David Riesbeck’s recent post on essentialism reminds me that I have a paper on a loosely related topic that I’ve been meaning (for eight years!) to revise and submit somewhere. As I’m teaching Edward Said’s Orientalism in the fall, I figured I’d make the time to revisit the book and the topic, and finally revise the paper. So here it is, in the interests of feedback from PoT readers, and potentially, for purposes of comparison and contrast with David’s post. Originally presented at the California Roundtable on Philosophy & Race, Hampshire College, October 2, 2009.
Orientalism, Racism, and Islam:
Edward Said Between Race and Doctrine
Edward Said’s Orientalism has gotten relatively little attention from philosophers in the Anglo-American analytic tradition. Arguably, though, the book has been at least as influential in contemporary political thought as has the work of say, Rawls, Nozick, or Dworkin, and has probably been more influential across the breadth of the humanities than the combined efforts of the sum total of analytic normative theorists. Widely regarded outside of philosophy as the foundational text of postcolonial studies, and as the touchstone of a progressive conception of comparative politics and area studies, Orientalism is also a pioneering contribution to race theory. Where English-speaking race theorists had, prior to Orientalism, devoted the bulk of their attention to anti-black racism and anti-Semitism, Said was one of the first academic writers to draw sustained attention to Western conceptions of the Arab/Muslim Oriental. As one early reviewer concisely summarized the book, “Professor Said uses [his] privileged vantage to observe the West observing the Arabs, and he does not like what he finds.”
In what way is Orientalism a contribution to race theory? The question leads to a bit of a conundrum. On the one hand, it is hard to deny that there is some such contribution. On the other hand, the contribution in question turns out to be surprisingly difficult to specify with any precision. I want to suggest that the conundrum arises from a systematic equivocation that runs throughout Said’s treatment of Orientalism—namely, his persistent conflation of claims about the essence of Oriental racial identity with claims about the essence of Islamic religious doctrine. Contrary to Said, a critique of the first sort of claim, however cogent and insightful, is not easily (or at all) transferable to claims of the second sort. The failure to distinguish race from doctrine undermines what is valuable about his account and abets serious confusion.
- ‘Orientalism’: the term and concept
It might help to begin by getting a grip on the concept of “Orientalism” itself. As Said notes, his account of Orientalism is revisionary in several respects. By the time Orientalism came out in 1978, the term “Orientalism” had more or less passed from scholarly discourse–even (or perhaps especially) in the disciplines that had previously been called “Orientalist”–to be replaced by more neutral designations like “Area Studies” or “Near East Studies.” When the term had been current decades earlier, it had had two distinct and apparently unrelated senses. In one sense, it referred to a school of nineteenth century romantic painting focused on “the East.” In another sense, it referred to a branch of scholarship whose practitioners saw themselves, in effect, as classical philologists specializing in “Eastern” languages and culture in the way that Hellenists or Latinists specialize in the language and culture of ancient Greece and Rome. To call x “Orientalist” in this older sense—where x denoted a person, institution, work of art or scholarship–was to identify x‘s aesthetic style or area of scholarly specialization, on par with calling x “Impressionist” or “Hellenic.” In neither sense was the term explicitly connected with race or racism.
Said changed this. For one, he exhumed the term from its previous obscurity and restored it to contemporary currency. He then revised its meaning by identifying what he took to be its foundational methodological assumption: Orientalism, he argued, was “a style of thought based upon an ontological and epistemological distinction between ‘the Orient’ and (most of the time) ‘the Occident'”–“a family of ideas the essence of which is the ineradicable distinction between Western superiority and Oriental inferiority” (pp, 2, 42). It was also, in accord with the earlier definition, “an academic tradition involving anyone who teaches, writes about, or researches the Orient in its specific or general aspects” (p. 2), but given the methodological assumptions he ascribed to it, Said’s Orientalism had a radically different character than, say, classical philology of the Hellenic or Roman variety. For (Said argued) scholars of Greece and Rome do not typically think of Greeks or Romans as metaphysically inferior to the scholars who study them, whereas it is (on Said’s view) an endemic feature of Orientalism that Orientalists see Orientals that way. It would be an eccentric Hellenist who spent a lifetime studying Homer out of racial animosity for Greeks, but (Said argued) the average Orientalist is willing to spend a lifetime on the Qur’an precisely out of such animus. Hence the old analogy between Hellenism and Orientalism is a spurious one: Orientalism is classical philology with a bad (racial) conscience.
But Orientalism is more than either painting or textual scholarship. It is, as Said puts it, “a generic term to describe the Western approach to the Orient; specifically, a collection of dreams, images, and vocabularies open to anyone who has tried to talk about what lies east of the dividing line” (p. 73, my emphasis). Understood in this way, Orientalism not only encompasses painting and philology, but encompasses all of the ‘Western’ arts and letters focused on the Orient–poetry, fiction, belles-letres, travel writing, philosophy, the social sciences, and music. So where the older conception of Orientalism had conceived of scholarship and painting as existing in unrelated compartments, and ignored the contributions of imaginative literature and art altogether, Said regards the entire “archive” of Western activity vis-a-vis the Orient as a single (Foucaultian) discourse, unified across the board by a common ontology and epistemology.
Finally, where the older conception of Orientalism had been relatively apolitical, Said explicitly politicizes it: at least since the eighteenth century, Orientalism has been a corporate institution for dealing with the Orient; “a Western style for dominating, restructuring, and having authority over the Orient” (p. 3), “a will to understand, control, manipulate, and incorporate” (p. 12); “a kind of Western projection onto and will to govern over the Orient” (p. 95); “an aspect of imperialism and colonialism” (p. 123); “a system of representations framed by a whole set of forces that brought the Orient into Western learning, Western consciousness, and later, Western empire” (p. 203). And along this “latent” (but omnipresent) political dimension, Orientalism is inherently racist: “It is therefore correct to say that every European, in what he could say about the Orient, was consequently a racist, an imperialist, and almost totally ethnocentric” (p. 204).
Though Said refers in this last passage specifically to nineteenth century Europeans, his point is that the ascription of racism to Orientalism characterizes Orientalism as such, and characterizes it well into the twentieth and twenty-first centuries–whether practiced by Europeans, by Americans, or by Arabs and Muslims themselves. Contemporary Orientalists may dilute or disguise their racist commitments; they may sincerely oppose racism and imperialism as individuals. But on Said’s view, their complicity in Orientalism renders them incapable of escaping its racist structures, whether conceptual or institutional, and thereby subverts whatever anti-racist decency they might otherwise have had.
To put the point somewhat crassly, then, the “cash value” of Said’s thesis–or at least one highly significant practical application–is its legitimation of the condemnation both of Orientalism as well as whatever is affiliated with, aligned to, or complicitous in it. Indeed, it’s not an exaggeration to say that insofar as it has any contemporary currency, the word “Orientalism” has now come to function less as a noun (as it once did) and more as an adjective involving the ascription to someone or something of a certain kind of racism. To call X “Orientalist” is ipso facto to call X racist, an inference with a fair bit of warrant in Said’s own writings, and with some rhetorical bite in contemporary polemics.
This epithetical use of “Orientalism” is a double-edged sword. On the one hand, it inspires vigilance, and puts racists in their place. On the other hand, it can just as easily induce paranoia, and serve as a cheap way of destroying reputations. For that reason, we need to clarify the conceptual relation between “Orientalism” and “racism.” What, in other words, facilitates the inference from ‘x is Orientalist’ to ‘x is racist’? This question turns out to be remarkably difficult to answer.
- From Orientalism to racism via demographic essentialism
On Said’s account, the connection between Orientalism and racism is supplied by what he calls “essentialism.” In general, then, the inference from ‘x is Orientalist’ to ‘x is racist’ goes by way of the suppressed premise that if x is Orientalist, x is essentialist. There is then some further connection, not entirely clear, between x’s essentialism and x’s racism. Our first order of business then, is to connect the conceptual dots. Why is essentialism racist? Why in particular is it racist when it comes to discussion of the Islamic Near East?
The specific conception of essentialism that emerges from Orientalism is what we might call demographic essentialism, that is, the quasi-biological idea of a population of organisms whose members characteristically (or even exceptionlessly) act true to type. We confront, let’s say, two distinct and basically exclusive populations of organisms, X and Y. We note that X’s characteristically φ, whereas Y’s do not. We want an explanation of this fact. To that end, we hypothesize that the X’s belong to some broader type F, and observe or infer that all F’s have some property or set of properties P that explains why any F characteristically φ’s. We then observe that Y’s do not belong to F, hence lack P, hence do not characteristically φ. With this explanatory scheme in hand, we embark on an investigation intended to confirm or disconfirm our hypothesis.
Suppose that our inquiry finds nothing but confirmatory evidence. In that case we conclude that we have successfully explained X’s characteristic mode of behavior, and have in consequence better understood the nature of X’s. If we ourselves happen to be Y’s, we inevitably become more conscious of the difference that divides us from X’s. We may then resolve to act accordingly, and encourage others who deal with X’s to take stock of our research.
Where ‘X’ denotes (say) the Serengeti lion, and ‘Y’ denotes zoologists, demographic essentialism may well be unexceptional. But where ‘X’ denotes (say) Arabs, and Y denotes Americans, things are more problematic. The Orientalist sees two radically different populations acting in radically different ways. He then hypothesizes that the difference between them owes its existence to the possession by Westerners and the lack of possession by Orientals of a set of population-specific properties. It turns out that Westerners are rational, virtuous, and liberty-loving, while Orientals are irrational, vicious, and slavish. These traits then figure in explanations of Western power and success on the one hand, and Oriental despotism, degradation and retrogression on the other. The success of one explanation abets others, and so a research program is born.
So far, we have an explanatory scheme that need not be racist. It could instead involve claims of civilizational superiority, where the superiority is not reducible to or explainable in terms of, race but of culture (or economics, politics, geography, climate, etc.) What makes the scheme racist on Said’s account is the nature of the distinction on which it is founded. Western superiority and Oriental inferiority are not, for the (Saidian) Orientalist, a merely contingent matter that might have been otherwise; they are expressions of the inherent nature of the two populations. And an inherent, ineradicable, ontological distinction between two groups of physically distinct human populations looks like a distinction of race. The attempt to explain their characteristic behaviors by race looks like racism.
The issue gets complicated, however. Most of Orientalism focuses on claims by Westerners of something like civilizational superiority over Orientals, where the superiority is not intended to be racial, where claims of racial superiority are flatly and vehemently denied, where even claims of superiority (as opposed to cultural difference) are flatly and vehemently denied, and where the putative topic of discussion is not race at all but religious doctrine. In many cases, it is not (at all) clear that the Western author that Said is discussing on any given page is committed to any kind of essentialism, or is making any large-scale ontological or epistemic claim about East versus West. Said insists in such cases that the surface appearance is misleading; the racism lurks beneath the surface. But in many cases, this claim inspires skepticism.
We could, in principle, take every claim of this sort on a case-by-case basis and have a useful debate about every one of them. But that would take more time than I have, and in any case to embark on such an inquiry, we need to know how to evaluate the evidence in the first place. What would count as racist? What wouldn’t?
Though Said apparently thought the answers obvious, I don’t think they are. Insofar as it means anything, a race in its paradigm sense is a (human) population differentiated from other populations by distinctive traits and behaviors whose basis is genetic. A race then is a human sub-population whose traits of character and intellect are determined by a gene pool particular to that group and expressive of the character and intellect of that group. In their paradigm sense, racist claims are somehow connected to race so conceived.
On the other hand, this paradigm conception of race can be a misleadingly narrow (or just plain misleading) basis for understanding racism. Genetics may allow the race theorist to give some precision to the concept of race, but it won’t necessarily give precision to the concept of racism. Racists are notoriously imprecise about the biological (or quasi-biological) presuppositions that lie behind their views. Since racism is a form of irrationality and neurosis, we cannot expect racists to observe fine-grained distinctions about the “basis” of racism. Racists can perceive or invent races where none exist (race itself may be such a phenomenon), and in sophisticated cases, they can express racism while denying that they have perceived or invented anything racial. They can do this sincerely or by deception, and in the latter case, they can deceive others or themselves.
Said does a creditable job at arguing that Orientalist racism takes such subtle forms. He marshals provocative (though not always conclusive) evidence to show that the Orientalist West has often depicted “the Arab” (or “the Muslim,” conceived covertly as a racial category) based on pseudo-empirical generalizations about racial essences. And he does a good job at unmasking the rationalizations by which racial categories masquerade as non-racial ones.
But some care is in order here. The connection between essentialism and racism is far from watertight. The idea at the heart of essentialism is that of an isolated population that acts true to type. This can be a racist idea, but it need not be. It is racist if the populations are isolated on racial grounds. It is not racist if they are not. Even allowing for the complexities described in the last few pages, the farther we get from the paradigm idea of traits that are perceived as or conceived of as heritable, the farther we travel from clear cases of racism. When it becomes clear that a trait neither is nor is in any sense conceived of as heritable, it is hard to see how claims of racism apply, or at least clearly apply.
The paradigm of non-heritability is voluntary adherence to a doctrine. Suppose S voluntarily affirms that p, voluntarily shapes his or her character and actions around p, and acts accordingly. If a whole population does this, the whole population will in some sense act according to type. If the population is relatively isolated or isolable from other populations, we will have an instance of demographic essentialism with respect to those who believe p. If someone outside of that population criticizes p, and the criticism reflects adversely on this population, the criticism may bear a surface similarity to racism. But it clearly is not racism. It is a criticism of p. And while we have to grant that criticisms of p may function obliquely as a cover for racism, we also have to grant the equal possibility that in the end, a criticism of p may just be what it says it is. This is the fact that Said systematically ignores when it comes to Orientalism about Islam.
- Said’s conflation of race and doctrine
Consider the following propositions, some nearly tautological, some relatively uncontroversial, others highly controversial:
(1) Racists act like racists.
(2) Imperialists engage in imperialism.
(3) Racism facilitates imperialism.
(4) Imperialism facilitates bloodshed.
(5) Racism facilitates bloodshed.
(6) Zionism is a form of sectarianism.
(7) Zionist sectarianism explains the rights-violative features of the Israeli settlers’ movement.
(8) Zionism is a form of racism.
(9) Orientalism is closely allied to racism, imperialism, and Zionism.
(10) Orientalists act like Orientalists.
Said is committed to all ten claims. In fact, the defense of these ten claims is practically the raison d’etre for his work. Another way of putting this is to note that Said is an unapologetic essentialist about racists, imperialists, Orientalists, and Zionists. All four categories of populations act true to type, and in all four cases, we explain what they do by invoking the type to which they belong. This clearly implies that not all essentialists are racist. It makes no sense to describe essentialism about racism as itself racist. Nor does it help here to insist that we ought to ‘contextualize’ racisms rather than speaking of racism ‘as such’. The variety of forms of racism doesn’t change the fact that racists are racists and act true to type. Any variation on racism is still bound to be racist.
The preceding claims may seem trivial or obvious, but they put Said in a difficult situation. When it comes to racism, imperialism, Orientalism, and Zionism, Said not only endorses essentialism, but tolerates no attempt to soften that commitment. A close reading of Said shows a steady commitment to the claim that no form of racism, imperialism, Orientalism, or Zionism is, has been, ever will be, or can be rationally or morally justified, and that adherence to the relevant doctrine determines what adherents do qua adherents. As far as Said is concerned, imperialism causes British, French, or American foreign policy in the Middle East, and Zionism caused (and continues to cause) the dispossession of the Palestinians. Etc.
Nor is Said shy about the historical scope of the causal claims to which he’s committed: the causal antecedents of Orientalism for him lie in the writings of Homer and Aeschylus, in the proselytizing mission of John of Damascus (676-749 AD), in the decisions of the Council of Vienne (1312), and in the Napoleonic invasion of Egypt (1798). It is not an exaggeration to say that on Said’s view, the Hebron Massacre of 1994, and the depredations of the Israeli settlers movement generally, can be explained by way of adherence to a text that precedes the settlers’ movement by more than 2500 years—the Books of Exodus and Esther.
By contrast, when it comes to Islam, Said argues that the attempt to talk about Islam as such—as a single, unitary, trans-historical and trans-contextual doctrine—is a form of Orientalism, hence of essentialism, hence of racism. His argumentative ‘strategy’ here is to pair doctrinal claims about Islam with instances of anti-Arab racism, to note the common commitment to essentialism in the doctrinal claim and the racist one, and then to treat the two as though they were morally and epistemically on par with each other. The underlying assumption here seems to be that doctrinal adherence to Islam is fundamentally analogous to race: Islam is to Muslims as an Arab genetic heritage is to an Arab. Just as it would be racist to invoke ‘Arab blood’ or genetic lineage to impugn, say, ‘characteristically Arab behavior’ or ‘the Arab mind’, so it is racist to invoke Islam to impugn the beliefs and practices of Muslims.
In a typical passage, Said writes that for the “Westerner,” the Oriental is “either a figure of fun or an atom in a vast collectivity designated in ordinary or cultivated discourse as an undifferentiated type, called Oriental, African, yellow, brown, or Muslim” (p. 252). In the 1996 Introduction to the Vintage Edition of Covering Islam, Said writes bitterly about the
strange revival of canonical, though previously discredited, Orientalist ideas about Muslim, generally non- white, people–ideas which have achieved a startling prominence at a time when racial or religious misrepresentations of every other cultural group are no longer circulated with such impunity. Malicious generalizations about Islam have become the last acceptable form of denigration of foreign culture in the West; what is said about the Muslim mind, or character, or religion, or culture as a whole cannot now be said in mainstream discussion about Africans, Jews, other Orientals, or Asians (xi-xii).
What is remarkable about these claims–and they come from a collection of nearly two dozen similar passages that I found in Orientalism alone–is the casual conflation they involve of racial stereotyping and doctrinal generalization. The Orientalist apparently treats all brown people as effectively the same, and all Muslims as effectively the same. But it is very far from obvious that these two maneuvers are themselves the same. To treat all brown people as the same is to assume that their complexion tracks facts about their beliefs, motivations, and character. To treat all adherent Muslims as the same is to assume that their doctrinal commitments track facts about their beliefs, motivations, and character.
It is obvious that the first method is racist, but it is not (at all) obvious that the second method is equivalent to the first, that it is in any reasonable sense racist, that it has anything to do with the first, or that there must be something wrong with it. There is no reason to think that complexion tracks belief, motivation, or character, but there is ample reason to believe that doctrinal commitments do. There is a good reason to discourage people from denigrating people of a certain skin color, but this has no application to the criticism of a doctrine, unless one simply equates doctrinal disagreement with racial denigration.
There are, in my view, two serious problems with Said’s conflation of race and doctrine, one (to my mind) rather obvious, the other more subtle. The obvious problem is that the analogy between race and doctrine simply does not hold. In making it, Said ignores the fact that unlike the racial stereotyping he criticizes, claims about the essence of Islam are not claims about traits or populations, but about the content of texts. In this sense, he ignores the fact that unlike “Arab personality” or “Semitic mind,” the term “Islam” has a genuine empirical referent with genuine explanatory power. Islam is a determinate body of thought and practice constituted by the Qur’an, the hadith collections, the holy law, and the commentaries on that law. While there is disagreement about the proper interpretation of these texts, there are also good grounds for identifying within them a core of authoritative beliefs and practices, and grounds as well for deriving second-order beliefs and practices from this core. Granted, not every Muslim acts true to the type Muslim, but it is not plausible to deny that every serious adherent of the faith qua adherent approximates action true to type.
An analogy might help convey this point. Consider the way in which ethical theorists talk about ethical theories—say, utilitarianism and Kantianism. Like Islam, utilitarianism and Kantianism are doctrines constituted mainly by texts. No single form of either doctrine has won the day, whether among adherents or critics. In that sense, one might say that there are many forms of utilitarianism or Kantianism. And yet the multiplicity of doctrinal forms in this context does not entail a ban on asserting what the true version of the doctrine says. Peter Singer and Milton Friedman might both be utilitarians, and utilitarians of very different sorts. But if Singer argues with Friedman, there is nothing racist about his assertion that Singerian utilitarianism is the truest or best exemplification of utilitarianism, whereas Friedmanite utilitarianism is not.
If this is right, it is a fundamental mistake to equate claims about the essence of Islam with claims about the racial nature of “the Arab.” Pace Said, generalizations about Islam do not have the same suspect moral or epistemic status as generalizations about, say, “the Arab mind” or “the Arab personality.” Again, pace Said, it is a mistake to equate classical Islamic studies in the Western Orientalist tradition with classical race theory as practiced by the likes of Gobineau or Schelegel. Indeed, I would argue that Said’s heavy reliance on such comparisons undermines what is most valuable about his account.
- The contextualist strategy
Said is sensitive to the preceding objection, and by my count, employs three distinct strategies for dealing with it. In a longer version of this paper, I call these the Nietzschean, Vichian, and contexualist strategies and discuss all three. For present purposes, I’ll confine discussion to the third.
In one of the last essays he published before his death in 2003, Said articulates an idea that had been latent in his work from the outset, but had never been made explicit. The claim here is that there are, as he puts it, irreducibly “many Islams” that make it impossible to generalize about “Islam as such.” If we are to speak of “Islam” at all–and he seems skeptical even of our doing this–we’re to do so contextually, referring to versions of Islam subdivided by specifically secular criteria–“different kinds of Islam, at different moments, for different people, in different fields,” e.g., “eleventh-century Islamic-Andalusian architecture,” or “eighteenth-century Yemeni religious controversy,” etc. The failure to contextualize or subdivide Islam in this way (Said argues) commits one to essentialism about Islam, which commits one to Orientalism, which commits one to racism. In what follows, I want to argue that despite its currency in certain quarters, Said’s argument fails in several overlapping and cumulatively fatal ways.
First of all, the question at issue is whether Islam as a whole has a doctrinal unity that permits generalizations about it of a sort differentiable from generalizations of a racist variety. Said’s opponent in this dispute claims that such generalizations are possible, and that these generalizations are not racist. They are not racist because unlike racist generalizations, they are not (necessarily) generalizations about the supposedly heritable traits of members of an organismic population but are generalizations about the features of a doctrine. To think of them as racist is just a category mistake. As for whether the generalizations are warranted, that depends on the generalization in question, and can only be settled on a case-by-case basis.
Contrary to first appearances, the contextualist strategy does not deal with any part of this claim. It merely changes the subject. Instead of dealing with evidence of doctrinal unity, or the disanalogy between demographic and doctrinal essentialism, it instructs us to focus all of our attention on the various differences we find among Muslims. But this is a red herring. The differences among Muslims are compatible with Islam’s being a unified doctrine; the differences in question could, after all, arise from disagreement among believers, not irreducible plurality in the doctrine. To the extent that the differences are still differences within Islam, considered as a unified doctrine, Said’s assiduous focus on difference is trivial. We might as well say that there is no such doctrine as utilitarianism because there are so many kinds of utilitarianism, or no such thing as racism because there are so many different forms of racism, or no such thing as Zionism because there are so many forms of Zionism. Unless we have an argument that shows that difference is incompatible with unity, Said’s argument (and all arguments of this form) are clear cases of ignoratio elenchi.
Second, and in partial confirmation of the first point, Said himself seems to concede the falsity of the “many Islams” thesis literally understood, for in the course of defending it, he concedes (albeit equivocally) that Islam is constituted by its “core beliefs…and the centrality of the Koran” (“Impossible Histories,” p. 70). This concession implies that the “many Islams” thesis must, even on Said’s view, be qualified by acknowledgement of the fact that the “many Islams” thesis involves the complex subdivision of a single doctrine. The question, then, is not whether such a unitary doctrine exists, but how exactly to characterize the variety subsumed by it.
There are disputes worth having about the scope and content of Islam’s core beliefs, how the Qur’an is to be interpreted, and how the superstructure of belief and practice is to be derived from it. There are likewise disputes worth having about the relation that obtains between the core of the doctrine as a textual phenomenon and the variety that obtains in the non-textual world. But however these debates turn out, if Islam is (as he admits) constituted by its “core beliefs,” it makes no sense to say (as Said does) that “Islam is not properly a subject at all.” It is, on the contrary, a subject constituted by its core beliefs.
Third, recall that Said’s thesis is meant, not as doctrinal apologetics for Islam, but as an attempt to defend Muslims against the racism that masquerades as a critique of Islam. This is a legitimate aim, but Said fails to see that as stated, the contextualist thesis cannot coherently be avowed by an orthodox Muslim believer, and leads to potential incoherence if avowed by a non-believer as well.
Consider the situation of the orthodox believer from a first-person perspective. If I am a believing Muslim, I am obliged to believe that there is one God and that Muhammad is His messenger. I am also obliged to believe that the Qur’an is the most authoritative (though not the only) source of the divine message. Given the presuppositions and imperatives of faith, I cannot possibly regard this message as something plural, inconsistent, or subordinate to secular concerns. The message comes from God, and God says explicitly that He is One, that His will is singular and self-consistent, that His message is intended for all of humanity, and that its claims override all secular concerns.
From this perspective, the important fact is that I am a Muslim, not that I am (say) “a first-generation male Pakistani-American Muslim born in Jersey City in 1969.” Nor is “my” Islam different from that of a ninth-century Andalusian architect or an eleventh-century Yemeni alim. There is no such thing as “my” Islam or “yours.” Islam isn’t private property; it’s God’s property, and our task as Muslims is to get it right, regardless of who we are, or where we are. If so, our identity qua Muslim overrides any secular descriptors by which we might differ. It follows that no Muslim qua Muslim can coherently affirm Said’s “many Islams” thesis.
Now consider the situation of the non-Muslim studying Islam. Suppose I am not a believer but wish to study Islam. One possibility (among others) is to view it from the perspective of its orthodox adherents, understanding them in an anthropological fashion, while denying in propria persona what they avow. The claims of Islam, I tell myself, are fictions, but my task in understanding Islam is to understand the structure of this fiction by seeing it (up to a point) as an orthodox Muslim would. Suppose I come to see this fiction as having a kind of internal coherence, just as an orthodox Muslim might. Now suppose that I am confronted with Said’s methodological demand: I must subdivide Islam according to historical and secular criteria. I cannot, on pain of being an “Orientalist,” study Islam as such. Surely I am entitled to ask why this demand must govern my inquiries. Islam may be a fiction, but if it is a coherent fiction, the “many Islams” thesis obscures precisely this fact about it. Why am I obliged to affirm a thesis that begins with an a priori dismissal of coherence that I may have discovered in Islam on empirical grounds? I don’t see that Said has an answer to this question. I know of no place in his work where he either raises the question or offers an answer.
At this point, Said compounds the difficulty by insisting that the non-Muslim inquirer into Islam take up “the first-person perspective of the Muslim believer,” presumably to encourage empathic identification with Muslims. But this further demand leads to further problems. For one thing, as I’ve argued, “the first-person perspective of the Muslim believer” is flatly incompatible with the thesis that there are many Islams. For another, as someone who is denying that there is a single true Islam, Said is in no position to be giving advice about “the perspective of the Muslim believer.” He is, after all, the one insisting that there is no such thing! If there is no one Islam, there is no one Muslim perspective. If there is no one Muslim perspective, the advice to take “it” is incoherent.
Finally, even if the thesis were true, its truth would not entail what Said takes it to entail. Suppose that there are many Islams but no single Islam, whatever that turns out to mean. Recall that the original worry was one about essentialism, and this worry is one about the illegitimacy of generalizations about members of some type acting true to type. Despite surface appearances, Said’s legitimate categories are no less essentialist than the categories he regards as illegitimate. The categories differ not in virtue of their essentialist character, but in their complexity and scope.
This implies that the many Islams thesis does not, as Said seems to think, eliminate essentialism. It implies instead that essentialism be narrowed in scope and multiplied in instances. At the end of the day, then, it’s not Muslims that act true to type in Said’s ontology but eleventh century Yemeni religious controversialists and ninth century Andalusian architects that do. At this point, however, Said is not rejecting essentialism so much as haggling about its proper application.
Having tried to prise apart what I take to be the conflation of race and doctrine in Said’s work, I want to conclude by suggesting that Said’s failure to distinguish ethnicity/race from religious doctrine explains one major and otherwise puzzling fact about the now three-decades-old reception to Orientalism—namely, the active hostility for or unsympathetic indifference to Said’s thesis from many in the Arab/Islamic world who might otherwise be thought receptive to it. Such hostility has often itself been explained away as a form of “native complicity” in imperialism and Orientalism, but my critique of Said suggests a more obvious, and to my mind, more plausible explanation: honest disagreement.
Consider two idealized but still possible readers of Said’s Orientalism in the Arab/Muslim world, one a believing Muslim, the other a militant apostate. Both would likely find something of value in Said’s account, and both might be awakened to the ways in which Orientalist discourse had dehumanized them and abetted injustice over the centuries.. But both might take similar exception to the letter of Said’s arguments.
The believing Muslim would note that while Said rightly takes Orientalists to task for their racism and their animus against Islam, he not only does nothing to defend Islam against their doctrinal misrepresentations, but manages in the end to confuse the real issues. Instead of presupposing a true Islam and measuring the misrepresentations against this standard, he denies that there can be such a thing as a true account of Islam. Further, having conflated race with doctrine, he makes it difficult to extricate the anti-racist polemic from the contextualist non-defense of Islam. What might have been an empowering book ends up producing a peculiar frustration.
Like the Muslim, the apostate agrees with Said’s critique of racism, but notes with discomfort that not all of the critique of racism is really about racism. Much of it concerns the problematic tendency of Westerners to castigate Arabs for their civilizational retrogression. The apostate resents some of this: there are too many facile generalizations and too much hypocrisy in the Western case to regard it as having been offered in good faith. There is also too much nonsense in it to regard it with the awe it invites, and too much complicity with power to trust it entirely. But if you scratch this apostate hard enough, he will tell you explicitly that the problem with Orientalism is the opposite of the one Edward Said raises. The problem is not that Orientalists are too harsh on Islam, but that they are not harsh enough.
It is an implication of my argument that both readers are in their own way correct. If so, Said’s work on Orientalism requires major revision before it can serve the anti-racist and anti-imperialist purposes to which so many have put it.
 By “basically exclusive,” I mean that while they can cross-breed as a matter of biological fact, as a matter of non-biological fact, they rarely do.
 It’s a separate task, but an important one, to identify what is valuable in Said’s work.
 Edward Said, “Impossible Histories: Why the Many Islams Cannot Be Simplified,” Harper’s (July 2002).
 One of the ambiguities of the essay is the question of who is doing the speaking. The context suggests that Said’s stricture applies only to non-Muslims speaking about Islam (not to Muslims speaking about Islam), but for reasons I indicate in this text, understood in this way, the stricture seems ad hoc.
 It makes matters worse for the “many Islams” thesis that the phrase “many Islams” has no grammatical equivalent in classical Arabic, the language of Islamic scripture, jurisprudence, liturgy, theology, and philosophy. There is not only no precedent for the concept within existing work, but no intelligible way of reading the concept backward onto the finished body of Islamic thought in Arabic so as to make sense of even the most elementary text. Unfortunately, I lack the space to develop this thought here.