Linda Zagzebski’s suggestion that if God existed God would be “omnisubjective” has generated a great deal of interest within the philosophy of religion.¹ The idea in brief is that God has not only all propositional knowledge but all experiential knowledge as well. In a completely different area of philosophy, Laurie Paul has developed the idea that experiences can be transformational in both an epistemic and a conative sense, that is, an experience can be such that one cannot know what it is like until one has it and it can change one’s preferences in ways that are alien to one’s current preferences. Both of these ideas, omnisubjectivity and transformative experience, have been quite fruitful in a relatively short period of time, rapidly producing academic literatures discussing the merits, models, and boundaries of each idea.²

What does not exist as yet is any work showing how the two ideas interact. That is what I take on in this essay. In particular, I will argue that epistemically transformative experiences are hard for at least some models of omnisubjectivity to accommodate. The dialectical context in which I will be bringing transformative experiences to bear on views about omnisubjectivity concerns certain lines of objection to perceptual or quasi-perceptual models of omnisubjectivity on the basis of their not re-duplicating the first-person perspective of subjects and thus seeming to be at too much of an epistemic remove.

Here is the argument I will be building toward.

In the next section, I will explicate Zagzebski’s position, in particular her reasons for favoring a model of omnisubjectivity grounded in imaginative projection over one grounded in perception. In section II, I will present the phenomenon of transformative experience, followed, in section III, with an exploration of how omnisubjectivity and transformative experience relate to one another.

Omnisubjectivity is a new divine property proposed and elaborated on by Linda Zagzebski (e.g. 2008, 2013, 2016, 2023). Whereas one might think of traditional descriptions of divine omniscience as concerned with God’s knowledge of all the facts, they do not typically draw attention to God’s epistemic standing as regards the realm of experience. Yet, as Frank Jackson’s Mary experiment has hammered home for all of us, there appears to be a difference between propositional and experiential knowledge, between something like knowing a fact about the color red and knowing what it is like to see red (cf. Zagzebski 2023, p. 7; see Jackson, 1986). Hence, it would seem that God’s knowledge would lack something, would fall short of the maximal, perfect knowledge God should have, if God were not aware of what it is like to have the various experiences we have.

In her 2013, she defines omnisubjectivity this way:

In her 2023, she introduces the idea as follows:

The phrasing in the latter quotation is slightly different. One possible explanation of this is that in her 2023 she weighs three models that all have some claim to explaining the phenomenon of interest —the total empathy model, the perception model, and the panentheism model—and one might think that the phrasing in the first quotation rules out a perceptual model from the get-go. In her 2023, she makes points for and against each of the three models, while still communicating that she thinks of the perceptual model as the weakest of the three.

The total empathy model represents Zagzebski’s first account of how God’s omnisubjectivity works. In her 2008 and 2013, she draws on simulationist accounts of the knowledge of other minds to suggest that God models the subjective experience of God’s creatures by way of God’s imagination.

Here is a description of the kind of process she has in mind.

This generates the model of total empathy as follows.

Yet, as of the 2013 version of this model, Zagzebski qualifies her view out of concern with the indirectness of simulation³ and does so through the use of perceptual language.

By the time she writes her 2023, Zagzebski has come to see the total empathy model and the perceptual model as rivals. For instance, the parallel passage to that above in her 2023 emphasizes the importance of directly grasping but does not express this contrast between the direct and the indirect through the language of perception (2023, p. 67).⁴ Moreover, despite her favoring the total empathy model over the perceptual one, she seems more inherently ambivalent about the adequacy of the total empathy model.

It is worth paying attention to one of Zagzebski’s sources of discontent with the total empathy model because, although she thinks it can be addressed, one can see why the total empathy model will have trouble with transformative experiences foreshadowed in this other case.

The problem at issue is not necessarily hosting a particular qualia—a color, a texture, an affect— but in certain additional features of the experience that may hold when it is a creature that is having it. It is possible for a creature to have their experience completely dominated by hosting a qualia (e.g. a despairing feeling). But it looks like that is one thing God cannot experience. God is never limited to hosting all and only some qualia like a feeling of despair that might “fill us up.”

Zagzebski, however, says that “perhaps this objection is not serious enough to ruin the empathy model” (ibid, p. 69). The reason she provides is that “sometimes we can imagine what something would be like if it filled up our consciousness, and we can then confirm it when it happens” (ibid, p. 69). The thought, then, is that if we can adequately learn what being filled up with an experience will be like without it having happened to us before, then surely God can do so too.

Whether one is satisfied with this response will hinge crucially on whether one thinks there is “something it is like” to be full, something it is like to be dominated by an experience above and beyond simply having the first-level qualia. After all, one can imagine, on the human level, someone who has been completely dominated by despair denying that a would-be human empathizer knows what this experience is like even if they have been partly despairing before. If there is something that it is like to not only be despairing but be completely despairing, then a simulation in which one does not reach full immersion in the experience may get one closer to the experience, perhaps even close enough to recognize the experience were it to happen to oneself, but it would not get one the experience itself. The exact same dynamics will be in play when we turn to transformative experience.

By contrast, Zagzebski, while acknowledging the pedigree of a perceptual account of God’s relation to our minds (e.g. in Aquinas), is more straightforwardly pessimistic about such accounts. She acknowledges that a perceptual account might be able to make sense of God’s knowledge of our sensory imagery and inner speech (ibid, p. 71), but she draws the limit at our feelings. Observing a pain and feeling it are two different things (ibid, p. 71-72) but to know what a feeling is like just is to have that feeling.

Something interesting about the evolution of Zagzebski’s views, then, is that in her 2013 perceptual language is added to the empathy model to assuage concerns about the indirectness of imaginative projection. In her 2023, she has distinguished empathic uses of the imagination and interpersonal perception and ranks the former over the latter because perception is thought to be too indirect a link to another person’s feelings to really understand them.

For our purposes, I will not be engaging the third model, that of panentheism, but will instead use transformative experience to put pressure on this comparison of imaginative projection and perception as rival means by which God might understand what we experience.⁵ I will do this with a mind to evaluate the claim that the perceptual model, but not a model grounded in imaginative projection, cannot account for the importance of knowing what it is like to have a feeling.

Laurie Paul’s work on transformative experience has its origins in a standard model for how we should think about making decisions. According to decision theory, to make a decision one needs to be able to identify different possible outcomes that hinge on one’s choice, their relative value vis-à-vis each other, and how likely they are.

Suppose, for instance, that someone is deciding between spending their money on a fancy meal or buying a raffle ticket for a prize. The two things it seems like one should know before deciding whether to pay for the meal or the ticket are how many tickets are in the lottery and whether one would like whatever the lottery prize is much more than a fancy meal. After all, if the odds of winning the prize are small or the prize just is a fancy meal, then entering the raffle might not make much sense. Suppose, though, that the only thing you know about the prize is that it is something you’ve never experienced before. In this situation, the standard machinery of decision theory appears to break down.

What Paul draws our attention to is that, in situations like the mystery raffle prize, one’s ability to make a rational decision frequently seems to hinge, among other things, on how much we value the experiences that we would have contingent on the different choices we can make. The problem is that we may not know what the experiences in question are like ahead of time and thus how much value to ascribe to those experiences.

There is always something unique about any new experience. After all, no matter how much pineapple one has eaten over the course of one’s life, one does not know what this next piece of pineapple tastes like until you taste it. This is not what Paul has in mind. Rather, she has in mind experiences that are different in kind (or at least to some extreme degree) from experiences that one has had (ibid, p. 36). The key is that they are not forecastable based on one’s current set of experiences. “[W]e can’t know what our lives will be like until we’ve undergone the new experience, but if we don’t undergo the experience, we won’t know what we are missing” (ibid, p. 3).

Tasting a durian fruit, by itself, is not a transformative experience. Someone who is used to eating durian might have all the same flavors hit their palate as someone who had never seen one before. Something’s being a transformative experience requires a contrast between the experience and ones with which one is familiar such that the new experience was not forecastable (e.g. a life of eating pineapple has not prepared one to forecast what eating durian will be like). If one does have an experience that is qualitatively interchangeable with eating a durian fruit, then eating durian cannot be a transformative experience for you. If you do not, then it can be.

One helpful way to put the point is to juxtapose transformative experiences with situations where one gets one’s mind around a choice by running a kind of simulation in one’s imagination (ibid, p. 27-28). Our ability to simulate depends on what we already have experienced. Simulation is primarily a combinatorial exercise, built off of our prior experiences and beliefs. While we engage in imaginative simulation to project what things we have not experienced are like, there is a limit to how much simulation can do. Borrowing from Nagel (1974), simulation is of limited utility in understanding what it is like to be a bat because those experiences are “fundamentally alien” (Paul 2014, p. 6). Yet, one does not have to change species to find a transformative experience; one only needs to not be in a position to tell what the new experience is like ahead of time, even by trying to imaginatively project through some kind of simulation. Notice, though, that this is exactly the kind of process that Zagzebski appeals to in order to explicate her total empathy model both in its earlier and latest incarnations.⁶

With this much background, let us turn back to the argument provided at the beginning of the essay. For ease of reference, here, once again, is that argument.

One can imagine Zagzebski replying on behalf of the imaginative simulations that lie at the heart of the total empathy account that there is a straightforward way in which God could come to know what it is like for a human to undergo a transformative experience. I can imagine reading a story about someone who has never experienced color, like the infamous Mary, and I can imagine Mary’s colorless experience giving way to a first-time experience of color. In doing so, it seems like I can gain some insight into what it’s like to see color for the first time, and seeing color for the first time would be a transformative experience. Just so, God could imagine Godself as a human in a state where the human doesn’t know what some experience e is like. For example, God might imagine Godself as a human that has never ingested psychedelic mushrooms or has never been a parent or has never seen the color red. God can then run an imaginative simulation forward from that starting point through the human’s having of a new experience of the relevant kind. God can simply note the surprise and/ or shock and move on. Recall, this is precisely the strategy Zagzebski uses to address the objection from experiences that fill up or dominate the whole of consciousness.

Our foray into the transformative experience literature, however, is instructive here. The whole reason why transformative experience causes trouble for decision theory is because, by definition, a transformative experience is not forecastable by the person who has it. If people could imaginatively project what it’s like to be a parent or become a vampire or eat a durian fruit, there would be no problem of transformative experience for Paul to puzzle over. One might think, of course, that God’s powers of imagination would be so much more powerful than our own that God would not have a problem simulating experiences that we cannot forecast. That is a fine thought, but it is actually a further reason for thinking that God cannot have transformative experiences. After all, if you can forecast what it’s like to have an experience, then having it is not a transformative experience for you.

Suppose we grant that God can perform imaginative simulations of humans undergoing transformative experiences. God can imagine being a human that cannot imaginatively project what an experience is like. This would indeed provide something relevant to experiential knowledge of what humans are going through when they have transformative experiences. This would amount to God simulating humans coming to know something God has always known and continues to know at each point of the simulation God performs. There is a similarity relation between imagining learning something you have always known and coming to know something you never knew, but they aren’t the same experience. For instance, when I imagine someone seeing red for the first time, the red qualia in my imaginative projection of the other person could be type identical to the red they experience. I could even have a kind of mental movie in which the first person perspective of the other person exactly matched their redless world and then, at the exact right time-matched moment, the mental movie I am running in my mind could include the exact shade of red they experience. I would not thereby have the first-person experience of coming to see red for the first time as a transformative experience. Likewise, if God could imaginatively model the moment to moment qualia of a human person, indeed, if God did so in a much more comprehensive way than I am capable of, then the result is something very experientially useful for understanding the person and the transformative experience they end up having. But God’s simulational experience is not a transformative experience because God already knows what the experience on the other end is like.⁷

One might think it would help to go atemporal here. The problem, one might think, is discussing the matter as if God is in time. If God could timelessly simulate a human psychology, then in one eternal moment God would hold the pre-transformed self and the post-transformed self of the human being simulated in the divine mind at the same moment. God would have, as it were, an imaginatively projected space-time worm in God’s mind. We do not need to make sense of how the divine mind could be confined, even imaginatively, to the simulated moments before the novel experience is had. Instead, God imaginatively projects all the moments as a sequenced spacetime worm at once. I’m afraid this move does not help, however. In fact, if anything, going atemporal would make matters worse, at least if one accepts Zagzebski’s critique of perceptual accounts of omnisubjectivity.

If God needs to have a transformative experience from the inside to know what our transformative experiences are like, then God would need to lack some experience e and what it’s like in order for that absence to give way to hosting e for the first time. Inhabiting all points of a simulation that includes e in some but not all of its segments would mean that both God’s divine mind as employed in an imaginative simulation and God’s mind independent of the simulation would both contain e. Rather than solving the problem, it looks like we violate the strictures on having a transformative experience twice over. Therefore, God cannot gain knowledge of what it’s like to have novel experiences simply by simulating our psychology. An atemporal being does not have a transformative experience by way of being able to imagine a complete spacetime worm’s psychology. Rather, such a being imaginatively depicts a psychology that can have a transformative experience.⁸ As Zagzebski says, a feeling is what it feels like, and so, if God does not undergo a transformative experience, God needs a way of knowing what it feels like “from the outside.”

Suppose the proponent of an imagination-based simulation account regroups as follows. I have granted that God could imagine a psychology that has a transformative experience even if God does not count as experiencing one thereby. Why not think that is good enough? Suppose God can’t literally experience a transformation where a new kind of experience, one heretofore alien to the divine life, occurs. Nonetheless, knowing everything else about a simulated human’s qualitative moment to moment experience and how it is sequenced ought to be good enough to know in a very robust sense what the individual who has had a transformative experience is going through. It is not as if one must say that there are any flavors of first-order qualia that God is not familiar with on this picture. Rather, it’s what it feels like to be subject to a certain pattern, a certain ordering of that qualia from the inside which is at issue. And, on top of that, God certainly can also know that a given sequence of experiential events would be transformative for a subject.

My answer to this suggestion I have inserted into my opponent’s mouth at this stage of the dialectic is that they would be absolutely right in making such a rejoinder…but that this is an understanding “from the outside.” I see no interesting difference between such a version of a simulation/ imagination view and a perceptual account. And the perceptual account is, in important respects, simpler.

Here is a model of how omnisubjectivity could work based on a kind of perception, that is, on God directly registering the existence and features of a thing because it exists and not by way of imagination or inference.

Let us call the following the perception+introspection account. God has perfect selfknowledge directly via the equivalent of introspection. God has an “insider” perspective on God’s own life. Everything God creates reflects God’s being and goodness to one degree or another. God also has experiential full access to God’s creatures—access to their location in time and space but also to the particular ways they are manifesting the being and goodness of God through their subjectivity.⁹ God is not imagining what they are doing. God has direct epistemic access to it and can match what God has direct access to in creatures to God’s experience of God’s own being, which, again, is what creatures reflects however imperfectly. God does not take on the inside perspective of each human but has direct access to it as a kind of perfect observer of thoughts, feelings, or any other mental state. There are various aspects of what someone feels that are observable by humans, and, even by Zagzebski’s lights, God would have access to more observable dimensions of the feeling person. God perceives these and is able to perfectly match the creature’s feeling to those aspects of God’s being and goodness they manifest, to which God has first-person access via introspection.¹⁰

When a human has a transformative experience, God sees each subjective moment of this experience, including the phenomenology of experiencing a transformation. God is able to match the qualia of human experience to the aspects of God’s being that it manifests and to which God does have an inside perspective qua psychological subject. God thereby knows the experiential content of what God’s creatures go through in a transformative experience but without literally having a transformative experience.

Furthermore, I do not think the advocate of an imagination based account like the total empathy model can take away any of the machinery I appeal to for the quasiperceptual account. Everything God creates still reflects God’s being and goodness to one degree or another. God still must have perfect self-knowledge. God also needs to have direct epistemic access to all parts of reality in all possible respects, including what goes on in our minds, and I do not see why an advocate of this account should deny that God can match what God has direct epistemic access to in our lives to God’s own infinitely rich and qualitatively encompassing experience of God’s own inner life.

The only question, then, is whether positing imaginative projection is necessary to get something unavailable on the model I have proposed. Zagzebski’s proposal, once again, was that the total empathy account but not a perceptual account can explain how God understands our feelings because feelings must be grasped from the inside. But, as argued above, some of the experiences that God needs to be able to appreciate epistemically to be omnisubjective, such as what it is like to undergo a transformative experience, can only be gotten from the outside. Either that’s not good enough to undergird omnisubjectivity and thus the total empathy account fails or else it is good enough and a perceptual account can succeed despite being, in a sense, a way of understanding what experiences are like from the outside.

So far as I know, I think both the perceptual model and the modified simulational model are possible, but I find the perceptual one more elegant. Once again, direct access to creatures in all ways possible and divine self-knowledge are going to have to be consistent with any model. If it is possible to get what we need without also positing imaginative simulations that the divine mind has to work through, then I don’t see why we should posit them. This is not to say that they are metaphysically impossible. They may sit awkwardly vis-à-vis a hypothesis like divine simplicity since imaginative simulations of different individuals would presumably be distinct mental events within the mind of God, but I am willing to countenance them. They may even be helpful to account for some of God’s counteractual knowledge.¹¹ As it relates to relating perfectly to the experience of individual, actual creatures, however, I find imaginative simulations unnecessary.

Moreover, the perceptual model puts God’s experiential knowledge of transformative experiences on a par with God’s knowledge of other human experiences. God has direct epistemic access to all of them (from the outside), and God can recognize the aspects of God’s being and goodness thereby reflected. If one thinks that God only knows what our experiences are like by way of simulating them from the inside, then the fact that transformative experiences aren’t amenable to that treatment, as has been shown, at best gives one reason to think that God’s knowledge of human transformative experiences would be epistemically inferior to God’s knowledge of other human experiences. This situation is rendered all the more untenable by the fact that any qualia that is experienced for the first time by a human being can be the subject of a transformative experience and thus the scope of God’s second-rate subjective knowledge would be quite wide. By contrast, on the perceptual account I have offered, all of our experiences are known precisely from the inside and the outside, from the outside by being perceived perfectly and from the inside by God’s knowing from God’s grasp of God’s own divine life the being and goodness they manifest.

By way of conclusion, in her 2023, after discussing the pros and cons of different models of omnisubjectivity, Zagzebski provides a philosophical meditation on a metaphor inspired by some of the properties of light and God’s role as cause and conserver of the universe (p. 80ff). She quotes with approval Aquinas’ line that “if light knew itself, it would know all colors” (ibid, p. 81). She elaborates the idea as follows.

What I have argued in this essay allows us to make good on this sentiment. Simulations or imaginative projection do not have a good corollary within this metaphor. To know that red is part of white light by running an experiment with a prism is not the same as grasping the nature of white light in such a way that one already understands what the prism shows. Instead, on the quasi-perceptual model I have articulated, God’s self-knowledge paired with God’s direct epistemic access to what aspect of God’s being creatures manifest allows God to relate perfectly to our subjective experience of ourselves epistemically. In God’s self-knowledge of divine light, as it were, God already knows each color and then recognizes it when it is on display on its own. At the end of the day, it is God’s recognition of the way our subjective world reflects the divine being and not hosting a token or copy of each of our experiences which matters.