Listen to this essay
22 minute listen
How tightly should you hold your beliefs? You face the question several times a week without noticing. Hold them too loosely and you don’t stand for anything, hold them too tightly and you become the sort of person no argument can reach. Most of us swing between the two and have no vocabulary for the ground in between. A Roman lawyer worked one out, under pressure far worse than anything we are likely to face, and it is still the best answer available.
The time is early 45 BCE, the place a villa near Lucrine Lake, northwest of Naples, in southern Italy. One of the most fascinating and controversial figures in the entirety of Western history, Marcus Tullius Cicero, is finishing writing a book that will be known as the Academica, where he explains the nature of sceptical philosophy and why it is so crucial not just to the advancement of human knowledge, but to living a good life.
It is a strange moment to retire into one’s villa to write. The Roman Republic, so beloved by Cicero, is about to be destroyed by aspiring autocrats: Julius Caesar, a longtime friend and now a political opponent, has declared himself dictator for life, a move that will soon lead to his assassination on the Ides of March 44 BCE. On the personal front, Cicero has recently divorced Terentia, his wife of 30 years, and a few months ago his beloved daughter Tullia died from complications following childbirth. You would think a man whose private and public worlds are collapsing so dramatically would hardly have the focus to write philosophy. Yet for Cicero such writing is a form of therapy, and the treatises he produces will influence medieval Christians, Renaissance authors, and the Founders of the American Republic.
Cicero managed all this because of his adopted philosophy, ‘Academic Skepticism’, so-called because it characterised, for the period from 266 to 90 BCE, the Academy originally founded by Plato. He used to be treated as a mere populariser, a gifted Roman relaying Greek ideas he did not really own, but scholars abandoned that uncharitable verdict decades ago as it became clear that Cicero argued and decided among his sources rather than simply transcribing them. He also built a Latin philosophical vocabulary from scratch, and much of the abstract terminology of European thought descends from his coinages. On the question that mattered most to his own school – whether a sceptic may hold opinions, and on what terms – his position was his own and differed from those of his teachers. What is true is that he philosophised in public, and tested his conclusions in the courts, in the Senate, and eventually against a death sentence. That is not popularisation: it’s philosophy done at street level.
Which is why he is worth your time now. What follows is partly the story of a man, but mostly a usable method: a way of holding beliefs that is neither the performed humility of the professional doubter nor the armoured certainty of the person who has ‘done their own research’. Cicero had a name for it – probabilia – and I think you will find it very useful any time you have to assess the validity of a notion and make a decision about it.
A genuine sceptic, confronted with compelling new evidence, updates – whereas the contrarian does not
‘Sceptic’ comes from the Greek skeptikos, meaning enquirer, and has a long and serious pedigree in philosophy. Yet, since the 1990s, the term has been hijacked by science-denial movements, giving us oxymorons like ‘climate sceptic’, ‘vaccine sceptic’, and ‘evolution sceptic’.
These movements understood that ‘sceptic’ has cultural prestige: it sounds rigorous, independent-minded, and anti-establishment, so they colonised it. The move was brilliant and corrosive, as it reframed denial as enquiry. The difference shows in practice: a genuine sceptic, confronted with compelling new evidence, updates – whereas the contrarian does not: they shift goalposts, raise new objections, or ignore what they cannot explain, all while citing studies selectively and appealing to the handful of dissenting experts cherrypicking the evidence. Their doubt is not the productive Socratic doubt that opens enquiry but a rhetorical shield that forecloses it. These self-styled sceptics are in fact dogmatists, having reached their conclusions in advance and thereafter simply engaging in rationalisations.
On the other side, a certain strand of scientism has overcorrected by demanding certainty as the price of admission for any belief worth holding, deploying ‘follow the science’ as a conversation-stopper rather than an invitation to enquiry. That is not how real science works either.
Both pathologies make the same mistake: they treat doubt and certainty as the only two options. Neither leaves room for what we actually need: calibrated confidence, the ability to say This is the best available account, I will act on it, and I remain open to being wrong. What is missing, then, is a philosophy of provisional but committed belief, a philosophy Cicero already worked out, in Latin, 2,000 years ago.
It began with a visit to Rome, in 87 BCE, by the then-head of the Platonic Academy, Philo of Larissa, having fled Athens amid political upheaval that eventuated, ironically, in the Roman sack of the city. Cicero, 19 years old and already a prodigy in oratory and law, went to hear Philo and was transfixed. What the visitor taught was not the abstract metaphysics of Plato but something more practical: a method based on the Socratic tradition of enquiry, the radical idea that wisdom begins in recognising the limits of one’s own knowledge.
The Academic Skeptics called their school the New Academy, to distinguish it from the Old one started by Plato. Arcesilaus and Carneades – the great sceptics of the 3rd and 2nd centuries BCE – had pushed Plato’s and Socrates’ dialogical method toward a radical conclusion: we are unlikely to achieve certainty, because human knowledge comes from our senses and from reason, both of which are fallible.
Crucially, they did not conclude that we should therefore believe nothing whatsoever. Carneades introduced the concept of the pithanon (a Greek word) – the persuasive, the plausible – as the proper standard for action. Cicero translated the word into Latin as probabile (plural probabilia, from which we get the English ‘probable’) and built his whole epistemology on it.
What Cicero deployed was not a generic mood of open-mindedness but a graded, checkable procedure
Cicero’s own formulation is plain. In the Academica, he separates his school from the dogmatists in a single sentence: others declare that certain things are settled and fixed, whereas ‘we hold many things to be probabilia, which we can easily follow but can scarcely affirm.’ Follow easily, affirm rarely. The entire method is encapsulated by that asymmetry.
The term probabilia does not mean ‘probable’ in the modern, quantitative statistical sense, but something closer to ‘worthy of approval’: a claim that has withstood scrutiny, coheres with the available evidence, and a reasonable person would assent to while remaining open to revision. It is enough to act on.
Nor is ‘probabilia’ a single, undifferentiated category, and this is the important part usually left out. Carneades graded impressions (ie, preliminary judgments) by how much scrutiny they had survived. At the first level, an impression simply strikes you as a bit more likely to be true than not. At the second level, it rises to being persuasive. At the third, it is ‘undiverted’, meaning that nothing else you believe conflicts with it. At the fourth and highest level, it is also thoroughly examined, essentially the result of scientific investigation. Trivial matters need only the first or second grade; a medical decision or a criminal verdict demands the third or fourth. What Cicero inherited and deployed, then, was not a generic mood of open-mindedness but a graded, checkable procedure.
That grading is where the Academic parts company with the other great sceptical school of antiquity, Pyrrhonism, which counselled suspending judgment on most issues and otherwise acting according to how things appear to be while following local customs. There are three differences between the two approaches: first, the Academic assents, in a weakened but real sense: they approve the claim, argue for it, and carry the responsibility if it proves wrong, whereas the Pyrrhonist withholds approval and merely goes along. Second, the Academic ranks: appearances are sorted by how much testing they have survived, and some genuinely turn out to be better than others, while to the Pyrrhonist appearances arrive with no comparative weight attached – one is just as good (or bad) as another. Third, the two schools want different things: the Pyrrhonist pursues ataraxia, meaning tranquility, and suspension of judgment is his means to it. The Academic is still after truth; they simply doubt they will ever be able to certify that they have reached it. One is a therapy for anxiety, the other is a research programme conducted by a fallible animal that nevertheless needs to act in the world.
Cicero is explicit in the Academica: the wise person acts on the best available view, not because they are certain it is true, but because suspending all judgment would make life impossible. We are not the kind of people who think nothing is true, say the Academics; we are the people who say that some falsehoods are mixed in with truths, so similar to them that we have no sure criterion for judging, so we can assess only provisionally.
Strip away the Latin and the toga, and the Academic method comes down to a handful of habits, all of which can be practised, starting today:
- Say how confident you are, not just what you think. ‘I believe X’ is nearly useless. ‘I believe X, at about 7 out of 10, and here is what is holding back the other 3’ gives your interlocutor something to work with and gives you somewhere to move in the light of new evidence or better arguments.
- Grade your beliefs by how hard they have been tested. Most of what any of us believes sits at Carneades’ second level: it merely struck us as persuasive. Reserve confident action for what has been through the third and fourth levels.
- Separate your values from your empirical beliefs. That children should not go hungry is a value and may be held firmly. That a particular school-meals programme reduces hunger is an empirical claim, to be held only as firmly as the evidence warrants. Conflating the two is what makes people defend failing policies as though their integrity depended on it.
- State in advance what would change your mind. If nothing would, you do not have a belief, you have an identity.
- Argue the other side properly. The Academy’s signature exercise was arguing in utramque partem, on both sides of a question. Carneades notoriously did it in Rome in 155 BCE, defending justice one day and demolishing it the next. The point was not cynicism but that you do not truly know what you believe until you have built the strongest case against it yourself.
- Act anyway. The sceptic who treats uncertainty as a reason to postpone action has missed the point entirely. Uncertainty is the permanent condition of humanity, deciding effectively within that is a skill we can all benefit from.
- Treat being wrong as data rather than disgrace. If you follow these sceptical rules of thumb, you will change your mind from time to time – or at least ought to. The question is whether you experience that as humiliation or as the method working.
None of this was merely theoretical for Cicero. It informed every major decision of his life, including the most contested one.
As Consul – the highest office in Republican Rome – in 63 BCE, he faced a conspiracy to overthrow the government led by a disgruntled senator, Lucius Sergius Catilina, who was in the field threatening Rome with an army. Cicero gathered what intelligence he could and acted on probabilia. He laid the evidence before the Senate, which passed the senatus consultum ultimum, an emergency decree authorising the consuls to take whatever measures the safety of the state required. Cicero then put the fate of the captured conspirators to a full senatorial debate, in which Caesar argued for imprisonment and Cato the Younger for death. On the constitutional understanding that Cicero and the Senate majority held, the executions that followed were lawful. That understanding was contested, then and since: citizens had a right of appeal, and several years later Cicero was driven into exile by a (retroactively applied) law framed precisely to assert that he had violated it. Nothing here settles whether he was right, and I am not arguing that he was. What is instructive is his method of arriving at a conclusion: he acted on a judgment he knew might be mistaken, claimed no certainty he did not have, and then spent the rest of his life defending the decision rather than disowning it. That is Academic Skepticism under fire – not paralysis, nor reckless certainty, but consequential commitment to the best available judgment, including its costs.
Cicero knew he might be signing his death warrant, but the probable good outweighed the probable cost
Once removed from public life by Caesar’s dictatorship, Cicero retreated to his estates and wrote philosophy at extraordinary speed: De natura deorum, De divinatione, Academica, De finibus bonorum et malorum and the Tusculan Disputations, among others. De divinatione applies Academic Skepticism to superstition and mysticism, and is now considered the first treatise ever written on the difference between science and pseudoscience. His method there is recognisable to any working scientist. He does not dismiss out of hand the augurs and haruspices who read entrails and interpret the flight of birds. He examines them: what is the mechanism by which an animal’s liver predicts a battle’s outcome? Is there a genuine pattern of successful prediction, or only cherrypicked successes remembered after the fact? On each count, the evidence fails to compel assent. Crucially, he leaves open the possibility that the gods communicate through omens, concluding only that the evidence falls short of the standard he would demand anywhere else. Two thousand years before Carl Sagan’s dictum that extraordinary claims require extraordinary evidence, Cicero was practising the same discipline and applying it to cherished Roman institutions. That is intellectual courage, not academic hedging.
The letters he exchanged with his friend Atticus in the months before his death are the most instructive documents Cicero left. Knowing that Mark Antony wanted him dead, he debated whether to flee, whether to return to Rome, whether to trust the young Octavian (the future emperor Augustus): reasoning under uncertainty, weighing probabilities, deciding anyway. When he returned and resumed attacking Antony in the discourses known as the Philippics, he knew he might be signing his death warrant, but the probable good outweighed the probable cost. He was murdered on 7 December 43 BCE. His last words, to the legionnaire sent to kill him, were: ‘I am stopping here. Come over, soldier. There is nothing right about what you are doing, but at least make sure you cut off my head properly.’
Is the Academic Skeptic approach worth revisiting and applying today? Very much so, at least in three areas.
First, science – contrary to what you may have heard – does not run on certainty. Scientific knowledge is not a collection of established facts, it is a provisional consensus of the best-tested theories, each held subject to revision. Thomas Kuhn showed that scientific paradigms shift over history, and Imre Lakatos argued that scientists rightly persist with research programmes despite anomalies they cannot explain, revising peripheral hypotheses before abandoning a theory’s core. What science practises is structurally identical to Cicero’s probabilia: the best current views, held with proportionate confidence, open to revision. Modern Bayesian probability theory (elaborating the theorem Thomas Bayes set out in the 18th century) formalises the same insight: priors updated by evidence, never certainty, always proportionality. The mathematics is new, but the wisdom is not.
This is not relativism, the view that theories are mere social constructs. Quite the opposite: modern philosophy of science describes a process by which science converges on progressively better accounts of reality, not by marching from certainty to certainty but by eliminating worse hypotheses and refining better ones. The scientist, like the Academic Skeptic, never claims final truth, but it doesn’t follow that anything goes. Interestingly, the sociologist Robert K Merton in 1942 described science as a kind of ‘organised scepticism’.
Second, democracy requires the willingness to be persuaded. Deliberation is possible only if citizens hold their political views as probabilia rather than as dogma, and the current political crisis is partly an epistemological one. Of course, there was never a golden age of reasonable citizens. But the measurements are unambiguous about the direction in which we have been moving: roll-call analyses of the US Congress show ideological overlap between the parties shrinking from the 1970s and effectively gone by the 2000s, while survey work by Pew and others documents a sharp rise since the 1990s in partisans holding consistently ideological positions and viewing the other side as a threat to the nation. Much of the change, moreover, is affective rather than programmatic: partisans have diverged on actual policy far less than they have come to dislike one another, which is exactly the sort of commitment that evidence cannot touch, and social media and the fake-news economy have accelerated it.
Developing a crystallised self that never updates is not integrity: it is rigidity dressed up as character
What would a genuinely probabilist citizen look like? They would hold strong political commitments, since probabilia are not wishy-washy and can be acted on with conviction. But they would distinguish their values from their empirical beliefs about how the world works. Suppose someone supports a universal basic income. The value underneath it – that no one should live in destitution in a wealthy society – is theirs to hold firmly. That a basic income is the best instrument for delivering it, however, is an empirical bet, and the trials now running in several countries might prove that bet ill-founded. The probabilist citizen would then be capable of saying, without shame: ‘I thought this would work; it does not appear to; therefore I am updating my view.’ That is not ‘flip-flopping’, it is exactly the cognitive posture that makes democratic persuasion possible. It is in fact a form of consistency: following the evidence wherever it may lead.
Third, on a personal scale, Cicero’s epistemology is an argument against the seductions of fixed identity. The person who has ‘found their truth’ and stopped questioning is not wise; they have surrendered to what Simone de Beauvoir, in The Prime of Life (1960), called ‘that conscious venture into lunacy known as a “philosophical system”,’ from which its adherents derive an obsessional attitude endowing their tentative patterns with universal insight and applicability. The well-examined life is one of continuous calibration and revision: strong-enough commitments to act with integrity, loose enough to update when the evidence changes.
This is countercultural in an age that prizes authenticity in the sense of having a fixed self, and urges us to own our identity. Cicero would have found that to be self-congratulatory. He changed his mind about friends, alliances and tactical judgments, and worried about it, as his letters make plain. But developing a crystallised self that never updates is not integrity: it is rigidity dressed up as character. The Ciceronian person is recognisable by a willingness to be embarrassed by earlier certainties, and by treating that embarrassment not as failure but as evidence that enquiry is working.
He was wrong about the Republic – it fell anyway. But he went on acting as though it could be saved, because that was the best available view on his analysis and worth acting on, and his legacy survived in the books he wrote in the shadow of Antony’s thirst for revenge.
The genuine sceptic is not the person who doubts everything, nor the person who demands proof before believing anything. It is the person who has learned to hold their best views with exactly the right strength of grip: firmly enough to act on them, loosely enough to let go when the evidence turns. That is what probabilia means. It is harder than dogmatism and harder than nihilistic doubt, but Cicero thought it the only honest stance available to a finite human being. He was probably right.