Solving the self-illness ambiguity: the case for construction over discovery

ABSTRACT Psychiatric patients sometimes ask where to draw the line between who they are – their selves – and their mental illness. This problem is referred to as the self-illness ambiguity in the literature; it has been argued that solving said ambiguity is a crucial part of psychiatric treatment. I distinguish a Realist Solution from a Constructivist one. The former requires finding a supposedly pre-existing border, in the psychiatric patient’s mental life, between that which belongs to the self and that which belongs to the mental illness. I argue that no such border exists, and that attempts to find it might even render the felt ambiguity worse. Instead, any solution must be constructivist; the patient (and others) should deliberate and discuss what to identify with or not. I further argue that psychiatric patients need not see their mental illness as wholly distinct from themselves to avoid ‘identifying with their diagnoses' in a problematic way. Finally, we can excuse problematic behaviour by mentally ill people – in fact, we can do so in a more nuanced and constructive way – while rejecting the view that the mental illness is wholly distinct from the patient’s self.


Introduction: constructivism, realism, and reification
Psychiatric patients sometimes ask where to draw the line between who they aretheir selvesand their mental illness. Is this really me? Do I really feel this way, or is it just my anxiety acting up? Do I have an assertive personality, or is my current behaviour a mania symptom? Friends and family can likewise say things like 'Right now, I don't know whether it's him or the illness talking' (Erler and Hope 2015;Karp 1992;Sadler 2007, 116). We want to know where our selves end and our illnesses start, both for therapeutic reasons and to escape guilt, shame, and blame. The latter is a common theme in texts written by both psychiatrists and sufferers, and frequently discussed in self-help groups (Karp 1992;1994;Wiener 2011;Miklowitz 2011, 7-9. For popular examples, see Sharpe 2020; Hedrick 2015; "It Is Simple: I Wasn't Me" 2014).
Psychiatrist John Z. Sadler (2007) calls said demarcation problem the self-illness ambiguity, and argues that solving said ambiguity is an important part of psychiatric treatment.
Sometimes talk of solving the ambiguity is itself ambiguous. Solving the ambiguity might mean that we draw a line and decide that traits, behaviours, thoughts, and emotions that fall on one side should be considered illness symptoms and those that fall on the other side part of my self. Call this a Constructivist Solution. A good example of this approach is provided by psychiatrist Mark Rego (2004). He writes that people with certain disorders often benefit from regarding them as external to their real selves, and draws on the works of philosopher Harry Frankfurt (which I will come back to in subsection 4.2.). He writes of patients deciding to regard their problems as external to their selves rather than part of their identities.
There is room for variety within Constructivism; there is a big difference between on the one hand my doctor telling me that she thinks it would be prudent to consider this or that a mere illness symptom, not a part of me, and on the other hand my engaging in deep deliberation and contemplationperhaps together with a clinician and/or my loved ones (De Haan 2020)trying to figure out what kind of person I truly want to be, and which thoughts, emotions, behaviours and so on I should subsequently either embrace or reject as mere symptoms. Nevertheless, in both instances, we assume that a border between self and illness is something we must construct. There is no pre-existing border to discover.
I will call the opposing view, that there is a pre-existing border to discover, Realism about the self-illness distinction, and solving the ambiguity this way a Realist Solution. In some self-management and psychoeducation literature, this idea is made fairly explicit (e.g. Miklowitz 2011). Other books and articles on self-illness ambiguity can easily be interpreted in a realist way, even though it is not the only way to read them (e.g. Sadler 2007), and we see psychiatric patients adopting this way of thinking in studies (Karp 1992, 155). Gerrit Glas laments that reification of self and illness and a scientistic view on the mental illness as a neurological dysfunction separable from the person are ubiquitous in psychiatry (Glas 2019, Ch. 1).
There are several reasons why Realism as well as reification hold a certain allure. If I think of my self as disordered, my prospects for self-management might look bleak. If I think, instead, that the illness is wholly distinct from my self, I might paint the latter as rational and ordered, and be optimistic about its ability to properly control the former (Wiener 2011, 452-453). Attempts to deal with guilt, shame, and blame might also push people towards this view. Mike Hedrick writes that he used to suffer from internalized stigma, and 'crazy hung over my head like some thick black fog'. He then came to see his schizophrenia as something 'chemical and biological' and therefore not him; this made him feel better about himself (Hedrick 2015). Similarly, 'Cara' writes in capital letters that she was not herself when she hurt people in the past, interestingly adding that she is fully aware that this sounds like a cowardly excuse ("It's Simple: I Wasn't Me" 2014). Why does it sound that way? Presumably because it seems like she tries to shirk responsibility. However, if the mental illness and the self are actually two distinct entities, and it was the illness rather than the self which acted on these occasionsif this is furthermore not how we choose to think of ourselves and our mental problems, but a fact we have discoveredthere is nothing cowardly about pointing out the facts. Finally, Sadler talks about the importance of having psychiatric patients realize that they can fight the illness without fighting themselves (Sadler 2007, 117; see Jeanne 2011 for a popular version of the idea). It is easier to picture yourself doing that if you also picture the mental illness, not as aspects of your mental life that you regard as destructive and choose to distance yourself from, but as its own separate thing.
In this paper, I will often talk of the mental illness as a thing when I discuss the Realist view. However, it might be possible to come up with a theory of the relation between self and illness that is Realist, and advocate for a Realist Solution to the self-illness ambiguity, without falling into any objectionable reification. Many of my arguments would be problematic for this imagined version as wellthe problem of how to find the border through introspection remains, as does finding a philosophical theory of the self which supports seeing the illness as distinct from my self.
The structure of this paper is as follows: In Sections 2 and 3, I argue that attempts to scrutinize every part of one's psyche to find out what is what are unlikely to succeed. Every questioning thought can itself be questioned; there is no place where the buck must stop. The problem of infinite questioning would make it hard to find a pre-existing border even if it were there, but I argue in Section 4 that we have no reason to think that such a border exists. For both these reasons, we should give up any hope for a Realist Solution to the self-illness ambiguity and opt for Constructivism instead. In Sections 5 and 6, I argue that we have some leeway in how to construct our narrative selves, and that doctors and other clinicians ought to have a permissive attitude when it comes to self-construction. Clinicians should accept the possibility that one patient's ego-alien mental illness symptom might be another patient's personality trait. I discuss some legitimate problems that identifying with one's diagnosis might lead to in Section 7, but argue that these can be avoided without constructing everything related to one's mental illness as external to the self.
Finally, in Section 8, I return to the question of moral responsibility and blame. As previously mentioned, talk of one's mental illness as separate from one's self is often used to escape guilt and blame. Does this mean that a generous self construct, big enough to incorporate much of the mental illness within its borders, leaves the sufferer with no excuse for problematic behaviour? Fortunately, this does not follow. On the contrary, seeing one's mental illnessor at least many of its symptomsas occurring inside one's self rather than attacking said self from outside, allows for a nuanced handling of excuses and responsibility.

Never-ending questioning
There is a fundamental difference between scrutinizing a part of one's bodysay, one's digestive systemand scrutinizing one's psyche, when trying to draw a line between illness symptoms and normal, healthy functioning.
Suppose that I suffer from Irritable Bowel Syndrome. Researchers studying this issue suggest that sufferers keep a diet low in certain kinds of carbohydrates (e.g. Magge and Lembo 2012). If I am an IBS sufferer trying to handle my problems through diet, I likely have to pay close attention to my stomach after eating, to see if it starts to swell up, if I feel any pain, or if the food I eat is digested without too much trouble. Managing a physical problem like this can be hard enough in itself; stomach pains might also bring down my mood, and thereby influence my thoughts and actions. Still, IBS only directly affects my digestive system, and it is to my digestion that I attend when trying to determine how well my diet works. The observing subject is thus separate from, even if affected by, the observed object.
Not so when I observe my own psyche, looking for mental illness symptoms. The observing subject and the judgment they just made can be turned into a new object of scrutiny the very next moment. This process can go on and on without end, forever drawing mental phenomena out of the subject and into the object analyzed, as in the following stream of thought: Maybe I should write a novel. Yes, I really should! I had an idea the other night, which would make for a really original and unique novel. No, wait; I have a mood disorder. So maybe this is just my disorder speaking? Maybe I don't want to write a novel, after all; it might have been my disorder giving me wild ideas. Well, good thing I caught it. Good thing I've learnt to selfmonitor and take control of my mental illness. But wait, am I really in control right now? Maybe I'm being over-confident, thinking that I'm in control. Maybe the fact that I think I'm in control, shows that I'm not in control, because only a person on the verge of mania would think they have amazing self-control despite their mental illness. Of course I can't control myself. The disorder does what it does. Or … maybe the previous thought, that I was in control, was the authentic one? Maybe that was really me, and the idea that I'm not in control is a depressed thought. I'm actually going into depression right now. That's why I immediately stopped myself, when I first thought it would be a good idea to write a novel. There's nothing wrong with writing a novel. Lots of people write as a hobby, even if they never get published. Also, people often say that even though the chances are slim, it makes sense to tell yourself that you're one of the few who will get published, because confidence increases your chances. That's it: I'm gonna write a novel, and it will get published, sell loads of copies and receive great reviews. I'll be a literary star! No, wait, that definitely seems like a manic thought, I guess I'm going manic after all … Every time I think the thought 'maybe my previous thought was just a mental illness symptom', this new thought can be doubted the same way. Wiener (2011), studying a self-help group for people with bipolar disorder, notes how they constantly get caught up in endless analyses of their own mental states, even emotions and reactions that seem perfectly commonplace, such as being excited about a job interview or angry and sad over one's apartment being burgled. 1 It should not come as a surprise that patients urged to analyze themselves to find the border between self and illness often get stuck in endless loops doing so. In a best-case scenario, you find some phenomenological distinction between that which feels like you and that which feels external, or you are struck by a strong intuition telling you what is what. But in the absence of such experiencesand the support group participants discussed by Wiener do not seem to have had such luckthere is no place for the buck to stop. It is always possible to keep going; to keep questioning your questioning, and analyze your analysis. There are no deliberative thoughts about what to do or to believe, no type of mood or emotion, that could not, in principle, be caused by your mental illness or constitute a mental illness symptom. Furthermore, the very process of scrutinizing and questioning every aspect of your mental life might itself undermine the possibility of simply feeling what is you and what is not.
In discussions about hedonism, it is frequently pointed out that if you want to be happy, it is better to pursue things you find interesting and worthwhile than to focus on achieving happiness. It might make sense to ask yourself which kind of life would make you the happiest and go for that, but scrutinizing every activity, every acquaintance, every little thing to determine how much happiness they produce inside you will not enable you to efficiently increase your happiness level; it will rather undermine this pursuit. Something analogous might be going on when people focus hard on which parts of their mental lives are truly them. In order to find ourselves (as well as to be happy) we need to look out at the world, how we live and what we do in that world, rather than getting stuck introspecting on our own mental lives. 2 Even if Realism were true, this never-ending questioning would pose a serious problem for the attempt to find a Realist Solution to the self-illness ambiguity. In the next section, I will show this with the help of classic philosophical manipulation scenarios.

Never-ending questioning in manipulation scenarios
The argument in Section 2 is not new. Philosophers have argued before that even if another person literally inserts thoughts into my head, it isunder certain conditionsfutile to try to find the border between what is me and what is inserted. A thought experiment by Christine Korsgaard (1996, 162-163) nicely illustrates why.
Suppose I agree to partake in an experiment, where scientists control my brain for one day. They will not make my arms and legs move as if I were a puppet, but rather work through my ordinary mental channels. I wake up on the day of the experiment. I have breakfast and get dressed as usual, and sit down at my desk to work on a paper. Suddenly, it strikes me that I only work on my paper because the scientists made me do it. It now seems silly to sit down and write. In a spirit of rebellion, I decide to ditch work and go shopping instead. But then I realize that the scientists made me think and do all of this too; they made think that they are the ones who made me write my paper, they put the rebellious feeling into me, and made me ditch my work and go shopping instead. So maybe I should just sit down and do nothing, refusing to cooperate … but wait, the scientists are making me think this as well! Korsgaard's point is that even in a science fiction scenario where other people make me think and act, this endless second-guessing gets me nowhere.
In Korsgaard's scenario, the scientists take complete control of my thoughts, but constant second-guessing can be as futile in a partial control scenario. Imagine that only fifty percent of my thoughts during the day of the experiment were generated by the scientists (once again, working through my ordinary mental channels), and fifty percent were my own. I know this much, but not which thoughts are which. There exists, in principle, a real border between my own thoughts and the implanted ones to discover. But unless I can ask the scientists and expect a straight answer, the inserted thoughts are clearly out of character for me, or there is a difference in phenomenology between my own thoughts and the inserted ones, trying to find this border by endlessly analyzing what I think is as pointless as before. Instead of thinking first that the scientists made me work, second that they made me question whether it makes sense to work, third that they made me go shopping instead, I would think first that it might be the scientists who made me work, and second that it might be them who made me question it … Still futile.
I might just as well ignore that I am part of an experiment, and go about my day as usual (and yes, this would also be something that the scientists made me think and do as I thought it and did it, but I do not ponder it anymore).
As I mentioned earlier, finding the border is hopeless unless I can ask the scientists, the implanted stuff is out of character, or there is a phenomenological difference. We can therefore imagine scenarios where the border is detectable. Suppose, for instance, that I am the victim of an unusually clumsy manipulatorlet us call him Charleswho wants to take over the world. He has somehow managed to insert a gizmo in my head that enables him to insert the occasional thought directly into my mind. As I sit down working on my latest paper, pondering whether an argument I just came up with is a good one or not, I suddenly think, out of nowhere 'Charles is the best! Charles should rule the world! I should print flyers for Charles and deliver them all over town!' Next, my thinking bounces back to normal. The implanted thoughts are radically out of character for me and disconnected from my regular stream of thought; perhaps they also feel inserted. In this scenario, I can easily draw a line between my real thoughts and the implanted ones, which I might do best to ignore as far as I can.
There are mental illness symptoms that have some similarity to the Charles case: thoughts that suddenly pop up, are radically out of character, sometimes with an 'inserted' feel or phenomenology to them as well. Since such thoughts are not literally implanted in my head by someone else, it is not obvious that my best option is to ignore them. It might be, or it might be better to think of them as expressions of hidden or repressed parts of my self, regardless of how alien they feel. Still, most psychiatric patients do not suffer from this specific symptom (so-called thought insertion), and will look in vain for a clear border between their own mental life and that which belongs to the illness.
Philosophers who discuss the futility of separating my self from the implanted thoughts and emotions in sci-fi scenarios do not suggest that I should, instead, uncritically accept and follow every thought, emotion, or impulse that pops up. Rather, their point is that I should make use of normal deliberative methods when deciding what to do (Korsgaard 1996; Jeppsson 2020).
If I worry about choosing the wrong path, I better conscientiously deliberate about what to believe and do, by evaluating my reasons in light of my values and empirical facts as far as I know them. This can certainly be difficult enough. There are no guarantees that the scientists (or the mental illness) have not, so to speak, gotten to me all the way down, and changed all my values and preferences in some problematic way. Deliberating together with others could, to some extent, alleviate that worry, but it does not go away. Nevertheless, regardless of those difficulties, carefully thinking through my options, evidence, reasons and so on, remains a better strategy than gazing inwards on my own mind. Setting the sci-fi scenarios aside to focus on mental illness, I might suffer from cognitive and emotional impairments that make it hard to deliberate about what to believe or to do. But such impairments will adversely affect introspective attempts to find the border tooin addition to the never-ending question problem. Even if my emotions and thinking are impaired, 'looking out' at my options and reasons might be a better strategy than 'looking in' at my own mind when trying to figure out what to do.

There is no good reason to believe in a pre-existing border
My arguments in Section 3 shows that searching for a pre-existing border between self and illness would often be futile even if said border existed. In this section, I further argue that we have no reason to believe that it exists in the first place. What belongs to my self and what lies outside of said self depends on which self-conception we use, but finding a conception that can support Realism about the self-illness distinction is not easy.

Sadler's five-aspect self
Some philosophers and scholars argue that the self is comprised of a number of different features. Sadler (2007, 114) lists the following five: 1. Agency: The sense that one is a purposeful actor, navigating the flux of life. 2. Identity: A distinction between oneself and others and the world. 3. A life trajectory, a sense of purpose and the future. Sadler concedes that it can be meandering and unfocused, but there is still an experienced life trajectory of some kind. 4. History, which is unique to the individual, and influences the other aspects in various ways. Once again Sadler writes that this can be messy, patchy, cloudy, and so on, but it is there. 5. Perspective: A unique viewpoint and experience of the world.
Clearly, the above five aspects are all important. Sadler is right to argue that patients' sense of agency, identity, life trajectory, personal history and perspective should not be overlooked in psychiatric treatment; psychiatrists should not be narrowly concerned with the presence or absence of easily defined symptoms, to the extent that they forget about the patients' selves (ibid, 114;Sadler 2008). Suppose that a patient with a psychosis disorder manages to get dramatic symptoms, like hallucinations and delusions, under control through medication, but ends up feeling devoid of agency, as if she just floats along with no control over her life, unable to picture a future for herself. This patient would have very serious problems still, even if they are less dramatic than her previous psychotic symptoms.
Nevertheless, as fruitful as it can be in certain contexts, this five-aspect self-conception fails to support a clear distinction between self and illness. In mental illness, one or more of these aspects are often negatively affected. Still, an illness can have a negative effect on my X, make my X deteriorate and so on, without therefore being a separate thing which can, at least in principle, be distinguished from my X. Comparing mental illness to physical illness can actually clarify this point.
It might be surprising to see me turn to physical illness, since physical-mental illness comparisons are normally used when people want to draw a sharp line between self and illness. For instance, Sadler writes that people with physical illnesses do not see infection-causing germs or tumours as part of their selves; the implication being that neither should people with mental illness see their illnesses in that way (Sadler 2007, 116, 118). Now, we should not generalize too much about how physically ill people see themselves. Some people with chronic and/or life-altering physical illnesses argue that these do shape who they are (see, e.g. Fraser [2015], on why he prefers to call himself 'a diabetic' rather than saying that he 'has diabetes'). But insofar as physically ill people do see their illnesses as separate from their selves, it might be because they do not see their bodies -and by extension, all that goes on thereas part of their selves to the same extent as their minds.
For the purposes of this paper, I take no stand on whether our bodies are part of our selves, and if so, to which extent, or if there even is an objective truth in the matter. I focus on people's views rather than actual metaphysics. Philosophers have defended the view that we are entire organisms, but also that we rather are our brains, or that we are persons more analogous to computer software than computer hardware (Olson 2007). Presumably, non-philosophers also have different views on how central (the non-brain part of) the body is to their personal identity. Thus, when physically ill people do see their illnesses as something outside their self, it need not be because they think there is a sharp distinction between self and illnessit might be because they, rightly or wrongly, do not think their body is identical to their self.
Seeing one's mental illness as distinct from one's self might be more analogous to seeing one's physical illness as distinct from one's physical body. For at least some illnesses, this view seems implausible. We might see infectious germs and even tumours as intruders that should be killed or taken out, but an illness like diabetes is a different matter. Suppose I have type 1 diabetes, and the Islets of Langerhans in my pancreas have permanently stopped producing insulin. Does it make sense to say, in this case, that 'the diabetes' is distinct from 'my body'? That 'the diabetes' is distinct from my pancreas? From the Islets? I do not think so. My diabetes is a state of primarily the Islets, then the pancreas, and more indirectly a state of my whole body. But it is not a separate thing, intruding on the body, the way viruses or even tumours at least arguably do.
Similarly, suppose that I have depression, and this negatively affects my agency, such that I have lost all motivation to do things and feel incapable of planning ahead. This does not imply that my depression is like a tumour, pressing on my agency and hindering it from doing its job, or like a viral or bacterial infection intruding on it. My depression could just as well be seen as a state of my agency and my self, like type 1 diabetes is a state of my Islets, pancreas, and body.
Thus, this kind of multi-aspect self-conception fails to support a distinction between on the one hand the self, and on the other hand the mental illness, a separate thing that attacks and intrudes on the self. How do other self-conceptions fare in this regard?

The deep self
In philosophy, there is a long tradition of more or less narrow self-conceptions, according to which one's true or deep self is identified only with those traits, habits, desires and so on that one endorses or values (Frankfurt 1971;Watson 1975;Wallace 2013). According to some theories, one's deep self is also aligned with what is truly valuable (Wolf 1987;Wolf 1990;Strohminger, Knobe, and Newman 2017). 3 It should be noted that none of these self-conceptions allow us to draw a line between on the one hand everything non-pathological, and say that this is part of the psychiatric patient's self, and on the other hand the mental illness, and say that it is not. Non-pathological bad habits, for instance, can fall outside of my deep self if I want to get rid of them and/or do not value them. Wolf brings objectively true values into the picture as well. The more demanding the view, the more non-pathological desires, preferences, feelings and so on will fall outside the deep self.
A realist about the self-illness distinction might accept that the mental illness has company outside the borders of the self. Much worse, from the perspective of the realist, is that some aspects of a person's mental illness might belong inside their deep self. Some people do identify with their madness, and deep self views do not imply that they are all wrong in doing so (Rashed 2019).
Suppose, for instance, that I suffer from schizophrenia. I have some cognitive difficulties, as well as frightening and uncanny experiences of derealization, depersonalization and threatening hallucinations. However, I think my schizophrenia has certain positive sides as well; my unusual experiences and way of thinking are creatively useful. I write and paint, inspired by my schizophrenic experiences. Eventually, I become better at coping with the more disruptive and frightening aspects of my disorder, to the point that I think of them as a price worth paying for being a unique kind of artist. Alternatively, I might detest my cognitive difficulties and frightening hallucinations, while seeing vaguer derealization and depersonalization experiences as more good than bad due to the inspiration they bring. I thus embrace and value at least some aspects of my mental illness. Wolf (1987) does write that my values do not count as truly mine if they are 'insane'. However, she illustrates her claim with the story of an evil dictator, whose son learns from early childhood that torture and genocide of one's opponents is perfectly fine. The son's problem is that he cannot see that these things are, in reality, horrible. (For the sake of discussion, I simply accept the underlying metaethics.) Whether he fulfils the criteria for any psychiatric disorder is beside the point. In my schizophrenia example, my strange experiences might be disconnected from reality, but I embrace them based on aesthetic and artistic values that cannot be summarily dismissed.

My counterfactual self or usual self
Finally, counterfactual statements such as 'If I had not been mentally ill, I would not have done A', or 'if I had not been mentally ill, my life would have been very different' can be true, but that is still no evidence that there is a pre-existing border between 'my self' and 'my mental illness' to discover. Counterfactual statements of the kind 'If I had not had A, I would not have done X' or 'If I had not been B, I would not have done Y' are frequently true for non-pathological, even fully endorsed, preferences, hobbies, and personality traits as well. (See Dings and Glas [2020, 342] for using counterfactual questions and statements in an effort to find oneself.) Miklowitz (2011, 64-66) advises patients to find their disorder-independent self by considering what they are 'usually' like, when they do not suffer from mood swings. This suggestion is somewhat undermined later in the same book, when patients are urged to scrutinize their moods and thoughts every day and conscientiously fill out symptom chartsthere are no 'usual' days during which such vigilance is not needed (ibid, Ch. 8-9). In any case, charts or no charts, there are many mentally ill people who are 'usually' disordered; if the mental illness has gone on for long enough, the way one was before its onset might not serve as a feasible baseline either (Kennett 2009, 103-104).

Summing up
We might conceptualize the self as a set of crucial features or capacities, and note that mental illness often has a negative impact on one or more of these. For instance, my mental illness might negatively affect my agency. Still, this kind of self-conception does not imply that mental illnesses are external to the self. It remains open whether the illness is external to my agency and self, infecting it like a germ or pressing on it like a tumour might press on an organ. We can regard it, instead, as a deteriorated state of my agency, the way type 1 diabetes is a deteriorated state of the Islets of Langerhans, or osteoporosis is a deteriorated state of the skeleton.
If we rather think of the self as the deep and true core of who we are, the seat of those desires and preferences we fully embrace and value, it is also an open question how much of a person's mental illness goes inside or outside the borders of this deep self. It might vary from person to person; one person's mental illness symptom could be another's cherished personality trait.
Counterfactual statements of the kind 'if I hadn't been A, I wouldn't have done X' can be just as true when substituting something innocuous like 'married' or 'a Star Trek fan' for A, as when sticking a mental illness label like 'bipolar' or 'schizophrenic' there; in more chronic cases, there might be no 'usual' self to compare with either. I doubt that it is possible to come up with a plausible conception of the self which is guaranteed to exclude every mental illness symptomat least not unless an ad hoc condition, according to which mental illnesses and their symptoms simply cannot be part of a person's self, is tacked on.

Narrative identity and constructing a self
Let us move on, then, from the idea of discovering a pre-existing self with pre-existing borders, to the idea of constructing one's self. 4 I do not argue that it is possible to radically invent oneself with few or any restrictions; however, I can still have some leeway in how I interpret my life story, and subsequently how I interpret the actions, reactions, emotions and so on of the main character (i.e. myself.) A self-narrative explains how events and experiences lead to action in a way that is intelligible and meaningful. It makes sense of my history, intentions, and plans for the future (e.g. Mackenzie 2014). For the purposes of this section, I need to distinguish between a subject's narrative life story and their narrative self as narrowly conceived. 5 I will use 'narrative self' here to refer to the narrator or the main character of the story of my life. A fictional storybe it a novel, movie, or something elsecentred around a single main character normally involves other characters too, an external environment, and various events that the main character reacts to and handles but did not initiate. The same is true of our own life stories. To take an everyday example: Suppose there is a rainstorm, and my basement gets flooded. The actual storm was external to me, and further beyond my control, albeit part of my life story. However, how I deal with and react to the consequences of the storm can influence how I see myself, i.e. how the main character of the story is portrayed. I might blame myself for not being better prepared and see myself as sloppy and irresponsible. Alternatively, I might think that it was unforeseeable, all that matters is how I deal with the problem here and now, and see myself as more of a responsible and conscientious person.
One way to think about my mental illness, is as analogous to an unforeseeable rainstorm and flooded basement. Both the onset of mental illness in the first place and all its symptoms might be considered things that just happen to me. They are included in the story of my life, just like the rainstorm and its consequences, but they are not part of me any more than the storm is. How I handle my mental illness through seeking treatment, taking medication, employing various self-management techniques, etc., can still be important for how I see myself. I might think that I am brave for admitting that there is something wrong with me and seeking treatment, that I am conscientious because I take my pills each day, and so on, but I do not see myself as, e.g. a melancholic person because I have depression, any more than I see myself as a wet person because my basement is flooded.
Let us look at a more elaborate example. Here are two stories we can tell about the same series of events, described in Bracken and Thomas (2001, 727).
Version 1: A woman had been hospitalized twice before during the past six years, and diagnosed with affective disorder. She was now back in hospital again, due to pressure of speech, a labile, irritable mood, a preoccupation with religion and past events, overactivity and over-spending. Her neurotransmitters were negatively affecting her thoughts, emotions and behaviour; they had to be stabilized by psychiatric drugs.
In this version of the story, the affective disorder and even events in her own brain are construed as external to her selfas rainstorms. She sees, in the words of Mark Rego, 'the problem as an event in his or her personal history but reject[s] the notion that it is part of his or her identity' (Rego 2004, 318). We might think that the way she deals with psychiatric appointments and medications reveal what kind of person she is, but the mental illness is external to her.
However, this is not the story the woman ended up telling herself. Rather, she ended up accepting the following one.
Version 2: In long conversations with a hospital nurse, the patient described a complicated family life. 6 Her relationship with her elderly mother-in-law, who lived in the same house, was particularly strained. On the one hand, the patient felt that her mother-in-law tried to usurp her role in the family, and boss everyone around; on the other hand, the mother-in-law was chronically ill, and the patient had to care for her. The patient eventually reinterpreted her alleged mood disorder symptoms as comprehensible reactions to a very tense situation: She had become overactive and spent too much money because she wanted to reclaim her role as mother, and as an important figure in the household. She talked too much and was irritable and hostile since she desperately tried to get her family's attention and make them see her point of view. She was preoccupied with the past and religion because she tried to put her tense and stressful situation in perspective; thinking about the past and religious matters was also a way for her to draw strength. This reinterpretation of her own behaviour in comprehensible rather than pathological terms eventually allowed her to talk things out with her husband, and achieve more long-term stability in her family situation.
In this version of the story, other people, like her mother-in-law, are obviously external problems that the woman has to deal with. Her mother-in-law living in her house and being irritating in various ways is part of her life story, but the mother-in-law is (obviously) not a part of herself, but a distinct person. Still, her intense speech and spending, her preoccupation with religion, etc., are now seen as ways in which she reacts to troubling circumstances, actions that she performs and interests that she has, rather than symptoms of an external illness that attacks her.
Bracken and Thomas promote a radical 'post-psychiatric' agenda, and end this story by noting that the main character had managed (at the time of their writing) without psychiatric drugs for twelve months. However, one might be more accepting of traditional psychiatry and still think that insofar as seeing her mental illness symptoms as comprehensible reactions rather than something external to her self helped this woman to recover, it was a good choice. In the next section, I will discuss better and worse ways to think of one's self and the mental illness. Since there is no fact of the matterno border between self and illness which is just there, independently of where we would like to put itwe should evaluate constructivist solutions to the self-illness ambiguity by ethical standards.

Different interpretations of self and illness
Although there are different conceptions of recovery in the literature on mental illnessresearchers tend to stress symptom remission, having a social life and ability to work/ study, whereas people from the patient side stress the importance of feelings of hope and purposeit is fairly uncontroversial that both wellbeing and function matter (Slade 2009, 35-38). For starters, we can thus use recovery, in terms of feeling better and/or regaining (social, job-related, etc.,) function, to measure whether a certain way to think of one's self and one's mental illness was beneficial or not.
There might be no one-size-fits-all solution here; what is most conducive to recovery likely varies between patients. There are some general considerations in favour of seeing the illness and its symptoms as more integral to oneself, and others that point in the opposite direction. Looking for intelligibility in psychiatric patients' obsessions, delusions and non-standard reactions is often helpful in therapy and improves patients' wellbeing and trust (Bortolotti 2015;Ritunnano, Humpston, and Broome 2021;Ritunnano, Stanghellini, and Broome 2021;Jeppsson 2021;Tekin 2011). There is a limit to how much we can externalize the mental illness and still see symptoms as intelligible and meaningful in a way that a tumour is not. Furthermore, if I have lived with a mental illness nearly all my life, it might be better to embrace at least some core aspects of it as parts of me, as the way I am, than to think of my self as essentially 'normal' and the illness as an external attacker. If the illness will be with me for the foreseeable future, the mentally healthy self I imagine might be a mere fantasy, and I might feel constant frustration and dissatisfaction with my real situation (Kennett 2009). On the other hand, Mark Rego (2004) argues that seeing the mental illness as external is often beneficial for patients with conditions like addiction or phobia; since the patient has so little direct control over their behaviour, they get better at controlling and treating themselves when they think of their condition as something external which must be controlled through indirect means.
We might have reasonable rules-of-thumb for how to weigh these (and other) considerations against each other with different patient groups, but I doubt that we can ever find stringent rules for when a patient is better aided by thinking of the illness as something external or the opposite. Furthermore, it is important for clinicians to respect the patients' own views on the matter.
Much has been written about the extent to which medical doctors either could or should be 'value neutral' when interacting with patients (e.g. Hoehner 2006). Nevertheless, I hope the following statement is weak enough to be widely agreed upon, at least as a pro tanto principle: Weak Neutrality: Doctors and other clinicians should not try to argue their patients into accepting controversial philosophical (or religious, ethical, etc.) positions.
Of course, we cannot draw a sharp line between the controversial and the uncontroversial. Still, some claims and positions fall clearly on the controversial side. The thesis that there is a pre-existing border between self and illness for psychiatric patients' and their doctors to discover is highly controversialat least by implication, since it is hard to find a philosophical self-conception capable of supporting this distinction. Clinicians, therefore, have a reason not to push it, even if they personally find this view appealing. 7 Furthermore, when clinicians openly talk about deciding to think of the illness in a more or less externalized way, they should respect the patient's own views as much as is possible while still suggesting the strategy that would be most conducive to recovery according to their experience.
This respect for the patients' own views should also be extended to the related topic of authenticity. It is controversial both how to conceptualize authenticity more precisely, and how much it matters. Erler and Hope (2015) write that most people find authenticity important, but a substantial minority do not (neither do philosophers agree on this topic). Among psychiatric patients who do care about being authentic, some believe psychiatric drugs reveal their authentic selves whereas others see them as suppressing it. In the latter group, some people do not want to take drugs for this reason, whereas others believe that their positive effects make reduced authenticity a price worth paying. Some patients believe that their mental illness is part of their authentic selves, even if they also suffer from it, whereas others see their authentic selves as mentally healthy. In clinical settings, there should be room for all these views, at least as long as they are not down-right destructive for the patient.
When assessing whether the way the patient thinks about their self and their illness is problematic or not, it is important for clinicians to be clear on what the implications are. Some important self-management skills can be mastered regardless of whether the patient sees the mental illness as part of them or as external. Think, for instance, of learning to recognize early signs of a relapse into psychosis. Suppose that feeling enthusiastic about everything and being unusually energetic and creative is a reliable precursor of psychosis in me, but whenever I enter this state of mind, I fail to see what it signals -I just think of myself as a naturally happy and creative kind of person. This is a problem. Still, it is not necessary for solving my problem that I come to think of this (in itself pleasant) mood as alien, and not truly part of me. There is no contradiction in seeing it as my mood, part of my personality, and simultaneously as a sign that trouble is on the way.

Avoiding the pitfalls of 'identifying with one's diagnosis'
In the previous section, I argued that it likely varies between psychiatric patients whether it is most conducive to their recovery and wellbeing to think of their mental illness as external or internal to their selves. However, 'identifying with one's diagnosis' is often held up as problematic in itself. In this section, I will argue that it can be, but it is not necessarily so. Clinicians might have reason to steer their patients away from problematic ways of identifying with their diagnoses, but they need not interfere with the views of patients who think of their mental illness as part of who they are and for whom this seems to work well.
It is problematic to identify hard with a particular diagnosis. Charland (2004) discusses psychiatric patients who spend much of their time in online support communities for people with a particular diagnosis; thus, both their identities and their (online) social lives come to revolve almost exclusively around the community members' shared condition. This way of identifying with one's diagnosis can lead to identity crisis if said diagnosis is changed. This can happen either because the patient gets a new doctor (psychiatric diagnoses have low reliability [Bentall 2004, 44-56;Aboraya 2007]), or if the diagnosis is removed from the DSM, as sometimes happens. For instance, Multiple Personality Disorder gave way to Dissociative Identity Disorder in 1994. They are largely similar, but the former diagnosis allowed for the possibility that the patient really had more than one personality, whereas the latter does not (Hacking 1995, 23-24). 8 Nevertheless, a changed diagnosis will only threaten one's sense of identity in case one identified with a particular diagnosis, rather than more broadly as a mentally ill or mad person. If I identify as mad, I need not fear a change of doctor or a DSM revisionbut I could still fear the prospect of turning 'normal', even if I feel conflicted and simultaneously long for it. Still, my problem in this situation is not so much that being mad is part of my identity, but that I hang my entire identity on this single peg. Doing so can be problematic even if the peg in question is not pathological.
Suppose I am a football player, and this is my entire identity. Football comes before relationships, day job and other hobbies, and I can see no other future for myself than succeeding in this sport and going professional. However, after a couple of knee injuries, I develop chronic problems and must give up the sport (a common enough fate). Suddenly, I do not know who I am anymore, what to do with myself, what to do with my time, whom to spend time with or be friends with, since I feel I have nothing in common anymore with my former teammates, and I had no other relationships to speak of.
There are important similarities between the former football player and a person whose entire identity and social life revolved around a mental illness that she later recovered from. Yet, we would not say that it is wrong and problematic in itself to identify as a football player. The problem is to do so solely, to the exclusion of all other identities (Dings and Glas 2020, 344).
Thus, to avoid identifying with my diagnosis or with my mental illness in a destructive way, I need not think of my mental illness as separate from my self. It is sufficient that I do not base my identity on a particular diagnosis, or base my entire identity on this. In the penultimate section, I will argue that seeing the illness as something separate from the self is not necessary, either, to excuse mentally ill people from problematic actions.

Mental illness and moral responsibility
So far, I have argued that we have no reason to believe in a pre-existing border between self and illness, and that even if it existed, attempting to find it through introspection and scrutinizing every aspect of one's mental life would often be futile. But what of the allure of Realism and reification that I mentioned in the beginning? If I find it useful for self-management and for 'fighting the illness' to think of it as distinct from my self, I can still do so. Guilt, blame, and shame are trickier matters; if I choose to see my mental illness as a separate thing, and choose to see the illness rather than myself as responsible for bad things that I have done, this can easily be perceived as a cowardly attempt to shirk responsibility for what I have done. A certain kind of complete exculpationit was (literally) not I who did all these things, it was the mental illness inside mebecomes unavailable when we reject Realism.
Nevertheless, there are other excuses available. Discussions about mental illness and moral responsibility occasionally proceeds from the assumption that mentally healthy people are harshly blamed almost every time they do something wrong, but this is not what our moral responsibility practices actually look like. Mentally healthy people are excused, fully or partially, all the timeand perfectly ordinary excuses can often be extended to cover mental illness cases too. The following subsection presents ideas that I develop in depth elsewhere, but albeit brief, it shows that mentally ill people (and their families, friends, and doctors) do not need the 'it was not me, it was the illness' exculpation to handle guilt, shame, and blame.
When philosophers discuss moral responsibility in general, or more specifically moral responsibility and mental illness, they tend to focus on necessary and sufficient conditions, and where to draw the line between responsibility and non-responsibility. However, our everyday interactions contain much more nuance; we frequently cut people slack because things were difficult for them, even when no disorders are involved. 9 Likewise, when looking back at bad things that we did under the influence of mental illness, we should remember how much we struggled and how serious our difficulties were and cut ourselves some slackperhaps lots of slack. We should urge others to see these things from our point of view and appreciate our difficulties as well. Nelkin (2016) and Wolf (1990, 86-87) explicitly discuss the role that difficulties play in diminishing a wrongdoers' responsibility. They argue that any plausible moral responsibility theory should make room for the following common-sensical principle: an agent is less blameworthy for doing wrong when abstaining from wrongdoing would have been very difficult. Nelkin discusses how difficulties come in at least two varieties: something can be difficult in the sense that it requires immense physical and/or mental effort, or it can be difficult because you have to make a sacrifice, i.e. give up something important, in order to do it.
Suppose that two people both fail to save drowning children. In the first case, a person simply looks on as a toddler drowns in a shallow pond. In the second case, a person fails to jump into a freezing cold river to save a drowning child. Depending on the details in the latter case, saving the child (or trying to) might seem like a supererogatory rather than morally obligatory act to many people. Nevertheless, let us assume that we have worked out the details of the scenario such that saving the child falls on the obligatory side of the obligatory-supererogatory divide. All else being equal, the person who fails to save a child from a cold river still strikes us as less blameworthy than the one who does not save a child from a shallow pond.
Someone for whom it requires intense mental effort to control their impulses, get a grip on their emotions, think things through, motivate themself to act, and so on, might thereby have serious difficulties when it comes to abstaining from certain wrongful actions or omissions. If the person does the wrongful thing, we can cite their difficulties as grounds for diminished responsibility and blameworthiness, without denying that it was still them who did it, not some separate illness entity. Now, mentally healthy people sometimes also use 'I wasn't myself' language when telling others that they are sorry for, e.g. previous rude and angry behaviour. 'I'm sorry, I wasn't myself, my sick toddler had kept me up all night'. This is really an appeal to difficultieswhen sleepless and exhausted it's difficult to control one's tempernot to be taken literally. The rude toddler-parent does not ask for a complete excuse either, just for some understanding and 'slack-cutting'. Similarly, I might talk about how I 'wasn't myself' because of my mental illness at the time I did wrong in this more metaphorical sense, asking for understanding. I must still accept, though, that appeals to difficulties cannot completely wipe away all responsibility on my part, because in a more literal sense, I still did the things I did, not some other entity. Nevertheless, this might be an advantage for maintaining healthy relationships to other people, since, as already noted, attempts to claim complete non-responsibility might be seen by others as 'cowardly'.
In addition, we also excuse each other all the time for ignorant acts, when the ignorance was not the wrongdoer's fault. Suppose my partner tells me not to take the car to work, since they need the car for a crucial errand that cannot be done by bus or train later. If they had failed to wake me up, if I did not hear anything because I was asleep, I am not to blame when I nevertheless drive off in the car. Psychosis but also extreme affective states might make me unable to understand crucial pieces of information or take it into account when I actin these cases, I should once again be excused because of my difficulties.
Summing up, we should not say 'I don't blame you, since it wasn't really you, it was the illness.' We should not think 'I should not blame myself, since I realize now it wasn't me.' But we can still say 'Well, I understand that this was extremely difficult for you'. We can still think 'Looking back, it would have required an enormous effort on my part, due to my condition' or 'looking back, I didn't know how to handle myself in that kind of situation, it's really not easy'.

Conclusion
Solving the self-illness ambiguity is arguably crucial to the treatment of mental illness. Still, there are different ways we can picture this solution. One is Realist: discover where the border between self and illness, which is supposedly there already, actually goes. Another is Constructivist: ponder and deliberatewith more or less help from other people, clinicians, friends and familyabout what to identify with.
I have argued that the attempt to find a pre-existing border suffers from two problems, each one serious enough to undermine that project on its own. First, observing my mental states is fundamentally different from observing my body. With mental states, the observer can always be observed in turn, and the questioning questioned in turn, in a neverending process, with no place for the buck to stop. Second, there is no good reason to believe that there even exists a border between on the one hand the mental illness and all of its symptoms, on the other hand the self.
Thus, if I am mentally ill, I have to rely on a more constructivist, interpretative effort to find out who I really am. Sometimes, thinking of the mental illness as something external to my self can be beneficial, but then again, sometimes thinking of mental illness symptoms as understandable reactions to events in my life can be more constructive and better aid recovery. Doctors and other clinicians can rightfully offer their expert opinions on what would be best here, but they should be open to the possibility that different approaches work for different patients, and respect their patients' own views.
'Identifying with one's diagnosis' can be problematic and detrimental to recovery, but psychiatric patients who think of their illnesses as part of who they are can avoid the pitfalls without externalizing the illness.
Finally, when it comes to exculpating mentally ill people for problematic behaviour, we can do this in a better and more nuanced way if we do not assume that the mental illness is something other than the person that sometimes takes over to run the show, but instead, see the person as an agent struggling with responsibility-diminishing difficulties.
6. See also De Haan (2020) for the social nature of constructing our identities. 7. Appealing, perhaps, for the reasons outlined in Section 1. They might feel more optimistic about the patient's self-management prospects and of fighting the illness without fighting the patient when they think of the distinction between patient and illness as concrete and real; they might also believe that they must ascribe any disruptive behaviour on part of the patient to a separate illness entity in order not to blame them. 8. Hacking quotes David Spiegel, a driving force behind the change. According to Spiegel, these patients do not suffer from having too many personalities, but from having less than one. 9. Perhaps we need hard and fast lines in the legal system; I leave that open. In interpersonal relationships, there is certainly no reason to assume that we must be able to divide actions into those the agent was one hundred percent responsible for and those that they could not help at all.