Narrative, addiction, and three aspects of self-ambiguity

ABSTRACT ‘Self-ambiguity’, we suggest, is best understood as an uncertainty about how strongly a given feature reflects who one truly is. When this understanding of self-ambiguity is applied to a view of the self as having both essential and shapable components, self-ambiguity can be seen to have two aspects: (1) uncertainty about one's essential or relatively unchangeable characteristics, e.g. one's sexuality, and (2) uncertainty about how to shape oneself, e.g. which values to commit to, actions to pursue, or essential features to identify with. We explain how a narrative account of agency can accommodate these forms of self-ambiguity and argue that such an account also reveals another kind of self-ambiguity, namely, (3) uncertainty about whether one's established self-narrative represents who one really is. We illustrate this third form of self-ambiguity in the context of addiction where people's established addiction self-narratives make it difficult to identify with recovery. We argue that recovery will require embracing, especially, our third form of self-ambiguity as a chance for positive self-transformation. Treatment for addiction should, therefore, support people in going through and ultimately narratively resolving the inevitable self-ambiguities of the recovery process.


Introduction
'Self-ambiguity' is a feeling of uncertainty about one's authentic self, that is, an uncertainty about which values, actions, desires, emotions, bodily features, and so on most reflect who one is; 'Self-illness ambiguity' is a sub-category of self-ambiguity where the experience of mental illness, diagnosis, and treatment raises questions about authenticity (cf. Dings and Glas 2020;Sadler 2007). Self-illness ambiguity is represented by questions of the kind: 'Does my illness reflect who I am? Are these my desires, actions and emotional responses or are they attributable in part or in whole to my illness? '. (cf. Dings and Glas 2020) This paper begins by resolving a potential source of confusion in the account of selfambiguity put forward by Dings and Glas (2020). We argue that self-ambiguity 1 does not arise from simply having partially authentic features, that is, features that fall in the middle of what Dings and Glas (2020) call the 'spectrum of "mineness"'(335), but rather from uncertainty about the position of any given feature on that spectrum. We then apply this understanding of self-ambiguity to our preferred view of the self as having both essential and shapable components. This view reveals that self-ambiguity has two aspects: (1) uncertainty about one's essential characteristics, and (2) uncertainty about how to authentically shape oneself, e.g. how to prioritise conflicting values, actions, career paths, relationships and so on.
We then show how a narrative account of agency not only accommodates these two aspects of self-ambiguity, but also makes room for a third aspect, namely, (3) uncertainty over whether one's self-narrative represents who one really is. This form of self-ambiguity is possible because it is easier to change one's judgment about who one is than it is to change one's established self-narrative of who one is. One can then be uncertain as to whether one's judgment about who one is or one's established self-narrative best represents who one really is. In self-illness ambiguity this creates questions of the kind: 'Has my illness distorted my judgment or my self-narrative, and if so, which best represents who I truly am?'.
Finally, we illustrate this third form of self-ambiguity in the context of addiction where people's established addiction self-narratives often make it difficult to identify with recovery. We argue that recovery from addiction will require embracing a period of self-ambiguity because, as with any self-transformation, it inevitably involves developing oneself in a way that clashes with the established self-narrative. This self-ambiguity can be resolved either by effortfully reconfiguring the self-narrative to align with the judgment that one can recover, or by reverting to the judgment that accords with the established self-narrative -'I am an addict after all'and relapsing. Therefore, we argue, the treatment of addiction should, in part, be designed to support people through the inevitable selfambiguities of the recovery process.

Self-ambiguity
In this section, we begin our discussion by clarifying self-ambiguity as an uncertainty over the degree of 'mineness' (Dings and Glas 2020) of essential and self-creative characteristics. Dings and Glas (2020) suggest that the conceptual space for self-ambiguity appears when we move beyond a Frankfurtian dichotomous view that sees actions, values, desires, emotions, essential features, et cetera as either entirely our own (internal) or entirely other (external) (cf. Frankfurt 1988): 2 The alternative view […] is different in its emphasis on the fact that we might experience degrees of whether an action feels as ours, and that within such a spectrum of "mineness," there might be ambiguous cases where we are not sure whether something feels as internal or external (Dings and Glas 2020, 335).

Disambiguating self-ambiguity
We think that this description conflates two spectraa 'spectrum of mineness' and a spectrum of certainty about where X (e.g. an action, desire, value, or essential feature) falls on the spectrum of mineness. To properly understand self-ambiguity, we need to get clear on the relationship between these two spectra.
If we conceive of self-ambiguity as a potentially problematic state that an agent might want to resolve, then uncertainty, which we assume comes in degrees, is a necessary ingredient, but degrees of mineness are neither necessary nor sufficient. To see this, assume the dichotomous view of mineness, on which X is either internal or external but never in-between; it would still be possible to be uncertain about where X fell. One might think, 'I thought X was mine but now I am unsure; it might be external after all', or vice versa. We take it that this person is suffering from a self-ambiguity that they may want to resolve. So, the dichotomous view can accommodate a form of self-ambiguity, but it limits the possible ways in which the ambiguity can be resolved -X must be classed either as fully internal or fully external. Now, assume that there are degrees of mineness, but the agent is always perfectly certain of the positions of all Xs on that spectrum. To modify an example given by Dings and Glas (2020, 334), imagine an ignorant referee who will reject your article unless you implement a misguided point of feedback. Imagine that you come to the settled decision to implement this feedback anyway because, all things considered, it is better to get the work published quickly. The act of editing the paper significantly reflects someone else's ideas but, in making the compromise, you are complicit. The act partly reflects your own reasons and partly reflects someone else's reasons. Yet presumably, when you are certain about where X falls on your spectrum of mineness, you don't experience this as a problem of self-understanding that needs resolution. So, if self-ambiguity is supposed to refer to a category of experiences that people may perceive as problematic, then degrees of uncertainty are necessary but degrees of mineness are not.
That said, we agree that there is a continuum of mineness, not because it is needed to explain self-ambiguity, but because such a continuum is true to our folk-psychology. Therefore, we think that self-ambiguity refers to cases where the agent is uncertain about where X falls on their continuum of mineness and not cases where X falls somewhere in the middle of that continuum, but the agent is certain about where it falls.
Even though the continuum of mineness isn't necessary for self-ambiguity, it does ratchet up the epistemic challenge. Rather than simply working out whether X is mine or not, one must figure out where it sits in relation to other Xs on the continuum. Presumably, the greater the uncertainty, that is, the further the possible position of X on the spectrum of mineness from the position one thought it had, the greater the potential distress. 3

Self-creation and self-acceptance
The scope of self-ambiguity, that is, the range of Xs one needs to concern oneself with, depends on how one views the self. We assume that typical adults can shape or create themselves within the limits set by their essential features, that is, features which are difficult or even impossible to change, such as, genetics, various physiological characteristics, and childhood socio-economic context. This means that developing certainty about one's authentic self is not only about settling on one's values and how to express them but also about accepting one's essential features and settling on an attitude towards them.
There is a close relationship between identifying with X, valuing X, and shaping X, but they do come apart. We tend to identify strongly with our values (i.e. place them at the internal end of the continuum of mineness). Indeed, studies show that people widely take moral values to be the most central aspects of who someone is (Strohminger, Knobe, and Newman 2017). But people also identify with and value various relatively unchangeable, essential aspects of themselves, such as, sexuality, physical features, family, nationality, et cetera. A tendency to identify with things we value supports the self-esteem and selfrespect that are necessary for autonomy (Mackenzie 2014b). Nevertheless, it is possible to identify strongly with disvalued aspects of oneself. One might come to think that a physical deformity significantly defines who one is and wish for a different body, or one might identify strongly as an addict despite wishing one was not. Although an excessively negative self-concept undermines agency and is indicative of poor mental health, a commitment to accepting the truth about oneself will usually entail identifying with some disvalued features to some extent.

Realism and the authentic self
The resolution of self-ambiguity depends on whether one takes a realist or anti-realist view of the authentic self. We assume a realist view of the self, according to which our self-concepts can be more or less accurate in tracking the truth about who we are and who we are becoming. This does not entail an essentialist view of the self on which the only goal of self-understanding is discovering one's static underlying nature (Kristjánsson 2010). Realism about the self is compatible with people being able to transform themselves entirely. The realist just believes that people can be mistaken about what they think they have transformed into. We prefer this to an anti-realist view which precludes the possibility of being mistaken about oneself (see Kristjánsson [2010] for a detailed argument).
Furthermore, we think that the constraints of reality should extend to claims of authenticity. Therefore, a person who claims that their authentic self is honest and fair, is wrong about themselves if they lie and cheat to get ahead. Similarly, a Type 1 diabetic who finds jazz aversive is wrong if they insist that their authentic self is a non-diabetic jazz-lover. Such inaccurate self-presentations might be deliberately deceptive but, because we are not perfectly transparent to ourselves, they can also arise from mistaken self-beliefs or, in some cases, outright self-deceptions motivated by wishful or negligent self-reflection. In sum, we believe (contra Strohminger, Knobe, and Newman 2017) that there is such a thing as a true or authentic self even if people often have inaccurate views of their authentic selves and even if it is not always easy to verify claims of authenticity.
On our view then, some Xs feature on one's continuum of mineness whether one wants them to or not; for example, if one has Type 1 diabetes, then one should identify, however weakly, as a diabetic. Reality also constrains the ease with which Xs can be moved along the continuum of mineness. Some people might easily identify as jazzlovers because they find the rhythms and melodies of jazz intuitively appealing, while for others it might be more challenging at first, because jazz rhythms and melodies initially strike them as unpleasant. To the extent that a given X's position on the continuum of mineness is beyond our control, we must settle for accurate self-discovery and self-acceptance over self-creation. To the extent that Xs can be moved, we can engage in self-creation, settling on where X would ideally feature on the continuum of mineness and working towards making that true. In other words, we assume what Pugh, Maslen, and Savulescu (2017) call a 'dual-basis framework' of authenticity which … … incorporates the essentialist notion that certain elements of an individual's character are more or less fixed, but also the existentialist notion that individuals can choose which of these more or less fixed elements to bring to the fore, and which to downplay in projects of self-creation (Pugh 2020(Pugh , 1667 The implication of this 'dual-basis' view for self-ambiguity is that some uncertainty will be resolved by finding X's accurate position on the continuum of mineness, and some uncertainty will be resolved by moving X to a position on that continuum that reflects our settled preference. The balance of self-acceptance and self-shaping will depend on the nature of the X in question.

Two aspects of self-ambiguity
Using the 'dual basis' view of authenticity we just outlined and the understanding of selfambiguity we proposed in Section 2.1. We can distinguish two aspects within Dings and Glas's (2020) notion of 'self-ambiguity': The first aspect of self-ambiguity is an uncertainty about one's essential characteristics. One might wonder about one's sexuality, family background, and natural talents. Are these characteristics genuinely essential or can they be changed? How strongly does one identify with or value these characteristics? This aspect of self-ambiguity will often come to one's attention as a clash between one's embodied, pre-reflective experience and one's cognitive representation of one's authentic self or 'self-concept' for short (see Dings and Glas [2020, 336], for the distinction between the 'unreflective' and 'reflective' level): Imagine a homosexual man raised as a fundamentalist Christian who has simply not considered the possibility that he is homosexual. His self-concept is consistent with his Christian views, so, when he feels attracted to a man, it strikes him as totally alien. At first, he assumes his self-concept best captures who he really is and not his embodied experience, but eventually he accepts that he is homosexual. He thus resolves his selfambiguity by adjusting his self-concept to fit his bodily reality. He might not, of course, come to value that aspect of his self-concept. We can imagine scenarios where he values and identifies strongly with being homosexual, abandons the church, and moves to a more 'gay-friendly' community. We can also imagine scenarios where he remains ashamed of being homosexual and does little to explore or develop his sexuality.
In other cases, our embodied experience can be shaped to fit with our valued self-concepts. For example, a soldier might feel physically and emotionally sick after killing people in war but deliberately try to identify with killing, so that he can do his job better and sleep at night. If he claimed at first that he killed wholeheartedly, he was lying or kidding himself but, over time, we can imagine that he becomes comfortable with such killing, thereby resolving his self-ambiguity. It is often unclear what the limits of self-shaping are, so part of the self-ambiguity the soldier faces is uncertainty over whether he can identify with killing as strongly as he hopes to.
The second aspect of self-ambiguity we identify can be characterised as an uncertainty about one's potential, shapable characteristics. What exactly does one value and how should one express those values? How should one prioritise conflicting values, plans, or relationships given one's essential characteristics? For example, a talented adolescent might want to become a concert pianist and a professional football player and be blessed with the talent to realistically succeed in either career. Unfortunately, however, there isn't enough time in the day to develop both talents to a professional standard and she doesn't want to settle for being merely a good pianist and a good footballer. Therefore, she must choose. Both ends might seem equally central to who she is right now, but practical rationality requires that one end be moved sufficiently far down the continuum of mineness that it no longer undermines the other. 5 Until this decision has been made, the talented adolescent experiences self-ambiguity because she is uncertain about where each of these ends fit on her continuum of mineness. We understand this ambiguity primarily as an uncertainty about how to settle on a diachronically stable hierarchy of values: In the long run, the talented adolescent will need to decide how to order the value she places on playing the piano in relation to the value she places on playing football.
We can note at this point that we don't need a specifically narrative account of selfconcept to understand these two aspects of self-ambiguity. Below, we will show how a self-narrative account accommodates these two forms of self-ambiguity but also makes conceptual space for a third form which can only be understood with reference to self-narrative structure. Before doing so, we outline our favoured account of selfnarration.

A narrative account of agency 6
Narratives are sequences of symbols that specify the causal, teleological, or thematic connections between events, rendering them meaningful (Schechtman 2007). As such, narratives are more than lists, but often include banal and extraneous material, and so, don't meet the aesthetic standards of novels or films. In self-narration, people select the meaningful connections between experienced, anticipated, and planned events that they judge to be relevant to their lives. This is an iterative process of self-interpretation and self-projection involving self-creation and self-discovery (McConnell 2016b, 30). In selfcreation, as 'inventor-enactors' (Velleman 2005, 221), agents imagine plausible valued futures for themselves and then choose to enact the projection they judge the most valuable, setting aside less valued projections. The agent monitors progress with selfinterpretation and, when discrepancies between self-projection and self-interpretation arise, she updates the project in light of those interpretations (McConnell 2016a, 310). To the extent that the projection is accurate, the narrator has a ready-made self-interpretation. The success of self-creation depends on sufficiently realistic projections based on sufficiently realistic self-interpretations. Self-narratives also incorporate self-interpretations of our more essential features, e.g. aspects of our physiology, family, or socio-economic context. Importantly, again, we have some control over how we self-narrate the essential characteristics that we cannot change: We can be proud or ashamed of them, emphasise or deemphasise them in our self-narrative. This is most clearly the case when we use our essential features to make sense of our self-creative projects. For example, I might pursue basketball because I am tall or campaign for providing children with free school lunches because I suffered from food poverty as a child.
Self-narration is embodied. Although our bodies have some relatively unchangeable features that place limits on accurate self-narration, we can also shape our bodies and our experiences through self-narration: 'Insofar as we are narrative agents, our experience of our embodiment and our agency are mutually interwoven in a complex, although not seamless, unity' (Mackenzie 2014a; see also Dings 2019). Self-ambiguity arises at these 'seams' and we will return to this below.
Self-narratives don't just describe; they also motivate. One will tend to perceive, think, and act in ways that make sense in light of one's established self-narrative and not in ways that would contradict it (McConnell 2016a, 314). This is because our self-narratives represent our current best attempt to understand what is going on in our lives. To move away from our self-narratives and go 'off-piste' would leave us reacting and planning in a suboptimal way (by our own lights). To act contrary to one's self-narrative would leave one relatively incomprehensible to oneself and self-narrators find this aversive.
Social contexts exert a strong influence on self-narration. Especially as children, but also as adults, we narrate based on cultural archetypes and more specific content that others, acting as co-authors, provide for our self-narratives (cf. MacIntyre 1984). We are also sensitive to the ways that others validate or challenge the self-narratives we tell. Those self-narrative threads that are consistently confirmed by others tend to become more deeply entrenched in our self-narrativeswe have increasingly good reason to believe they are true and relevant to how others see us (McConnell and Snoek 2018, 33). Co-authoring generally provides a shortcut to accurate and empowering self-narration, but we also need to recognise and protect ourselves from uncharitable, and malicious co-authoring, which can entrench harmful content in our self-narratives (McConnell 2016b). On our view, a self-narrative is not typically one unified and exhaustive story of someone's life, but rather, a collection of narrative threads (cf. Lloyd 1993;Wollheim 1984). Some threads cluster around particular foci (e.g. one's career, specific relationships, a traumatic accident), while other threads are more general and make sense of the relationship between foci (e.g. 'I began my career as a nurse after seeing what good work nurses did during my rehabilitation from the accident'). Some threads detail closed chapters (e.g. one's experience of high school), while others are open-ended (e.g. promoting social justice); yet others might be left permanently incomplete (e.g. a home improvement project left unfinished).
Two considerations about narrative will be especially important for our discussion of narrative and self-ambiguity below: First, we need to understand the relation between self-narration and one's judgment of who one is: Usually, the two go hand-in-hand because we choose to enact the narrative self-projections we most value. Nonetheless, sometimes self-narrative and judgment come apart. When this happens, our judgment does not get completely cut loose from the self-narrative, since, at the very least, one can identify oneself as the person making this divergent judgment. Yet, there are fewer self-narrative threads making sense of this judgment in comparison to the much larger part of one's self-narrative that clashes with the judgment.
Second, we suggest that narrative structure can be understood as another continuum of mineness: Some narrative threads become more central or deeply internalised in the overall self-narrative than others. The centrality of a feature (or 'X') in a self-narrative is a function of how many narrative threads include that feature and how many other narrative threads assume the truth of that first set of threads. For example, to discover late in life that one was adopted is likely to radically upset one's deeply established self-understandings, while to discover that one was merely mistaken about which hospital one was born in, probably will not.
The more central a thread is to the overall self-narrative, the more motivating it will be, because the more alienating it will feel to act contrary to it. If one takes a narrative thread to represent the truth about who one is, acting in ways that require implausible continuations of that narrative will feel alienating or even seem impossible (McConnell and Snoek 2018, 34). In other words, once self-narrative threads have become deeply internalised, they feel like essential aspects of self, even though, for the most part, they began as optional and could (with the narrative work described below) be re-interpreted. The self-narrative structure, therefore, operates as a kind of continuum of minenessthe central threads being at the more internal end of the continuum and the peripheral threads being at the external end. With this in mind, we turn to the implications of our narrative view for self-ambiguity.

Self-Narrative and three aspects of self-ambiguity
In this section, we show how the above self-narrative account of agency can accommodate the two aspects of self-ambiguity we identified earlier, and crucially, make room for a third kind where the narrative structure instantiates a continuum of mineness. This yields a tripartite view of authenticity with three distinct continua of mineness: (1) The embodied continuum (that features in the first aspect of self-ambiguity described above) refers to the physical structure of our bodies: Some experiences of bodily interactions with the world will feel more comfortable and more mine than others because they fit more naturally with one's bodily reality, e.g. feelings of sexual attraction to some people and not others or the experience of listening to jazz being pleasant or unpleasant. (2) The evaluative continuum (that features in the second aspect of self-ambiguity described above) refers to a hierarchy of values: One identifies one's values and hierarchically orders them according to how centrally they reflect who one is, e.g. prioritising pursuing excellence in piano over excellence in football. This still allows for the possibility of identifying with disvalued characteristics as a function of being committed to the value of truth or honesty to oneself. (3) The newly added narrative continuum concerns the centrality of narrative threads. The more central a thread is in one's narrative, the more internal it will be on this continuum of mineness. Now, we are suggesting that self-ambiguity can arise if one faces uncertainty about the position of a feature on each of these continua, and we can be thrown into such uncertainty through mismatches between the continua. What our narrative structure indicates is most mine can diverge from our evaluative judgment of who we most are, and from our essential characteristics which place their own limits on who we really are and can be. X will feel most like mine when it aligns well with all three continua of mineness, but we get a messier picture when it aligns well with some but not others. With this in mind, we will now redescribe the two aspects of self-ambiguity we sketched earliernow adding the narrative dimensionand then introduce another self-ambiguity that characteristically arises when the narrative continuum misaligns with the evaluative continuum. What we will now call 'Type 1 self-ambiguity', described in narrative terms, is an uncertainty about how to self-narrate what appears to be an essential aspect of oneself, or an uncertainty over whether an established self-narrative accurately captures an essential aspect after all. The homosexual fundamentalist Christian will be uncertain about how to self-narrate his attraction to men within his wider self-narrative. He will note that his established self-narrative fails to accommodate his bodily reality and try to make sense of this.
What we will now call 'Type 2 self-ambiguity' is an uncertainty about how to settle on a hierarchy of values and how to express those values while maintaining a sufficiently coherent self-narrative. To avoid future narrative incoherencies, the talented adolescent must settle on whether she ultimately values playing the piano or football more highly. If we fail to settle on a diachronically stable hierarchy of values, then our self-ambiguity at the evaluative level begins to infect our self-narrative and we get incoherence within the self-narrative. If the talented adolescent fails to settle her judgment, her 'concert pianist' self-narrative focus will develop in parallel to her 'professional footballer' self-narrative focus. She might decide late that she is going to be a footballer after all. But then, she has to deal with the greater disappointment of 'retiring' her partly developed pianist narrative focus. On this view then, Type 2 self-ambiguity is primarily an uncertainty within one's hierarchy of values that can lead to an incoherent self-narrative. It is an uncertainty on the evaluative continuum that comes to be reflected in an uncertainty on the narrative continuum, rather than a clash between narrative and judgment. 7 The new, 'Type 3 self-ambiguity', we introduce can arise if the content of one's self-narrative clashes with one's judgment of who one is. One can identify with X more or less wholeheartedly at the evaluative level without having (yet) deeply internalised it in one's self-narrative. Conversely, one might have internalised X deeply in one's self-narrative but that doesn't necessarily say anything about how wholeheartedly one now endorses it as 'mine' at the evaluative level. One might feel that X has been made overly central to one's self-narrative or excessively marginalised due to misleading coauthoring or one's own misguided self-narration. When there is a conflict between one's evaluation of how well X reflects who one is and how centrally X is incorporated in the self-narrative, the agent may be unsure what best represents who she really is: the self-narrative or the occurrent judgment. Such a clash between narrative and judgment is especially likely to occur when our judgment of who we are changes. Due to its inert structure, narrative is more difficult to change, so it will lag behind more agile changes of judgment, leaving one in a state of self-ambiguity.
Consider again the example of the homosexual fundamentalist Christian. When we first introduced the example, we described him as facing a clash between the cognitive and the embodied level. Now, at the cognitive level, we can divide again: into narrative and evaluative judgment, which can clash with each other or together clash with embodied experience. In the earlier version of our example, the man's self-narrative and judgment of being a fundamentalist Christian aligned with each other but clashed with his embodied experience of attraction to men. He resolved this kind of self-ambiguity by adjusting his judgment to fit his bodily experience: He judged that he is, in fact, homosexual.
The potential split between narrative and judgment now adds that, despite this resolution, the homosexual Christian may still face another kind of self-ambiguity: Judgment and embodiment may align with each other but clash with his established self-narrative: 8 Having grown up as a fundamentalist Christian, the man likely has a self-narrative as a fundamentalist Christian that has been established throughout his life, by co-authoring and validation from his community. This brings with it certain projections of how his life is meant to continue. Homosexuality is unlikely to feature as even a possible way of being in such a self-narrative. Coming to judge that one is homosexual may itself take a lot of time and effort in such a setting. Yet, adapting the self-narrative accordingly will take even longer, due to its inert structure. The Christian may then judge himself to be homosexual but remain uncertain whether his Christian self-narrative or his judgment of being homosexual better represents who he really is. Even after judging that he is homosexual, the Christian may struggle to change other established aspects of his self-narrative, for example, he might find it almost impossible to revise his narrative projection of marrying a woman and raising a family with her.
A clash between narrative and judgment of what a good (future) life entails can be a serious problem. We are motivated to enact our disvalued self-narratives because they make sense to us and to act otherwise would seem alien and unrealistic (McConnell and Snoek 2018, 34). 9 We will return to this when we consider addiction below.
To summarise this section, we have distinguished between three continua of mineness: embodied, evaluative, and narrative. We have shown how self-narrative can accommodate and refine the two identified kinds of self-ambiguity and make room for a third one. There are then three interrelated forms of self-ambiguity: (1) Uncertainty about our essential characteristics, typically revealed by a mismatch between our embodied response to a situation and either or both of our self-narrative and judgment about who we are; (2) Uncertainty in the evaluative stance over what counts as a good life and/or what aspects of oneself truly represent who one is. This is typically revealed by unexpected embodied responses to the world or self-narrative developments that bring existing judgments into question or change which valued futures are realistically accessible. If unresolved, uncertainty at the evaluative level will go on to create incoherence in the self-narrative; (3) Uncertainty about whether our self-narrative represents who we really are, typically revealed by a mismatch between self-narrative and evaluative stance (or embodied response to stimuli).
We have indicated that the clash between narrative and evaluative judgment is especially problematic when it keeps the agent from pursuing more valued projections. However, in the final section, we will suggest that such an ambiguity can also be seen as a chance for positive self-transformation. Before doing so, in the following subsection, we will suggest how distinguishing forms of self-ambiguity makes a difference to how one should resolve them.
4.1. Implications for resolving self-ambiguity: self-narrative work Dings and Glas (2020, 336) suggest that self-ambiguity can be resolved as follows: The proposed solution then is to re-negotiate oneself (or "one's self"); to reflect on one's selfconcept or deliberate on one's self-narrative. When this self-concept or self-narrative is made coherent, the phenomenological ambiguity is resolved in that one has either accommodated or endorsed the ambiguous factor (i.e. internalized) or discarded it (i.e. externalized). In other words, resolving self-ambiguity, according to the current proposal, entails achieving a form of congruence between one's reflective self-understanding (e.g. one's self-concept) and the bodily and affective feedback one receives while unreflectively acting on the world.
We would first note that, as discussed at the beginning of the paper, we think that selfambiguity doesn't have to be resolved by internalising or externalising X, but rather by reducing uncertainty about X's position on the continuum of mineness. That issue aside, Dings and Glas actually refer to two different goals here: achieving self-narrative coherence and achieving congruence between the self-narrative and bodily feedback. We agree that both processes are important but would add that, because evaluative stance and self-narrative can come apart, sometimes the resolution of self-ambiguity will also require achieving coherence among one's values and congruence between evaluative stance, self-narrative, and bodily feedback. More specifically, resolving Type 1 self-ambiguity will require working out what one's essential characteristics are. One will need to self-narrate more accurately (self-discovery) or, where possible, find ways to get pre-reflective experience to fall in line with a valued self-narrative (self-shaping). This can happen through introspective reflection on how one (re)acts in given situations over time. The soldier might discover that he gradually gets more used to killing or he might find that he does not have it in him to kill ruthlessly after all.
Resolving Type 2 self-ambiguity will require identifying which valued ends are available and determining who one wants to become out of the various valuable possibilities. There may be no one way to do this, but likely one will have to try out different options and engage in imaginative projections. The talented adolescent might reflect on whether she likes playing football or piano better, but she might also consider whether the long career of a pianist might outweigh the more widely appreciated but shorter career of a footballer. One should try to make one's imaginative projections as accurate as possible, for instance, by seeking advice from those who have taken the path one is considering.
Resolving Type 3 self-ambiguity requires 'self-narrative work' . In self-narrative work, we can increase the centrality of an aspect by consciously increasing the narrative threads in which it features, or decrease its centrality by replacing narrative threads in which it features with alternatives. When a self-narrative thread (that would align with one's judgment) feels alien, one doesn't necessarily have to take that pre-reflective feeling at face value and reject the thread. Often self-narrative threads that initially seem alienating can still be integrated in the self-narrative, but it takes reflective work to do so. Self-narrative work might involve relatively creative and detailed means-ends planning ('Perhaps I have the means to achieve this goal after all') and/or trying to think of ways that the narrative trajectory makes sense ('Now that I think about it, I've always wanted to try something creative'). Agents can reinterpret parts of their established self-narratives, so that these become more compatible with the valued projection. The greater the disparity between the potential narrative thread and the established self-narrative, the more narrative work will be required (see McConnell and Snoek [2018] for a detailed discussion).
As an example of narrative work, we can imagine the homosexual fundamentalist Christian reflecting on whether he should marry a woman or change this narrative projection to align more closely with his new judgment of who he is. This will require narrating a plausible connection between his current self-narrative as a fundamentalist Christian and a self-projection of living out his sexuality. One could imagine him embracing his trajectory and becoming a mentor for young people with similar struggles. Or perhaps he could manage to find an interpretation of his faith that makes room for his sexual orientation.
Another case in which such narrative work is crucial is addiction -We will now turn to applying the foregoing analysis of self-ambiguity types and their resolution to addiction.

Self-ambiguity in addiction
Addiction is a self-destructive pattern of habitual substance use involving some agential impairment. Criteria for diagnosing addiction (see American Psychiatric Association 2013) include: not managing to do what one should at work, home or school because of substance use; continuing to use, even when it causes problems in relationships; giving up important social, occupational or recreational activities because of substance use; taking the substance in larger amounts or for longer than one meant to; and wanting to cut down or stop using the substance but not managing to.
We have yet to identify a distinctive aetiology that occurs in all and only in cases of addiction. Nevertheless, most addiction cases share mutually reinforcing neurological, environmental, and psychological causal features. Dysregulation of the dopamine system makes one unusually vulnerable to temptations to use drugs. This might be experienced as an unusually strong craving despite both disvaluing drug use and not expecting to enjoy it (cf. Holton and Berridge 2013; see also Holton 2009). Addictive temptations are also thought to cause 'judgment shifts', so that under temptation, the addicted person temporarily comes to see drug use as the best option (Levy 2014).
In addition, addicted people have often grown up in, or cultivated, environments where cues for drug use abound. Over extended periods of drug use, opportunities for alternative lifestyles become increasingly limited. Furthermore, long-term addicted persons often suffer from intense shame (Flanagan 2013), low self-esteem and selfrespect, and have developed disvalued self-narratives as 'hopeless addicts'. These psychological features, reinforced by social stigma, undermine the belief that one can, or deserves to recover (cf. Kennett, McConnell, and Snoek 2018). Nevertheless, drugseeking andusing behaviour remains sophisticated, reasons-responsive, and contextsensitive. So, although addicted agency is impaired, it remains under significant control, and recovery depends on the addicted person's efforts (Pickard 2017).
In the following, we outline how addiction becomes embedded in self-narrative and how the three discussed types of self-ambiguity arise in addiction. We argue that, especially, going through Type 3 self-ambiguity will be crucial for recovery.
In the early stages of drug use, there is unlikely to be any self-ambiguity: Regular drug use begins like any other activityit is valued for its own sake and/or for other values it facilitates (Pickard 2021). There is no self-ambiguity because drug-users will typically feel sufficiently certain of how strongly they identify with drug-related actions, desires, values, beliefs, embodied experiences, et cetera. Their judgment about who they are, self-narrative structure, and bodily reality are all well-aligned. For some, these drug-related aspects of self might be very central: I had established myself as a druggie. My friends and family knew me as such, and in a way I was proud of my varied life experiences and my street-smarts (Kate 2013). For Kate, drug use strongly reflected who she was; she valued it highly, embedded it deeply in her self-narrative and, presumably, it also aligned with her embodied, pre-reflective sense of who she was. For others, drug use might be more instrumental, e.g. to help manage pain or social anxiety, and so, not valued so highly or incorporated so deeply in the self-narrative. But, as long as these individuals are sufficiently sure of the position of drug use on their continua of mineness, they won't experience self-ambiguity. Indeed, at this stage, there is no guarantee that a serious addiction will develop; some people manage to balance drug use with other values indefinitely.
When drug use becomes problematic, it tends to become more internalised in the selfnarrative, at the exclusion of other threads. For many, drug use can begin to dominate their lives. Marc Lewis (2012) describes this process as the 'hourglass shape of addiction': Each person begins with his or her own specific culture, family environment, level of education, personality, social network, personal secrets, and all the rest. But then, when addiction takes hold, these lives start to look exactly the same. Regardless of whether it's cocaine, opiates, alcohol, or even food, that wide range of individual differences shrinks to a narrow tubethe middle of the hourglass.
In the 'middle of the hourglass', the non-drug using facets of life are minimised or set aside entirely. As a result, the self-narrative threads one had developed around hobbies, work, and many close relationships are no longer developed. Some threads might come to a premature end, e.g. if one is fired from one's job or rejected by a friend. The narrative threads related to sourcing and using drugs, however, are the dominant focus and steadily built upon.
Friends and family often withdraw, so the addicted person loses her supportive coauthoring. Addicted individuals no longer play their part as son or daughter, sibling, parent, team member, colleague et cetera (cf. Kennett, McConnell, and Snoek 2018). The co-authors available tend to be other drug-users who aren't typically interested in supporting the person's non-drug-using facets. Other co-authoring takes the form of stigma and even family can despair: Oh I know the person that's using the drugs is not the son I had, I understand that, that's a shell of the person I knew … I mean all I see is an addict, I don't see my son anymore because I know that's not my son, definitely isn't my son (parent Mr Bell in Barnard 2007, 50).
Once addiction has taken hold, drug-related aspects of one's life have become central to the self-narrative while non-drug-using features that had previously been central have been marginalised. From here, there are three general paths: (1) Denialone doesn't have a problem; (2) Resignationone has a problem but it can't be solved; (3) Hopeone has a problem and recovery is possible but difficult.
Each of these three paths, we suggest, has a different self-ambiguity profile: (1) Denial Denial sets the scene for Type 1 self-ambiguity, as embodied reality increasingly clashes with self-narrative (and judgment): An addicted person in denial turns a blind eye to the mismatch between what is really happening in her life, and her valued but inaccurate self-narrative. This doesn't initially cause self-ambiguity but, typically, the mismatch between reality and self-narrative eventually becomes so great that the deception cannot be maintained. The addicted person then feels strong self-ambiguity when they finally realise that their self-narrative is inaccurate, but they do not yet know how to narrate the reality they have been ignoring: I was angry when I got into treatment. I didn't relate to the fact that drugs were the problem. I thought God was punishing me. I thought, I am not an addict, others are addicts. I can successfully do drugs. That adventure failed. I am an addict. (a nurse in Hutchinson 1987, 341).
Typically, the dishonesty with oneself required for denial is supported by (sometimes unwitting) dishonesty with others who might challenge the self-deceptive self-narrative. Once the addiction becomes serious, others tend to see through this dishonesty and may then contribute to ending the addicted person's self-deception by calling it into question. On our view, achieving some level of sincerity with oneself and others is a precondition for successful self-transformation and resolution of self-ambiguity. Although one will never be perfectly transparent to oneself, seriously inaccurate understandings of the basis of self-ambiguity will prevent one from recognising the ambiguity, misdirect efforts aimed at resolving it, or lead to merely apparent resolution.

(2) Resignation
Resignation is a way of avoiding or resolving Type 2 and 3 self-ambiguity: Resigned 'addicts' identify with their addiction and so give up on values and narrative projections that would clash with an addicted lifestyle and would thereby cause self-ambiguity. This way of avoiding self-ambiguity comes at the cost of excluding the non-drug-using aspects of the self: I went to places I used to swear I would never go. I did things I could not imagine myself doing. I hung out with people that at one time I would cross the street to avoid. There came a time when, looking into the mirror, I honestly did not know just who was looking back at me ( Kennett, McConnell, and Snoek 2018). (3) Hope Those who do harbour hope of being able to recover, live with some Type 2 and 3 selfambiguity and will initially experience this more strongly in their attempts to recover: Type 2 self-ambiguity occurs when the increasingly adverse effects of drug use lead addicted people to question the value they place on drug use and to wonder how to prioritise it in relation to other values. But people are often reluctant to demote drug use sufficiently far in their hierarchy of values. As a result, they struggle with self-ambiguity over what they want to value: Well a couple of times I did try to stop for different reasons like I tried to stop because ma mum was goin' to throw me out of the house or because ma sister wasn't talking to me or because ma girlfriend had left me or because I was going to get the sack (from) ma job. But basically what was happening there was I was doing it all for the wrong reasons. I wasn't doin' it for maself and it is the only way it works if you do it for yourself because all the answers lie within (Jo in McKeganey and McIntosh 2001, 55-56).
Jo struggled to recover because what he deemed his more authentic values and reasons still spoke in favour of drug use.
Those who increasingly disvalue drug use and judge themselves to be someone who can recover, will struggle with Type 3 self-ambiguity. By the time the addicted person decides that she values recovery, her established self-narrative as a 'hopeless addict' tends to make recovery-directed actions feel alien (McConnell 2016a). These addicted people now value recovery over drug use but struggle to internalise that change of values deeply on their narrative continuum of mineness. They can then face an ambiguity between an evaluative stance that supports recovery and a narrative that makes recovery seem impossible and alienating: One reason why it is difficult to establish a diachronically stable evaluative stance in which recovery has more value than drug use is that drug use remains central to the self-narrative. The new narrative threads articulating one's commitment to recovery are relatively marginal and they clash with that established addiction narrative. The dominant addiction self-narrative can continue to draw one back to drug use: I felt not so much like I missed the druggie lifestyle, but that I was starting to lose my grip on who I was, and was finding it hard to function. I was tempted to return to old habits for it was all I knew. I felt like I was a sell-out and was disloyal to my past. I felt like if I moved on I'd have nothing, just a big void on my CV, where a whole bunch of jobs and study should have been … When you've lived all that to such an intensity, it's hard, and feels quite disloyal to move on and forget it. You fear forgetting it, in all its realness and richness (Kate 2013). In this self-narrative context, it is recovery that can feel alien.
To make room for change, these addicted persons must ask themselves: 'Am I better represented by my evaluative stance or by my self-narrative of addiction?' 'Am I really capable/worthy of recovery, or does my addiction self-narrative represent who I really am?' If one wants to recover, one has to hope that who one really is can be better represented by one's evaluative stance and nascent recovery self-narrative than by one's established addiction self-narrative.
Recovery from addiction, we suggest, will then first require largely overcoming Type 2 self-ambiguity because one must consistently value recovery highly in order to motivate the lengthy and difficult process of recovery. Because judgment changes more readily than self-narrative, one can make progress in resolving Type 2 self-ambiguity prior to changing one's established self-narrative of addiction. Therefore, to recover, one will then have to embrace a period of Type 3 self-ambiguity, as the mismatch between the more valued judgment of who one wants to be, and the disvalued addiction self-narrative grows increasingly large.
This means that therapists should indeed carefully elicit Type 3 self-ambiguity by encouraging patients to value aspects of recovery that explicitly clash with the established 'addict' self-narrative, e.g. having a valued job or good relationships with family. Encouraging such judgments will also point the individual towards the kinds of activities through which those judgments can begin to be embedded in supportive narrative threads, e.g. embarking on teacher training or setting up social meetings with family.
In the context of addiction, this Type 3 self-ambiguity will initially cause one to feel alienated from one's recovery, a feeling we might call self-recovery ambiguity. Addicted individuals may ask themselves whether the value they place on recovery and their nascent developing recovery self-narrative threads are really theirs or instead an imposition by society. This means that to recover, addicted persons must not only realign their narrative with their evaluative stance, but also weave recovery into their self-narrative such that the recovery becomes plausibly their own. As suggested earlier, we think this can be achieved through narrative work. 10 Kate's experience shows how this can be done, in her case, by narratively connecting the recovery to aspects of the established self-narrative: Around my 24th birthday I had a big poppy design tattooed on my thigh. It's my way of remembering and respecting what I went through … Over the years I found I didn't forget the druggie life as I had feared. My role is now more about being an 'ex-user,' and I'm comfortable with that. I also made some progress on professionalising my past deviances, by using my experience with drugs to help others-recently becoming a board member at an organisation that provides needle exchange services and also studying trauma, loss and grief. This has further cemented my new identity and filled the void I felt (Kate 2013).
Getting the poppy tattoo and helping others as an ex-user helped Kate connect her addicted past and recovered present. The individual narrative she develops around her recovery makes recovery plausible within her wider self-narrative, in which addiction remains part of the story (McConnell 2016a). 11 Kate's new professional setting also provides constructive co-authoring as to the role addiction can play in her current self-narrative without her continuing to engage in self-destructive drug use.
Others' narrative work may include connecting with a past self (that may or may not have really existed): When I was younger before I used drugs, I was such a lovely caring person and thoughtful person and a loving person and I used to be very thoughtful and do things for other people without even being asked … I feel like that's who I am becoming again you know (Nicole in McConnell and Snoek 2018).
Therapists can facilitate such narrative work by suggesting and encouraging helpful narrative threads that will embed the recovery within the self-narrative and by offering reinterpretations of the established self-narrative, such that drug use becomes less plausible (cf. McConnell and Snoek 2018). Once the positive evaluation of recovery and the developing recovery narrative are sufficiently embedded in an increasingly verified selfnarrative, the (former) 'addict' will feel that who they really are is someone who can recover, and that this recovery is their own.
Indeed, when the self-narrative structure has been sufficiently aligned with one's evaluative stance, it begins to serve as protection against drug-using temptations because drug use doesn't make sense in that self-narrative context: Oh it's not a real craving. It's more just like I think, oh that would be nice but it's not something that's a part of my life. (…) I remember what that feels like. That's … but it's not something that I really have to work through because it's just not a part of my life. It doesn't fit into my life now. (…) I guess what I'm saying is that it hasn't really come up for me where I've had a craving and thought that, oh I could actually go and do that (R40C in Snoek, Levy, and Kennett 2016, our emphasis).
For this person, there is no ambiguity over whether drug use could be back in her life. Drug use just 'doesn't fit' on the continuum of mineness represented by the self-narrative structure anymore.
An important milestone in recovery is, therefore, to reach and pass a 'peak' of selfambiguity, beyond which it will be easier to identify with the judgment that one really is someone who can recover and that this self-narrative is one's own, than it will be to revert to the old addiction self-narrative. Maintaining this therapeutic direction of change in Type 3 self-ambiguity can be difficult outside the clinic because temptations and unhelpful social influences can reverse the process. Successful recoveries, therefore, typically depend on supportive environments where family and friends (and ideally, society at large) provide constructive co-authoring that helps the recovering person identify increasingly strongly with the self-narrative, value, and embodied experience of recovered life.

Conclusion
To conclude, we have suggested that self-narrative structures represent another continuum of mineness, which can yield an additional, third form of self-ambiguity when it clashes with one's judgment of who one is. We have illustrated how this and other forms of self-ambiguity we identified occur in addiction and have argued that going through self-ambiguity, especially of the third form we introduced, is a necessary component of recovery from addiction. The implication of this argument is that treatment programmes should encourage addicted persons to go through and narratively resolve their self-ambiguities. Notes 4. See Pugh (2020) for references on essentialist self-discovery and existentialist self-creation approaches to authenticity. 5. This kind of conflict in the second aspect of self-ambiguity doesn't necessarily involve a clash between two mutually exclusive ends. One might also be unsure about which of two selfinterpretations best reflects who one is. For example, in the aftermath of being scammed out of one's life savings, one might wonder whether to see oneself as a victim or a survivor. 6. For a more detailed description of this narrative view, see McConnell 2016a. 7. Additional difficulties arise within the evaluative continuum when people struggle to accurately identify their values, i.e. they are not just uncertain about how to rank their values but also lack clarity about exactly what their values are in the first place. Our primary aim in this paper is to illustrate (three kinds of) self-ambiguity understood as an uncertainty over where a given X falls on one's continuum of mineness. Therefore, for the purposes of illustration, we focus on cases where X itself is transparent to the agent, while acknowledging that this may not always be the case. When there is unclarity about the elements themselves the ambiguity of which one is trying to resolve, this surely can complicate self-ambiguity and its resolution. 8. Three continua of mineness that can (mis)align yield more possible combinations: The fundamentalist Christian could have a Christian self-narrative and corresponding embodied experience, but a conflicting judgment that he is homosexual. This is theoretically possible but practically unlikely in this case because such a judgment will usually only be triggered by a change in embodied experience. 9. We don't deny that it also feels alien to act contrary to one's conception of a good life when opportunities to pursue that life are available. We simply point out that sometimes that alienation can be outweighed by the greater alienation from acting contrary to one's deeply established self-narrative. 10. For a more detailed account of narrative work in addiction, see McConnell and Snoek 2018. 11. For an extended discussion of Kate's case, see McConnell 2016a.