A few days ago, someone on a philosophy forum asked whether it is possible to know the truth about God without going through any religion or sacred scripture. He wanted arguments, evidence, and reliable sources, not personal beliefs. It is an honest question, and it has the virtue of honest questions: it is harder than it looks.

The quick answer is yes, since philosophy has been doing exactly that for twenty-five centuries. But that answer conflates two things. One is to think about God without religion, which is perfectly possible and has an enormous tradition behind it. The other is to know God, which demands something more: a path that lets us tell what is true from what we would like to be true.

The thesis I defend is uncomfortable for both camps. Reason alone gets us to a fairly precise point. It can show that the question is serious, that there are weighty arguments and weighty objections, and that neither side has a proof. What it cannot do is hand us a God with a face. And when we try to strip away the face entirely, as Spinoza did, a new problem appears: it is no longer clear that we are still talking about God.

Three Words to Take Apart

The question has three terms, and each one conceals a decision. “Truth” can mean proof, as in mathematics, or the best available explanation, as in the sciences. “God” can be the personal God of the Abrahamic religions, a first cause without a biography, or nature as a whole. And “without religion” can mean without scripture, or also without religious experience, which is not the same thing.

The discipline that deals with this has a name: natural theology. It is the attempt to find out what can be known about God using only reason and what we know of the world, setting revelation aside (SEP, Natural Theology and Natural Religion). It is neither a modern invention nor a skeptics’ preserve. Aristotle, Thomas Aquinas, Leibniz, and Hume all worked in it, each reaching different conclusions.

It is worth setting a criterion at the outset. An argument in natural theology is not judged by who made it or by the tradition it comes from, but by its premises and by the validity of the step it takes. That criterion will return at the end, when the temptation arises to dismiss ideas because of where they came from.

What Science Cannot Weigh

Science works under a ground rule called methodological naturalism: it explains natural phenomena by natural causes. This is not a thesis about what exists, but a decision about how to investigate. That is why a laboratory can neither confirm nor refute God, just as a thermometer cannot measure justice.

Put differently, science does not settle the question, but it supplies philosophy with the data it works on. Cosmology tells us that the observable universe had a hot, dense initial state; physics tells us that certain constants fall within narrow ranges compatible with complex chemistry. What those data mean is no longer physics. It is metaphysics, and that is where the work begins.

Methodological naturalism must be kept distinct from ontological naturalism, which is a thesis: that nothing exists beyond nature. Many scientists hold it, but it does not follow from the method. It is a philosophical position that has to be defended with arguments, just as theism does.

The Arguments Still Standing

None of the classical arguments cites scripture. All of them start from premises anyone can examine. That is their strength, and it is also where they are attacked.

The argument from contingency. Everything we see might never have existed. Leibniz, in the Monadology (§§ 32–38), grounds the argument in the principle of sufficient reason: nothing is without a reason why it is so and not otherwise. A chain of contingent things, even an infinite one, does not explain why the chain as a whole exists. What is needed is a necessary being, one that carries within itself the reason for its own existence. Alexander Pruss, in The Principle of Sufficient Reason (2006), is today its most careful defender (SEP, Cosmological Argument). The classic objection comes from Hume, and Russell repeated it in his 1948 radio debate with Copleston: to explain each member of a series is already to explain the series, and to ask for more is to commit the fallacy of composition. The other way out is to deny the principle of sufficient reason in its strong form. The argument lives or dies with that principle.

Fine-tuning. The constants of physics appear to be calibrated within very narrow ranges that allow for stable atoms, stars, and organic chemistry. For Richard Swinburne, in The Existence of God (2004), this is more probable if there is a designer than if there is not. There are three replies. The multiverse: if there are a great many universes, one of them was bound to be habitable. The selection effect: we can only observe universes in which observers can exist. And the most technical, from Timothy McGrew, Lydia McGrew, and Eric Vestrup in Mind (2001): if the possible range of a constant is infinite, the principle of indifference would require a uniform distribution that cannot be normalized, and without one there is no way to say that the life-permitting range is “improbable.” The technical crux is the normalization problem. Assigning equal probabilities over an unbounded interval violates Kolmogorov’s axioms: a constant density over an infinite range integrates to zero or to infinity, never to one. And if one forces the calculation anyway, the result becomes absurd, because any finite range, however wide, receives probability zero, so that a universe only “coarsely tuned” would be just as improbable as a finely tuned one. One can use non-uniform distributions, such as a Gaussian or an exponential, but fixing any of them requires choosing a mean or a scale, and that choice already presupposes the very plausibility the argument set out to establish. It is precisely this arbitrariness that the argument cannot justify (SEP, Fine-Tuning). My impression is that this last objection is the least known and the most serious.

The ontological argument. Anselm formulated it in the Proslogion, and Descartes took it up again in the fifth of his Meditations: if God is the most perfect being conceivable, and existence is a perfection, then God exists. Kant replied that existence is not a real predicate: to say that something exists does not add a property to it, but asserts that the concept has an object. Alvin Plantinga recast the argument in modal logic in The Nature of Necessity (1974): if it is possible that a maximally great being exists, then that being exists. The trouble is that the possibility premise is exactly what the skeptic will not grant (SEP, Ontological Arguments).

The moral argument. If there are objective moral obligations, something must ground them, and a world of purposeless particles does not seem up to the task. Robert Adams defends this in Finite and Infinite Goods (1999). The reply comes from secular moral realism, which holds that moral facts can be objective without a lawgiver, and from the old dilemma of the Euthyphro: is something good because God wills it, or does God will it because it is good? (SEP, Moral Arguments for the Existence of God).

The Objections That Won’t Go Away

On the other side there are arguments of equal seriousness, and they deserve to be presented with equal fairness.

The problem of evil. J. L. Mackie, in “Evil and Omnipotence” (Mind, 1955), argued that an omnipotent and perfectly good God is logically incompatible with the existence of evil. Plantinga answered with the free will defense, and most philosophers today accept that the logical version does not succeed. But William Rowe, in 1979, replaced it with a probabilistic version that is still very much alive: evil need not be incompatible with God; it is enough that there be suffering that seems to serve no purpose at all. His example was a fawn dying in agony for days in a forest fire, unseen by anyone (SEP, The Problem of Evil).

Divine hiddenness. J. L. Schellenberg, in Divine Hiddenness and Human Reason (1993), put forward a subtler argument. If a perfectly loving God existed, he would want a relationship with every person capable of having one. That requires that such a person be able to believe God exists. Yet there are people who search sincerely and find nothing. Therefore, he concludes, such a God does not exist (SEP, Hiddenness of God). The argument carries an irony: the forum question is itself an instance of what Schellenberg describes.

Naturalism as the better theory. Graham Oppy, in Arguing about Gods (2006), does something more modest and harder to evade. He does not try to prove that God does not exist. He shows that no argument, on either side, ought to persuade a reasonable person who starts from the opposite position. He then compares theism and naturalism as two theories of the world: if they explain the same data, the one that posits less wins. For Oppy, that is naturalism, because the universe can be the final brute fact without anything further behind it.

The genealogy of belief. The cognitive science of religion proposes that the human mind comes equipped with a hypersensitive agency detector: when there is a rustle in the grass, assuming a predator was cheaper than ignoring it. That would be the source of our tendency to see intentions behind nature (SEP, Cognitive Science of Religion). So far, this might read as a refutation. It is not. Explaining why we believe something does not show that it is false; if it did, the evolutionary psychology of perception would also refute our belief in the external world. Justin Barrett, one of those who did most to develop the idea of the detector, is a theist. What the genealogy does accomplish is to call for suspicion: if a belief can be explained without its object, the burden of proof shifts.

Kant: The Limit and the Side Door

If any philosopher took the forum’s question seriously, it was Kant. In the Critique of Pure Reason (1781/1787), in the chapter on the ideal of reason, he reviews the three classical proofs: the ontological, the cosmological, and the argument from design. His diagnosis is that the latter two ultimately depend on the first, and that the first fails because existence is not a predicate. Our understanding can know only what appears in space and time. God, by definition, does not appear.

It is tempting to call this agnosticism, and people often do. But it is an oversimplification. Kant says in the preface to the second edition that he had to “deny knowledge in order to make room for faith” (B XXX). In the Critique of Practical Reason (1788), God returns as a postulate: if the moral law obliges us to pursue the highest good, and that good includes happiness proportioned to virtue, we must presuppose a guarantor of that proportion (SEP, Kant’s Philosophy of Religion). This is not knowledge. Kant calls it rational faith.

In other words, Kant closes the door of theoretical reason and opens a side door, that of practical reason. For someone asking the question “without religion,” the lesson is twofold. Proof is not on offer. But that does not leave the question to arbitrariness: it leaves open the possibility of a reasoned belief, one that knows it is not knowledge.

Spinoza: A God Without a Face

If the problem with the gods of the religions is that they look too much like us, Spinoza offers the most radical solution. In the Ethics (1677), God is the one substance, “a being absolutely infinite,” consisting of infinite attributes (I, def. 6). He has no will that might change its mind; he neither rewards nor punishes; he has no preferences. Deus sive Natura: God, that is, Nature. In 1929, when a rabbi asked Einstein whether he believed in God, Einstein replied that he believed in Spinoza’s God, the one revealed in the harmony of all that exists, and not in one concerned with the deeds of human beings.

That disposes of dogma and anthropomorphism in a single stroke. But it raises an uncomfortable question: without personality or intentions, how does that substance differ from what we today simply call the cosmos? Schopenhauer had already remarked that pantheism is only a polite form of atheism.

Spinoza has two answers. The first is the distinction between natura naturans and natura naturata (I, 29, scholium): God is not the collection of things produced, but nature as producing, the activity from which everything follows. The second concerns the attributes. Physics studies extension; the human mind, which for Spinoza is the idea of the body (II, 13), knows only thought besides. But God has infinitely many attributes. Therefore God exceeds everything physics will ever be able to describe.

Here lies the point that interests me. If we have access to only two attributes, the infinitely many others are, for all practical purposes, unverifiable. Doesn’t that reduce the difference from modern naturalism to a merely formal one?

The specialists disagree, and the disagreement is instructive. Edwin Curley, in Spinoza’s Metaphysics (1969) and Behind the Geometrical Method (1988), proposed reading the dependence of things on God not as that of a property on its subject, but as causation. On his reading, the infinite modes are the ontological expression of the laws of nature, and Spinoza’s God comes very close to the lawful structure of the universe (SEP, Spinoza’s Modal Metaphysics). This is the reading that vindicates Einstein. Behind the Geometrical Method is, moreover, the book that best unravels the Ethics: it explains why Spinoza chose the geometrical order of definitions, axioms, and propositions, and reads the work as a critical dialogue with Descartes and Hobbes. From Descartes, Spinoza takes the vocabulary of substance, attribute, and mode, and uses it to reach conclusions Descartes would have rejected; from Hobbes, he takes mechanistic rigor, but without the materialism.

Michael Della Rocca, in Spinoza (2008), puts the principle of sufficient reason at the center: for Spinoza there are no brute facts; everything must be explicable. From that standpoint, the infinite attributes are not an empirical hypothesis awaiting confirmation, but a demand of reason: there would be no reason for substance to have some attributes and not others. That we cannot see them says something about our minds, not about reality. Curiously, Della Rocca also calls Spinoza a naturalist, in the sense that everything in nature obeys the same principles (review of the Oxford Handbook of Spinoza). The difference from Curley, then, is not naturalism versus mysticism, but two naturalisms: one grounded in the laws that science discovers, and another grounded in what reason demands. Yitzhak Melamed, for his part, argues that Della Rocca pushes the principle further than Spinoza himself ever did (Melamed, Spinoza’s Metaphysics: Substance and Thought, 2013).

There is one more thing, and it is what tends to be forgotten. Spinoza does offer a way of knowing God. It is the third kind of knowledge, scientia intuitiva, which grasps things insofar as they follow from the nature of God (V, 25), and which culminates in the intellectual love of God (V, 32–36). It is neither revelation nor mystical experience. It is understanding something so thoroughly that one sees it in its necessity. My suspicion is that, for someone asking the question “without religion,” this is the most honest answer in the entire tradition, at a price: the God one comes to know this way does not hear prayers.

What Remains Standing

Let me return to the forum’s question with what we have gathered. Is there convincing evidence? There are weighty arguments on both sides and no proof; anyone who says otherwise, in either direction, is selling something. Can science settle it? No, though it gives philosophy the material it argues over. Can God be separated from culture? Only in part: the more the concept is purified, the fewer human features it retains, and at the limit, Spinoza appears.

There is a temptation worth resisting, and that is to settle everything by origins. “That idea of God is a cultural invention” or “it is the product of a cognitive bias” does not prove it false, any more than “this argument was written by an atheist” proves it bad. That is the genetic fallacy, and it is committed just as comfortably from faith as from skepticism. The only honest filter is the one I proposed at the outset: premises and validity.

Can we know God without religion? My impression is that the question changes shape once one has traveled its full length. Kant says God cannot be known, but can be believed in for good reasons. Spinoza says God can be known, but the God one comes to know is not the one we were looking for. Between the two lies the place of the one who asks. Perhaps the pertinent question is not whether God exists, but what we would still be willing to call “God” after stripping away everything we ourselves put there.

Sources

Classical Works

  • Anselm of Canterbury, Proslogion (1078), chs. 2–4.
  • Descartes, Meditations on First Philosophy (1641), Fifth Meditation.
  • Spinoza, Ethics, Demonstrated in Geometrical Order (1677), Parts I, II, and V.
  • Leibniz, Monadology (1714), §§ 32–38.
  • Hume, Dialogues Concerning Natural Religion (1779).
  • Kant, Critique of Pure Reason (1781/1787), “The Ideal of Pure Reason”; Critique of Practical Reason (1788), “Dialectic.”

Contemporary Works

  • Adams, Robert M., Finite and Infinite Goods, Oxford University Press, 1999.
  • Barrett, Justin L., Why Would Anyone Believe in God?, AltaMira, 2004.
  • Curley, Edwin, Spinoza’s Metaphysics: An Essay in Interpretation, Harvard University Press, 1969.
  • Curley, Edwin, Behind the Geometrical Method: A Reading of Spinoza’s Ethics, Princeton University Press, 1988.
  • Della Rocca, Michael, Spinoza, Routledge, 2008.
  • Mackie, J. L., “Evil and Omnipotence,” Mind 64 (1955); The Miracle of Theism, Oxford University Press, 1982.
  • McGrew, Timothy; McGrew, Lydia; Vestrup, Eric, “Probabilities and the Fine-Tuning Argument: A Sceptical View,” Mind 110 (2001).
  • Melamed, Yitzhak Y., Spinoza’s Metaphysics: Substance and Thought, Oxford University Press, 2013.
  • Oppy, Graham, Arguing about Gods, Cambridge University Press, 2006.
  • Plantinga, Alvin, The Nature of Necessity, Oxford University Press, 1974.
  • Pruss, Alexander, The Principle of Sufficient Reason: A Reassessment, Cambridge University Press, 2006.
  • Rowe, William, “The Problem of Evil and Some Varieties of Atheism,” American Philosophical Quarterly 16 (1979).
  • Schellenberg, J. L., Divine Hiddenness and Human Reason, Cornell University Press, 1993.
  • Swinburne, Richard, The Existence of God, 2nd ed., Oxford University Press, 2004.

Stanford Encyclopedia of Philosophy

  • [Natural Theology and Natural