What God Is Claimed to Be

A physicist and a theologian argue about whether God exists. The physicist is denying that an invisible agent intervenes in the causal order, adjusting outcomes that physics would otherwise settle. The theologian is asserting that everything which exists depends for its existence on something that does not depend on anything. Neither claim contradicts the other. Both parties go home satisfied.

Nothing in this subject can start until the disputed claim has been fixed, and fixing it is more work than it looks. The word carries a specific package of commitments in the philosophical tradition this subject works in, and the package has parts that pull against one another.

The claim this subject is about

Several features of that formulation are doing work.

It is singular. Whatever else the claim is, it is not the claim that there are powerful beings somewhere, which almost nobody disputes in any interesting sense and which no argument in this subject is designed to establish.

It is about a being on which everything else depends, rather than the most impressive item within an inventory of things. That is why establishing that some very powerful entity exists inside the universe would not be establishing this.

And it attributes perfections rather than large quantities. The claim is not that God is very knowledgeable but that God's knowledge cannot be exceeded, and the difference matters because most of the technical difficulty in this subject comes from the maximality rather than from the magnitude.

That conception took its shape from a specific historical joining, in which the God who acts in the history recorded by scripture was identified with what Greek philosophy had described as perfect and ultimate. Almost everything argued about in this subject inherits its terms from that joining, including the problems.

One warning about the phrase. Some writers use classical theism in a stricter sense, for the doctrine that God has no parts, no temporal location, and no states brought about from outside, in explicit contrast to a conception of God as one very impressive being alongside others. On that stricter use, a great deal of contemporary philosophy of religion is not classical theism at all. The advanced tier takes that dispute up directly. Until then the broader use is intended.

Where the attributes come from

The list is not arbitrary and it is not simply read off scripture. It is generated by a method, and knowing the method explains both the list and its instability.

Anselm, writing in the Proslogion around 1077, characterised God as that than which nothing greater can be conceived. Taken as a method rather than as a proof, this yields a procedure: for any property, ask whether a being that has it would be greater than an otherwise similar being that lacks it, and if so, attribute it to God. Power, knowledge, and goodness come out immediately. So do a range of properties the beginner tier sets aside, including having no parts, having no temporal location, and being incapable of change.

The method has two weaknesses that are worth seeing at the start rather than at the end.

It assumes there is a single ordering of greatness on which any two candidate properties can be compared. Whether being incapable of change is greater than being responsive to what happens is not obvious, and different traditions have answered it differently.

It assumes the resulting properties can all be held at once. That the greatest conceivable being has every property it is better to have does not follow if some of those properties exclude others, and much of the intermediate and advanced work in this subject consists of arguments that some of them do.

Maximal power

The immediate question is what maximal power includes, and the standard answer is narrower than the word suggests.

Aquinas held that what implies a contradiction does not fall within omnipotence at all, and his reason was not that God's power runs out at that point. It was that a self-contradictory description does not pick out anything to be done, so there is nothing there for power to fail at. On this account it is more accurate to say that such things cannot be done than that God cannot do them. Maimonides took the same line. It remains the majority view.

Descartes appears to have taken the other side. In letters to Mersenne in 1630, beginning with one of 15 April, he held that God freely established the eternal truths, including those of mathematics, so that they depend on God's will rather than constraining it. The position is genuinely more radical, and its cost is that it makes it very hard to say what is being asserted when the impossible is said to be within God's power.

The best-known puzzle here asks whether God can create a stone too heavy for God to lift. Answer yes and there is something God cannot do, namely lift it. Answer no and there is something God cannot do, namely make it. The standard reply distinguishes a being that happens to be omnipotent from a being that is omnipotent necessarily. For the second, a stone unliftable by an omnipotent being is not a possible object, so failing to be able to make one is not a limitation any more than failing to be able to make a round square is. Whether that reply succeeds, and what it costs, belongs to the intermediate tier.

A separate question is whether maximal power includes the power to act wrongly. If it does, then a being that cannot act wrongly is not omnipotent. If it does not, then omnipotence has been defined partly in terms of goodness, and the attributes are less independent than the list makes them look.

Maximal knowledge

That definition is clean and it generates two pressures immediately.

The first concerns the future. If God knew eighty years ago that you would do what you will do tomorrow, then a fact about the past, namely what God believed, settles what you will do. The past is not something anyone can now affect. So it looks as though tomorrow's action is already unavoidable, which is not what anyone means by a free action. Nelson Pike set this out precisely in "Divine Omniscience and Voluntary Action" in The Philosophical Review in 1965, and the argument is stronger than the older versions because it does not rest on any confusion about necessity. Every available response gives something up: the fixity of the past, or the inference from God's belief to its content, or the claim that God knows the future free acts at all. The advanced tier works through them.

The second concerns knowledge that seems tied to a position. Norman Kretzmann argued in the Journal of Philosophy in 1966 that knowing what time it is now requires changing as the time changes, so that a being who never changes cannot know it, and an omniscient being must therefore change. The same structure applies to first-person knowledge. What you know when you know that you are the one in the hospital is not obviously the same thing anyone else can know, and if it is not, then either omniscience does not require it or God's knowledge includes something that looks impossible for a being who is not you.

Neither pressure is a refutation. Both show that the clean definition is doing more than it appears to.

Perfect goodness

The hard question here is what kind of goodness is meant, and it is easy to miss because the word looks familiar.

On one reading, God is a moral agent of the same general kind as a human being, subject to the same standards, and simply meets them perfectly. This is the reading that makes the problem of evil bite, since a perfectly good agent who could prevent a child's suffering and does not has some explaining to do. It is also the reading that makes the word informative. Mill pressed the point that if the goodness ascribed to God is not what we mean by goodness, then calling God good says nothing.

On another reading, associated with much medieval theology, God is not a moral agent at all. God is the standard of goodness rather than a very good instance of it, and welfare-directed obligations of the sort that bind us do not apply to a being who is not one of us and owes us nothing. This reading dissolves some problems and creates others, the most obvious being that it makes the praise offered to God difficult to interpret.

Contemporary work has not settled which reading the tradition is committed to, and a good deal of argument in this subject turns out to depend on the answer without saying so.

A God with a mind

This attribute is easy to overlook because it is presupposed rather than argued for, and it is what most of the arguments in the intermediate tier actually try to establish or refute.

Its significance shows up by contrast. Several major religious and philosophical traditions posit an ultimate reality that is not a person. Brahman as understood in Advaita Vedanta and the Dao of the Daoist texts are both explicitly not agents with plans, and describing either as having intentions misdescribes them. So the question of whether the ultimate is personal is a real fork, not a detail, and the last article in this tier returns to what happens to the standard arguments when the target changes.

The attribute also generates the deepest internal strain in the classical package. Push the perfections hard and the result grows less person-like. A being with no parts has no distinct beliefs and desires to combine. A being outside time does not deliberate, since deliberating takes time and ends in a decision that was not made before. A being incapable of being affected does not respond to anyone. Each of those attributes was reached by the method described above, and each subtracts something from personhood. The tradition's answer is that our words apply to God analogically rather than in the sense they bear of us. Whether that answer preserves the content of the claim, or quietly removes it, is one of the live disputes in the advanced tier.

Next

This article has fixed what is being claimed. It has not said anything about whether the claim is true, or about what the alternatives to accepting it are. The next article separates three positions that are routinely confused, and shows that the disagreement between them is not a disagreement about how confident anyone feels.