Epistemic status: a position I hold confidently in its weak form, less confidently in its strong form, and with a great deal of remaining uncertainty about how the strong and weak forms come apart. The piece is partly an attempt to figure out where the line is.

I. The Febvre Puzzle

There is a question raised by Lucien Febvre in a book published in 1942 which I have been thinking about. The question is whether it was possible, in sixteenth-century France, to be an atheist.

The question sounds odd. Of course it was possible! Right? People believe and disbelieve things in any century, and the sixteenth century was full of heretics, Protestants, occultists, libertines and a great many people who said things in private that would have got them in serious trouble if said in public. Febvre was a careful historian and he knew all this. Indeed he spent four hundred pages working through the available evidence, which is considerable. He goes on to examine the polemical literature, the trial records, the libertine poets, the philosophical underground that flourished around the universities, the Italian materialists who had circulated their texts into France, and the various people who had been accused of atheism by their enemies. He was looking for one well-attested case of an actual sixteenth-century atheist, in something like the modern sense of the word, and his conclusion at the end of four hundred pages is that he could not find one. Not a single one. There were freethinkers and anti-clericals and heretics of every variety. There were people who said God did not care about specific things or that the priests were liars or that the resurrection was an allegory. There were not anywhere in the available evidence any people who had successfully disbelieved in God in the sense in which a contemporary atheist disbelieves in God.

This is a strange finding, isn’t it? It wouldn’t be so surprising if the finding was about the sociology of religion or the persistence of belief under pressure. But the fact that it is a finding about something more basic, which is the availability of a particular position qua position, I think has a lot of consequences on epistemology. Again, Febvre’s wasn’t claiming that no one in 1500 dared to be an atheist, because of some sort of fear, and remained closeted. The conceptual schema within one’s cognition that is necessary to hold the position (to hold it firmly with conviction over time and against the press of one’s daily life) was utterly absent in the linguistic and intellectual environment of the time. The vocabulary simply did not exist so as to map one’s thought onto. There were no arguments to build upon. If we consider the existence of sceptics at the time, as opposed to the sceptical frame of reference, then there were many. But their scepticism was always about specific theological claims and not about the framework within which those claims were made. The word atheist was in circulation yes but it did not mean what we mean by it; it was a term of abuse, applied to people who held the wrong religious views and not a self-description anyone would take up. To be an atheist in 1500 you would have had to assemble the position from nothing against a culture in which the equipment for assembling it was not lying around, and against an inner life that had been shaped, from cradle to grave, by the bells and the calendars and the rituals and the pervasive theological vocabulary in which all reflection was conducted. Nobody managed it. So far as Febvre could tell nobody seriously tried.

Marc Bloch in Feudal Society makes a similar point about the medieval period more broadly. The whole picture of the world that medieval Europeans inhabited was theological in a way that is hard for us to even begin to reconstruct. It was not the case that they had a theology and a worldview, with the theology enveloped loosely on top, as how we often think of present religious communities. It was rather that the theology was the worldview, the substrate within which all thinking happened. Even the doubts that did surface “lacked any rational basis”, in that there was no scaffolding of independent inquiry to which the doubts could attach themselves and become serious. The doubts were sparks struck against air.

II. Objections

Here are three responses one might give to Febvre’s claim, in roughly the order I have heard them.

Response 1: People were too scared. On this view, there were plenty of atheists in 1500 but the Inquisition was really scary and the punishments for heresy were severe, so atheists obviously kept their views to themselves. Febvre is just looking at the public record and missing the private reality.

We cannot get into people’s hearts, sure, but there’s always going to be some sort of evidence. The libertine and freethinking literature of the sixteenth century is large and quite candid. People said many shocking things and got away with most of them, especially in private circles and especially in writing intended for private circulation. The historical record contains an enormous quantity of almost-atheism - e.g., anti-clericalism, materialism, scepticism about miracles, arguments that God might be conceived in radically non-Christian ways, but it does not contain actual atheism. An almost-atheist material would be exactly the kind of thing we would expect to find if Febvre is right and exactly the kind of thing we would not expect to find if Response 1 is right. (If everyone were a closet atheist you would expect more aggressive almost-atheism in the underground literature, with the few public expressions being the tip of a hidden iceberg. Instead the underground literature is itself almost-atheist and specifically not atheist, so the iceberg’s tip seems to be the iceberg.)

Response 2: The category is anachronistic. On this view, modern atheism is a particular cultural product of the post-Enlightenment world, and asking whether sixteenth-century people were atheists is like asking whether they were existentialists. The question is not interesting because the category does not apply.

I am very sympathetic to this response but it also seems to concede the point Febvre is making. If modern atheism is a particular cultural product that did not exist in 1500, then we have given up the picture in which there is a steady transhistorical question (does God exist?) to which different ages have given different answers, with some answers being correct and others incorrect. We have replaced this picture with one in which the question itself in the form in which we ask it is a product of historical conditions that did not obtain in 1500, and therefore the correctness of any particular answer is also indexed to those conditions. This is, after all, roughly the position I want to defend.

Response 3: The question is about psychology and not philosophy. On this view, sixteenth-century people had the same cognitive capacities as we do but they just lived in a different cultural environment that made certain positions hard to articulate. Though atheism was psychologically available, it was not specifically socially available.

I am most sympathetic to this view and I think it almost works in a way that’s informative about what we actually mean by availability.

The intuition behind Response 3 is correct in the limited sense that nobody is denying that sixteenth-century brains and modern brains are made of roughly the same neural structure. The question is whether psychological availability, abstracted from the social and linguistic environment in which a person actually thinks, is even a coherent concept. I am not sure it is. To hold a position in any robust sense seems to need more than the bare neural capacity to entertain the relevant proposition. Holding a position, to my mind, means the capacity to formulate the position in some vocabulary, to defend it against counterarguments one has actually encountered, relate it to other things one believes, resist the press of ordinary life in a way that requires constant ongoing maintenance and re-description, and to have access to enough other people who are doing something similar, to such an extent that, over the long course of an actual life, it does not simply dissolve under the weight of being unsharable. The hypothetical sixteenth-century atheist would have had to do all of this without models or shared vocabulary or community of fellow practitioners or philosophical traditions to draw on, and against a cultural environment in which every bell and every calendar and every casual reference at the dinner table reaffirmed the framework he was supposed to be rejecting. This would be rather extra-ordinary.

I think it is at least an open question whether anyone of any era can hold a complex philosophical position under those conditions for the full length of an adult lifetime. I am not sure I could. (I notice that I have many positions I currently hold which I am fairly sure I would not have held had I been born in different conditions. Some of these are positions I think are correct. The ones I think are correct, I think are correct in part because of the conditions that allow them to be held.) If we take Response 3 seriouslyt then we would have to hold that positions float free of the conditions that make them holdable, which I think is a picture I want to reject.

III. Anti-Foundationalism

The position I want to defend is that there is no view from nowhere. The sentences that count as true in any given linguistic community count as true relative to a set of practices, vocabularies, criteria of evidence and institutional structures that have themselves been produced by a particular history. The history could have gone differently and the sentences would then have been different. There is no further fact beyond the practices and the history that determines which set of sentences correctly corresponds to the world. Though the world causally pushes against our sentences, the world does not dictate which sentences to use, and the sentences we end up using are tools we have developed for the work we are trying to do and not transcriptions of an antecedent reality. ‘Truth’ is a property of sentences, sentences are tools, and the choice of tools is a question for us.

IV. Foucault’s Anti-Foundationalism

The position I have just stated is more or less what Richard Rorty had argued his entire career, and Rorty got most of it from Wittgenstein, Sellars, Davidson, and Quine, with some Dewey at the bottom. None of this is Foucault. Why does Foucault figure?

Foucault figures because the philosophical idea I have just described is, on its own, somewhat thin in its empirical content. Rorty tells us that vocabularies are tools and that we choose between them on pragmatic grounds but he does not, in his own work, do very much of the actual work of showing how this happens in particular cases. Foucault however did. Foucault spent his career producing extremely detailed historical case studies of how particular vocabularies came into being and how they affected us once they existed and what happened to the people whose lives were reorganised by their existence. Whatever one thinks of the specific Foucault interpretations (and many of them have been disputed by careful historians on points of fact, and I do not want to be taken to be a Fou-CULTist), the kind of work Foucault did is the kind of work the Rortyan position needs as empirical oomph.

History of Madness is Foucault’s first major book and, in my view, still his strongest. The basic claim is that ‘madness’, as a unified category of mental disorder requiring medical treatment, did not exist in any robust form before the late eighteenth century. There were people with what we would now call schizophrenia, depression, mania; there were people considered possessed by demons; there were people kept by their families as harmless eccentrics; there were people who lived as fools or holy fools or village idiots. What there was not, until quite recently, was a single category that grouped all of these as instances of a unified phenomenon called mental illness, distinct from physical illness, requiring the specific intervention of a specific class of professional. The category was produced in a particular historical moment by particular institutional pressures (including the rise of the asylum, the professionalisation of medicine, the bureaucratic needs of the modern state), and once produced it began altering the lives of the people it now classified.

The standard objection to this kind of analysis is that it sounds like a denial that mental illness is real. Foucault sometimes encouraged this reading and his anti-psychiatry admirers ran with it, but I think the strong version of the reading is wrong, and I want to be clear about what the analysis is and is not committed to. It is not committed to the claim that the brains of people we now diagnose with schizophrenia have something distinctive going on. They do. The neurological and biochemical correlates of schizophrenia are well-documented, the heritability is substantial, the cross-cultural prevalence is more uniform than for most psychiatric conditions, and the underlying disturbance is very much real.

What the analysis is committed to is the claim that the category of schizophrenia, which is to say the way the disturbance is grouped, named, classified, treated, distinguished from other things, integrated into legal and welfare and family structures, and experienced from the inside by the person to whom it applies, is a historical product. The category could have been carved differently. It has, in fact, been carved differently in different periods and different societies, and such difference shows up in the diagnosis rates, in the treatments offered, in the patients’ own experiences of their conditions, and in the social fates that follow from being categorised one way rather than another. Philosopher Ian Hacking gives an example of the way the diagnostic criteria for schizophrenia in mid-twentieth-century European psychiatry shifted as Kurt Schneider’s First Rank Symptoms became the standard list. Once auditory hallucinations were near the top of the list, more people were diagnosed, the people diagnosed were a different population, the patients themselves began to take their hallucinations more seriously, the public response shifted, and the trajectory of the diagnosis therefore was altered. There is a real underlying disturbance but how this underlying disturbance gets taken up into a system of categories alters our social world is what’s at issue.

The the point I want to draw out is that categories are not passive descriptions of an antecedent reality as much as they alter the things they classify, sometimes by altering how those things are treated and sometimes by altering how the classified things conceive of themselves. Hacking calls this the ‘looping effect’. Once you have a classification (autistic, schizophrenic, fat, gay, introvert, gifted, traumatised), the people who fall under it begin to live, partly, in the conceptual space the classification opens up. They make sense of their behaviour in its terms. They organise their lives around it. They join communities of others similarly classified. They take on therapies, prognoses, identities. And the classification, in turn, shifts as a result of how the classified people behave because the diagnostic criteria are calibrated against the population, and the population has been altered by the diagnosis.

V. The Christendom Case Revisited

Now we can go back to the Febvre puzzle and say what is going on there.

Christendom, in Febvre’s analysis, was not a body of doctrine that medieval Europeans happened to accept and that they could in principle have rejected. It was as we have desctibed a substrate within which the capacity to accept or reject doctrines had been formed. The bells, the calendar, the guilds, the schools, the family rituals, the legal vocabulary, the categories of the self (sinner, penitent, saved, damned), the cosmological scheme, the political theory, the medical theory, the moral psychology, the theory of language - all of this was theological through and through, in such a thorough way that there was no neutral place from which one could survey the theology and decide whether to accept it. The mental tools with which one would have evaluated the theology were themselves products of the theology.

We said that this is not an account of indoctrination of repression, however much such things existed. The account is rather of how the conditions of intelligibility for a position is instantiated. To be an atheist in any robust sense you need a vocabulary in which God is one entity among others rather than the substrate of intelligibility itself; you need a public conception of evidence that is independent of testimony from religious authorities; you also need a community of others holding similar views or at least a tradition of texts you can read in private. You moreover need the conceptual distinction between reasons and causes to be established and stable and you need at least a few of the key sceptical arguments to have been formulated by someone, somewhere, in a form you can engage with. None of this was lying around in 1500.

Positions are holdable relative to the conditions of holdability that obtain in a given time and place. We do not, as historical agents, have unlimited access to all possible positions. We merely have access to the positions our circumstances have made available and the positions our circumstances have made available are a function of the entire prior history of how those circumstances came to be the way they are.

We are not exempt. The same kind of analysis that we apply to sixteenth-century Christendom can in principle be applied to twenty-first-century secular liberalism, in which there are positions the contemporary mind cannot easily hold, because the conditions of holdability are not present. Various forms of pre-modern religiosity are in this category for most contemporary Westerners. So are various forms of communalism that take the family or the tribe rather than the individual as the basic unit of moral concern. So, increasingly, are the strict materialism of certain mid-twentieth-century left-wing politics, and the Christian humanism of the early Cold War, and a number of other positions that were once the default for intelligent people and are now nearly impossible to occupy with full conviction. It is simply the case that the conditions of these positions’ holdability are not present in our environment and one would have to do an enormous amount of reconstructive work to hold them, of the kind that the sixteenth-century atheist would have had to do in reverse.

VI. Relativism?

The standard objection at this point is that the position I have just laid out fails because it implies relativism. If positions are holdable only relative to the conditions of holdability that obtain in a given time and place then what is the difference between a position being true and a position being currently holdable? Has not the whole apparatus of truth been replaced by the apparatus of cultural fashion?

The rejoinder is this: the world causally pushes against our sentences even though it does not semantically dictate them. There are sentences we cannot say with a straight face because the world makes them too hard to maintain; that is to say, most people would find it a not very useful sentence. ‘The cat is on the mat’ survives because if there is a cat and there is a mat and the cat is on the mat, the sentence is supported by the kind of evidence ordinary cat-mat speech-communities have organised themselves around. ‘The Earth is six thousand years old’ does not survive in any community that has organised itself around the kinds of evidence geology and astronomy provide, because that evidence pushes back against the sentence too hard for it to remain a load-bearing claim within the community’s web of belief. TThe world is the thing that prevents us from being able to say just anything. What it does not do is select among the sentences that survive its causal pressure because there are many such sentences and the choice between them is up to us.

For questions that are heavily constrained by the world’s causal pushing (most of empirical science, most of ordinary perceptual judgement, most of well-formed mathematics) the question of which sentences to use is overdetermined by the constraints, and there is something close to convergence over time across many different starting points. The sciences are useful for predicting future states of affairs, so scientific statements are true insofar as most people across time find such predictions useful to attain their goals. For questions that are weakly constrained by the world’s causal pushing (most of normative ethics, most of social ontology, most of the questions about how to organise human life) the question of which sentences to use is genuinely up to us, and the differences between traditions and historical periods are correspondingly larger. The interesting questions are mostly of the the second kind. Some questions are heavily underdetermined and some are not and the heavily underdetermined ones include most of the ones we actually fight about.

VII. More Objections Answered

I want to spend a few paragraphs on the strongest objections to the position I have just sketched.

Objection: anti-foundationalist position is self-defeating. If all positions are products of the conditions of holdability that obtain in a given time and place, then so is the anti-foundationalist position. So why should anyone hold it?

This is an objection I take seriously and the one I am least confident I have a fully satisfying answer to. Yes, the anti-foundationalist position is itself a product of conditions, and one should be appropriately humble about it. We are not at the end of history. There may be ways of thinking, currently inaccessible to us, from the vantage of which the anti-foundationalist position will look like a mistake of our particular cultural moment, the way nineteenth-century positivism now looks like a mistake of its moment. I cannot rule this out and I do not see how I could.

What I can say is that the position is available to me now, in a way that the alternative is increasingly not, and that the work of holding it is the work of sustaining a particular vocabulary against the press of competing vocabularies in the present cultural moment. I hold it because it seems to me to make better sense of more of the things I have to make sense of than the alternatives do, and because the costs it imposes on me seem lower than the costs the alternatives impose. This is not a foundationalist defence of anti-foundationalism. It is an anti-foundationalist defence of anti-foundationalism, which is the only kind of defence the position permits and which has the appropriate kind of humility built into it.

I notice that this answer will not satisfy a hard-line foundationalist, and I am not sure it should. The argument between foundationalism and anti-foundationalism has, to a non-trivial extent, the structure of an argument over which language one prefers to conduct the argument in. There is no neutral language in which to settle the meta-question. This is awkward but I do not think it is a defect of the position; I think it is a feature of the situation we are actually in.

VIII. What This Is Good For

I have been going on for some thousands of words now and I want to end with something concrete. What’s the point in all this?

For one, it lets us historicise. The Foucauldian work of figuring out where particular categories came from, what they did, who they helped and who they hurt, and what the alternatives were - this work is much harder to do if one is committed to the categories being descriptions of a fixed reality. The anti-foundationalist position licenses a particular kind of investigation that the foundationalist position does not, and the investigation tends to be illuminating about a wide range of things (mental illness, sexual identity, criminality, race, nationality, the family, the modern state) that are otherwise hard to think about clearly.

It also lets one disagree without metaphysics. A great many disputes that present themselves as disputes about reality turn out to be disputes about how to use a word. The dispute about whether trans women are women is one example I have elaborated on elsewhere; the dispute about whether international law is really law is another one I have elaborated on elsewhere; the dispute about whether Pluto is a planet is a third. The foundationalist framing produces a particular kind of doomed discussion in which each side accuses the other of misdescribing the reality, while the actual disagreement is about which use of the term would best serve which purposes. The anti-foundationalist position lets one see this clearly and have the actual argument rather than the proxy argument.

Lastly, it lets one accept that we might be wrong, that that our own current views are appropriately provisional. If one’s vocabulary is a tool that has been developed under particular conditions for particular purposes, then one’s vocabulary is also revisable, and the project of revising it is part of the ongoing work of being a person who thinks. This is the kind of intellectual humility that lets one hold one’s positions firmly while remaining genuinely open to the possibility that they are mistaken in ways one currently cannot see.

IX. Coda

Going back, finally, to the Febvre puzzle. The reason no one in 1500 could be an atheist is, on the analysis I have just offered, the same reason no one in 2025 can hold any number of positions that may, in 2500, look like obvious truths that everyone of the period was somehow blind to. We are all, in this sense, the medievals of someone’s future.

References

David Couzens Hoy, Foucault: A Critical Reader, 1986.

Ian Hacking, Historical Ontology, 2002.

Karen Armstrong, A History of God, 1994.

Lucien Febvre, The Problem of Unbelief in the Sixteenth Century: The Religion of Rabelais, 1942.

Michael, Foucault, The Order of Things, 1966.

Marc Block, Feudal Society, 1961.

Rayomnd Guess, Genealogy as Critique, 2014.

Talal Asad, Genealogies of Religion, 1993.

1

This is not to “deny truth”, but to deny Truth construed as a substantive metaphysical entity, as opposed to intersubjective agreement.

2

For example, it was not simply the testimonial report of the empty tomb by the disciples that caused Europe to believe that Jesus was resurrected and that God was Jesus incarnate, but this report must be put into context with, for example, the rhetorical genius of St. Paul, the embodiment of his rhetoric in institutions, and the predisposition of a Greek community to take that rhetoric seriously.