Paimana

Seventh of Muharram, 1448 AH

The three essays before this one were about the people of Karbalāʾ, the one who acted, the ones who suffered and died, the one who survived to see. This one is about a difficulty in our knowing of all of them, and it is a difficulty I have been putting off, because pressed too hard, and in the wrong spirit, it can seem to dissolve the whole edifice, and pressed in the right spirit it turns out to do nothing of the kind. I ended the previous essay by flagging one report, the story of the small daughter said to have died on being shown her father’s head, as late and weakly attested, and promising to say something about the standing of such reports. This is that something, and it is more general than the one report, because the problem reaches, in principle, the whole narrative, and through the narrative a question that belongs as much to the theory of knowledge as to the science of ḥadīth.

Let me state it at its sharpest, which is the only way to take its measure. The tradition that mourns al-Ḥusayn is in possession of one of the most rigorous instruments of source-criticism that any premodern civilization produced. It grades its reports by the soundness of their chains of transmission, the isnād. It has a whole science, ʿilm al-rijāl, devoted to assessing the reliability and probity and accuracy of each named transmitter in a chain. On the strength of these gradings it sorts reports into the sound, the fair, the weak, and the forged; and it will not, in the ordinary case, derive a point of law from a report whose chain it judges weak. A jurist who tried to ground an obligation on a broken or single-stranded or disreputable chain would be laughed out of the seminary. And now the difficulty. The detailed narrative of Karbalāʾ — the thing this community is more certain of than almost anything, the event it builds its identity upon and mourns with the whole of itself — reaches us, in its detail, through exactly the kind of transmission the same community would not accept for a ruling about the timing of a prayer. The earliest systematic compiler of the Karbalāʾ reports was Abū Mikhnaf, a Kūfan of the second century, and his Kitāb Maqtal al-Ḥusayn is lost. What we have of it survives at one remove, in the recension of his student Hishām al-Kalbī, embedded in the universal history of al-Ṭabarī more than two centuries after the event. Al-Ṭabarī gives some of the most affecting material through a single strand, and some of it with no chain at all. Abū Mikhnaf himself is graded, as a transmitter of ḥadīth, as weak, his partisanship counted against him. And the book that circulates in the bazaars under his name, the one from which the preachers often read, is a later forgery that no serious scholar accepts.[^1] [^2] So here is the shape of the thing. The community is most certain of what it has least authenticated. By the tradition’s own standards of authentication, the event at the centre of its devotional life is, in its detail, among the more weakly grounded things it transmits.

I want to handle, first, the answer that the jurists themselves give, because it is correct as far as it goes and it is essential to see exactly how far that is, which is not far enough. The answer is that history and law are different genres, held to different evidentiary thresholds, and rightly so. One does not need a sound chain for history in the way one needs it for law, because history is not in the business of imposing obligations, and the cost of believing a slightly shaky historical report is low, whereas the cost of grounding an obligation on a false one is high. The principle is formalized in the jurists’ own leniency about the evidence for supererogatory and devotional matters, the tasāmuḥ fī adillat al-sunan, the toleration of weak reports where nothing binding hangs on them.[^3] All of this is sound, and I have no wish to dispute it. But notice that it answers a different question from the one I have asked. It explains why we may tolerate weak evidence for low-stakes history, why we need not hold the chronicle to the standard of the law-book. It says nothing about why we should feel certainty, and certainty of the highest and most absolute kind, about an event whose detail is weakly attested. These are not the same posture at all. Toward an ordinary battle of the early period, transmitted through the same kind of chains, the historically careful person feels a provisional and revisable confidence, holding the details lightly, ready to be corrected by a better source. Toward Karbalāʾ the same community feels nothing of the sort; it feels an unshakable, identity-constituting, grief-bearing certainty, the certainty one has of the central facts of one’s own life. The genre distinction accounts for our tolerance of weak evidence. It does not account for our certainty on weak evidence, and it is the certainty, not the tolerance, that is the puzzle.

Here is the diagnosis I want to propose, and it inverts the apparent order of things. The isnād paradigm presents itself, and is ordinarily understood, as the ground of religious certainty: one is certain of a report because, and to the degree that, it has a sound chain back to its source. The case of Karbalāʾ shows that this cannot be the actual ground of the deepest certainty, since here the certainty is maximal and the chain is weak, and the certainty does not so much as waver at the weakness of the chain. The relation between the certainty and the chain runs, I think, in the opposite direction from the official one. The community is certain of Karbalāʾ first, on grounds that have nothing to do with chains, and the isnād is the form in which it afterward dignifies, systematizes, and defends a certainty it already securely has. The historical scholarship bears this out, in its own idiom: the work of Modarressi and others on the formative period shows a tradition whose memory and identity consolidated through the partisan transmission and communal rehearsal of its foundational events well before the antiseptic apparatus of chain-grading was elaborated and then projected backward over the inherited material.[^4] The isnād, on this view, is not the foundation under the certainty. It is the retrospective justification erected over a certainty whose real foundation lies elsewhere.

And now I want to say where the real foundation lies, and to do it I need a frame from outside the tradition entirely, from the later Wittgenstein, who was worrying about the structure of certainty in general and never heard of Abū Mikhnaf. Every chain of justification, Wittgenstein saw, has to end somewhere. The child’s question, “and how do you know that?”, repeated of each answer in turn, cannot be answered forever; at some point one reaches propositions for which one has no further justification and needs none, propositions that are not themselves the conclusions of any inquiry but are rather the fixed points around which all inquiry turns. He called them hinges. The questions we raise and the doubts we entertain, he wrote, depend on certain propositions being exempt from doubt, hinges on which the door of inquiry swings; at the foundation of well-founded belief lies belief that is not itself founded.[^5] A hinge is not a poorly justified belief; it is a belief of a different logical kind, one whose job is to be the ground against which other things are tested, and which therefore cannot itself be tested without bringing down the whole structure that rests on it. Karbalāʾ, as this community holds it, is a hinge in exactly this sense. It is not a conclusion the community arrived at by following an isnād to its end; it is the fixed point around which the community’s other inquiries, including its inquiries into chains of transmission, are conducted. And a hinge, by its very office, cannot be authenticated, because authentication is the operation of testing a proposition against some firmer ground, and there is no ground firmer than the hinge to test it against. To demand the authentication of the hinge is to have misunderstood what a hinge is. So the thing that looked like a scandal, that the most certain matter is the least authenticated, turns out to be very nearly a tautology: the ground of authentication cannot itself be authenticated, on pain of an infinite regress, and the deepest certainties of any system are therefore always, and necessarily, the ones it has not derived from anything else.[^6]

This is the same point that Kripke, reading Wittgenstein on the following of a rule, put in terms of bedrock. When I have given all my reasons for going on as I do, the reasons give out, and I am left not with a further reason but with what I simply do; my spade, as Wittgenstein put it, is turned, and I act without further justification.[^7] What lies at the bottom is not a deeper proof but a practice, a thing the community does, a form of life. And this gives us, at last, the positive answer to the question of what grounds the certainty of Karbalāʾ, if not the chains. What grounds it is the practice. The event is held fast not by the isnād but by the majlis, by the annual and unbroken rehearsal of the event in the gathering, by a transmission that runs from mother to daughter and assembly to assembly in the form of a life lived around this memory. I said something of this in the first essay, when I called it an isnād of affect. That transmission is in one plain respect more robust than any chain of named transmitters, because it does not depend on the reliability of any single link. A chain of rijāl is only as strong as its weakest narrator; one disreputable name and the whole chain fails. The transmission of Karbalāʾ by the form of life has no single load-bearing narrator who could fail; it is carried by everyone and by the practice itself, and it has in fact carried the event, intact in its core, across fourteen centuries with a fidelity that no merely textual chain has matched.

I should say, before going on, that the tradition is not without resources of its own for the point I have been making with imported tools, and the resource it has is worth naming, because it shows that the structure I am describing was seen, in its own idiom, long ago. The science of transmission distinguishes two kinds of report. There is the solitary report, the khabar al-wāḥid, which comes by one or a few chains, yields at best probable assent, and must be graded transmitter by transmitter. And there is the mutawātir report, which comes by so many independent chains that the supposition of collusion or common error becomes absurd, and which therefore yields not probability but certainty, indeed what the theorists call ʿilm ḍarūrī, necessary or compelled knowledge, the kind one cannot withhold assent from however one tries.[^11] The decisive feature of tawātur, for my purposes, is that its certainty does not come from grading the transmitters at all. One does not assess the probity of each narrator in a mass-transmitted report, because the certainty arises from the multiplicity of independent paths and not from the reliability of any one of them; a mutawātir report can be carried by transmitters not one of whom would pass muster individually, and be certain anyway. Here, in the tradition’s own epistemology, is a mode of certainty that bypasses isnād-grading entirely. And the core of Karbalāʾ is certain in exactly this way. The bare event, the killing of al-Ḥusayn and his people on that ground in that year, is not a solitary report resting on Abū Mikhnaf’s contested chain. It is mass-transmitted, by Shīʿa and Sunnī and the merely curious, by so many independent lines that to doubt it would be to doubt the method by which we know anything at all about the period. What the theorists call tawātur maʿnawī, the mass-transmission of a general sense across many reports that differ in their particulars, is the precise technical description of how the core comes down to us: the speeches vary and waver and are sometimes invented, but the event they all report, in their disagreeing ways, is fixed by their very disagreement. The certainty of the core was never hostage to the grading of any chain, and the tradition has always known this, and has had, in tawātur, the name for it.

I have now to be very careful, more careful than I have been, because the argument to this point can be misheard as a license, and the license would be a betrayal of the very scholarship that let me see the problem in the first place. If the certainty of Karbalāʾ does not rest on the chains, someone will say, then the weakness of the chains is no objection to anything, and we may believe the whole narrative, the latest and most florid embellishment along with the firmest core, with the same untroubled conviction. This does not follow, and the reason it does not follow is the most important thing in this essay. Two quite different things have been run together under the single word “Karbalāʾ,” and the hinge argument applies to one of them and not the other. There is the core event, that al-Ḥusayn and the men of his house were killed at Karbalāʾ in the year 61, having refused allegiance to Yazīd and been cut off from water, the women and children taken captive. And this core is not only a hinge of the community’s life but, quite independently, among the best-attested events of early Islamic history, transmitted by friend and foe, by many independent paths, with a unanimity about the skeleton that no source-critic seriously disturbs. And then there is the detailed narrative, the particular speeches, the verbatim dialogues, the individual scenes, the story of the dying child, much of which is late, some of which is single-stranded, some of which is the invention of preachers, and all of which is the proper object of exactly the source-critical scrutiny the tradition is so good at. The mistake the devotional culture makes is to let the hinge-certainty that belongs to the core flow, undiminished and unexamined, onto the embellishments, so that the latest fabrication of a moving scene comes to be mourned with the same absoluteness as the killing itself. The hinge does not immunize the details. The bedrock under the core gives no warrant whatever to the accretions built up around it, and the scholar’s spade, which strikes rock at the core and rightly stops, has a great deal of honest digging still to do among the details.

And I want to insist, because it bears directly on how this essay should be received, that this critical labour is not an intrusion of modern or foreign suspicion into a pious tradition that would otherwise believe everything. It is itself one of the tradition’s own disciplines, practiced by its most rigorous men. The great Akhbārī traditionist Mīrzā Ḥusayn Nūrī collected reports with famous and even excessive generosity. Yet he was so disturbed by the fabrications that the maqtal-reciters and the pulpit-preachers had woven into the story of Karbalāʾ that he wrote a separate book against them, Luʾluʾ wa Marjān, to expose the invented scenes and the baseless embellishments and to demand that the reciters confine themselves to what could be supported.[^8] The sifting of the detail from the core, in other words, is not the orientalist’s contribution; it is the tradition policing the boundary of its own memory, and it is continuous with the very science of rijāl whose seeming failure at Karbalāʾ set this essay going. The science does not fail at Karbalāʾ. It simply turns out to have a different office from the one it advertised. It cannot manufacture the certainty of the core, which it never did and never could, that certainty being a hinge and grounded in the practice. But it remains exactly the right tool for the second task, the disciplining of the detail, which is real work and is never finished.

There is an objection to all of this that I take seriously, the more so because it is the objection a good philosopher would press, and it runs roughly so. To call the foundational certainty a hinge, exempt from the demand for evidence and untouched by the weakness of the chains, is precisely the move by which every dogma in history has shielded itself: declare the cherished belief a hinge, set it beyond the reach of evidence, and you have licensed yourself to hold it in the teeth of anything. If hinges are exempt from doubt, what stops a community from nominating as a hinge whatever it most wishes to keep, and so insulating its favourite errors from correction forever?[^12] The objection is fair, and the answer to it is the whole discipline of the position, without which the position really would collapse into the fideism the objector fears. The answer is that hinge-status is not something a community confers on a proposition by wishing it; it is something a proposition has, or lacks, by virtue of its function, by whether the practice in fact turns on it. A genuine hinge is identifiable by a test: try to doubt it, and see whether the surrounding practice stays intelligible or falls apart. Doubt that al-Ḥusayn was killed at Karbalāʾ, and the entire form of life of this community, its mourning, its calendar, its self-understanding, ceases to make sense; the proposition is load-bearing, the door swings on it, and that, not the community’s attachment, is what makes it a hinge. Now apply the same test to the embellishments. Doubt a particular late and unsupported scene, the exact wording of a speech, the story of the dying child, and the practice does not falter for a moment; one may disbelieve the scene and mourn Karbalāʾ exactly as before. Those propositions are not hinges, whatever the fervour with which they are recited, because nothing turns on them; they are ordinary historical claims wearing devotional dress, and they remain fully exposed to the evidence and fully the business of the source-critic. So the hinge framework, far from licensing the indiscriminate dogmatism the objector fears, is the very instrument that separates the genuinely foundational, which is rightly certain, from the merely cherished, which is not, and the test that does the separating is available to anyone, believer or not. The fideist’s real move is to claim hinge-status for the embellishments, to mourn the fabrication with the certainty owed only to the core, and that move the framework exposes rather than excuses.

There is, finally, a way of seeing all this that the social epistemologists have made available, and that states the matter with a precision I find clarifying. Philosophers who study testimony divide, roughly, into two camps. The reductionists, whose patron is Hume, hold that a report gives me warrant to believe only to the extent that I have independent evidence of the reliability of the reporter, that testimony is credible only as backed by my own inductive knowledge of how far testimony of that kind has matched the facts.[^9] The anti-reductionists, whose ancestor is Reid and whose modern champions are writers like Coady and Burge, hold that testimony is a basic source of warrant in its own right, that we come with a default entitlement to accept what we are told. And this entitlement is not derived from any prior evidence of reliability, nor, in fact, derivable, since one could never establish the general trustworthiness of testimony without already relying on testimony to do so.[^10] Now, the isnād paradigm is, in its official self-understanding, a thoroughly Humean and reductionist instrument: it warrants a report precisely by independent assessment of the reliability of each transmitter, the ʿadāla and the ḍabṭ, the probity and the accuracy, of every name in the chain. And the lesson of Karbalāʾ is that the tradition is not, at its foundation, the Humean reductionist it takes itself to be. Its certainty about its own constitutive event does not rest, and given the weakness of the chains could not possibly rest, on an independent reductionist assessment of Abū Mikhnaf and al-Kalbī. It rests on something much closer to Reid’s default entitlement, an anti-reductionist communal trust in a memory the community has always simply had, over which the reductionist apparatus of rijāl was later laid like a mask, and which at the foundational point that apparatus cannot underwrite and does not need to. It is worth adding that the tradition’s own category of tawātur sits athwart this very dichotomy, and instructively so. It grounds certainty in the multiplicity of independent reports, which sounds reductionist, an inference from accumulated evidence. And yet the certainty it is said to yield is ḍarūrī, compelled and immediate. It is not drawn from the reports as a conclusion is drawn from premises, but forced by them all at once, which is the signature of the non-inferential. Tawātur is thus a hybrid the modern dichotomy did not anticipate, evidential in its source and immediate in its result. And the certainty of the core of Karbalāʾ, grounded as I have argued in the form of life, lies further still toward the non-inferential pole, a thing the community knows in the way it knows the ground under its feet.

So the chains are weak, and it does not finally matter, and it is worth being exact about why it does not matter, because the reason is not that we are free to believe what pleases us. The reason is that the chains were never the thing carrying the weight. A hinge does not hang from the chain of justifications that swings on it; the practice does not rest on the apparatus that was built to dignify it; the certainty of a community about the event at the centre of its life is not the conclusion of an argument and was never going to be overturned by the weakness of one. When we gather in these days and weep, we are not asserting a well-attested historical proposition, and we do not need to be, and the person who imagines that the whole edifice stands or falls with the chain of Abū Mikhnaf has mistaken the retrospective justification for the foundation. The tears are not evidence, and they were never asked to be evidence. They are the transmission itself, the form of life by which this event has reached us more surely than any chain of names could carry it, and which has not failed, and will not, whatever the source-critic, doing his necessary and unfinished work on the details, finds among the chains.

---

[^1]: Abū Mikhnaf (Lūṭ b. Yaḥyā al-Azdī al-Kūfī, d. 157/774) was the earliest systematic compiler of the Karbalāʾ reports and the principal, at points almost exclusive, source for al-Ṭabarī’s account of early Iraqi history. His Kitāb Maqtal al-Ḥusayn is lost; the material survives through the recension of his pupil Hishām b. Muḥammad al-Kalbī (d. 204/819) as preserved in al-Ṭabarī (d. 310/923), Taʾrīkh al-rusul wa-l-mulūk. As a transmitter of ḥadīth he is graded weak by the Sunnī rijāl critics (al-Dhahabī, Yaḥyā b. Maʿīn, al-Dāraquṭnī, Abū Ḥātim), his tashayyuʿ counted against his reliability; and al-Ṭabarī transmits some of the Karbalāʾ material through a single strand or with no chain at all. On the reconstruction of the genuine fragments from al-Ṭabarī, see Muḥammad Hādī Yūsufī Gharawī, Waqʿat al-Ṭaff.

[^2]: The printed Maqtal Abī Mikhnaf in popular circulation, frequently drawn on by preachers, is not Abū Mikhnaf’s lost work but a later compilation (probably sixth/twelfth century, and possibly connected to Ibn Ṭāwūs), whose content diverges sharply from al-Ṭabarī’s citations and which scholars do not accept as authentic. See the entry “Maqtal al-Ḥusayn (by Abū Mikhnaf)” in the standard reference literature, and E. Kohlberg, A Medieval Muslim Scholar at Work: Ibn Ṭāwūs and His Library.

[^3]: On al-tasāmuḥ fī adillat al-sunan, the principle by which weak reports are tolerated in matters of supererogatory devotion and ethics where nothing legally binding turns on them, see the standard works of Imāmī and Sunnī uṣūl al-fiqh; the principle is precisely what licenses a lower evidentiary threshold for non-obligatory and historical material than for the derivation of aḥkām.

[^4]: Hossein Modarressi, Crisis and Consolidation in the Formative Period of Shīʿite Islam (Princeton: Darwin Press, 1993), on the consolidation of Imāmī doctrine and communal identity through the formative period; and Tradition and Survival: A Bibliographical Survey of Early Shīʿite Literature, vol. 1 (Oxford: Oneworld, 2003), on the transmission and shaping of the early corpus. I am drawing the general point, that communal memory and identity consolidated through partisan transmission before the formal apparatus of chain-criticism was elaborated and applied to the inherited material, rather than attributing to Modarressi any specific claim about the Karbalāʾ reports in particular.

[^5]: Ludwig Wittgenstein, Über Gewissheit (On Certainty, 1969), §§341–343 (the “hinge” passages: that our doubts depend on some propositions being exempt from doubt, like hinges on which the door turns) and §253 (“At the foundation of well-founded belief lies belief that is not founded”); cf. the riverbed image of §§96–99, in which some propositions function as the hardened bed in which the river of inquiry runs.

[^6]: The point that the ground of justification cannot itself be justified on pain of regress is the ancient one (Agrippa’s trilemma); Wittgenstein’s contribution is to argue that the regress does not end in an arbitrary stipulation or a self-evident axiom but in a practice, a way of acting. I am applying it to say that the un-authenticated status of the foundational memory is not a defect peculiar to this tradition but the universal condition of any system’s deepest certainties.

[^7]: Saul Kripke, Wittgenstein on Rules and Private Language (1982), on the rule-following considerations and the “skeptical solution” in which the regress of justifications terminates not in a fact that constitutes meaning but in communal practice; the image of the turned spade is Wittgenstein’s, Philosophical Investigations §217 (“If I have exhausted the justifications I have reached bedrock, and my spade is turned”).

[^8]: Mīrzā Ḥusayn Nūrī al-Ṭabrisī (d. 1320/1902), the Akhbārī traditionist and compiler of Mustadrak al-Wasāʾil, wrote Luʾluʾ wa Marjān fī ādāb ahl al-minbar expressly to attack the fabrications and baseless embellishments that the maqtal-reciters and pulpit-preachers had introduced into the narrative of Karbalāʾ, and to insist that reciters confine themselves to supportable reports. That so generous a collector of traditions drew the line here is the clearest evidence that the sifting of detail from core is an internal discipline of the tradition and not an external imposition.

[^9]: David Hume, An Enquiry concerning Human Understanding (1748), §10 (“Of Miracles”): our assurance in testimony is “derived from no other principle than our observation of the veracity of human testimony, and of the usual conformity of facts to the reports of witnesses,” i.e. from independent inductive evidence of reliability. This is the reductionist position about testimonial warrant.

[^10]: Thomas Reid, An Inquiry into the Human Mind (1764), on the innate “principle of veracity” and “principle of credulity” as original constituents of the mind, not products of reasoning; C. A. J. Coady, Testimony: A Philosophical Study (Oxford, 1992), the major modern anti-reductionist work, arguing that the Humean reduction is unworkable because the reliability of testimony in general cannot be established without recourse to testimony; and Tyler Burge, “Content Preservation,” Philosophical Review 102 (1993), on the a priori entitlement to accept what is intelligibly presented as true. The application of the reductionist/anti-reductionist distinction to the self-understanding of isnād-criticism is my own.

[^11]: On al-tawātur and the contrast with khabar al-wāḥid, and on the doctrine that mass-transmission yields ʿilm ḍarūrī (necessary, compelled knowledge) independent of the assessment of individual transmitters, see the standard works of dirāyat al-ḥadīth and uṣūl al-fiqh; the further category of al-tawātur al-maʿnawī, in which a general meaning is mass-transmitted across reports differing in detail, is the one I am claiming describes the core of the Karbalāʾ narrative.

[^12]: This is the long-standing worry that Wittgenstein’s hinge epistemology licenses a relativism or fideism in which any community’s framework-certainties are immune to criticism, a charge leveled at “Wittgensteinian fideism” (the phrase is Kai Nielsen’s) and much discussed in the recent literature on hinge epistemology (Duncan Pritchard, Annalisa Coliva, and others). My reply, that hinge-status is fixed by the functional test of what the practice actually turns on and is therefore not in the community’s gift, is meant precisely to block the slide to fideism.