very struggle over its inclusion. Ideology therefore still plays a part in the framing of constitutional language in the Arab world, but it is doubtful whether now or in the past ideology could be counted as the major goal of constitution writing. Brown’s second domestic purpose for the nonconstitutional constitution, the organization and augmentation of state authority, is more important. He is at pains to point out that constitution writing, even in liberal states, often takes the organization and extension of state authority as one of its primary purposes (see pp. 97–104). This is incontrovertibly true, even if it is far more widely acknowledged by constitutional theorists than Brown seems to think. Brown is right in this context to invoke Stephen Holmes’s provocative analogy to grammar, in which “far from simply handcuffing people, linguistic rules allow them to do many things they would not otherwise have been able to do or even have thought of doing” (p. 99).6 Writing a constitution can, in principle, organize public governance by creating institutions through which power may be exercised. Holmes’s analogy gives rise to a range of concrete questions about the relationship between constitutional rules in the abstract and what constitutional actors “have been able to do” in practice. How, exactly, have the constitutional rules in place in the Arab states actually shaped constitutional practice there? How would things have been different if the rules had differed? Brown takes up these questions obliquely in his chapters on parliamentarism and judicial review in the Arab world. In his chapter on Arab parliaments, Brown proposes that these have been weak in part because of constitutional rules that permit heads of state to issue legal decrees under emergency conditions. Although such an emergency decree must typically be approved by a parliament at its next sitting, “this procedure allows rulers the initiative” (p. 115). Brown also mentions the existence of upper houses of parliaments as checks on the power of more democratic lower houses, and the phenomenon of ministers who are responsible to kings or presidents rather than to parliaments directly. These elements of constitutional structure seem far less important, however, in curbing the capacities of Arab parliaments, than the dominance of single national 6 Quoting Stephen Holmes, Precommitment and the Paradox of Democracy, in CONSTITUTIONALISM AND DEMOCRACY 227 (John Elster & Rune Slagstad eds., Cambridge Univ. Press 1988). 398 BOOK REVIEWS parties, maintained through elections that themselves are rigged to ensure the dominance of the presidential party. Executive initiative in proposing laws is hardly the sole preserve of systems that proceed by emergency decree. Upper parliamentary houses can coexist with robust legislative power even when they are unelected. Furthermore, as Brown himself notes, many European constitutions long “lacked provisions for ministerial responsibility to parliament” even as they developed significant constitutional authority (p. 103). One is left with the feeling that written constitutional rules have been far less important than unwritten customs of single-party dominance in determining the weakness of Arab parliaments. In the context of judicial review, Brown suggests that the independence of Arab constitutional courts has been compromised by constitutional rules allowing for judges to be selected by some combination of executive and legislative bodies, and allowing for periodic changes in court personnel. But, of course, it is not these procedures, many of which can be found in Europe and indeed in state supreme courts in the United States, that are at fault in compromising judicial independence: it is the fact that the Arab legislatures are under the thumb of dictatorial executive direction, and that individual judges can easily be intimidated, regardless of how they are appointed. Brown ends up making a somewhat parochial argument in favor of the self-selection practiced by the Egyptian Supreme Constitutional Court, even as he admits that presidential influence remains available even under those conditions. He also suggests that the weakness or silence of constitutional texts has impeded judicial review in the Arab world; this claim, too, remains unconvincing when one observes that the High Court of Israel has plunged itself into extremely aggressive judicial review in the absence of any single constitutional text at all, relying on a handful of vague and thinly sketched Basic Laws that never mention judicial review. The explanation for the difficulties of judicial review in most of the Arab world is surely the obvious one, namely the de facto dictatorial power of presidents and kings. Once again, Brown’s analysis leaves the distinct impression that written constitutional rules have actually been much less important in shaping outcomes in Arab states than the underlying political fact of single-party dominance under presidential or monarchic dictatorship. So Brown’s answers to the first layer of Holmes-inspired questions about the relationship between constitutional rules and constitutional practice suggest that, in the Arab context, the rules have had relatively little effect on practice compared to the effects of extraconstitutional realities of dictatorship. But what of that? Does it undercut Brown’s general argument that constitutional rules matter? To glimpse the beginnings of an answer to this question, one must take a step back, and ask about Brown’s second major claim in the book: that nonconstitutional constitutions can serve as a useful groundwork for future Arab constitutionalism. On the surface, it would seem to matter very much for Brown’s argument about the usefulness of nonconstitutional constitutions that these constitutions have played an important role in establishing the governing realities of the Arab world. Surely the reason that Brown insists that Arab constitutions are not routinely violated, and the reason he argues that the organization of state authority is one of their main (successfully accomplished) purposes, is that he wants to convince us that Arab constitutions matter. If they matter, then they might have the capacity to introduce reforms into the Arab world. If, on the other hand, Arab constitutions do not matter much, then perhaps they lack the capacity to play a meaningful role in the expansion of what Brown calls constitutionalism. 399 BOOK REVIEWS Considered solely in these terms, Brown’s argument would not ultimately be convincing. Brown’s account does not convincingly show that Arab written constitutions have played a determinative role in shaping the actual deployment of public power in Arab states. If this descriptive claim were a necessary precondition for the argument that Arab constitutions have the capacity to play a valuable role in Arab political development going forward, then one would be forced to conclude that written Arab constitutions will not be a major factor in whatever changes in political governance occur in Arab states in the foreseeable future. 5. Fortunately, it is not necessary to ascribe unrealistically great importance to written Arab constitutions in order to think of ways in which they might prove useful in the processes of governmental reform in the Arab world. The key is to understand written constitutional texts as one element in the interplay of law, custom, institutional culture, domestic politics, and international engagement that together make up the constitution of any state. Written constitutions, on this view, are not trivial or epiphenomenal, but they have meaning only when considered within the set of political practices in which they are embedded. Most importantly, although written constitutions purport to direct and order processes of government and the exercise of public power, only very rarely can one identify a unidirectional relationship of the kind that the written constitutions purport to create. One way to see the interplay of constitutional text and constitutional practice is to use a metaphor that Brown mentions: the ship of Theseus, rebuilt board by board until all the original boat has been replaced, yet maintaining its identity nonetheless. Brown says that this nautical metaphor has been “recently introduced” (p. 102),7 but in fact Matthew Hale introduced it in the context of constitutional governance as early as 1739, in his celebrated History of the Common Law of England.8 Hale’s point was much the same as Brown’s: that constitutional rule develops by a slow and contingent process in which successive generations make use of the materials at hand in shaping their constitutional norms. Notice, though, that for Hale, operating under English constitutional norms that never purported to be reducible to a single text, and writing fifty years before the U.S. Constitution, written texts such as the Magna Carta or the English Bill of Rights were just one part of the construction of a constitutional order; one type of board, as it were, in the ship that is being constantly rebuilt. It would be a great mistake to think that the ship in the metaphor is the written constitution. No, the ship is the ship of state, composed not only of law and authority but of the multifarious social, political, and legal practices and customs that constitute it. We need to return to such a view today if we are to develop an improved understanding of the relationship between constitutional text and constitutional practice. Specifically, we need to return to a terminology that would have been familiar to Hale, in which the word “constitution” does not refer to constitutional text alone, but to the 7 Citing INSTITUTIONAL DESIGN IN POST-COMMUNIST SOCIETIES: REBUILDING et al. eds., Cambridge Univ. Press 1998). 8 THE SHIP AT SEA (John Elster HISTORY OF THE COMMON LAW OF ENGLAND 40 (Charles M. Gray ed., Univ. of Chicago Press 1971). 400 BOOK REVIEWS complex interplay of social, political, and legal forces that describe and order political governance. Once we have done this, we can make relatively modest claims for the effect of constitutional texts while still recognizing that these texts play an important part in the creation of a constitutional order. Making the shift in our sense of what we mean by speaking of the constitution of a particular state should move us away from the prevalent terminology in which the word “constitutionalism” is often deployed to mean something like “government limited by a written constitution.” Brown wants to offer a different approach; he defines constitutionalism to mean “ideologies and institutional arrangements that promote the limitation and definition of means of exercising state authority” (p. 8). This is an improvement, in that it does not depend on text alone. But the definition is also equivocal on the question of whether ideologies and institutions that define state authority without limiting it are properly “constitutionalist.” If they are, then Arab constitutions would not properly belong to the “nonconstitutional world” of Brown’s title, because, for Brown, those Arab constitutions define and augment the exercise of state power without limiting it. The very title of Brown’s book therefore suggests that he uses the term constitutionalism to refer to ideologies and institutional arrangements that limit government. One could, with Brown, usefully broaden the term “constitutionalism” to include a range of laws, institutions, and customs, including written constitutional texts, that together effectively limit government. But the bottom line is that the employment of the word “constitutionalism” is frequently more confusing than it is illuminating. The practice of adopting written constitutions can be described as exactly that, and with greater precision. The objective of limiting government through a series of techniques, one of which is writing a constitution, also requires no special term. Democracy is, as Brown acknowledges, logically independent of the existence of a written constitution, although liberal democracy surely cannot be achieved without limited and accountable government (pp. 9–10). Beyond terminological confusion, the common use of the term constitutionalism to mean government limited by a written constitution misleadingly suggests that it is the written constitution that by itself has the effect of limiting government. This is never, or almost never, the case. Rather, the written constitution is one tool in the process of making government limited and accountable. The written text cannot interpret, enforce, or execute itself. The written text cannot on its own guarantee that people will either appeal to it or develop mechanisms for applying it in practice. It is one typical feature of the written constitution that it claims for itself absolute authority over structures of governance, even to the point of purporting to specify mechanisms for its own amendment. But there is no good reason for our terminology, as it seeks to describe the relation of constitutional text to public governance, to subscribe to the written constitution’s unavoidable advertisement for itself: no reason, that is, unless our descriptive practices are intended primarily to have a normative effect. 6. There is a further, fascinating side effect of conceptualizing the phenomenon of the written constitution as one element in an interplay of social, political, and legal forces that together produce what we might usefully call a constitution. This reconceptualization BOOK REVIEWS 401 should deepen our understanding of the relationship between political theory and political practice—or if you prefer, constitutional theory and constitutional practice. To see how this is so, let us return to Stephen Holmes’s analogy between constitutional rules and grammatical rules, which “allow [people] to do many things they would not otherwise have been able to do or even have thought of doing” (p. 99). If we take Holmes to be comparing written constitutional rules and prescriptive rules of grammar, then the analogy produces an important insight about the constitutive power of rules as well as useful questions about how rules are played out in practice. But what if we push the analogy further? What if we understand Holmes to be referring not only to written constitutional rules but to unwritten ones? And what if these unwritten rules are compared to the multiple layers of grammar with which contemporary linguistics is concerned, namely the fundamental, universal grammar that underlies all language and the surface grammar that constructs and governs the particular human languages that we speak? Now the analogy may bear still richer fruit. In Noam Chomsky’s terms, the unspoken, objective, deep rules of universal generative grammar underlie all languages. They are the fundamental conditions and rules of language, hard wired into human biology. We express them, however, only through the contingent grammatical rules of particular human languages. We never speak solely in deep grammar, but we never speak without it, either. The same may perhaps be said for the social, political, and legal forces that are at work in shaping a constitutional order. Perhaps, to begin with, there exist some fundamental or hard wired rules of political governance, universally observable in every constitutional system. Classical political theory was once very interested in identifying such universal phenomena and relating them to human nature—Aristotle’s polis, for example, a necessary place for the flourishing of the political animal, or Hobbes’s state, which sometimes sounds as though it must be present wherever there is political governance. Today such speculations are rather rarer, more likely to be found in the murky realms of sociobiology or optimization-based game theory. But perhaps comparative constitutional studies might more modestly reorient itself in the direction of seeking out some such deep structures of constitutional governance. Such a deep grammar of constitutional governance would reflect fundamental constitutional necessities: for example, that a written constitution must be interpreted by somebody; or that a constitutional order must produce some account of its own legitimacy. These necessary generative rules of constitutional governance, if they exist, would shape unwritten and written constitutions, but they would not entirely determine their form. Indeed, like deep grammatical rules, we will never observe them on their own in the world. What we will see, instead, are the manifestations of many particular, contingent systems of political governance, each with its own peculiarities, beauties, and pitfalls. These rules, comprising the written and unwritten laws, institutions, and customs that make up particular constitutions, correspond to surface grammar, the grammar of a particular language. Surface grammar would reflect the way a particular constitutional order has worked out a universal rule in practice: for example, how a constitutional court interprets the constitutional text; or whether legitimacy is based upon custom, or consent of the governed, or some combination. Considering constitutional rules both as deep and surface grammar does not exhaust the value of Holmes’s analogy, though: there is a further insight to be derived from the analysis of surface grammar itself. One of the great revolutions of modern 402 BOOK REVIEWS linguistics was the realization that we can get a much better sense of the surface grammar of a language by looking at the things people actually say than we can by focusing exclusively on the grammatical rule books that state what is technically permitted. The grammar books tell us something, of course, but they speak always in normative, prescriptive terms, and so provide a limited picture of the complexity of linguistic practice. The same is true of written constitutional text; it must speak normatively, and so on its own cannot provide more than a partial account of constitutional phenomena. A robust study of constitutions, whether in comparative terms or not, requires one to look far beyond prescriptive texts alone for the full range of social, political, and legal phenomena that truly constitute political governance. It requires us not only to ask what constitutional texts say but what constitutional actors, in fact, do. To imagine that constitutional rules consist only in the rules recorded in written constitutional texts would be just as myopic as looking at prescriptive grammar books to the exclusion of the actual operative rules of grammar gleaned from a thick description of linguistic practice. Holmes’s analogy thus also calls for a descriptive constitutional social science that understands the operation of constitutions in catholic terms. Just as descriptive grammar has given rise to a new terminology, in which calling an expression “ungrammatical” means not that a technical rule of prescriptive grammar has been violated, but rather that competent users of the language would identify the phrase in question as outside the norms of their linguistic practice, so too constitutional studies may need to develop, or redevelop, a terminology in which unconstitutionality is measured by the views of the relevant constitutional actors, not merely by reference to a text taken in the abstract. Such a development would require careful delineation of the boundary between description and prescription. As I suggested at the outset, the tone of normativity that pervades constitutional discourse, even in its comparative guise, makes this line hard to draw. But the line must be drawn, to the extent possible, if our goal is truly to deepen our understanding of constitutional phenomena. This is so even if our ultimate goal is the design of institutions that serve their citizens more faithfully and with greater respect for their basic rights. A renewed descriptive approach to characterizing constitutional relations can help us overcome the danger of relativistically legitimating unjust regimes as in some sense “constitutional.” We can begin by acknowledging that comparative constitutionalism will generally involve us in the study of the partial: the partial rule of law, and the partial existence of constitutional government. Once we have laid this conceptual groundwork, it will not seem as threatening to say that the ruling National Democratic Party in Egypt has a “constitutional” role despite the fact that it is never mentioned in the written constitutional text, just as it should not seem threatening to say that political parties play a crucial constitutional role in the United States despite being absent from the text of the Constitution and most of the constitutional case law. Not that every political society has a “constitution”; some political societies are so profoundly in flux that no constitutional order can be said to exist. But in most places where political arrangements are settled, we will be able to speak of the presence of a constitutional order of some kind. Just as the presence of a written constitution in almost every country today is no assurance of political justice, so referring to the existence of a “constitution” in a broader sense will not lead us to assume that justice prevails there. Comparing constitutional rules to grammar opens a final direction that should be of great interest to students of comparative constitutional phenomena. One crucially important effect of the existence of deep grammatical structures is that one can learn BOOK REVIEWS 403 to speak other languages, sometimes flawlessly, sometimes with particular quirks or accents. Not immutability, but malleability is the hallmark of human linguistic production. Even within a single language, the linguistic order emerging from deep grammar is far more flexible than one that depends on written rules. The poet or the phrasemaker can always try something new, and even if, on first hearing, other users of the language are skeptical or disapproving, there always exists the possibility that a new usage will be sufficiently appealing that others will begin to use it. When that happens, the possibilities for change are virtually limitless. The theme of significant constitutional change emerges in Brown’s excellent final chapter on the phenomenon of Islamic constitutionalism. To many Westerners, it has come to seem that constitutional or democratic structures cannot coexist with religious theories of law or politics. Yet as Brown suggests, and as I have argued at length elsewhere,9 the possibilities for productive engagement between Islamic ideas of political governance and ideals of constitutional democracy are, in fact, very rich. The reasons lie in the flexibility of both democratic theory and Islamic political thought. And as Brown also shows, a surprising range of Muslim political theorists today is increasingly interested in that interaction. It is no easy task to convince a skeptical reader that Islamic constitutionalism, or as I would prefer it, Islamic democracy, is a viable option for political governance in the Arab world. If it can be done, however, it must come through convincing those familiar with the constitutional rules of the West that there are other constitutional languages available, languages that can benefit from an encounter with Western constitutional ideas but that can also have a reciprocal influence upon them. To change the constitutional structures of the Arab world for the better will require much more than incrementalism derived solely from the texts of Arab constitutions. It will require the efforts of Arabs themselves, and a change in policy on the part of Western countries, particularly the United States, that support autocratic Arab governments. When and if Arab constitutional reform occurs, it will come in the context of the reordering of political, social, and legal relations in the Arab states. This may not require violent revolution, but it will be revolution of sorts all the same—a fundamental change in the ordering of power, away from autocratic government and to the people. When that happens, Islam is sure to play a role. There is, then, some warrant for Brown’s optimism, notwithstanding the great difficulty of showing that constitutional texts in the Arab world provide on their own the seeds for accountable or democratic government. In recent months, for a variety of complex reasons, more and more voices are being heard in the Arab world and in the West calling for the democratization of Arab states. There are no guarantees, and the road ahead is long. The change, if and when it comes, will require political impetus, legal reform, and the transformation of the social phenomena that have so far undergirded autocracy. There is a place for written constitutions in that process of transformation: not the only place, perhaps not even the leading place, but a place nonetheless, within the structures of custom, law, and language that can be described, in a larger sense, as the constitution. 9 NOAH FELDMAN, AFTER JIHAD: AMERICA AND THE STRUGGLE FOR ISLAMIC DEMOCRACY (forthcoming April 2003).