1. THE PROBLEM OF TRANSFORMATIVE EXPERIENCE
Most people at some point in their lives face transformative decisions that could result in experiences that are radically different from any that they have had, and that could radically change their personalities and preferences. For instance, most people make the conscious decision to either become or not become parents. In a recent but already influential book, L. A. Paul (2014) argues that transformative choices cannot be rational – or, more precisely, that they cannot be rational if one assumes what Paul sees as a cultural paradigm for rational decision-making. Paul arrives at this surprising conclusion due to her understanding of transformative experience as being both epistemically and personally transformative. An experience is epistemically transformative if it ‘teaches [a person] something she could not have learned without having that kind of experience’ (11), but it is personally transformative if it changes the person's point of view and her fundamental preferences (16).
Epistemic and personal transformations pose rather different problems for the possibility of rational decision-making. Choices that result in personal transformations raise the question of which self to consult: the self that exists prior to the transformation or the self that the transformation creates. Many parents, for instance, report that some of their core preferences and values changed as a result of becoming parents, a typical example being a newfound willingness to put their family ahead of their career. When deliberating about whether to become a parent, should you consult your current and career-driven self, or the family-oriented person that you expect you will become after you have your child? This is the problem of personal transformation.
On the other hand, the problem of epistemic transformation, applied to the choice of becoming a parent, is this: Since one cannot, Paul thinks, know what it is like to be a parent before one has actually experienced it, one can neither rationally decide to become one, nor decide to stay childfree, by ‘mentally simulating’ the experience of being a parent. But such a simulation of what it would be like for one to experience being a parent is required by the ‘predominant cultural paradigm’ (25). More precisely, when rationally deciding whether to become a parent, one estimates, by performing such a simulation, the ‘subjective value’ of experiencing the possible outcomes of the choice to become a parent, and similarly for the option of living a child-free life, and chooses the alternative that has the highest expectation of subjective value. But since one cannot possibly know – or even entertain – what it is like to be a parent before actually experiencing it, one cannot mentally simulate the experience. Hence, she concludes that a transformative choice, such as choosing to become a parent, cannot be made rationally (10–11, 31; Reference Paul2015 b: 491).
Moreover, Paul thinks that the subjective value of transformative alternatives can neither be determined by consulting people who already are parents, nor by reading all the available literature or examining all the evidence on the matter. For such information will not determine what ‘it is like’ to be a parent in phenomenological or experiential terms.Footnote 1 She compares the predicament of a prospective parent to Frank Jackson's (Reference Jackson1982) Mary, who has lived her whole life without ever seeing the colour red. Mary could not possible know what it is like to see red, Jackson argues, despite knowing all the relevant science, before experiencing seeing red herself. The same holds for prospective parents and for others who are faced with an epistemically transformative alternative: no evidence or information can tell them what it is like to experience the outcomes of their choices obtaining before actually experiencing them. Therefore, they cannot, Paul thinks, make such choices rationally.
So how, then, should one make this life-changing decision? Paul's proposal is that when it comes to this and other transformative choices, we ‘reformulate’ or ‘reconfigure’ the decision-problem so that it is seen as a choice between having and avoiding a revelation (113–115). In particular, we should frame the parenthood decision-problem in terms of whether we want to discover what parenthood would be like for us, rather than framing it in terms of whether we want the lives we expect to lead as parents.
The aim of this paper is to examine Paul's argument that transformative choices pose a general problem for the possibility of rational decision-making and her solution to this problem. We will focus on the transformative choice of becoming a parent. This is only one of the transformative choices Paul discusses. Other examples include choosing to become a vampire, deciding to taste vegemite for the first time, and choosing to have a new sensory ability. The reason we focus on the choice to become a parent is that, first, this is the decision-problem that started the discourse we are entering (see Paul Reference Paul2015 a), and, second, this decision is important (unlike the option to taste vegemite) and one that most people actually face (unlike the option to become a vampire or to have a new sensory ability).
We will for most parts set aside the problem of personal transformation, and focus on the question of whether the epistemically transformative nature of parenthood makes it impossible to rationally decide whether to become a parent. The reason we set aside the problem of personal transformation is that there already is a quite extensive literature on that problem, which Paul unfortunately does not address.Footnote 2 In contrast, Paul's argument and the work she has inspired does raise the new and rather intriguing question of whether epistemic transformations pose a general problem for the very possibility of rational decision-making. We shall argue that it does not. However, there is a particular type of agent for whom the problem identified by Paul does arise. With this agent in mind, we examine Paul's suggestion for how to approach a transformative decision problem, namely, that one's decision should be based on whether one would like to come to know what the experience in question will teach one. We conclude that Paul's suggestion does not solve this agent's problem. In other words, Paul's solution fails for the only type of agent for whom the problem she has identified arises.
2. A PROBLEM FOR RATIONAL CHOICE?
As previously mentioned, Paul thinks that when considering a possible object of choice, one should try to determine what it would be like to experience its possible outcomes, by running a mental simulation of them. It is quite unclear to us, however, why Paul thinks that one should be rationally required to approach a decision-problem in this way. Nor do we recognize a cultural paradigm that requires people to do so. Rather, it seems to us that to the extent that there is a generally accepted view on how one should approach personal decisions, the view is that one should make the choices that best serve one's desires and values (given one's beliefs). This is certainly the view embodied in orthodox rational choice (or decision) theory.
As we see it, the problem Paul identifies only arises for a very special type of person, i.e. for someone who cannot decide whether to become a parent before he knows ‘what it is like’ to be a one, in Paul's sense; by which she does not mean simply how pleasurable or painful it will be, but rather the precise phenomenology,Footnote 3 or, as we shall call it, the texture of the experience. To see why not all decision-makers need the aforementioned information before making up their mind, let's consider another way of approaching the question. How people generally approach this question is, of course, an empirical matter.Footnote 4 But being parents ourselves, we can at least confirm that some people approach this decision-problem by considering the risk of complications, how parenthood will affect their career, whether they would like to be the person they expect to become as a parent, what their partner desires, how happy they can expect to be as parents, how pleasurable or painful they expect the experience to be, how good a life they can expect their child to lead, whether there are moral reasons for or against becoming a parent, and so on. Since we think that we are pretty normal in this respect, we hope the reader forgives us for calling this the normal approach to the parenthood decision-problem.
The above matters can all be decided without knowing exactly what it would be like to be a parent. While Paul might be right in claiming that one cannot know for sure how parenthood will affect one's desires and beliefs (see e.g. 81, 91), one can at least form informed expectations by talking to those who have already gone through the transformation and by reading the relevant empirical literature.Footnote 5 The same can be said about the impact of parenthood on career prospects and happiness (or subjective well-being). Paul is of course correct in pointing out that these statistics won't tell anyone for sure how happy they will be as a parent (or what that happiness or misery will ‘feel like’), nor how it will affect their career. But by complementing data about the average effects on people sufficiently similar to oneself with information gained by asking people one trusts how parenthood affected them – and, perhaps more importantly, why it affected them the way it did – one can certainly form informed and reasonable expectations about the effects parenthood will have on one's career and subjective well-being. These expectations can then help one decide whether one's present desires and values are best served by becoming a parent or by staying childfree. Hence, for people who approach this decision-problem the way we did it, the story of Mary seeing red is irrelevant to the question of whether one can rationally solve the problem.Footnote 6
To make the discussion a bit more precise, we can represent any possible outcome as a vector of aspects that the agent of interest cares about. Typical agents for instance care about morality, wellbeing, happiness, pleasure and displeasure, both their own and others’, and objective features of the outcomes such as freedom, knowledge, friendship, etc., both for themselves and for others. Each outcome can then be represented by a vector that shows how the relevant aspects are realized in the outcome.
Exactly how the vector will look depends on which aspects are deemed relevant. Moreover, some aspects can be sub-aspects of others. Perhaps the pleasure aspect is a sub-aspect of the happiness aspect, or perhaps the happiness aspect is a sub-aspect of the wellbeing aspect, and so on. We want to stay neutral on these questions, so we will just use the notation a 1, a 2,. . ., an for the relevant aspects. So a possible outcome x of an alternative action can be represented as:
$$\begin{equation*}
< {{{a}}_1}\left( {{x}} \right),{{ }}{{{a}}_2}\left( {{x}} \right),{{ }} \ldots ,{{ }}{{{a}}_{{n}}}\left( {{x}} \right) >
\end{equation*}$$
These functions, a 1(x), a 2(x), etc., need not be numerical. They can output relevant descriptions of the outcomes, or properties of the outcomes. For example, the morality function could yield the description ‘does morally wrong action’ or the property of doing something wrong, or something even more specific.
As we understand Paul, she wants to add to this list an aspect that essentially has to do with the phenomenology of experiencing a life (or a part of a life) – or, as we have called it, the texture of the experience, t.Footnote 7 The function t does not output a numerical value or a description; it outputs a phenomenological feel or qualia property. So we get a vector like:
$$\begin{equation*}
< {{{a}}_1}\left( {{x}} \right),{{ }}{{{a}}_2}\left( {{x}} \right),{{ }} \ldots ,\,{{{a}}_{{n}}}\left( {{x}} \right),{{ t}}\left( {{x}} \right) >
\end{equation*}$$
As already discussed, Paul claims (10–11, 32) that this t-aspect is not entertainable for some outcomes, i.e. it is not entertainable by the agent for the outcomes of epistemically transformative alternatives that she has not yet chosen and experienced. Let us symbolize this with ‘?’. So, a possible outcome of a transformative choice will look like this:
$$\begin{equation*}
< {{{a}}_1}\left( {{x}} \right),{{ }}{{{a}}_2}\left( {{x}} \right),{{ }} \ldots ,{{ }}{{{a}}_{{n}}}\left( {{x}} \right),? >
\end{equation*}$$
Now, Paul's claim that transformative choice creates incomparability means that the agent of interest can neither rationally form preferences over the pair
$$\begin{equation*}
< {{{a}}_1}\left( {{x}} \right),{{ }}{{{a}}_2}\left( {{x}} \right),{{ }} \ldots ,{{ }}{{{a}}_{{n}}}\left( {{x}} \right),{{ t}}\left( {{x}} \right) > {{\rm and}} < {{{a}}_1}\left( {{y}} \right),{{ }}{{{a}}_2}\left( {{y}} \right),{{ }} \ldots ,{{ }}{{{a}}_{{n}}}\left( {{y}} \right),? >
\end{equation*}$$
nor over the pair
$$\begin{equation*}
< {{{a}}_1}\left( {{x}} \right),{{ }}{{{a}}_2}\left( {{x}} \right),{{ }} \ldots ,{{ }}{{{a}}_{{n}}}\left( {{x}} \right),? > \,{{\rm and}}\, < {{{a}}_1}\left( {{y}} \right),{{ }}{{{a}}_2}\left( {{y}} \right), \ldots ,{{ }}{{{a}}_{{n}}}\left( {{y}} \right),? > .
\end{equation*}$$
However, it seems to us that many people can rationally form such preferences. For instance, suppose an agent is certain that outcome <a 1(x), a 2(x), . . ., an (x), ?> will be the result of choosing to become a parent, whereas outcome <a 1(y), a 2(y), . . ., a n (y), t(y)> will result if she chooses to stay childfree. We are assuming for now that the agent knows the texture of living a childfree live, which is not unrealistic since that is the life she has been living so far. And suppose the agent has a view on how aspects a 1(x)-an (x) compare with a 1(y)-a n (y), and that she finds the child-free life to be vastly more desirable when it comes to these aspects. Finally, suppose she very much welcomes the texture aspect of the child-free life. Then unless she thinks that there is no limit to the positive value of the unknown texture aspect of the life as a parent, it is quite possible that she will be able to compare the alternatives in question, despite being unable to determine the subjective value of the texture aspect of being a parent (and despite being unable to entertain what parenthood will be like phenomenologically).
Now Paul assumes that in most transformative decision-problems – in particular, in the choice between becoming and not becoming a parent – this kind of trumping is not (or, at last, should not be) a feature of the agent's preferences (Reference Paul2015 b: 479). Nevertheless, she does in some cases allow for trumping: she claims that when the difference in the other factors is great enough, then texture features will not tip the balance either way. For instance, she claims that we can know that being attacked by a shark is something to be avoided, even though we cannot know what it is like to be attacked by a shark before experiencing it (27–28). But since she admits that some transformative outcomes are such that one can know, before experiencing them, that their non-texture aspects outweigh any possible value of their texture aspects, it seems a bit puzzling that she does not admit that this may be the case for less extreme outcomes than being attacked by a shark.
3. THE TEXTURE FETISHIST
For the reasons outlined above, we find it safe to conclude that the epistemically transformative nature of parenthood does not make it impossible, for someone who approaches the parenthood decision-problem in the normal way, to rationally solve the problem (i.e. to solve it in accordance with the orthodox view in rational choice theory). Other people commenting on Paul's book have reached a similar conclusion.Footnote 8 But there is a particular type of agent who arguably cannot, for the reasons given by Paul, rationally choose whether to become a parent. Paul's critics may be forgiven for failing to notice that there are such unfortunates, since the type of attitudes that are definitive of such agents have not received much attention in the rational choice literature (or anywhere else).
Consider such a person who finds it absolutely impossible to compare outcomes unless he knows exactly the texture aspects of the outcomes. This might not be because the person only cares about the texture aspect. However, for him, the value of an outcome is to a very large extent determined by the texture of experiencing the outcome. Moreover, there is, according to him, neither an upper nor lower bound on the possible subjective values that the texture aspect might have, nor does he have any probability distribution(s) over these possible values. In particular, for a transformative decision like that of becoming a parent, the person thinks that it is possible (without being able to assign the possibility even a range of probabilities) that the value of the texture aspects outweighs any benefit that one alternative has over another in terms of non-texture aspects. Therefore, he has no rational way of forming a preference between transformative alternatives (nor between a transformative alternative and a non-transformative one).
To further illustrate the attitudes of the person we have in mind, we can imagine that he consults a crystal ball, that informs him exactly how happy and wealthy he will be if he becomes a parent, how it will affect his career, how pleasant or unpleasant it will be, how happy the child will be, and so on. More generally, the crystal ball can give him any possible information about the effects of becoming a parent, except the exact texture of the experience. The same holds for the option of not becoming a parent. Now, even with all this information, the person we are imagining cannot compare the value of the option of becoming a parent with the value of the option of not becoming one. So, his preference between these options remains incomplete even with all this information. For he intrinsically cares deeply about – let us say, for vivacity, fetishizes – the mere texture of experiences; he thinks that this texture has the capacity to make parenthood either more or less desirable than the alternative, and that this holds almost irrespective of other features of the alternatives.
4. REVELATORY VALUE
Suppose such a texture fetishist is faced with the choice between becoming a father and staying childfree. He must make a choice – even inaction constitutes a choice, namely, the choice to stay childfree. But how should he make his choice? Here is what Paul suggests:
[T]he proposed solution is that, if you are to meet the normative rational standard in cases of transformative choice, you must choose to have or to avoid transformative experience based largely on revelation: you decide whether you want to discover how your life will unfold given the new type of experience. (120)
In other words, to solve his decision-problem, a texture fetishist should not frame it in terms of choosing between becoming and not becoming a parent, but rather between having and not having the revelation that he will have if he becomes a parent.
Using the framework introduced above, Paul's solution consists in introducing a new aspect: revelatory value (r). So, the outcome-vector now looks like this:
$$\begin{equation*}
< {{{a}}_1}\left( {{x}} \right),{{ }}{{{a}}_2}\left( {{x}} \right), \ldots ,{{ }}{{{a}}_{{n}}}\left( {{x}} \right),{{ t}}\left( {{t}} \right),{{r}}\left( {{x}} \right) >
\end{equation*}$$
Or, when the texture aspect is undefined, it looks like this:
$$\begin{equation*}
< {{{a}}_1}\left( {{x}} \right),{{ }}{{{a}}_2}\left( {{x}} \right), \ldots ,{{ }}{{{a}}_{{n}}}\left( {{x}} \right),?,{{r}}\left( {{x}} \right) >
\end{equation*}$$
However, we strongly doubt that the introduction of this new aspect will solve the problem of incomparability. Suppose we follow Paul and simply ignore all factors except t and r. (Recall that Paul thinks that, for the cases at hand, if only one outcome has a defined t-value the incomparability prevails, and nothing is thus gained by adding another factor that is defined for both outcomes.) Then all outcomes can be represented by severely pruned vectors. When the t-factor is defined we can represent an outcome x as <t(x), r(x)>. Since you cannot come to know something you already do know, the value of r(x) is either zero (assuming that the measure is zero-normalized around outcomes that are neither desirable nor undesirable) or gives the description ‘one does not come to know how it feels like living x’ when the t-factor is defined. For simplicity, we will adopt the former alternative in the following, and represent such outcomes as <t(x), 0>. When the t-factor is undefined, an outcome x can be represented by <?, r(x)>, where r(x) is ‘coming to know how it feels like living the life in x’.
So in a transformative choice between the ‘phenomenologically familiar’ and the ‘phenomenologically unfamiliar’ the choice is between <t(x), 0> and <?, r(y)>. How is the introduction of the r-factor supposed to help the decision-maker? One option – trumping – is that the r-factor trumps the t-factor. But then Paul has first stipulated away good candidates for trumping factors (e.g. wellbeing, happiness, morality) just to ‘stipulate in’ what seems to be a dubious candidate for trumping factor.
Another option – deletion – is that we represent outcomes simply as <r(x)>. In other words, the idea is that when faced with a transformative decision, we delete all aspects other than the revelation – including the texture aspect. But then one wonders why the revelation aspect r was introduced in the first place. We could just as well have solved the incomparability problem by representing the outcomes in terms of some sub-vector of the original vector of outcomes, for instance, by focusing on aspects such as happiness, wellbeing and morality while excluding the texture aspect.
No matter which of these options we choose, there is also the question of how to compare r-factors of different outcomes. It should be evident why this problem arises if we adopt the second version of Paul's solution, i.e. if we assume that a transformative choice should be made solely on the basis of the r-factor. But the problem also arises if we adopt the first version of Paul's solution, i.e. if we assume that r-factors trump t-factors, since there are transformative choices between two unfamiliar outcomes, both of which will have undefined t-values.
In order to answer this last question, we need to get a firmer grip on the nature of revelatory value. It is clear that it somehow has to do with the value of coming to know what it is like to live a certain life. But what exactly is it that has this value? We see two main options. What has value is either:
-
(a) the revelation of a life-experience,
-
(b) the experience of revelation of a life-experience.
On (a), value is assigned to an epistemic state, that of coming to know what it is like to live a certain life. On (b), value is assigned to an epistemic experience, the experience of coming to know how it feels to live a certain life. So, by the revelatory value of being a parent we can either mean the value of the epistemic state of coming to know the experience of being a parent, or the value of the epistemic experience of coming to know the experience of being a parent.
For each of these possible value bearers, there is also a question of whether the value is supposed to be objective, in the sense that if something has this value, then no matter whether the agent in fact cares about it, she is rationally required to care about it; or subjective, in the sense that it is up to the agent whether it has value: if she cares about it, it has value, if not, not.
This gives us the following menu of choice for Paul:
Options (1) and (3), according to which we should make transformative choices based on the objective value of revelations or experiences of revelations, does not seem to fit well with Paul's general argument. These options would go against one aspect of the aforementioned ‘predominant cultural paradigm’. Recall that the purported problem that transformative choices raise is that they cannot be made in accordance with this cultural paradigm, according to which we should choose the alternatives that maximize subjective value, ‘where the subjective value we assign to an outcome depends upon what we care about, whatever that might be’ (25). Suggesting another decision-method that also violates this paradigm is hardly a solution to the problem in question.
Moreover, when formulating the problem of transformative experience, Paul insists on setting aside all non-subjective values, ‘in order to focus on the subjective value of outcomes’ (25). But if we are not allowed to appeal to e.g. objective moral or prudential considerations when solving the problem of transformative choice, it would seem odd if Paul allowed herself to appeal to some purported objective value of revelation. Why first ban objective values and then bring them back later on, when they could all along have helped the agent make a rational decision?
So it would seem that Paul couldn't favour any of the objective interpretations of revelatory value in her solution to the problem of epistemic transformation. What about option (2), according to which we should simply tell the agent to decide whether she wants to come to know what it feels like living a certain life – i.e. to decide whether or not she wants to be in a particular epistemic state, rather than having an epistemic experience? This option is also very problematic.
First, this option goes against another aspect of the ‘cultural paradigm’ Paul adopts, namely that we should make decisions by performing mental simulations about what the future would be like experientially. To ask the agent to decide whether they prefer to be in a certain epistemic state or not is not to ask her to imagine how things would feel.
Second, recall that the only type of person for whom the epistemically transformative nature of parenthood raises a serious problem is someone for whom the relevant values are to an important extent determined by the texture of the experiences of being, and of not being, a parent. In other words, the problem in question arises because this person has certain values: he intrinsically cares about the texture of experiences. Hence, it is not a solution to this problem to point out another type of good the valuing of which doesn't create the same problem. That would be to try to solve the problem that a texture fetishist is faced with by pointing out that another type of agent, with different values, does not face the same problem. Indeed, we already knew that the problem would disappear if the agent cared more about non-texture factors such as well-being, pleasure and morality. As Ruth Chang puts it (albeit when raising a slightly different objection to Paul's solution): ‘it isn't a solution to a problem to find a related phenomenon in the neighbourhood that doesn't raise the problem’ (Chang Reference Chang2015: 261).
That brings us to the final option, that of assigning subjective value to epistemic experiences (option 4).Footnote 9 This option will not go against the ‘cultural paradigm’, since this option assumes that the values are subjective and the value bearer is a certain kind of experience, which the agent can try to mentally simulate. One obvious question that this raises, however, is whether there is a distinct feel and flavour of being in the epistemic state in question. But let this worry pass.
Another problem is that this solution will only work for a texture fetishist that exclusively cares about epistemic experiences when choosing between two radically different lives, neither of which he is now phenomenologically familiar with. If in addition to caring about the subjective values of epistemic experiences, he also cared about the experience of living these future lives, then he would not be able to form an overall preference for one life over another, since he would be clueless about the subjective values of the experiences of the lives themselves.
Even if we put this troubling issue aside, the question remains how a person can determine the subjective value of the experience of a revelation. Since according to the texture fetishist, the texture of an experience determines, to a large extent, the value of an experience, Paul's solution does not work unless the person is able to know the texture of experiencing the revelation in question. But why should we think that people can know the texture of the experience of discovering what it is like to be a parent, if they cannot determine the texture of the experience of being a parent? After all, people who are deciding whether to become parents have never experienced this particular revelation before.
In some places, Paul seems to suggest that the texture and thus the subjective value of revelatory experiences is sufficiently invariant across different revelations such that one can use one's experience of past revelations to make decisions between transformative options (113–114). Note, however, that if this is to be true for a texture fetishist, then it has to be the case that he can draw on his experience of past revelations caused by certain experiences to determine the texture of experiencing possible future revelations caused by very different experiences. However, as far we have a handle on the notion of epistemic experiences, we cannot see why this must be true. The texture of the experience of learning how a certain experience feels like seems to depend to a large extent on that experience itself. For us who have tried both, coming to know how vegemite tastes seems phenomenologically quite different from coming to know what it is like to be a parent. Indeed, it is difficult for Paul to deny this possible variance, since she adopts an expansive account of phenomenology, according to which the phenomenology of an experience depends on its content. So two epistemic experiences with very different contents – i.e. coming to know very different experiences – cannot be assumed to have the same phenomenology (12, fn. 16).
Furthermore, if the phenomenology of revelations is invariant, so is the subjective value of such revelations. But this means that much of the significance of revelatory value is lost, since agents should now assign the same subjective value to <r(x)> and <r(y)>, for all alternatives x and y. In order to tip the balance of reason for or against an option, other non-texture aspects would have to be invoked.
If, on the other hand, the phenomenology of such revelations is not sufficiently invariant, then, since a texture fetishist needs to know what the experience feels like in order to assess its value – i.e. he needs to know the texture of the experience – it would seem that such a person cannot know the value of experiencing a revelation before he has had the revelatory experience. Another way to put this, is that if we take seriously what Paul says about the unique and transformative experiences of, say, having a child, and her claims about the unpredictable nature of the phenomenology of such experiences – which is meant to undermine the possibility of using evidence from past transformative experiences to estimate the subjective value of potential future transformative experiences – then it would seem that the same should hold for the experience of coming to know what it is like to be a parent. That is, her arguments for the hypothesis that one cannot know the texture of the experience of being a parent until one has had the experience, would, if successful, seem to suggest that one cannot know the texture of the experience of coming to know what it is like to be a parent. Hence, very little is gained, in terms of helping a texture fetishist decide whether to have a child or not, by reformulating the choice as one between having and foregoing the experience of the revelation in question. For just as he cannot determine the subjective value of being a parent, so he cannot determine the subjective value of experiencing the revelation that parenthood brings with it.Footnote 10
In conclusion, it seems to us that however we interpret revelatory value, Paul's suggestion for how to solve transformative decision problems is of no help to a texture fetishist. But recall that a texture fetishist is the only type of person for whom the problem Paul identifies arises. That is, a texture fetishist is the only type of person for whom the epistemically transformative nature of parenthood makes it impossible to rationally choose between becoming a parent and staying childfree. So, Paul's solution is no solution at all for the only type of person for whom the problem arises.
