Feats of engineering have made, and continue to make, life better for us. We cannot help but marvel at their intellectual and physical ingenuity and give thanks for their efforts. Yujin Nagasawa's Maximal God: A New Defence of Perfect Being Theism is a feat of engineering, at least in my view. Not in terms of either its literary or physical qualities, though it is both well written and well produced, as is characteristic of both Nagasawa and of Oxford University Press, respectively, but rather in terms of its central aim, that of conceptual engineering. Now, Nagasawa himself does not describe the aim of the book in this way, and so I hesitate to call it as such, but I think that this is the best way to understand and evaluate the book. Let me explain.

I will begin with Nagasawa's stated aim of the book, namely, to defend perfect being theism, as the subtitle of the book might suggest. Now, perfect being theism, according to Nagasawa, ‘is a form of theism based on Anselm's concept of God. According to perfect being theism, God, as the being than which no greater is conceivable or metaphysically possible, exists’ (2). Given this conception of perfect being theism, Nagasawa writes:

To carry out this aim, in Part I of the book, Nagasawa gives a detailed analysis of perfect being theism, including a sophisticated account of just how God might be the greatest metaphysically possible being in various forms of great chains of being. In Part II of the book, Nagasawa takes on a variety of arguments against perfect being theism, helpfully categorizing them into three types:

• • Type A: arguments that purport to show the inconsistency of a single property of a perfect being (e.g. the paradox of the stone against the consistency of omnipotence)

• • Type B: arguments that purport to show the joint inconsistency of multiple properties of a perfect being (e.g. the argument that no being could be both omnipotent and omnibenevolent)

• • Type C: arguments that purport to show the inconsistency of the properties of a perfect being and the properties of the world (e.g. the argument from evil against the existence of a being with the omni-properties) (82–87)

Nagasawa then notes that many scholars have addressed many instances of these types of argument, and some with great success, in a case-by-case approach, but, he then writes: ‘[t]he most obvious problem with the case-by-case approach is that, whether or not the individual responses to these arguments ultimately succeed, it is not very efficient as a defence of perfect being theism’ (89). To provide such a defence, one that is ‘radically new and more economical, . . . that aims to eliminate the force of the existing arguments against [perfect being theism] all at once and block any further arguments’ (90; emphasis in the original), he provides what he takes to be the argument that underlies all objections to perfect being theism:

1. 1. If perfect being theism is true, then the perfect being thesis is true (namely, that God is the greatest metaphysically possible being (9) ).

2. 2. If the perfect being thesis is true, then the omni God thesis is true (namely, that God is necessarily: omniscient, omnipotent, and omnibenevolent (81) ).

3. 3. If the omni God thesis is true, then God is an omniscient, omnipotent, and omnibenevolent being.

4. 4. There cannot be an omniscient, omnipotent, and omnibenevolent being.

5. 5. Therefore, the omni God thesis is false.

6. 6. Therefore, the perfect being thesis is false.

7. 7. Therefore, perfect being theism is false. (90–91)

As Nagasawa notes, defenders of perfect being theism have concentrated on denying (4), arguing that the purported properties of a perfect being, which they take to be necessary omniscience, necessary omnipotence, and necessary omnibenevolence, are individually and jointly consistent, as well as being consistent with the properties of the world. Instead, Nagasawa suggests that premise (2) is the one to deny, and the remainder of this part of the book is dedicated to arguing that a perfect being need not be omniscient, omnipotent, and omnibenevolent, but could, rather, be a maximal God, that is, ‘a being who has the maximal consistent set of knowledge, power, and benevolence’ (92). And in Part III of the book, after reviewing various ways of formulating and attacking the ontological argument for the existence of God, he argues that a version of the modal ontological argument is successful in establishing such a conception of God: if it is possible that God exists, then God exists; and it is indeed possible that God exists, since he has the maximal consistent set of knowledge, power, and benevolence. (I want to record my appreciation for Nagasawa's careful presentation of Peter Millican's (2004) objection that Anselm's ontological argument equivocates on various interpretations of the superiority of existence, which I take to be the best objection to this perennial argument, along with Nagasawa's own reply (2007) to Millican, where he disputes that Anselm does, or needs to, endorse a principle about the superiority of existence, and Millican's rejoinder to Nagasawa (2007), where he continues to maintain that Anselm does and needs to maintain this principle. I found, and continue to find, this exchange highly illuminating, and what is of particular value is that Nagasawa takes the opportunity to reply to Millican's rejoinder in this book. Future students of Anselm's ontological argument could well make this a starting point for further progress on this ingenious argument for God's existence.)

There is much for a reviewer to comment on in such a rich and provocative book. Nagasawa characteristically presents the philosophical problems he addresses clearly, concisely, and accurately, such that this would be a good text to use for undergraduate classes in the philosophy of religion. Moreover, he presents novel solutions to these problems, such that this text makes a valuable contribution to the literature on perfect being theology. However, I think that this contribution could be better appreciated if it were understood in a way that Nagasawa does not explicitly endorse, namely, as an instance of conceptual engineering in perfect being theology. To explain why I think this, I begin with a brief account of conceptual engineering.

Conceptual engineering concerns ‘the activity of assessing and improving our representational devices’ (Herman Cappelen, ‘Conceptual engineering: the master argument’, in Alexis Burgess, Herman Cappelen, and David Plunkett (eds) Conceptual Ethics and Conceptual Engineering (Oxford: Oxford University Press, forthcoming), 1; in the following discussion I am indebted to this work by Cappelen). It sounds a rather technical activity, but it is not; it is something that many, if not most, if not all, of us do regularly when we do philosophy. Think of it like this.

A lot of the time, we do philosophy with words. But we did not create the vast majority of the words we use, the meanings they have, and the concepts they refer to. Rather, we have inherited them. Now, these words could have had other meanings and referred to different concepts. Some of these meanings and some of these concepts will be better than others for particular purposes. When we do philosophy, we want our words to have the best meanings and refer to the best concepts for what we aim to do when we do that bit of philosophy. Now, it's likely that our words do not already have these best meanings and refer to these best concepts. So, when we do philosophy, we should try to make our words have these best meanings and refer to these best concepts, relative to our particular purposes at the time. When we do this, we do conceptual engineering.

Conceptual engineering is probably something that philosophers have always done (though I am no historian), but it is clearly something that Rudolf Carnap (in Meaning and Necessity) was doing, whose method of explication is often cited as a forerunner of current conceptual engineering. In employing this method, Carnap aimed to improve meanings by removing indeterminacy and vagueness, so that they are a better fit for science. (Indeed, it was Richard Creath who coined the term in his edited edition of the correspondence between Carnap and Quine.)

In recent times, Sally Haslanger's work on race and gender (in Resisting Reality) has become a paradigm example of conceptual engineering, as she aims to improve these concepts for social and political aims. Another example, which will be relevant for understanding Nagasawa's text, at least in my view, is Kevin Scharp's argument in Replacing Truth that the concept of truth is inconsistent, and so we need a better one, or multiple concepts to do the work the concept of truth is meant to do. More generally,

• • A conceptual engineer in metaphysics will try to improve our metaphysical concepts, for example, philosophers engaged in giving an account (or objecting to an account) of causation.

• • A conceptual engineer in epistemology will try to improve our epistemic concepts, for example, philosophers engaged in giving an account (or objecting to an account) of knowledge.

• • A conceptual engineer in moral philosophy will try to improve our moral concepts, for example, philosophers engaged in giving an account (or objecting to an account) of moral goodness.

And so on. Nagasawa in this book, I think, is a conceptual engineer in philosophy of religion, in particular, in perfect being theology. Here is why.

Let us go back to the argument (1)–(7) Nagasawa takes to underlie all objections to perfect being theism. For premises (1) and (3), he says: ‘Premise [(1)] expresses merely the commitment of perfect being theism to the perfect being thesis . . . Premise [(3)] imparts the content of the omni God thesis.’ But, of premise (2), he says, ‘The claim that the perfect being thesis entails the omni God thesis seems, however, ungrounded’ (93), because, he thinks, there are no good arguments for it, and, ‘because there are many epistemically possible scenarios in which, while the perfect being thesis and the maximal God thesis are true, the omni God thesis is false’, it is not analytically true that the perfect being thesis entails the omni God thesis (94). Consequently, according to Nagasawa, the omni God thesis is ‘an unnecessary commitment’ for the perfect being theist. This opens the door for Nagasawa to give an alternative account of what a perfect being is, namely, as maximal God instead of omni God. Or, semantically, while keeping the lexical item ‘perfect being’ constant, along with its extension, God, constant, Nagasawa alters its intension from ‘omni God’ to ‘maximal God’. This is conceptual engineering at work. Now how does he do it?

Nagasawa engineers the concept of a perfect being in the following way. He begins with the great chain of being, and offers two models, a linear model and a radial model. For reasons of space, I concentrate on the linear model because that is most relevant for Nagasawa's engineering the concept of a perfect being. According to the linear model, ‘God is the being than which no greater is metaphysically possible by virtue of occupying the top link in the great chain of being’ (51). This, of course, assumes value-commensurability, of which Nagasawa is well aware, and which we will assume without discussion, again for reasons of space. Now, a being's place in the great chain of being is determined by how many great-making properties they have (the extensity of their greatness) and to what degree they have them (the intensity of their greatness). And the greatest being, the being that occupies the top link in the great chain of being, is the one that has, overall, the maximal consistent extensity and intensity of great-making properties.

Now, a consequence of this account is that it allows for something to be a perfect being yet have less knowledge, power, or benevolence than another being. As Nagasawa writes,

So, if God is a perfect being, it may well be that someone else knows more than God, can do more than God, or is morally superior to God. It may strike you, as it did me, that this is a reductio ad absurdum of Nagasawa's account. For, it is just intuitively obvious that a perfect being is the most knowledgeable, most powerful, and most morally superior being. It may be that there is no good argument for this claim, or that it is not analytically true, but it is intuitively true, nonetheless. And so, it might be concluded, Nagasawa's account must be false. But I think this would be too quick. For it misses the point of conceptual engineering, which is to make our concepts better, which presupposes that our intuitions might not always be true. And so such an objection to Nagasawa's account would be too quick. A better objection would be something like this.

Nagasawa realizes that a maximal-God account of a perfect being gives rise to a worry that there could be two (or more) perfect beings, something perfect being theists typically deny. Nagasawa writes:

In response, Nagasawa thinks this possibility poses no more a problem for the maximal God account than for the omni God account of a perfect being, for just as the maximal God account entails that there is a second greatest possible being, so does the omni God account. But, in reply, there is good reason to think that the omni God account of a perfect being can block the possible coexistence of two perfect beings, while the maximal God account cannot. And this would be a problem for perfect being theists who think necessary uniqueness a great-making property. Here is the argument (adapted from Robert Merrihew Adams's ‘Primitive thisness and primitive identity’, Journal of Philosophy, 76 (1979), 5–26).

As Nagasawa writes above, on the maximal God account, it is possible that there is a perfect being, a, who has the maximal consistent set of knowledge, power, and benevolence, and an almost perfect being, b, who, let us say, has the same degree of power and benevolence, but just a little bit less knowledge: say, it knows one fewer thing, p, than a. Let us also stipulate that neither being is omnipotent. It seems possible that these beings, a perfect being a, and an almost perfect being b, coexist. (Here I am assuming a presumption of possibility. For discussion and criticism, see Robert Merrihew Adams, ‘Presumption and the necessary existence of God’, Noûs, 22 (1988), 19–32.)

It also seems possible that, while coexisting with a, b comes to know p. Perhaps a tells it to b. Given these assumptions, it then follows that the maximal God account of a perfect being allows for the coexistence of multiple perfect beings. But there is good reason to think that the omni God account of a perfect being does not allow for this, since, as some have argued, it is impossible for multiple omnipotent beings to coexist, as it is possible for their wills to conflict. (For discussion, see James Baillie, and Jason Hagen, ‘There cannot be two omnipotent beings’, International Journal for Philosophy of Religion, 64 (2008), 21–33; Aldo Frigerio and Ciro Florio, ‘Two omnipotent beings?’, Philosophia, 43 (2015), 309–324; and Jonathan L. Kvanvig, ‘Anselmian adversities’, Religious Studies (forthcoming).)

So, it seems that there is good reason to think that, while the omni God account of a perfect being includes the necessary uniqueness of a perfect being, the maximal God account does not.

Now, I do not take this, in any way, to be a conclusive objection against the maximal God account Nagasawa advocates, since this argument, too, involves intuitions, intuitions which might be conceptually engineered. And this might indeed be the right way to go. But what we see is that when we start engineering the concept of a perfect being, it is likely that we will have more concepts to engineer. Which means, then, that this is not really an objection, but rather signals the start of a project, one that might begin as follows.

Consider the following ‘paradigmatically philosophical’ intuitions, given by Matti Eklund:

1. 8. That assuming the watery stuff around us is H₂O, watery stuff with a different chemical composition found on a different planet is not water.

2. 9. That modus ponens is valid.

3. 10. That the subject in a Gettier case does not have knowledge.

4. 11. That it is wrong to push the fat man in the relevant trolley case.

5. 12. That there is no object composed of Rand Paul's nose and the Eiffel Tower.

   (Matti Eklund, ‘Intuitions, conceptual engineering, and conceptual fixed points’, in Christopher Daly (ed.) Palgrave Handbook of Philosophical Methods (London: Palgrave Macmillan, 2015), 364–365)

Intuitions (8) and (9) Eklund calls ‘competence intuitions’ (but as he notes, he does not mean a great deal by ‘competence’ here), where their reliability is explained by our dispositions to use ‘water’ and ‘modus ponens’ in accord with (8) and (9), and intuitions (11) and (12) he calls ‘rational intuitions’ (again, not meaning a great deal by ‘rational’ here), where their reliability is not so explained. Intuition (10), according to Eklund, seems hard to classify in either category: it does seem that we have the disposition to use ‘knowledge’ in accord with this intuition, which seems to pull it towards being a competence intuition, but the normative content of ‘knowledge’ seems to push it towards being a rational intuition, hence the Gettier wars.

I say all this because it seems to me that the intuition that a perfect being is omnipotent, omniscient, and omnibenevolent seems to be like intuition (10). I do think we have a disposition to use ‘perfect being’ in accord with this intuition, hence why it seems intuitively obvious that, if a being a is more knowledgeable, more powerful, or morally superior to being b, b cannot be a perfect being, and so pulling it towards being competence intuition, but the normative content of ‘perfect’ seems to push it towards being a rational intuition. This makes the concept of a perfect being, and related concepts as discussed above, ripe for conceptual engineering, and, it seems to me, this is what Nagasawa does in this book. If this is right, then the way to evaluate Nagasawa's account of a perfect being as maximal God, as opposed to omni God, is to ask whether he has improved the concept of a perfect being. Has he made our intellectual and/or religious lives better?

As I see it, Nagasawa gives what he takes to be two advantages of his account: taking a perfect being to be maximal God, as opposed to omni God, means that all of the arguments against the existence of a perfect being can be refuted economically, and the possibility premise of the modal ontological argument is straightforwardly true. I am happy enough to grant that the maximal God account of a perfect being does have these consequences. But I wonder if they are benefits. For it just seems too easy (not that I want to diminish in any way Nagasawa's achievement in this book, which I take to be showing the conceptual connection between the maximal God account of a perfect being and these two consequences, something that is a considerable achievement). Let me explain.

One way of understanding what we do when we do philosophy is that we untie (conceptual) knots. In his review of Brian Loar's Consciousness and Meaning: Selected Essays, Joseph Levine writes:

I find this an appealing way of doing philosophy, not least because I enjoy solving puzzles. It is much the way Robert Nozick characterizes giving an explanation to a philosophical problem:

For example:

• • How could we be free, if we are physical creatures subject to physical laws?

• • How could we have perceptual knowledge, if we could be brains in a vat?

Or, more relevantly for our present concerns:

• • How could a being be unable to lift a stone if it is omnipotent?

• • How could a being be omnipotent if it is omnibenevolent?

• • How could an omnipotent, omniscient, and omnibenevolent being exist, if evil exists?

These philosophical problems are like conceptual knots, knots that many people have tried very hard to untie over the centuries. Perhaps they have not been successful. Maybe the reason that these knots have stayed tied despite our best efforts to untie them is that perfect being theology is really a big conceptual Gordian knot. If this way of understanding perfect being theology is right, then what Nagasawa does in this book, it seems to me, is like what Alexander the Great does in the legend, namely, he cuts the knot of perfect being theology with the sword of conceptual engineering by giving an account of a perfect being not as omni God but rather as maximal God. The benefits of doing so certainly include efficiency, as Nagasawa says frequently in his book, but is it effective?

Well, just as Alexander untied the Gordian knot, in a way, so did Nagasawa. But, in another way, he did not, just as Alexander did not, for, strictly speaking, he cut the knot, and did not untie it. Perhaps Alexander's cutting the knot was a stroke of genius, as it may have been that the knot was simply impossible to untie, the knot being something of a Greek Kobayashi Maru. We will never know, for, putting aside this being a legend, once a physical knot is cut, there is no tying it back again. Fortunately, being a conceptual knot, we still have the knot of perfect being theology. And, I for one, want to keep trying to untie it. Perhaps it is impossible to untie, in which case, the task is even better, for, as Ethel Lina White once wrote, ‘Lost causes are the only causes worth fighting for’ (Ethel Lina White, The Wheel Spins (New York: Harper, 1936), 270).

That may give us a clue to another way to cut the knot, as it were, if we were of a mind to do so. Look back at the argument Nagasawa takes to underlie all objections against perfect being theism, (1)–(7) above. He quickly dismisses objecting to premises (1) and (3) because he thinks they give the content of perfect being theism and the omni God thesis, respectively. But could it not be the case that those concepts could be engineered? Perhaps, for premise (3), the concept of omni God could not, as it seems it could only be the concept of an omniscient, omnipotent, and omnibenevolent being. But what of, in premise (1), taking a perfect being not to be one that is the greatest metaphysically possible? That is, might we engineer the concept of a perfect being so that (1) is false? Perhaps we might rethink what it is to be a perfect being in accord with the New Testament's portrayal of Christ, as one ‘who, though he was in the form of God, did not regard equality with God as something to be exploited, but emptied himself, taking the form of a slave, being born in human likeness. And being found in human form, he humbled himself and became obedient to the point of death – even death on a cross’ (Philippians 2.6–8), and counsels his disciples, who were squabbling among themselves as to who among them was the greatest, ‘Whoever welcomes this child in my name welcomes me, and whoever welcomes me welcomes the one who sent me; for the least among all of you is the greatest’ (Luke 9.48). Perhaps a perfect being is not one who is metaphysically greatest but rather is one who is metaphysically humble. Thinking of a perfect being in this way might help not only our intellectual and religious lives, but also our social and political ones, too.